Re: [time-nuts] Frequency Counter Choice

[Oz-in-DFW]

On 10/22/2020 11:52 AM, jimlux wrote:
>  I wonder if one could implement the protocol using GPIO pins on
> something like a Beagleboard or Arduino clone?
>
> (yes, as a product, with the right connector and line
> driver/receivers, etc. - it would cost the same as a Prologix... but
> as a hack...)
>
> After all GPIB is a very old standard, and was developed back in the
> day of 1 MHz logic. I think you could do the handshaking in software -
> punch the bits one by one, assert DAV, etc.
> Listener reads bits one by one, then asserts the ACK or NAK
Actually it's bytes.  It's byte-serial not bit serial.

Most modern micros can pretty easily keep up with the bus - especially
if you use GPIOs that can generate interrupts. The only real challenges
are finding a reliable source of the appropriate connector - and driving
the 5V lines at the current levels required for a loaded bus.  Most
modern micros do not have 5V tolerant I/O but that's pretty easily dealt
with. Connectors can be had, NORCOMP's 112 series should be ideal.
https://www.norcomp.net/series/112-series and are available through
DigiKey, Mouser, Farnell, etc. TI's SN75160/SN75161 bus transceivers
still seem to be in production as SOICs as well.

I've noodled about building a board that was intended to connect to a
single instrument, but not really gotten serious about it.  I wanted
everything in the shop on Ethernet.

Oz (in DFW)

___
time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe, go to 
http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.

Re: [time-nuts] Frequency Counter Choice

[Elio Corbolante]

> >While I love my Prologix adapters, I wonder if one could implement the
> >protocol using GPIO pins on something like a Beagleboard or Arduino clone?
> >(yes, as a product, with the right connector and line driver/receivers,
> >etc. - it would cost the same as a Prologix... but as a hack...)
>

@jimlux:
Take a look at the following project:
https://github.com/Twilight-Logic/AR488 and
https://www.eevblog.com/forum/projects/ar488-arduino-based-gpib-adapter/
https://oshpark.com/profiles/artag
For more details or if you need a PCB for the Arduino Micro (32u4-v3-gbr)
please send me a private message

_Elio.
___
time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe, go to 
http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.

Re: [time-nuts] Frequency Counter Choice

[Bob kb8tq]

Hi

…. and now I have a copy :)

Bob

> On Oct 22, 2020, at 6:12 PM, Attila Kinali  wrote:
> 
> On Thu, 22 Oct 2020 13:29:07 -0400
> Bob kb8tq  wrote:
> 
>> There is a paper running around behind the IEEE paywall by Collins that 
>> details
>> one approach to designing multi-stage limiters. There are other ways to do 
>> it. 
>> The key is to not go to crazy all in one stage. 
> 
> 
> You can find a modern, more accurate version of Collins result
> in [1]. While most formulas in there give the noise up to some 
> proportionality factor, all the information to calculate the noise 
> level is in there. The big difference to Collins here is that this 
> result is exact if you use all harmonics up to infinity (which is 
> easy) or can break up with a known error if you don't want to go 
> that far. Collins uses a trapezoid approximation throughout and thus
> over-estimates the noise contribution of low-gain stages, which
> are the first stages thus contribute to a relatively large error
> in the calculation.
> 
> 
> 
>   Attila Kinali
> 
> [1] "A Physical Sine-to-Square Converter Noise Model"
> by yours truly, IFCS 2018
> http://people.mpi-inf.mpg.de/~adogan/pubs/IFCS2018_comparator_noise.pdf
> 
> -- 
>   The bad part of Zurich is where the degenerates
>throw DARK chocolate at you.
> 
> ___
> time-nuts mailing list -- time-nuts@lists.febo.com
> To unsubscribe, go to 
> http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
> and follow the instructions there.


___
time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe, go to 
http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.

Re: [time-nuts] SMPS or conventional?

[Angus via time-nuts]

   I suspect that one of the main issues these days is the quality of
the module - there is a lot of unbranded stuff out there of unknown
quality.
   Cooling a rubidium is probably a relatively easy life for a peltier
since there are limited temperature changes, but I would still make
sure that I had a spare to drop in if I did have an issue.
   Since you tend to rate cooling for worst case, and it's usually a
lot less efficient there, the peltier will usually spend most of its
time at a fraction of full power.
   I don't think that they like unfiltered PWM, but there are plenty
of other reasons to avoid that in this case anyway.

Angus

On Wed, 21 Oct 2020 21:36:41 -0700, you wrote:

>> like with TEC heat/cool capability
>
>What's the typical MTBF of TEC coolers?  How much does it depend on how much 
>power you put through them?

___
time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe, go to 
http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.

Re: [time-nuts] 5065A Rb Cavity RX - Varactor repair?

[Gerhard Hoffmann]

Am 22.10.20 um 23:54 schrieb Richard (Rick) Karlquist:

I suspect that the stock 5065 chain has better
phase noise than this chip.  I know for sure
that you can get much better phase noise than
this chip by using conventional architectures.
Of course they are more complicated, etc.  Just
wanted to put this chip in perspective.

Rick N6RK


On 10/22/2020 12:12 PM, Gerhard Hoffmann wrote:


This is a ADF5356 with 2 LT3042 to clean up the supply voltage.
I measured the successor to this board yesterday on 10 GHz out
with my new 1/8 prescaler and the SR620 counter and some


As long as you can measure the noise with a 89441A vector signal

analyzer instead of a signal source analyzer, it cannot be leading edge.

Close-in, it seems quite ok. Enough to guide a 5 MHz crystal oscillator, 
and


also to guide a (in my case) 10 GHz dielectric osc with a Q of 20 to 30K.

Sorry, I must stay diffuse at this point.


My question for the phase noise in the Engineer Zone was also kinda 
"scheinheilig",


German word, (hypocritical, falsely innocent, ...lost in translation). I 
know that I could


go to a phase detector frequency of 150 MHz with this chip (instead of 
20*2), and it


would help for sure.  But I'm still struggling to make a connection between

the Windows evaluation software results and the data sheet.  My software 
driver is


1200 lines of C code (but including lots of comments.) ~400 bits to be 
set up.



And the 100th harmonic of 60 MHz won't be a phase noise wonder either, with

or without a SRD. You can see the effect in the well-known pictures of 
the old


HP spectrum analyzers that display a staircase with many a step, 
depending on


the LO harmonic.


Sky and MA/COM still make step recovery diodes, whether they fit is a 
different


question. But I cannot think of a special SRD failure mode. If you can 
do that,


solder it out and measure it with a normal Ohm meter. If it still 
behaves diode-ish,


it is very probably a working SRD. Maybe without the soldering, 
depending on the circuit.


Repairing will always be cheaper than transplanting a completely new heart.


Cheers, Gerhard





___
time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe, go to 
http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.

Re: [time-nuts] 5065A Rb Cavity RX - Varactor repair?

[ed breya]

I doubt that the SRD is the problem. It's just the last thing in the 
chain, and a number of things ahead of it all have to work right first.


I'd look first at any connections and wiring between various sections, 
and jiggle things around to see if the symptoms can return. Next, if ICs 
and transistors and other parts are in sockets, jiggle them too. For 
ICs, it's easy to pry one end out slightly - with power off - then press 
it back in, and repeat at the other end if it seems necessary. Re-wiping 
the contacts can do wonders sometimes. I don't know if HP ever used the 
notorious TI IC sockets, but they are a big cause of problems in Tek 
gear. The same for TI brand ICs that have silver plated leads - whether 
socketed or soldered in, these can suffer connection failure. This is 
well known and discussed a lot over the years.


Sometimes an intermittent can be a short instead of an open, with 
various causes, like errant braid strands on coax, wire clippings, 
solder gobs, and loose hardware that happen to eventually land in just 
the right spot. I have found this generally the second most likely 
fault, after open or poor connections. This is also a type that can 
happen once, if say a loose item caused the problem, then gets knocked 
out of place (sometimes to a different place where it can cause a 
different problem later) by moving the unit or opening it up.


The problem in diagnosing an intermittent that came and went only once, 
is that you'll never know for sure what caused it, unless it can be 
forced to fail again and stay that way for a while, and tracked down. 
Or, if you happen to find something rattling around inside that can be 
blamed. Short of that, the next best thing is to check and clean up 
anything suspicious, and hope that you hit the actual culprit in the 
process.


Ed

___
time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe, go to 
http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.

Re: [time-nuts] SMPS or conventional?

[Bruce Griffiths]

If the drive current ripple is too high fatigue failure from cyclic 
thermomechanical stress can be significant.

Bruce
> On 23 October 2020 at 10:37 ed breya  wrote:
> 
> 
> Regarding TEC life, in my experience, what wrecks them the most is 
> cooling applications where the cold side is below the dew point, and 
> water condenses out onto and into the TEC Peltier array. This eventually 
> rots the elements from corrosion and electrolysis, until something craps 
> out. If the TEC will be exposed to ambient air, and run below or even 
> near the expected dew point, the sides should be sealed up with silicone 
> goop for longer life. Even then, it's not foolproof, since it will never 
> be a perfect seal.
> 
> If the TEC is in a closed, benign environment or vacuum, then it should 
> last indefinitely, unless abused power-dissipation-wise, or mechanically 
> by thermal stress. When the temperature cycles, the whole part must be 
> allowed to change dimensions slightly, so the mounting needs to have 
> some compliance.
> 
> Ed
> 
> ___
> time-nuts mailing list -- time-nuts@lists.febo.com
> To unsubscribe, go to 
> http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
> and follow the instructions there.

___
time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe, go to 
http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.

Re: [time-nuts] Frequency Counter Choice

[Attila Kinali]

On Thu, 22 Oct 2020 13:29:07 -0400
Bob kb8tq  wrote:

> There is a paper running around behind the IEEE paywall by Collins that 
> details
> one approach to designing multi-stage limiters. There are other ways to do 
> it. 
> The key is to not go to crazy all in one stage. 


You can find a modern, more accurate version of Collins result
in [1]. While most formulas in there give the noise up to some 
proportionality factor, all the information to calculate the noise 
level is in there. The big difference to Collins here is that this 
result is exact if you use all harmonics up to infinity (which is 
easy) or can break up with a known error if you don't want to go 
that far. Collins uses a trapezoid approximation throughout and thus
over-estimates the noise contribution of low-gain stages, which
are the first stages thus contribute to a relatively large error
in the calculation.



Attila Kinali

[1] "A Physical Sine-to-Square Converter Noise Model"
by yours truly, IFCS 2018
http://people.mpi-inf.mpg.de/~adogan/pubs/IFCS2018_comparator_noise.pdf

-- 
The bad part of Zurich is where the degenerates
throw DARK chocolate at you.

___
time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe, go to 
http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.

Re: [time-nuts] 5065A Rb Cavity RX - Varactor repair?

[Richard (Rick) Karlquist]

I suspect that the stock 5065 chain has better
phase noise than this chip.  I know for sure
that you can get much better phase noise than
this chip by using conventional architectures.
Of course they are more complicated, etc.  Just
wanted to put this chip in perspective.

Rick N6RK


On 10/22/2020 12:12 PM, Gerhard Hoffmann wrote:


This is a ADF5356 with 2 LT3042 to clean up the supply voltage.
I measured the successor to this board yesterday on 10 GHz out
with my new 1/8 prescaler and the SR620 counter and some



___
time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe, go to 
http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.

Re: [time-nuts] SMPS or conventional?

[ed breya]

Regarding TEC life, in my experience, what wrecks them the most is 
cooling applications where the cold side is below the dew point, and 
water condenses out onto and into the TEC Peltier array. This eventually 
rots the elements from corrosion and electrolysis, until something craps 
out. If the TEC will be exposed to ambient air, and run below or even 
near the expected dew point, the sides should be sealed up with silicone 
goop for longer life. Even then, it's not foolproof, since it will never 
be a perfect seal.


If the TEC is in a closed, benign environment or vacuum, then it should 
last indefinitely, unless abused power-dissipation-wise, or mechanically 
by thermal stress. When the temperature cycles, the whole part must be 
allowed to change dimensions slightly, so the mounting needs to have 
some compliance.


Ed

___
time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe, go to 
http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.

Re: [time-nuts] 5065A Rb Cavity Rx - SRD repair

[Bob kb8tq]

Hi

The SRD is not easily accessible. Digging in to get it is going to 
involve a bit of work. 

The power level is going going to be a function of the cavity design.
Apparently that design is “non-standard” so any number will be a guess.
The 5065 manual shows typical RF voltage levels into the cavity.

Bob

> On Oct 22, 2020, at 4:31 PM, Ulf Kylenfall via time-nuts 
>  wrote:
> 
> 
> I see.
> Is the SRD in the Rx module accessible at all? 
> What type of package is it? Is it possible 
> to dismantle the Rx unit without risk of cracking 
> the glass enevlope? 
> 
> Does anybody know the power level at 6 GHz? 
> 
> BR  
> 
> Ulf Kylenfall  
> SM6GXV  
> 
> 
> 
> 
> 
> 
> 
> ___
> time-nuts mailing list -- time-nuts@lists.febo.com
> To unsubscribe, go to 
> http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
> and follow the instructions there.


___
time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe, go to 
http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.

[time-nuts] 5065A Rb Cavity Rx - SRD repair

[Ulf Kylenfall via time-nuts]

I see.
Is the SRD in the Rx module accessible at all? 
What type of package is it? Is it possible 
to dismantle the Rx unit without risk of cracking 
the glass enevlope? 

Does anybody know the power level at 6 GHz? 

BR  

Ulf Kylenfall  
SM6GXV  







___
time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe, go to 
http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.

Re: [time-nuts] 5065A Rb Cavity RX - Varactor repair?

[Richard (Rick) Karlquist]

I would tend to agree with the assessment that the
SRD is not a high failure rate item.  Also, it
could experience a faux failure where the fixturing
of it corroded such that it was not making good
electrical contact any more, but the underlying diode
was still good.  Cleaning up the contact area could
fix that.

If the SRD truly failed, the only way to get a replacement
would be to cannibalize another 5065.

I also agree with the other comment that replacing
the whole RF chain would be a major undertaking, having
designed several RF chains.

Rick N6RK

On 10/22/2020 10:06 AM, cdel...@juno.com wrote:

Ulf,

I have never had a 5065A with a bad or malfunctioning step  recovery
diode in the physics package.

MANY other problems are more likely the cause.

Cheers,

Corby


___
time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe, go to 
http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.



___
time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe, go to 
http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.

Re: [time-nuts] Frequency Counter Choice

[Bob kb8tq]

Hi

One approach to a mixer setup is the DMTD (dual mixer time difference). 
It has a “little cousin” that also does very will running one mixer instead of 
two. The limit there being that you need to be able to tune one or both the
devices you are testing to get a beat note in the 2 to 10 Hz range. 

There are lots and lots of posts on DMTD’s so I’lll leave that to others. Single
mixer sort of works like this:

Grab a MiniCircuits RPD-1 and drive both ports into saturation ( about +7 dbm).

What comes out will be a triangle wave. To get the max voltage out ( which is
a good thing in this case) terminate the output into >= 5K ohms at audio / DC. 
You will need some RF filtering on the output to knock down the 2X input 
frequency
RF on the output. 

Any of a number of op-amps from the good old OP-37 on will do a fine job of 
acting as a preamp for the audio. My preference is to amplify it up pretty far
(like 10X) and then feed it into a multi stage limiter. 

There is a paper running around behind the IEEE paywall by Collins that details
one approach to designing multi-stage limiters. There are other ways to do it. 
The key is to not go to crazy all in one stage. 

The output of the last limiter stage feeds your counter. With a good limiter
you might get 6 fairly solid digits on your 5335 (off of a pair of 10 MHz 
OCXO’s).
Without a good limiter, you are doing well at three solid digits on the 5335. 

Fun !!!

Bob

> On Oct 22, 2020, at 11:00 AM, Giorgio Barinetti  wrote:
> 
> Hi Bob,
> 
> 
> 
> Nice explanation. I'll see what I can do to find a 53132.
> 
> 
> 
> Maybe, in the meantime, if I'll be lucky to find a 5370, I'll play on it.
> 
> 
> 
> Regarding the mixer based setup, I'm already in progress for setting up a 
> suitable test jig.
> 
> Please point to some information, if you have it handy.
> 
> 
> 
> Regards,
> 
> Giorgio.
> 
> 
> 
> 
> 
> Hi
> 
> 
> 
>> On Oct 22, 2020, at 7:50 AM, Giorgio Barinetti > barinetti.it>
>>  wrote:
> 
>> 
> 
>> Hi,
> 
>> 
> 
>> As a newbie in the field, I've collected, by chance, some frequency 
>> standards.
> 
>> 
> 
>> Now is the time to measure them, and see how they perform.
> 
>> 
> 
>> I've inherited a 5371, but something tells me that is not the right 
>> instrument - and maybe is even faulty.
> 
>> 
> 
>> So, I'm in the need to buy a counter to use togheter with TimeLab.
> 
>> 
> 
>> Choices are many, but I'll try to avoid the "older" machines lile 5370 or 
>> 5335. The 531xx series seems nice ( money apart )
> 
>> But again : which one between the 3 ? 53131, 53132 or 53181 ?
> 
> 
> 
> 53181 = single channel 400 ps counter
> 
> 53131 = dual channel 400 ps counter
> 
> 53132 = dual channel 100 ps counter
> 
> 
> 
> 5370 (when working) = dual channel 20 ps counter
> 
> 
> 
> The gotcha is that none of them are good enough to properly measure 1
> 
> second ADEV on a good ( but still could be cheap ) OCXO. A mixer based
> 
> setup is a cheap way to get things done, even with a 5335.
> 
> 
> 
> The 531xx counters all share a couple of issues:
> 
> 
> 
> 1) The power supply was made by who knows who and their quality is not
> 
> the best. Good news is that the power supply out of a (cheap) 53181 will
> 
> swap over to revive an expensive 53132.
> 
> 
> 
> 2) They have a fan, it plugs with dust. When that happens it gets hot inside
> 
> the cabinet. Things (like the power supply) die as a result.
> 
> 
> 
> 3) They have a VFD display. It's a nice one, but they do wear out. Replacing
> 
> them is problematic.
> 
> 
> 
> 4) They have a fancy setup to add digits to a frequency measurement. It
> 
> can lead you astray. It also tends to go deaf right at 10 MHz. There are app
> 
> notes out there that explain the details.
> 
> 
> 
> 5) With *any* counter, a good external reference is the way to go for "Time 
> Nut"
> 
> grade measurements.
> 
> 
> 
> None of that is to say they are a bad counter, far from it. I have a number
> 
> of them and have been using them here and at work for decades. The 532xx
> 
> counters are the latest and greatest. They are on eBay, but not cheap.
> 
> 
> 
> Bob
> 
> 
> 
> 
> 
> 
> 
>> 
> 
>> Can somebody shed some light, and maybe help even to found a baseline for us 
>> beginners ?
> 
>> 
> 
>> Regards,
> 
>> IZ2JGB
> 
>> Giorgio
> 
>> 
> 
>> My NTP servers: (ntpd on FreeBSD - PPS in via Serial)
> 
>> 
> 
>> HP/Symmetricom 55300A  https://www.ntppool.org/scores/93.41.196.243
> 
>> 
> 
>> Efratom Rb/Xc GPSDO https://www.ntppool.org/scores/95.255.136.126
> 
>> 
> 
>> 
> 
>> 
> 
>> ___
> 
>> time-nuts mailing list -- time-nuts at 
>> lists.febo.com
> 
>> To unsubscribe, go to 
>> http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
> 
>> and follow the instructions there.
> 
> ___
> time-nuts mailing list -- 

Re: [time-nuts] Frequency Counter Choice

[Andy Talbot]

In days of yore + DOS, I implemented a GPIB control by bit-banging a
parallel printer port.
Those were the days

Andy
www.g4jnt.com




Virus-free.
www.avg.com

<#DAB4FAD8-2DD7-40BB-A1B8-4E2AA1F9FDF2>

On Thu, 22 Oct 2020 at 17:59, jimlux  wrote:

> On 10/22/20 8:29 AM, Tom Van Baak wrote:
>
> >
> > 3) Instead of replacing the front panel readout, instead do all your
> > comms over SCPI, to the point where you have created a virtual
> > instrument, not unlike some of the logic analyzer or 'scope PC apps.
> > This solution would be very welcome by lots time-nuts. It means we could
> > fully operate our 53131/53132 counters with the VFD disabled. The
> > downsize is that it requires a GPIB (Prologix) adapter.
> >
> > /tvb
> >
> >
> While I love my Prologix adapters, I wonder if one could implement the
> protocol using GPIO pins on something like a Beagleboard or Arduino clone?
>
> (yes, as a product, with the right connector and line driver/receivers,
> etc. - it would cost the same as a Prologix... but as a hack...)
>
> After all GPIB is a very old standard, and was developed back in the day
> of 1 MHz logic. I think you could do the handshaking in software - punch
> the bits one by one, assert DAV, etc.
> Listener reads bits one by one, then asserts the ACK or NAK
>
>
>
> (Aieee.. Wikipedia says it was supported by the Commodore PET... there's
> a challenge for you)
>
> ___
> time-nuts mailing list -- time-nuts@lists.febo.com
> To unsubscribe, go to
> http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
> and follow the instructions there.
>
___
time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe, go to 
http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.

Re: [time-nuts] Frequency Counter Choice

[Attila Kinali]

On Thu, 22 Oct 2020 11:50:08 +
Giorgio Barinetti  wrote:

> Choices are many, but I'll try to avoid the "older" machines lile 5370 or 
> 5335. The 531xx series seems nice ( money apart )
> But again : which one between the 3 ? 53131, 53132 or 53181 ?

Maybe try to get hold of one of the Philips (later licensed to
Fluke) PM6680 or PM6681? These are more common in Europe than
in the US, so the big US dominated websites/forums/.. don't
mention them that often. Solid devices that can be had as low
as 300€ if you are willing to wait, 500-800€ is the usual going
price. The SR620 is the workhorse that drives a lot of the
time and frequency metrology worldwide and can be had new and
used (new on http://thinksrs.com goes for 800-2000€ used).

If you go for a new one, I would consider looking at the
Pendulum CNT-90 and CNT-91. (Pendulum is the company that
took over Philips frequency counter business and the CNT-90
is the continuation of the PM668x line, also sold as PM6690
by Fluke)

Attila Kinali
-- 
The bad part of Zurich is where the degenerates
throw DARK chocolate at you.

___
time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe, go to 
http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.

[time-nuts] 5065A Rb Cavity RX - Varactor repair?

[cdelect]

Ulf,

I have never had a 5065A with a bad or malfunctioning step  recovery
diode in the physics package.

MANY other problems are more likely the cause.

Cheers,

Corby


___
time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe, go to 
http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.

Re: [time-nuts] Frequency Counter Choice

[jimlux]

On 10/22/20 8:29 AM, Tom Van Baak wrote:



3) Instead of replacing the front panel readout, instead do all your 
comms over SCPI, to the point where you have created a virtual 
instrument, not unlike some of the logic analyzer or 'scope PC apps. 
This solution would be very welcome by lots time-nuts. It means we could 
fully operate our 53131/53132 counters with the VFD disabled. The 
downsize is that it requires a GPIB (Prologix) adapter.


/tvb


While I love my Prologix adapters, I wonder if one could implement the 
protocol using GPIO pins on something like a Beagleboard or Arduino clone?


(yes, as a product, with the right connector and line driver/receivers, 
etc. - it would cost the same as a Prologix... but as a hack...)


After all GPIB is a very old standard, and was developed back in the day 
of 1 MHz logic. I think you could do the handshaking in software - punch 
the bits one by one, assert DAV, etc.

Listener reads bits one by one, then asserts the ACK or NAK



(Aieee.. Wikipedia says it was supported by the Commodore PET... there's 
a challenge for you)


___
time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe, go to 
http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.

Re: [time-nuts] Frequency Counter Choice

[Tom Van Baak]

Mark,

Yes, the VFD on those counters do seem to have a limited lifetime. I 
have some 20 years old that are getting hard to read in bright light.


If the VFD dies, yes, the counter will continue to measure just fine, 
not only all the serial talk-only data output, but the full capability 
of SCPI over GPIB. TimeLab won't know the difference. The downside is 
that the custom little annunciation icons inside the VFD won't be 
visible and those are sometimes quite useful, for example, to know the 
status of "ext ref", or even the M/Hz/u/s units.


Here are some other ideas:

1) Just replace the dead / dying VFD. You can find NOS on eBay. Taka has 
done this.


2) Use an Arduino or similar to capture the digital signals to the 
display board containing the failed VFD to create a mirror of the 
display. A multi-digit LCD module should work. Or try one of those nice 
looking programmable OLED pixel displays.


Or, for a bit of nostalgia, not to mention a top spot on Hackaday, use 
Nixie tubes for the readout and NE-2 bulbs for the status icons. Or get 
modern and output the display over BLE and write a phone app as the display.


3) Instead of replacing the front panel readout, instead do all your 
comms over SCPI, to the point where you have created a virtual 
instrument, not unlike some of the logic analyzer or 'scope PC apps. 
This solution would be very welcome by lots time-nuts. It means we could 
fully operate our 53131/53132 counters with the VFD disabled. The 
downsize is that it requires a GPIB (Prologix) adapter.


/tvb


On 10/22/2020 7:22 AM, Mark Spencer wrote:

I am just curious how viable would it be to use a 531xx series counter with 
Timelab if the counter display has failed ?

Thanks in advance

Mark Spencer
m...@alignedsolutions.com
604 762 4099


On Oct 22, 2020, at 6:36 AM, Bob kb8tq  wrote:

Hi


On Oct 22, 2020, at 7:50 AM, Giorgio Barinetti  wrote:

Hi,

As a newbie in the field, I've collected, by chance, some frequency standards.

Now is the time to measure them, and see how they perform.

I've inherited a 5371, but something tells me that is not the right instrument 
- and maybe is even faulty.

So, I'm in the need to buy a counter to use togheter with TimeLab.

Choices are many, but I'll try to avoid the "older" machines lile 5370 or 5335. 
The 531xx series seems nice ( money apart )
But again : which one between the 3 ? 53131, 53132 or 53181 ?

53181 = single channel 400 ps counter
53131 = dual channel 400 ps counter
53132 = dual channel 100 ps counter

5370 (when working) = dual channel 20 ps counter

The gotcha is that none of them are good enough to properly measure 1
second ADEV on a good ( but still could be cheap ) OCXO. A mixer based
setup is a cheap way to get things done, even with a 5335.

The 531xx counters all share a couple of issues:

1) The power supply was made by who knows who and their quality is not
the best. Good news is that the power supply out of a (cheap) 53181 will
swap over to revive an expensive 53132.

2) They have a fan, it plugs with dust. When that happens it gets hot inside
the cabinet. Things (like the power supply) die as a result.

3) They have a VFD display. It’s a nice one, but they do wear out. Replacing
them is problematic.

4) They have a fancy setup to add digits to a frequency measurement. It
can lead you astray. It also tends to go deaf right at 10 MHz. There are app
notes out there that explain the details.

5) With *any* counter, a good external reference is the way to go for “Time Nut”
grade measurements.

None of that is to say they are a bad counter, far from it. I have a number
of them and have been using them here and at work for decades. The 532xx
counters are the latest and greatest. They are on eBay, but not cheap.

Bob




Can somebody shed some light, and maybe help even to found a baseline for us 
beginners ?

Regards,
IZ2JGB
Giorgio

My NTP servers: (ntpd on FreeBSD - PPS in via Serial)

HP/Symmetricom 55300A  https://www.ntppool.org/scores/93.41.196.243

Efratom Rb/Xc GPSDO https://www.ntppool.org/scores/95.255.136.126



___
time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe, go to 
http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.


___
time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe, go to 
http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.

___
time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe, go to 
http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.



___
time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe, go to 
http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.

Re: [time-nuts] Frequency Counter Choice

[Dave Miller]

As a newbie to nuttery I would like to throw my brief experiences into the
mix.
When I started down this slippery slope I was using a HP5386A frequency
counter because I had one.
It was entertaining for a short period of time.
I then acquired a HP5372A which for hooking up to Timelabs and running Adev
plots did not yield the results I was hoping for as the resolution is
rather coarse. I think it does some neat things standalone but not really
the solution to run Timelabs and make Adev plots and have fun.
So then I found a 5370B on fleabay.
I managed to get one that seems to work well. It is measuring frequency
differences of 30-40 ps as far as I can tell.
So I am happy.
Eventually I will figure all this stuff out.
Good luck and have fun.

Dave





On Thu, Oct 22, 2020 at 8:25 AM Magnus Danielson  wrote:

> Hi,
>
> Small correction, 5371A and 5370A are different beasts. The 5370A has 20
> ps resolution and the 5371A has 200 ps resolution. So, the 5371A is
> about the same performance as the 53132A at 150 ps resolution. I work
> with 5372A which is essentially a 5371A with hardware histogram and a
> few other things.
>
> I have both 5370A and 5371A.
>
> The 5371A and 5372A actually can do linear regression processing for
> frequency estimation, a historic note that predates this capability of
> the Philips/Fluke/Pendulum counters to the best of my knowledge.
>
> PS. I also have a 5370C, which is a flat-bed scanner from HP.
>
> Cheers,
> Magnus
>
> On 2020-10-22 15:33, Tom Van Baak wrote:
> > Avoid the 53181A; too limited. Both the 53131A / 53132A are
> > "universal" counters so they do frequency, period, time interval, and
> > lots more.
> >
> > They are very useful counters. The rs232 talk-only mode makes it very
> > easy to collect data on a PC. No programming required. The 53132A has
> > a bit higher resolution than a 53131A so that's a plus. But if you
> > plan to use the external reference the 53131A accepts 5/10 MHz while
> > the 53132A, last I checked, will only accept 10 MHz.
> >
> > A lower cost alternative is the TAPR/TICC. If you already have a
> > commercial counter I would recommend looking at the TICC as your next
> > counter. As this is your first counter I would recommend a 53131/53132
> > over the TICC. The TICC is "headless" and has only hard-coded CMOS
> > digital inputs. The hp units have all the physical front panel
> > buttons, status LED's and live display. Plus the comprehensive input
> > conditioning features. And useful frequency gate time or digits modes,
> > etc. Lots of cool features. Built-in statistics too.
> >
> > The 5371A is a beast. But it you can get it working it has 10x better
> > performance than a 53132A.
> >
> > /tvb
> >
> >
> > On 10/22/2020 4:50 AM, Giorgio Barinetti wrote:
> >> Hi,
> >>
> >> As a newbie in the field, I've collected, by chance, some frequency
> >> standards.
> >>
> >> Now is the time to measure them, and see how they perform.
> >>
> >> I've inherited a 5371, but something tells me that is not the right
> >> instrument - and maybe is even faulty.
> >>
> >> So, I'm in the need to buy a counter to use togheter with TimeLab.
> >>
> >> Choices are many, but I'll try to avoid the "older" machines lile
> >> 5370 or 5335. The 531xx series seems nice ( money apart )
> >> But again : which one between the 3 ? 53131, 53132 or 53181 ?
> >>
> >> Can somebody shed some light, and maybe help even to found a baseline
> >> for us beginners ?
> >>
> >> Regards,
> >> IZ2JGB
> >> Giorgio
> >>
> >> My NTP servers: (ntpd on FreeBSD - PPS in via Serial)
> >>
> >> HP/Symmetricom 55300A  https://www.ntppool.org/scores/93.41.196.243
> >>
> >> Efratom Rb/Xc GPSDO https://www.ntppool.org/scores/95.255.136.126
> >>
> >>
> >>
> >> ___
> >> time-nuts mailing list -- time-nuts@lists.febo.com
> >> To unsubscribe, go to
> >> http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
> >> and follow the instructions there.
> >>
> >
> >
> > ___
> > time-nuts mailing list -- time-nuts@lists.febo.com
> > To unsubscribe, go to
> > http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
> > and follow the instructions there.
>
>
> ___
> time-nuts mailing list -- time-nuts@lists.febo.com
> To unsubscribe, go to
> http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
> and follow the instructions there.
>


-- 
72 de Dave
VE7HR
___
time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe, go to 
http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.

Re: [time-nuts] Frequency Counter Choice

[Bob kb8tq]

Hi

Provided you can get a few key parameters set up (= enable the serial 
port / set baud rates), it should run just fine without the display. You would
need to do some GPIB fiddling to set various modes. 

Bob

> On Oct 22, 2020, at 10:22 AM, Mark Spencer  wrote:
> 
> I am just curious how viable would it be to use a 531xx series counter with 
> Timelab if the counter display has failed ?  
> 
> Thanks in advance
> 
> Mark Spencer
> m...@alignedsolutions.com
> 604 762 4099
> 
>> On Oct 22, 2020, at 6:36 AM, Bob kb8tq  wrote:
>> 
>> Hi
>> 
>>> On Oct 22, 2020, at 7:50 AM, Giorgio Barinetti  wrote:
>>> 
>>> Hi,
>>> 
>>> As a newbie in the field, I've collected, by chance, some frequency 
>>> standards.
>>> 
>>> Now is the time to measure them, and see how they perform.
>>> 
>>> I've inherited a 5371, but something tells me that is not the right 
>>> instrument - and maybe is even faulty.
>>> 
>>> So, I'm in the need to buy a counter to use togheter with TimeLab.
>>> 
>>> Choices are many, but I'll try to avoid the "older" machines lile 5370 or 
>>> 5335. The 531xx series seems nice ( money apart )
>>> But again : which one between the 3 ? 53131, 53132 or 53181 ?
>> 
>> 53181 = single channel 400 ps counter
>> 53131 = dual channel 400 ps counter
>> 53132 = dual channel 100 ps counter
>> 
>> 5370 (when working) = dual channel 20 ps counter 
>> 
>> The gotcha is that none of them are good enough to properly measure 1 
>> second ADEV on a good ( but still could be cheap ) OCXO. A mixer based
>> setup is a cheap way to get things done, even with a 5335.
>> 
>> The 531xx counters all share a couple of issues:
>> 
>> 1) The power supply was made by who knows who and their quality is not 
>> the best. Good news is that the power supply out of a (cheap) 53181 will 
>> swap over to revive an expensive 53132.
>> 
>> 2) They have a fan, it plugs with dust. When that happens it gets hot inside
>> the cabinet. Things (like the power supply) die as a result.
>> 
>> 3) They have a VFD display. It’s a nice one, but they do wear out. Replacing
>> them is problematic.
>> 
>> 4) They have a fancy setup to add digits to a frequency measurement. It
>> can lead you astray. It also tends to go deaf right at 10 MHz. There are app
>> notes out there that explain the details. 
>> 
>> 5) With *any* counter, a good external reference is the way to go for “Time 
>> Nut”
>> grade measurements. 
>> 
>> None of that is to say they are a bad counter, far from it. I have a number
>> of them and have been using them here and at work for decades. The 532xx
>> counters are the latest and greatest. They are on eBay, but not cheap. 
>> 
>> Bob
>> 
>> 
>> 
>>> 
>>> Can somebody shed some light, and maybe help even to found a baseline for 
>>> us beginners ?
>>> 
>>> Regards,
>>> IZ2JGB
>>> Giorgio
>>> 
>>> My NTP servers: (ntpd on FreeBSD - PPS in via Serial)
>>> 
>>> HP/Symmetricom 55300A  https://www.ntppool.org/scores/93.41.196.243
>>> 
>>> Efratom Rb/Xc GPSDO https://www.ntppool.org/scores/95.255.136.126
>>> 
>>> 
>>> 
>>> ___
>>> time-nuts mailing list -- time-nuts@lists.febo.com
>>> To unsubscribe, go to 
>>> http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
>>> and follow the instructions there.
>> 
>> 
>> ___
>> time-nuts mailing list -- time-nuts@lists.febo.com
>> To unsubscribe, go to 
>> http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
>> and follow the instructions there.
> 
> ___
> time-nuts mailing list -- time-nuts@lists.febo.com
> To unsubscribe, go to 
> http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
> and follow the instructions there.


___
time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe, go to 
http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.

[time-nuts] Frequency Counter Choice

[Giorgio Barinetti]

Hi Bob,



Nice explanation. I'll see what I can do to find a 53132.



Maybe, in the meantime, if I'll be lucky to find a 5370, I'll play on it.



Regarding the mixer based setup, I'm already in progress for setting up a 
suitable test jig.

Please point to some information, if you have it handy.



Regards,

Giorgio.





Hi



> On Oct 22, 2020, at 7:50 AM, Giorgio Barinetti  barinetti.it>
>  wrote:

>

> Hi,

>

> As a newbie in the field, I've collected, by chance, some frequency standards.

>

> Now is the time to measure them, and see how they perform.

>

> I've inherited a 5371, but something tells me that is not the right 
> instrument - and maybe is even faulty.

>

> So, I'm in the need to buy a counter to use togheter with TimeLab.

>

> Choices are many, but I'll try to avoid the "older" machines lile 5370 or 
> 5335. The 531xx series seems nice ( money apart )

> But again : which one between the 3 ? 53131, 53132 or 53181 ?



53181 = single channel 400 ps counter

53131 = dual channel 400 ps counter

53132 = dual channel 100 ps counter



5370 (when working) = dual channel 20 ps counter



The gotcha is that none of them are good enough to properly measure 1

second ADEV on a good ( but still could be cheap ) OCXO. A mixer based

setup is a cheap way to get things done, even with a 5335.



The 531xx counters all share a couple of issues:



1) The power supply was made by who knows who and their quality is not

the best. Good news is that the power supply out of a (cheap) 53181 will

swap over to revive an expensive 53132.



2) They have a fan, it plugs with dust. When that happens it gets hot inside

the cabinet. Things (like the power supply) die as a result.



3) They have a VFD display. It's a nice one, but they do wear out. Replacing

them is problematic.



4) They have a fancy setup to add digits to a frequency measurement. It

can lead you astray. It also tends to go deaf right at 10 MHz. There are app

notes out there that explain the details.



5) With *any* counter, a good external reference is the way to go for "Time Nut"

grade measurements.



None of that is to say they are a bad counter, far from it. I have a number

of them and have been using them here and at work for decades. The 532xx

counters are the latest and greatest. They are on eBay, but not cheap.



Bob







>

> Can somebody shed some light, and maybe help even to found a baseline for us 
> beginners ?

>

> Regards,

> IZ2JGB

> Giorgio

>

> My NTP servers: (ntpd on FreeBSD - PPS in via Serial)

>

> HP/Symmetricom 55300A  https://www.ntppool.org/scores/93.41.196.243

>

> Efratom Rb/Xc GPSDO https://www.ntppool.org/scores/95.255.136.126

>

>

>

> ___

> time-nuts mailing list -- time-nuts at 
> lists.febo.com

> To unsubscribe, go to 
> http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com

> and follow the instructions there.

___
time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe, go to 
http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.

Re: [time-nuts] 5065A Rb Cavity RX - Varactor repair?

[Bob kb8tq]

Hi

The device in the cavity that takes the 60 MHz up to 6.x GHz is a 
step recovery diode ( SRD ). The actual device used (apparently)
was unique to the 5065. 

One proposed “fix” is to eliminate the whole synth / multiply empire
and replace it with a very high resolution DDS generating the 6.x GHz. 
My guess is that will turn out to be a lot of work ….

Bob

> On Oct 22, 2020, at 9:59 AM, Ulf Kylenfall via time-nuts 
>  wrote:
> 
> 
> Strolled into my little lab only to find the 5065A's Red Integrator Limit  
> light on. Control negative fsd. PhotoCurrent OK. No 2nd harmonic. 
> 
> Was just about to power off and start to disassemble the unit. 
> All of a sudden, the Integrator light went out. Control went to zero 
> and by pressing Reset, the unit locked with previous 2nd harmonic 
> level indication. 
> 
> What ever the problem was, I do not know. I guess I can fix (redesign) 
> the synthesizer and replace some semiconductors that have failed 
> but if the problem is in the Rb Cavity RX unit - Varactor - there is a 
> problem. From a similar unit (but without the TED) I can see that 
> some part of the varactor assembly can be dismantled, but then 
> there are a few tiny screws that seems hard to reach. 
> 
> Has anybody repaired the Rx part with the varactor assembly? 
> 
> Regards 
> 
> Ulf Kylenfall 
> SM6GXV 
> 
> 
> 
> ___
> time-nuts mailing list -- time-nuts@lists.febo.com
> To unsubscribe, go to 
> http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
> and follow the instructions there.


___
time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe, go to 
http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.

Re: [time-nuts] Frequency Counter Choice

[Magnus Danielson]

Hi,

Small correction, 5371A and 5370A are different beasts. The 5370A has 20
ps resolution and the 5371A has 200 ps resolution. So, the 5371A is
about the same performance as the 53132A at 150 ps resolution. I work
with 5372A which is essentially a 5371A with hardware histogram and a
few other things.

I have both 5370A and 5371A.

The 5371A and 5372A actually can do linear regression processing for
frequency estimation, a historic note that predates this capability of
the Philips/Fluke/Pendulum counters to the best of my knowledge.

PS. I also have a 5370C, which is a flat-bed scanner from HP.

Cheers,
Magnus

On 2020-10-22 15:33, Tom Van Baak wrote:
> Avoid the 53181A; too limited. Both the 53131A / 53132A are
> "universal" counters so they do frequency, period, time interval, and
> lots more.
>
> They are very useful counters. The rs232 talk-only mode makes it very
> easy to collect data on a PC. No programming required. The 53132A has
> a bit higher resolution than a 53131A so that's a plus. But if you
> plan to use the external reference the 53131A accepts 5/10 MHz while
> the 53132A, last I checked, will only accept 10 MHz.
>
> A lower cost alternative is the TAPR/TICC. If you already have a
> commercial counter I would recommend looking at the TICC as your next
> counter. As this is your first counter I would recommend a 53131/53132
> over the TICC. The TICC is "headless" and has only hard-coded CMOS
> digital inputs. The hp units have all the physical front panel
> buttons, status LED's and live display. Plus the comprehensive input
> conditioning features. And useful frequency gate time or digits modes,
> etc. Lots of cool features. Built-in statistics too.
>
> The 5371A is a beast. But it you can get it working it has 10x better
> performance than a 53132A.
>
> /tvb
>
>
> On 10/22/2020 4:50 AM, Giorgio Barinetti wrote:
>> Hi,
>>
>> As a newbie in the field, I've collected, by chance, some frequency
>> standards.
>>
>> Now is the time to measure them, and see how they perform.
>>
>> I've inherited a 5371, but something tells me that is not the right
>> instrument - and maybe is even faulty.
>>
>> So, I'm in the need to buy a counter to use togheter with TimeLab.
>>
>> Choices are many, but I'll try to avoid the "older" machines lile
>> 5370 or 5335. The 531xx series seems nice ( money apart )
>> But again : which one between the 3 ? 53131, 53132 or 53181 ?
>>
>> Can somebody shed some light, and maybe help even to found a baseline
>> for us beginners ?
>>
>> Regards,
>> IZ2JGB
>> Giorgio
>>
>> My NTP servers: (ntpd on FreeBSD - PPS in via Serial)
>>
>> HP/Symmetricom 55300A  https://www.ntppool.org/scores/93.41.196.243
>>
>> Efratom Rb/Xc GPSDO https://www.ntppool.org/scores/95.255.136.126
>>
>>
>>
>> ___
>> time-nuts mailing list -- time-nuts@lists.febo.com
>> To unsubscribe, go to
>> http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
>> and follow the instructions there.
>>
>
>
> ___
> time-nuts mailing list -- time-nuts@lists.febo.com
> To unsubscribe, go to
> http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
> and follow the instructions there.


___
time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe, go to 
http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.

Re: [time-nuts] Frequency Counter Choice

[Mark Spencer]

I am just curious how viable would it be to use a 531xx series counter with 
Timelab if the counter display has failed ?  

Thanks in advance

Mark Spencer
m...@alignedsolutions.com
604 762 4099

> On Oct 22, 2020, at 6:36 AM, Bob kb8tq  wrote:
> 
> Hi
> 
>> On Oct 22, 2020, at 7:50 AM, Giorgio Barinetti  wrote:
>> 
>> Hi,
>> 
>> As a newbie in the field, I've collected, by chance, some frequency 
>> standards.
>> 
>> Now is the time to measure them, and see how they perform.
>> 
>> I've inherited a 5371, but something tells me that is not the right 
>> instrument - and maybe is even faulty.
>> 
>> So, I'm in the need to buy a counter to use togheter with TimeLab.
>> 
>> Choices are many, but I'll try to avoid the "older" machines lile 5370 or 
>> 5335. The 531xx series seems nice ( money apart )
>> But again : which one between the 3 ? 53131, 53132 or 53181 ?
> 
> 53181 = single channel 400 ps counter
> 53131 = dual channel 400 ps counter
> 53132 = dual channel 100 ps counter
> 
> 5370 (when working) = dual channel 20 ps counter 
> 
> The gotcha is that none of them are good enough to properly measure 1 
> second ADEV on a good ( but still could be cheap ) OCXO. A mixer based
> setup is a cheap way to get things done, even with a 5335.
> 
> The 531xx counters all share a couple of issues:
> 
> 1) The power supply was made by who knows who and their quality is not 
> the best. Good news is that the power supply out of a (cheap) 53181 will 
> swap over to revive an expensive 53132.
> 
> 2) They have a fan, it plugs with dust. When that happens it gets hot inside
> the cabinet. Things (like the power supply) die as a result.
> 
> 3) They have a VFD display. It’s a nice one, but they do wear out. Replacing
> them is problematic.
> 
> 4) They have a fancy setup to add digits to a frequency measurement. It
> can lead you astray. It also tends to go deaf right at 10 MHz. There are app
> notes out there that explain the details. 
> 
> 5) With *any* counter, a good external reference is the way to go for “Time 
> Nut”
> grade measurements. 
> 
> None of that is to say they are a bad counter, far from it. I have a number
> of them and have been using them here and at work for decades. The 532xx
> counters are the latest and greatest. They are on eBay, but not cheap. 
> 
> Bob
> 
> 
> 
>> 
>> Can somebody shed some light, and maybe help even to found a baseline for us 
>> beginners ?
>> 
>> Regards,
>> IZ2JGB
>> Giorgio
>> 
>> My NTP servers: (ntpd on FreeBSD - PPS in via Serial)
>> 
>> HP/Symmetricom 55300A  https://www.ntppool.org/scores/93.41.196.243
>> 
>> Efratom Rb/Xc GPSDO https://www.ntppool.org/scores/95.255.136.126
>> 
>> 
>> 
>> ___
>> time-nuts mailing list -- time-nuts@lists.febo.com
>> To unsubscribe, go to 
>> http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
>> and follow the instructions there.
> 
> 
> ___
> time-nuts mailing list -- time-nuts@lists.febo.com
> To unsubscribe, go to 
> http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
> and follow the instructions there.

___
time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe, go to 
http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.

[time-nuts] 5065A Rb Cavity RX - Varactor repair?

[Ulf Kylenfall via time-nuts]

Strolled into my little lab only to find the 5065A's Red Integrator Limit  
light on. Control negative fsd. PhotoCurrent OK. No 2nd harmonic. 

Was just about to power off and start to disassemble the unit. 
All of a sudden, the Integrator light went out. Control went to zero 
and by pressing Reset, the unit locked with previous 2nd harmonic 
level indication. 

What ever the problem was, I do not know. I guess I can fix (redesign) 
the synthesizer and replace some semiconductors that have failed 
but if the problem is in the Rb Cavity RX unit - Varactor - there is a 
problem. From a similar unit (but without the TED) I can see that 
some part of the varactor assembly can be dismantled, but then 
there are a few tiny screws that seems hard to reach. 

Has anybody repaired the Rx part with the varactor assembly? 

Regards 

Ulf Kylenfall 
SM6GXV 



___
time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe, go to 
http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.

Re: [time-nuts] Frequency Counter Choice

[Tom Van Baak]

Avoid the 53181A; too limited. Both the 53131A / 53132A are "universal" 
counters so they do frequency, period, time interval, and lots more.


They are very useful counters. The rs232 talk-only mode makes it very 
easy to collect data on a PC. No programming required. The 53132A has a 
bit higher resolution than a 53131A so that's a plus. But if you plan to 
use the external reference the 53131A accepts 5/10 MHz while the 53132A, 
last I checked, will only accept 10 MHz.


A lower cost alternative is the TAPR/TICC. If you already have a 
commercial counter I would recommend looking at the TICC as your next 
counter. As this is your first counter I would recommend a 53131/53132 
over the TICC. The TICC is "headless" and has only hard-coded CMOS 
digital inputs. The hp units have all the physical front panel buttons, 
status LED's and live display. Plus the comprehensive input conditioning 
features. And useful frequency gate time or digits modes, etc. Lots of 
cool features. Built-in statistics too.


The 5371A is a beast. But it you can get it working it has 10x better 
performance than a 53132A.


/tvb


On 10/22/2020 4:50 AM, Giorgio Barinetti wrote:

Hi,

As a newbie in the field, I've collected, by chance, some frequency standards.

Now is the time to measure them, and see how they perform.

I've inherited a 5371, but something tells me that is not the right instrument 
- and maybe is even faulty.

So, I'm in the need to buy a counter to use togheter with TimeLab.

Choices are many, but I'll try to avoid the "older" machines lile 5370 or 5335. 
The 531xx series seems nice ( money apart )
But again : which one between the 3 ? 53131, 53132 or 53181 ?

Can somebody shed some light, and maybe help even to found a baseline for us 
beginners ?

Regards,
IZ2JGB
Giorgio

My NTP servers: (ntpd on FreeBSD - PPS in via Serial)

HP/Symmetricom 55300A  https://www.ntppool.org/scores/93.41.196.243

Efratom Rb/Xc GPSDO https://www.ntppool.org/scores/95.255.136.126



___
time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe, go to 
http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.




___
time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe, go to 
http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.

Re: [time-nuts] Frequency Counter Choice

[Bob kb8tq]

Hi

> On Oct 22, 2020, at 7:50 AM, Giorgio Barinetti  wrote:
> 
> Hi,
> 
> As a newbie in the field, I've collected, by chance, some frequency standards.
> 
> Now is the time to measure them, and see how they perform.
> 
> I've inherited a 5371, but something tells me that is not the right 
> instrument - and maybe is even faulty.
> 
> So, I'm in the need to buy a counter to use togheter with TimeLab.
> 
> Choices are many, but I'll try to avoid the "older" machines lile 5370 or 
> 5335. The 531xx series seems nice ( money apart )
> But again : which one between the 3 ? 53131, 53132 or 53181 ?

53181 = single channel 400 ps counter
53131 = dual channel 400 ps counter
53132 = dual channel 100 ps counter

5370 (when working) = dual channel 20 ps counter 

The gotcha is that none of them are good enough to properly measure 1 
second ADEV on a good ( but still could be cheap ) OCXO. A mixer based
setup is a cheap way to get things done, even with a 5335.

The 531xx counters all share a couple of issues:

1) The power supply was made by who knows who and their quality is not 
the best. Good news is that the power supply out of a (cheap) 53181 will 
swap over to revive an expensive 53132.

2) They have a fan, it plugs with dust. When that happens it gets hot inside
the cabinet. Things (like the power supply) die as a result.

3) They have a VFD display. It’s a nice one, but they do wear out. Replacing
them is problematic.

4) They have a fancy setup to add digits to a frequency measurement. It
can lead you astray. It also tends to go deaf right at 10 MHz. There are app
notes out there that explain the details. 

5) With *any* counter, a good external reference is the way to go for “Time Nut”
grade measurements. 

None of that is to say they are a bad counter, far from it. I have a number
of them and have been using them here and at work for decades. The 532xx
counters are the latest and greatest. They are on eBay, but not cheap. 

Bob



> 
> Can somebody shed some light, and maybe help even to found a baseline for us 
> beginners ?
> 
> Regards,
> IZ2JGB
> Giorgio
> 
> My NTP servers: (ntpd on FreeBSD - PPS in via Serial)
> 
> HP/Symmetricom 55300A  https://www.ntppool.org/scores/93.41.196.243
> 
> Efratom Rb/Xc GPSDO https://www.ntppool.org/scores/95.255.136.126
> 
> 
> 
> ___
> time-nuts mailing list -- time-nuts@lists.febo.com
> To unsubscribe, go to 
> http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
> and follow the instructions there.


___
time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe, go to 
http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.

Re: [time-nuts] SMPS or conventional?

[Bob kb8tq]

Hi

The magnetic field sensitivity is dependent on a lot of things. 
Strangely enough, the higher the C field, the more sensitive 
the physics package is ( yes, that’s weird …). All Rb’s incorporate
shielding to reduce the external field impact. Since the shielding
is rarely perfect, there are asymmetries in the response. 

All that turns this into a “field from here at level X” matters on this
Rb and not on that one sort of thing. 

===

The “easy” way to get a feel for what’s what is to rotate the device 
in the horizontal plane. The earth’s mag field (unless you are close
to the magnetic poles) will change in a predictable manner. If you see
a measurable frequency change … you have an answer.

The more complicated approach uses Helmholtz coils. It’s more complicated
mainly because they take time to fab …..

Both approaches are only approximate unless you go a bit crazy on 
a bunch of minor details. 

Bob

> On Oct 22, 2020, at 6:17 AM, Angus via time-nuts  
> wrote:
> 
> Hi,
> 
>   The magnetic field issue has bothered me a little too, but I don't
> know whether it has a practical effect. Maybe putting a fan in a
> plastic or aluminium box close to the Rb might show something - unless
> someone already knows the effects.
> 
>   I have a compass on the bench beside the LPRO, and it does drift a
> bit. What surprised me most was how much the field around the mains
> transformers changed each time they are switched on or off.
> 
> Angus.
> 
> On Wed, 21 Oct 2020 10:04:11 -0400, you wrote:
> 
>> Hi
>> 
>> The only issue with the “controlled fan” approach is that you
>> have a variable magnetic field as a result. That and the vibration
>> both can impact the stability of the Rb. Some means of isolating
>> the fan from the immediate vicinity of the Rb sounds like a good idea. 
>> 
>> Bob
>> 
>>> On Oct 21, 2020, at 9:31 AM, Luiz Alberto Saba  wrote:
>>> 
>>> I am thinking in an black aluminium finned heatsink, the size of the base 
>>> plate, and a fan, controlled by an electronic thermostat (sensor direct 
>>> coupled to the heatsink). Is that ok?
>>> 
>>> Em 21/10/2020 09:40, Wannes Sels escreveu:
 There are conflicting requirements regarding temperature in Rb's:
 
 - For best performance, the rubidium and quartz oscillators must be kept at
 a high and stable temperature.
 - For reliability, the supporting electronics must be kept at a lower
 temperature.
 
 The heater takes care of the high temperature. The stability of the
 temperature can be improved by increasing the thermal mass, i.e. adding a
 block of aluminium or copper.
 The electronics can be kept cooler with a heatsink, and forced ventilation
 if needed.
 
 When you stick both the electronics and physics in a small package, this
 becomes more difficult. For the PRS-10 some form of heatsink is pretty much
 required to keep the electronics cool enough. Although I seem to remember
 reading somewhere that the "benchtop" heatsink they offer now is lighter
 than older units.
 
 If the room temperature is not stable enough, you might want to increase
 the size of the heatsink and add a fan. This reduces the temperature swing
 inside the unit, while still cooling the electronics, at the cost of
 increased power draw for the heater.
 
 
 
 
 On Wed, Oct 21, 2020 at 5:38 AM Hal Murray  wrote:
 
>> I spent a lot of years buying Rb’s and putting them on small heatsinks.
> I
>> always was disappointed in their reliability. That continued to be the
> case
>> up to the point that the baseplate temp’s got into the 40C region.  In my
>> case, that took a fan 
> .
> How well did it work if the heat sink wasn't small?  What is your version
> of
> small?
> 
> Do you have any data (or vague memories) of how much it helps to orient
> the
> heat sink so the fins are vertical so they encourage warm air to flow up
> past
> the fins?
> 
> --
> These are my opinions.  I hate spam.
> 
> 
> 
> 
> ___
> time-nuts mailing list -- time-nuts@lists.febo.com
> To unsubscribe, go to
> http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
> and follow the instructions there.
> 
 ___
 time-nuts mailing list -- time-nuts@lists.febo.com
 To unsubscribe, go to 
 http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
 and follow the instructions there.
>>> 
>>> 
>>> ___
>>> time-nuts mailing list -- time-nuts@lists.febo.com
>>> To unsubscribe, go to 
>>> http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
>>> and follow the instructions there.
>> 
>> 
>> ___
>> time-nuts mailing list -- time-nuts@lists.febo.com
>> To unsubscribe, go to 

Re: [time-nuts] SMPS or conventional?

[Bob kb8tq]

Hi

TEC’s are indeed prone to wear out. That said, so is a compressor …..

Bob

> On Oct 22, 2020, at 12:36 AM, Hal Murray  wrote:
> 
>> like with TEC heat/cool capability
> 
> What's the typical MTBF of TEC coolers?  How much does it depend on how much 
> power you put through them?
> 
> -- 
> These are my opinions.  I hate spam.
> 
> 
> 
> 
> ___
> time-nuts mailing list -- time-nuts@lists.febo.com
> To unsubscribe, go to 
> http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
> and follow the instructions there.


___
time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe, go to 
http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.

[time-nuts] Frequency Counter Choice

[Giorgio Barinetti]

Hi,

As a newbie in the field, I've collected, by chance, some frequency standards.

Now is the time to measure them, and see how they perform.

I've inherited a 5371, but something tells me that is not the right instrument 
- and maybe is even faulty.

So, I'm in the need to buy a counter to use togheter with TimeLab.

Choices are many, but I'll try to avoid the "older" machines lile 5370 or 5335. 
The 531xx series seems nice ( money apart )
But again : which one between the 3 ? 53131, 53132 or 53181 ?

Can somebody shed some light, and maybe help even to found a baseline for us 
beginners ?

Regards,
IZ2JGB
Giorgio

My NTP servers: (ntpd on FreeBSD - PPS in via Serial)

HP/Symmetricom 55300A  https://www.ntppool.org/scores/93.41.196.243

Efratom Rb/Xc GPSDO https://www.ntppool.org/scores/95.255.136.126



___
time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe, go to 
http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.

Re: [time-nuts] IEEE 1588 PTP support on raspberry pi 4 compute module

[Wojciech Owczarek]

On Thu, 22 Oct 2020 at 01:53, Magnus Danielson  wrote:

> Hi,
>
> Yeah, poor CPU time-stamping can be worse than network biases, but
> hardware time-stamping and careful use of good time-stamps handles most
> of the host issues. Then network biases becomes the dominant problem
> unless one does not have other implementation issues, and well, those
> are common, including dropping PTP packets aimed for processing. In one
> instance there was a rate-limit internal to the equipment that did the
> wrong thing. Slowly the implementation issues is being ironed out. I
> learn more of this than I want to.
>
>
The story of my life pretty much. Early PTP transparent clock switches were
two-step and software would handle PTP message corrections and PTP messages
indeed went to the CPU, so this was somewhat error-prone. Pretty much all
new higher-end Broadcom switching ACICs (OK, hardly home switch material)
have had PTP TC support completely in hardware for a number of years now,
with zero software control required. First TCs were mostly industrial
automation kit like Hirschmann, Moxa and what have you. Since Broadcom's
TCs (and BCs) hit the mainstream, PTP on the network side has been more
attainable than ever. But on the host side, it's still a minefield.


-- 
-

Wojciech Owczarek
___
time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe, go to 
http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.

Re: [time-nuts] IEEE 1588 PTP support on raspberry pi 4 compute module

[Wojciech Owczarek]

Hello Bob,

On Thu, 22 Oct 2020 at 01:53, Bob kb8tq  wrote:

> Hi
>
> Simply that you can get into the 10’s of microseconds (done in many times)
> with
> NTP. If the local LAN is streaming video, it needs 1588 switches / routers
> to do
> better than that.
>

That is all true, however not every single frame will hit those worst
latencies. With careful filtering and high message rates, lucky packets can
be picked out from the rest. The result shows more jitter than NTP, but
overall tighter bounded and with significantly less wander. Also, a 10 us
offset reported does not necessarily mean it's been passed to the clock
discipline and will cause the clock to swing by that, but consume enough of
those and it will, yes.

Well. your lab and my home LAN produce *very* different results …..
>
> Bob
>

Again - I can only nod until I produce some results :-) I tend to ignore
self-reported offset figures from software, so unless a thing has a 1PPS
output, I am very sceptical. The NICs I test with, do.
___
time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe, go to 
http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.

Re: [time-nuts] IEEE 1588 PTP support on raspberry pi 4 compute module

[Wojciech Owczarek]

(this message is half redundant now since there were some more posts since
I typed it)

On Wednesday, 21 October 2020, Hal Murray  wrote:
>
>
> Even if the network is lightly loaded, there can be asymmetry in CPU
> processing delays.


And that is why software timestamps (in a non-realtime OS) are the worst
thing that can happen to synchronisation.


> In traditional NTP, the transmit time stamp is grabbed as late as possible
> before sending a packet..  That cancels out if both CPUs are the same
> speed.
> Adding authentication (crypto type computations) increases delay from grab
> to
> send.  That will increase the offset if the CPUs are not matched.


That is all only a part of the problem. NTP(/d) can't do anything about the
fact that even once the socket send call is made, the kernel will
essentially send the packet when it feels like it. Nothing can cancel out
the unpredictable latency of software-driven interrupt handling either.
Crypto or anything else in userspace after grabbing a timestamp is a killer
to sync - PTP (2019) is attempting to do crypto via side channels. NTP's
focus from its inception was on sync over the Internet, where the
timestamping uncertainty component was negligible compared to the network
component, and it was a similar case over early LANs. With today's network
hardware this uncertainty can no longer be ignored. Now when I say NTP I
mean the protocol, not it's implementations.

>
> There is work in progress at the IETF working group on an "interleaved"
> mode
> that would send the time stamp from one packet in the next packet.  That
> adds
> a delay in getting the data which can't be good but I don't know how bad
> it
> is.  I've been thinking that a system could measure that delay and add it
> to
> the time stamp.  No code yet.


Interesting - this is pretty much PTP's two-step mode. The only way to
attempt at grabbing a TX timestamp sensibly in software is by the kernel,
just before it passes the packet to hardware, and long after the
application "sent" it, and Linux supports this. PTP's Follow-Up message is
transmitted as soon as the TX timestamp is grabbed. This extra dead time is
minimal. No way can that timestamp be transmitted in the original packet
itself - that would be like PTP's one-step mode, but only hardware can do
it. FPGA- or otherwise hardware-based NTP servers do this today.



>
>
> --
> These are my opinions.  I hate spam.
>
>
>
>
> ___
> time-nuts mailing list -- time-nuts@lists.febo.com
> To unsubscribe, go to
> http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
> and follow the instructions there.
>
___
time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe, go to 
http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.

Re: [time-nuts] SMPS or conventional?

[Angus via time-nuts]

Hi,

   The magnetic field issue has bothered me a little too, but I don't
know whether it has a practical effect. Maybe putting a fan in a
plastic or aluminium box close to the Rb might show something - unless
someone already knows the effects.

   I have a compass on the bench beside the LPRO, and it does drift a
bit. What surprised me most was how much the field around the mains
transformers changed each time they are switched on or off.

Angus.

On Wed, 21 Oct 2020 10:04:11 -0400, you wrote:

>Hi
>
>The only issue with the “controlled fan” approach is that you
>have a variable magnetic field as a result. That and the vibration
>both can impact the stability of the Rb. Some means of isolating
>the fan from the immediate vicinity of the Rb sounds like a good idea. 
>
>Bob
>
>> On Oct 21, 2020, at 9:31 AM, Luiz Alberto Saba  wrote:
>> 
>> I am thinking in an black aluminium finned heatsink, the size of the base 
>> plate, and a fan, controlled by an electronic thermostat (sensor direct 
>> coupled to the heatsink). Is that ok?
>> 
>> Em 21/10/2020 09:40, Wannes Sels escreveu:
>>> There are conflicting requirements regarding temperature in Rb's:
>>> 
>>> - For best performance, the rubidium and quartz oscillators must be kept at
>>> a high and stable temperature.
>>> - For reliability, the supporting electronics must be kept at a lower
>>> temperature.
>>> 
>>> The heater takes care of the high temperature. The stability of the
>>> temperature can be improved by increasing the thermal mass, i.e. adding a
>>> block of aluminium or copper.
>>> The electronics can be kept cooler with a heatsink, and forced ventilation
>>> if needed.
>>> 
>>> When you stick both the electronics and physics in a small package, this
>>> becomes more difficult. For the PRS-10 some form of heatsink is pretty much
>>> required to keep the electronics cool enough. Although I seem to remember
>>> reading somewhere that the "benchtop" heatsink they offer now is lighter
>>> than older units.
>>> 
>>> If the room temperature is not stable enough, you might want to increase
>>> the size of the heatsink and add a fan. This reduces the temperature swing
>>> inside the unit, while still cooling the electronics, at the cost of
>>> increased power draw for the heater.
>>> 
>>> 
>>> 
>>> 
>>> On Wed, Oct 21, 2020 at 5:38 AM Hal Murray  wrote:
>>> 
> I spent a lot of years buying Rb’s and putting them on small heatsinks.
 I
> always was disappointed in their reliability. That continued to be the
 case
> up to the point that the baseplate temp’s got into the 40C region.  In my
> case, that took a fan ….
 How well did it work if the heat sink wasn't small?  What is your version
 of
 small?
 
 Do you have any data (or vague memories) of how much it helps to orient
 the
 heat sink so the fins are vertical so they encourage warm air to flow up
 past
 the fins?
 
 --
 These are my opinions.  I hate spam.
 
 
 
 
 ___
 time-nuts mailing list -- time-nuts@lists.febo.com
 To unsubscribe, go to
 http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
 and follow the instructions there.
 
>>> ___
>>> time-nuts mailing list -- time-nuts@lists.febo.com
>>> To unsubscribe, go to 
>>> http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
>>> and follow the instructions there.
>> 
>> 
>> ___
>> time-nuts mailing list -- time-nuts@lists.febo.com
>> To unsubscribe, go to 
>> http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
>> and follow the instructions there.
>
>
>___
>time-nuts mailing list -- time-nuts@lists.febo.com
>To unsubscribe, go to 
>http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
>and follow the instructions there.

___
time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe, go to 
http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.

Re: [time-nuts] SMPS or conventional?

[Hal Murray]

> like with TEC heat/cool capability

What's the typical MTBF of TEC coolers?  How much does it depend on how much 
power you put through them?

-- 
These are my opinions.  I hate spam.




___
time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe, go to 
http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.

Re: [time-nuts] IEEE 1588 PTP support on raspberry pi 4 compute module

[Poul-Henning Kamp]

Wojciech Owczarek writes:

>This topic was bound to hit time-nuts - great to see Rpi finally going for
>a MAC/PHY with hardware timestamp support and hopefully future full-sized
>models will follow suit. Unfortunately for most chips other than Intel and
>enterprise comms vendors like Solarflare, Mellanox etc, a IEEE-1588
>compliant MAC/PHY means that the silicon physically only timestamps PTP
>packets, whereas with Intel and others, it can timestamp all packets,
>making for example hardware-backed NTP possible.

So a couple of footnotes from somebody who have actually used this kind
of hardware.

What the "timestamping" hardware gives you, is a snapshot of a
counter running a the PHY frequency.

That PHY frequency and its quality depends on what is on the other
end of the cable.

Rule of thumb: If your switch does not have a metal cabinet and
does not speak ssh(1), you will be disappointed.

You need to tie that PHY counter to your kernels timekeeping.

If you do it by appointing it "timecounter", any event affecting
PHY frequnecy, downnegotiation, unplugged cable, power-cycled switch,
EMC, Lightning, shitty hardware will screw up your computer's
time-keeping.

If you do not appoint it "timecounter", you need to poll it from
the kernel often enough to precisely model it detect all of the
above disturbances.

On "real" PTP hardware, this probing uses a separate signal which
is timestamped by the PHY counter *and* the CPU timecounter
you depend on.  This may be easier, and will certainly be cheaper
in magic smoke, on an RPi, but only possible given chip-docs.

(When I did this on Intels '599 chip ten years ago, my questions
regarding these details caused the "datasheet" to grow by nearly
250 pages.)

Absent access to such a signal, you have to do the probing entirely
in software, which is tricky, expensive, and due to the memory-hierarchy
and clock boundaries, very noisy process.

So, yeah... Perfect time-nuts project :-)


-- 
Poul-Henning Kamp   | UNIX since Zilog Zeus 3.20
p...@freebsd.org | TCP/IP since RFC 956
FreeBSD committer   | BSD since 4.3-tahoe
Never attribute to malice what can adequately be explained by incompetence.

___
time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe, go to 
http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.