Re: More circuit help required please

[Adrian Graham]

On 08/01/2017 18:20, "Tony Duell"  wrote:

>> However, two of us bought units at the same time and the other one has a
>> little bridgeboard containing a 74LS244 that sends the RGB signals to a DIN
>> socket on the back of the case for an external monitor. I also can't see a
> 
> And in a sense there was...

Yep. I'm pleased we found that mod in the other unit because I made my own
bridgeboard and hooked the display up to a Microvitec CUB, I got exactly the
same picture (in colour) on that as I did in mono on the 5" TV so I knew the
analogue board and TV were OK.
 
>> luminance line on the connection to the TV board on mine, just
>> red-green-blue with sync and inverse sync which go to the horizontal driver
> 
> Do you mean that the common cathode connection of those diodes goes to
> the monitor connector but nowhere then on the monitor PCB?

No, in trying to trace some of the lines along the bottom of the board I
took the monitor cable off and traced the pins. They are:

1 - 5V
2 - NC
3 - looped to 5
4 - Sync out
5 - looped to 3
6 - emitter of V326 via R336
7 - GREEN
8 - RED
9 - Inverse sync (which also goes via R332 to the base of V333)
10 - BLUE
11 - GND
12 - GND
13 - 12V
14 - 12V
 
>>> Is the MR9735 doing anything? Is it accessing memory, for example?
>>> I still wonder if it has to be initialised (say to vewdata rather than
>>> teletext mode) and if this is not occuring.
>> 
>> Its address and data buses are active as is the pair of 2114s that are
>> acting as a page store. There's a 74LS240 buffering the data bus and that's
>> active.
> 
> OK.. My dislike of 2114s is legendary, but probably not the problem here

They're OK too, surprisingly. I took them off the board and put them in a
PET8032 as video RAM.

>> You could be right though, I'm not convinced the unit is actually running
>> any code yet despite the ROMCS lines working and I can see activity in all
>> 16 RAM chips. Data bus and address bus at CPU and ROM chips looks happy from
> 
> What are the RAMs? If they're more 2114s I wonder if they have problems
> that are confusing the CPU (perhaps corrupting the stack?)

4116-2 wired as 2 banks of 8.

>> a 'doing something' point of view. All the control lines on the CPU are
>> pulsing with the exception of IO/M but if all the IO the chip is doing is to
>> memory then that's to be expected.
> 
> What's the CPU? 8085? I've had those fail in odd ways.

Yeah, 8085A. Thanks to list members I have some spares to try.

>> I've disassembled the ROM code so I'll have a look through that for an init
>> routine, also the datasheet should tell me what its expecting at powerup.
> 
> Personally, I'd stick a logic analyser on it and start tracing the code. See
> if it is executing anything correctly.

I've only got a cheap-ish Saleae compatible 16 channel analyser. Watching
code is a bit above my current level of self-paced learning :)

-- 
Adrian/Witchy
Binary Dinosaurs creator/curator
Www.binarydinosaurs.co.uk - the UK's biggest private home computer
collection?

Re: More circuit help required please

[Tony Duell]

On Sun, Jan 8, 2017 at 6:11 PM, Adrian Graham
 wrote:
> On 08/01/2017 17:21, "Tony Duell"  wrote:
>
>>> I photographed both sides of the board and merged the pictures together to
>>> make a front/back composite. The trace for the cathodes of the 3 diodes
>>> links together on the back of the board and doesn't go anywhere else. The
>>> front trace from V302 cathode goes straight to R312. The anodes DO link with
>>> the monitor driver board through the connection marked MON though.
>>
>> OK, probably some kind of luminance signal. Was there ever a monochrome
>> version of this unit?
>
> The whole unit is monochrome, it only has a little 5" B TV as its output
> device -

OK, was there ever a colour version :-)

>
> http://www.binarydinosaurs.co.uk/STCExecutel01.jpg
>
> However, two of us bought units at the same time and the other one has a
> little bridgeboard containing a 74LS244 that sends the RGB signals to a DIN
> socket on the back of the case for an external monitor. I also can't see a

And in a sense there was...

> luminance line on the connection to the TV board on mine, just
> red-green-blue with sync and inverse sync which go to the horizontal driver

Do you mean that the common cathode connection of those diodes goes to
the monitor connector but nowhere then on the monitor PCB?

>> Is the MR9735 doing anything? Is it accessing memory, for example?
>> I still wonder if it has to be initialised (say to vewdata rather than
>> teletext mode) and if this is not occuring.
>
> Its address and data buses are active as is the pair of 2114s that are
> acting as a page store. There's a 74LS240 buffering the data bus and that's
> active.

OK.. My dislike of 2114s is legendary, but probably not the problem here

>
> You could be right though, I'm not convinced the unit is actually running
> any code yet despite the ROMCS lines working and I can see activity in all
> 16 RAM chips. Data bus and address bus at CPU and ROM chips looks happy from

What are the RAMs? If they're more 2114s I wonder if they have problems
that are confusing the CPU (perhaps corrupting the stack?)


> a 'doing something' point of view. All the control lines on the CPU are
> pulsing with the exception of IO/M but if all the IO the chip is doing is to
> memory then that's to be expected.

What's the CPU? 8085? I've had those fail in odd ways.

>
> I've disassembled the ROM code so I'll have a look through that for an init
> routine, also the datasheet should tell me what its expecting at powerup.

Personally, I'd stick a logic analyser on it and start tracing the code. See
if it is executing anything correctly.

-tony

Re: More circuit help required please

[Adrian Graham]

On 08/01/2017 17:21, "Tony Duell"  wrote:

>> I photographed both sides of the board and merged the pictures together to
>> make a front/back composite. The trace for the cathodes of the 3 diodes
>> links together on the back of the board and doesn't go anywhere else. The
>> front trace from V302 cathode goes straight to R312. The anodes DO link with
>> the monitor driver board through the connection marked MON though.
> 
> OK, probably some kind of luminance signal. Was there ever a monochrome
> version of this unit?

The whole unit is monochrome, it only has a little 5" B TV as its output
device - 

http://www.binarydinosaurs.co.uk/STCExecutel01.jpg

However, two of us bought units at the same time and the other one has a
little bridgeboard containing a 74LS244 that sends the RGB signals to a DIN
socket on the back of the case for an external monitor. I also can't see a
luminance line on the connection to the TV board on mine, just
red-green-blue with sync and inverse sync which go to the horizontal driver
chip and vertical driver chip respectively. You may remember I was asking
questions about these chips a couple of months ago because that TV board
suffered some component rot and I thought the wavy picture was because of
that.

>> One of the reasons I'm asking these questions is I'm trying to see if
>> there's another source of composite sync given that the MR9735 itself
>> doesn't do it, even though it should.
> 
> I doubt it Why would there be?

Clutching at straws :)
 
> Is the MR9735 doing anything? Is it accessing memory, for example?
> I still wonder if it has to be initialised (say to vewdata rather than
> teletext mode) and if this is not occuring.

Its address and data buses are active as is the pair of 2114s that are
acting as a page store. There's a 74LS240 buffering the data bus and that's
active. 

You could be right though, I'm not convinced the unit is actually running
any code yet despite the ROMCS lines working and I can see activity in all
16 RAM chips. Data bus and address bus at CPU and ROM chips looks happy from
a 'doing something' point of view. All the control lines on the CPU are
pulsing with the exception of IO/M but if all the IO the chip is doing is to
memory then that's to be expected.

I've disassembled the ROM code so I'll have a look through that for an init
routine, also the datasheet should tell me what its expecting at powerup.

Cheers,

-- 
Adrian/Witchy
Binary Dinosaurs creator/curator
Www.binarydinosaurs.co.uk - the UK's biggest private home computer
collection?

Re: More circuit help required please

[Tony Duell]

> I photographed both sides of the board and merged the pictures together to
> make a front/back composite. The trace for the cathodes of the 3 diodes
> links together on the back of the board and doesn't go anywhere else. The
> front trace from V302 cathode goes straight to R312. The anodes DO link with
> the monitor driver board through the connection marked MON though.

OK, probably some kind of luminance signal. Was there ever a monochrome
version of this unit?


> One of the reasons I'm asking these questions is I'm trying to see if
> there's another source of composite sync given that the MR9735 itself
> doesn't do it, even though it should.

I doubt it Why would there be?

Is the MR9735 doing anything? Is it accessing memory, for example?
I still wonder if it has to be initialised (say to vewdata rather than teletext
mode) and if this is not occuring.

-tony

Re: More circuit help required please

[Adrian Graham]

On 08/01/2017 16:18, "Mouse"  wrote:

 http://www.binarydinosaurs.co.uk/STCExecutelSyncCircuit.jpg
>>> I don't believe it's correct as drawn.  For one thing you have V333,
>>> an NPN transistor, with the collector grounded (and no -ve supplies
>>> on the circuit.
> 
> Bipolar transistors are to some extent symmetric under interchange of
> collector and emitter.  Simplified pictures showing a bar of
> semiconductor with the ends being emitter and collector and a thin base
> layer in the middle are, well, simplified, but there is _some_ truth
> lurking in them.
> 
> However, the only reason I could see that being done deliberately is if
> the circuit is analog and the transistor's behaviour is relevantly
> different from the normal way around.  This circuit doesn't look like
> that to me.  Redrawing it a bit less confusingly with V333 E/C swapped
> makes it look as though V333 and V326 are being used in their switching
> regions, not their linear regions.

Yes, I got the emitter and collector swapped. The software I'm using
(Fritzing) is excellent given it costs nothing, but occasionally you can
make a change then alter something else in the properties of the item and
the change is lost.
 
> The lack of negative supplies is hardly conclusivein itself; negative
> voltages could be developed in any of many ways.  However, V325, R322,
> and V326 would make it difficult for the R335/V325/V333 junction point
> to get too far below 5V (admittedly this is much less true if R322's
> value is actually significantly higher).
> 
> I also question the way R311/R310/R309 are all different.  I would
> expect the red, green, and blue circuits to be electrically more or
> less identical, and different pulldown values does not fit with that.

There's been a couple of resistor failures with this board so I'm at the
stage of starting with a DMM, swapping probes to hopefully get a consistent
reading then reading the bands and using the calculator here for ones I
don't instantly recognise:

http://www.allaboutcircuits.com/tools/resistor-color-code-calculator/

If the two don't match I take it out of circuit.

-- 
Adrian/Witchy
Binary Dinosaurs creator/curator
Www.binarydinosaurs.co.uk - the UK's biggest private home computer
collection?

Re: More circuit help required please

[Adrian Graham]

On 08/01/2017 14:06, "Tony Duell"  wrote:

> On Sun, Jan 8, 2017 at 2:00 PM, Adrian Graham
>  wrote:
> 
>> You were right, I'd drawn out R335 wrongly and flipped V333. All the anodes
>> go via R312 to the collector of V333 along with R309-R311. Base for V333 is
>> the inverted sync output of the MR9735. R322 is 6.8k
>> (blue-grey-black-brown). All diodes are 1N4148.
>> 
>> http://www.binarydinosaurs.co.uk/STCExecutelSyncCircuit2.jpg
> 
> That makes more sense. I am surprised that the anodes of the diodes don't go
> to a +ve power line (maybe +12V), though. It could be right as drawn,
> but I would check it.

I photographed both sides of the board and merged the pictures together to
make a front/back composite. The trace for the cathodes of the 3 diodes
links together on the back of the board and doesn't go anywhere else. The
front trace from V302 cathode goes straight to R312. The anodes DO link with
the monitor driver board through the connection marked MON though.

> When you power up, what voltage and/or signal (use a 'scope) do you see on the
> anodes of said diodes?

Looks to be 4V on all 3 lines for RGB:

http://www.binarydinosaurs.co.uk/STCExecutelRGBsignal.jpg

One of the reasons I'm asking these questions is I'm trying to see if
there's another source of composite sync given that the MR9735 itself
doesn't do it, even though it should.

Cheers!

-- 
Adrian/Witchy
Binary Dinosaurs creator/curator
Www.binarydinosaurs.co.uk - the UK's biggest private home computer
collection?

Re: More circuit help required please

[Tony Duell]

On Sun, Jan 8, 2017 at 4:18 PM, Mouse  wrote:

> I also question the way R311/R310/R309 are all different.  I would
> expect the red, green, and blue circuits to be electrically more or
> less identical, and different pulldown values does not fit with that.

Well, I wonder if it's producing a weighted sum of R,G,B to make a
monochrome luminance signal. Not at all uncommon.


-tony

Re: More circuit help required please

[Mouse]

>>> http://www.binarydinosaurs.co.uk/STCExecutelSyncCircuit.jpg
>> I don't believe it's correct as drawn.  For one thing you have V333,
>> an NPN transistor, with the collector grounded (and no -ve supplies
>> on the circuit.

Bipolar transistors are to some extent symmetric under interchange of
collector and emitter.  Simplified pictures showing a bar of
semiconductor with the ends being emitter and collector and a thin base
layer in the middle are, well, simplified, but there is _some_ truth
lurking in them.

However, the only reason I could see that being done deliberately is if
the circuit is analog and the transistor's behaviour is relevantly
different from the normal way around.  This circuit doesn't look like
that to me.  Redrawing it a bit less confusingly with V333 E/C swapped
makes it look as though V333 and V326 are being used in their switching
regions, not their linear regions.

The lack of negative supplies is hardly conclusivein itself; negative
voltages could be developed in any of many ways.  However, V325, R322,
and V326 would make it difficult for the R335/V325/V333 junction point
to get too far below 5V (admittedly this is much less true if R322's
value is actually significantly higher).

I also question the way R311/R310/R309 are all different.  I would
expect the red, green, and blue circuits to be electrically more or
less identical, and different pulldown values does not fit with that.

>> For another, R322 is ridiculously low.

Indeed.

I don't know how much you know about electronics at the relevant level,
so it's possible what I'm about to say is unnecessary or has already
been considered.

Did these resistor values come from reading bands, or measuring
in-circuit, or what?  Note that R322 is in parallel with the C-B diode
of V326; a simple ohmmeter put across it may end up measuring the diode
drop instead.  (To test this theory, switch ranges.  A diode drop will
usually give you approximately the same digits in each range - eg, 6.7
ohms on the 0-10 range, 67 ohms on the 0-100 one, 670 ohms for the 0-1K
range, etc.  Also, swapping the probe leads will give a different
reading if that's what's behind this.)

>> And I would expect the anodes of the 3 diodes on the RGB outputs to
>> go somewhere other than a resistor to ground.

At least with R312 as high as 10K.  If it were much lower, the diodes
could be clipping diodes; a diode drop is not too far from about the
right swing for a video signal.

/~\ The ASCII Mouse
\ / Ribbon Campaign
 X  Against HTMLmo...@rodents-montreal.org
/ \ Email!   7D C8 61 52 5D E7 2D 39  4E F1 31 3E E8 B3 27 4B

Re: More circuit help required please

[Tony Duell]

On Sun, Jan 8, 2017 at 2:00 PM, Adrian Graham
 wrote:

> You were right, I'd drawn out R335 wrongly and flipped V333. All the anodes
> go via R312 to the collector of V333 along with R309-R311. Base for V333 is
> the inverted sync output of the MR9735. R322 is 6.8k
> (blue-grey-black-brown). All diodes are 1N4148.
>
> http://www.binarydinosaurs.co.uk/STCExecutelSyncCircuit2.jpg

That makes more sense. I am surprised that the anodes of the diodes don't go
to a +ve power line (maybe +12V), though. It could be right as drawn,
but I would
check it.

When you power up, what voltage and/or signal (use a 'scope) do you see on the
anodes of said diodes?

-tony

Re: More circuit help required please

[Adrian Graham]

On 08/01/2017 07:31, "Tony Duell"  wrote:

> On Sat, Jan 7, 2017 at 11:37 PM, Adrian Graham
>  wrote:
>> Evening all,
>> 
>> I wish I had the ability to take a board layout and turn it into a logically
>> laid out schematic but as yet I don't. Video sync on my Executel 3910 is
>> still running me round in circles so could one of you fine folk take a look
>> at this board layout drawn as best I can:
>> 
>> http://www.binarydinosaurs.co.uk/STCExecutelSyncCircuit.jpg
>> 
>> ...and let me know what it does please? The chip on the left is a Plessey
> 
> 
> I don't believe it's correct as drawn. For one thing you have V333, an NPN
> transistor, with the collector grounded (and no -ve supplies on the circuit.
> For
> another, R322 is ridiculously low. And I would expect the anodes of the
> 3 diodes on the RGB outputs to go somewhere other than a resistor to
> ground.

You were right, I'd drawn out R335 wrongly and flipped V333. All the anodes
go via R312 to the collector of V333 along with R309-R311. Base for V333 is
the inverted sync output of the MR9735. R322 is 6.8k
(blue-grey-black-brown). All diodes are 1N4148.

http://www.binarydinosaurs.co.uk/STCExecutelSyncCircuit2.jpg

Thanks for looking!

-- 
Adrian/Witchy
Binary Dinosaurs creator/curator
Www.binarydinosaurs.co.uk - the UK's biggest private home computer
collection?

Re: More circuit help required please

[Adrian Graham]

On 08/01/2017 07:31, "Tony Duell"  wrote:

> On Sat, Jan 7, 2017 at 11:37 PM, Adrian Graham
>  wrote:
>> Evening all,
>> 
>> I wish I had the ability to take a board layout and turn it into a logically
>> laid out schematic but as yet I don't. Video sync on my Executel 3910 is
>> still running me round in circles so could one of you fine folk take a look
>> at this board layout drawn as best I can:
>> 
>> http://www.binarydinosaurs.co.uk/STCExecutelSyncCircuit.jpg
>> 
>> ...and let me know what it does please? The chip on the left is a Plessey
> 
> 
> I don't believe it's correct as drawn. For one thing you have V333, an NPN
> transistor, with the collector grounded (and no -ve supplies on the circuit.
> For
> another, R322 is ridiculously low. And I would expect the anodes of the
> 3 diodes on the RGB outputs to go somewhere other than a resistor to
> ground.

Hm, ok, that's today's task then. For V333 it could be that I just have the
pin numbers reversed but I'll double check. The diodes bothered me too and
I'm trying to think of a better way of tracing unknown lines on the bottom
of the board. My sponge-wrapped-in-tin-foil method isn't infallible.

Cheers!

-- 
Adrian/Witchy
Binary Dinosaurs creator/curator
Www.binarydinosaurs.co.uk - the UK's biggest private home computer
collection?

Re: More circuit help required please

[Tony Duell]

On Sat, Jan 7, 2017 at 11:37 PM, Adrian Graham
 wrote:
> Evening all,
>
> I wish I had the ability to take a board layout and turn it into a logically
> laid out schematic but as yet I don't. Video sync on my Executel 3910 is
> still running me round in circles so could one of you fine folk take a look
> at this board layout drawn as best I can:
>
> http://www.binarydinosaurs.co.uk/STCExecutelSyncCircuit.jpg
>
> ...and let me know what it does please? The chip on the left is a Plessey


I don't believe it's correct as drawn. For one thing you have V333, an NPN
transistor, with the collector grounded (and no -ve supplies on the circuit. For
another, R322 is ridiculously low. And I would expect the anodes of the
3 diodes on the RGB outputs to go somewhere other than a resistor to
ground.


-tony