Re: TRS80 Model 4P CRT contrast fail

[Adrian Graham via cctalk]

> On 26 Nov 2017, at 21:06, Adrian Graham  wrote:
> 
> I replied to this thread a while back but managed to only reply to Tony and 
> not the list, oops! Rather than using my multi-transistor tester for TR103 I 
> tested it with a DMM in diode mode and it’s almost a dead short from B-E, I 
> get the same reading in both directions so I’m going to order a new one in a 
> bit.

Replying to myself here but to my surprise the replacement 2SC2228 arrived 
today and the machine is running nicely now, albeit with some screen wobble. I 
also tested the replacement in my transistor tester and it was correctly 
identified as a BJT NPN so I should’ve questioned it at the beginning but hey. 
Live and learn.

It’s currently running the built-in RAM test which is a lot more than it’s done 
for the last 15 years :)

Cheers folks!

> —
> Adrian/Witchy
> Binary Dinosaurs - Celebrating Computing History from 1972 onwards
> 

Re: TRS80 Model 4P CRT contrast fail

[Adrian Graham via cctalk]

> On 26 Nov 2017, at 19:31, Brent Hilpert via cctalk  
> wrote:
> 
> 
> With a scope the objective would be to look at the levels when pixels are 
> 'on' in the video signal.
> For white/on pixels, you need conductivity through TR102.E-C and TR103.E-C,
> to allow electrons to flow from GND, through those transistors, into the 
> cathode of the CRT,
> and be emitted from the cathode towards the screen.
> 
> For a white pixel then:
>   - the video signal (and to a lesser level, the base of TR102) must be 
> going positive, to turn TR102 on.
>   - TR103 collector should be going 'lower' during the pixel period, the 
> black level is approx. 60 V,
> the white level should drop well below that.
> 
> You probably won't see much voltage variation at CRT PCB point H / TR102.C / 
> TR103.E
> as it's just current-switching there with no collector pull-up.
> 
> You could try connecting CRT PCB point H briefly to GND through a 150-300 ohm 
> R, or short TR102.C-E (not 103).
> That should bring up a white screen.
> If it doesn't it would suggest the problem is around TR103.
> If it does the problem is likely around TR102.

I replied to this thread a while back but managed to only reply to Tony and not 
the list, oops! Rather than using my multi-transistor tester for TR103 I tested 
it with a DMM in diode mode and it’s almost a dead short from B-E, I get the 
same reading in both directions so I’m going to order a new one in a bit.

> If your multimeter has an hfe mode, you could try pulling TR102 & 103 and see 
> if they show gain.

It does but the legs on TR103 weren’t long enough to reach the contacts in the 
BCE holes which is why I went for diode mode instead :)

Thanks for the explanation, if the new transistor doesn’t make a difference 
then I know where else I can look.

Cheers!

—
Adrian/Witchy
Binary Dinosaurs - Celebrating Computing History from 1972 onwards

Re: TRS80 Model 4P CRT contrast fail

[Brent Hilpert via cctalk]

On 2017-Nov-26, at 7:42 AM, Tony Duell via cctalk wrote:
> On Sun, Nov 26, 2017 at 3:02 PM, Adrian Graham via cctalk
>  wrote:
> 
>>> What happens if C101 is removed? Also try running it with TR101
>>> removed (or at least the emitter lead of that transistor disconnected).
>> 
>> Nothing happens, in that the picture is the same as it is with both those 
>> components fitted.
> 
> That suggests the problem is not in the contrast control or the clamp
> transistor TR101.
> I would suspect TR102, TR103 and associated components.
> 
> I don't have the schematic with me at the moment, but as I remember it
> you can look
> at the connection between TR102 (common emitter stage) and TR103 (common  base
> stage on top of it) on one of the connections to the CRT base PCB. And of 
> course
> the output of the video amplifier on the CRT cathode pin. It would be
> worth looking
> at the signals on those points with a 'scope, see what amplitude of
> video signal (not
> the 'standing' DC level).



With a scope the objective would be to look at the levels when pixels are 'on' 
in the video signal.
For white/on pixels, you need conductivity through TR102.E-C and TR103.E-C,
to allow electrons to flow from GND, through those transistors, into the 
cathode of the CRT,
and be emitted from the cathode towards the screen.

For a white pixel then:
- the video signal (and to a lesser level, the base of TR102) must be 
going positive, to turn TR102 on.
- TR103 collector should be going 'lower' during the pixel period, the 
black level is approx. 60 V,
  the white level should drop well below that.

You probably won't see much voltage variation at CRT PCB point H / TR102.C / 
TR103.E
as it's just current-switching there with no collector pull-up.

You could try connecting CRT PCB point H briefly to GND through a 150-300 ohm 
R, or short TR102.C-E (not 103).
That should bring up a white screen.
If it doesn't it would suggest the problem is around TR103.
If it does the problem is likely around TR102.

If your multimeter has an hfe mode, you could try pulling TR102 & 103 and see 
if they show gain.

The monitor board schematic on page pdf.173 of

https://amaus.net/static/S100/tandy/systems/model%204/Tandy%20Model%204P%20Service%20Manual.pdf
presents DC voltage levels to be expected at the transistors and other points.
That's for the green screen monitor, looks to be the same as the white-screen 
(pdf.172) except for some minor component value diffs.

Re: TRS80 Model 4P CRT contrast fail

[Tony Duell via cctalk]

On Sun, Nov 26, 2017 at 3:02 PM, Adrian Graham via cctalk
 wrote:

>> What happens if C101 is removed? Also try running it with TR101
>> removed (or at least the emitter lead of that transistor disconnected).
>
> Nothing happens, in that the picture is the same as it is with both those 
> components fitted.

That suggests the problem is not in the contrast control or the clamp
transistor TR101.
I would suspect TR102, TR103 and associated components.

I don't have the schematic with me at the moment, but as I remember it
you can look
at the connection between TR102 (common emitter stage) and TR103 (common  base
stage on top of it) on one of the connections to the CRT base PCB. And of course
the output of the video amplifier on the CRT cathode pin. It would be
worth looking
at the signals on those points with a 'scope, see what amplitude of
video signal (not
the 'standing' DC level).

-tony

Re: TRS80 Model 4P CRT contrast fail

[Adrian Graham via cctalk]

> On 26 Nov 2017, at 06:15, Tony Duell  wrote:
> 
> On Sat, Nov 25, 2017 at 11:02 PM, Adrian Graham via cctalk
>  wrote:
>> Evening!
>> 
>>> OK... Firstly, measure the voltage on pin 8 of the monitor PCB. Can you get
>>> it to swing from 0 to around 12V?
>> 
>> Yes, it tops out at 11.82V.
> 
> Can you get it down to 0V?

Yep, it runs from 0-11.82V.

>> 
>>> Scope the video output from the CPU board (on pin 2 of the monitor PCB). Is
>>> it a good TTL level
>>> signal?
>> 
>> Yes. Given that it’s TTL and a 3-wire connection (composite plus H sync) I
> 
> You mean 'video', not 'composite'. Composite is an analogue signal with the
> video information and the syncs all combined

Indeed. My brain was saying ‘video’ but my fingers typed ‘composite’ :)

>> should be able to feed that into a PET monitor shouldn’t I? Wish I hadn’t 
>> given
>> away my ‘big’ TRS80 Model 4 :/
> 
> I have no idea if the sync polarities are right, but you can try it with a Pet
> monitor. A TRS80 Model 3 has the same monitor as the Model 4 of course.

I’ll make up an extension cable I think, got more PETs than I know what to do 
with. I used to have multiple Model 3s AND a Model 2 but they had to go when I 
moved to a smaller house back in 2010.

> (And I wish I could find a 4P over here….)

Perhaps embarrassingly I can’t remember where mine came from but given I’ve had 
it since the early noughties that’s not surprising. It may even have been from 
a car boot sale.

> 
> What happens if C101 is removed? Also try running it with TR101
> removed (or at least the emitter lead of that transistor disconnected).

Nothing happens, in that the picture is the same as it is with both those 
components fitted.

Cheers

A

Re: TRS80 Model 4P CRT contrast fail

[Tony Duell via cctalk]

On Sat, Nov 25, 2017 at 11:02 PM, Adrian Graham via cctalk
 wrote:
> Evening!
>
>> OK... Firstly, measure the voltage on pin 8 of the monitor PCB. Can you get
>> it to swing from 0 to around 12V?
>
> Yes, it tops out at 11.82V.

Can you get it down to 0V?

>
>> Scope the video output from the CPU board (on pin 2 of the monitor PCB). Is
>> it a good TTL level signal?
>
> Yes. Given that it’s TTL and a 3-wire connection (composite plus H sync) I

You mean 'video', not 'composite'. Composite is an analogue signal with the
video information and the syncs all combined.

> should be able to feed that into a PET monitor shouldn’t I? Wish I hadn’t 
> given
> away my ‘big’ TRS80 Model 4 :/

I have no idea if the sync polarities are right, but you can try it with a Pet
monitor. A TRS80 Model 3 has the same monitor as the Model 4 of course.

(And I wish I could find a 4P over here)

>
>> Measure the voltage on pin A of the CRT base PCB. Expect around 60V here
>
> 60.3V

Fine.

>
>> Check that C101 is not shorted. You could try the machine with it removed
>> to see what happens.

What happens if C101 is removed? Also try running it with TR101
removed (or at least the emitter lead of that transistor disconnected).

>>
>> Have you checked the TR103 on the CRT base PCB?
>
> It tests as 2 diodes which is correct for a bipolar transistor I believe.

Well, it's not open or shorted which is good. Doesn't mean it has
any gain, but for the moment let's look elsewhere

-tony

Re: TRS80 Model 4P CRT contrast fail

[Adrian Graham via cctalk]

Evening!

> OK... Firstly, measure the voltage on pin 8 of the monitor PCB. Can you get
> it to swing from 0 to around 12V?

Yes, it tops out at 11.82V.

> Scope the video output from the CPU board (on pin 2 of the monitor PCB). Is
> it a good TTL level signal?

Yes. Given that it’s TTL and a 3-wire connection (composite plus H sync) I 
should be able to feed that into a PET monitor shouldn’t I? Wish I hadn’t given 
away my ‘big’ TRS80 Model 4 :/

> Measure the voltage on pin A of the CRT base PCB. Expect around 60V here

60.3V

> Check that C101 is not shorted. You could try the machine with it removed
> to see what happens.
> 
> Have you checked the TR103 on the CRT base PCB?

It tests as 2 diodes which is correct for a bipolar transistor I believe.

Cheers Tony,

A

Re: TRS80 Model 4P CRT contrast fail

[Adrian Graham via cctalk]

> On 24 Nov 2017, at 08:44, Adrian Graham via cctalk  
> wrote:
> 
> 
>> On 24 Nov 2017, at 08:41, Brent Hilpert via cctalk  
>> wrote:
>> 
>> On 2017-Nov-23, at 5:07 PM, Adrian Graham via cctalk wrote:
 On 23 Nov 2017, at 21:18, Brent Hilpert via cctalk  
 wrote:
 ...
 You might check the contrast control.
 I can't find the actual schematic for the contrast control connections, 
 but I 'm guessing the ends are connected across 12V (video board pins 5 & 
 1/4/10) and the wiper to video board pin 8.
 The contrast control can be expected to take pin 8 somewhere + to get 
 pixels fully on.
>>> 
>>> You’re right with pin 8, tracing the circuit I can see why someone 
>>> recommended I check Q101 and Q102 which look OK in my tester.
>>> The contrast control itself tests OK resistance wise, it goes from pretty 
>>> much 0ohms to 610M.
>> 
>> 610M ?
>> I hope that M is a typo.
>> 
> 
> Morning!
> 
> It is, 610k or 0.6M :)
> 

Gah, ignore this. 

—
Adrian/Witchy
Binary Dinosaurs - Celebrating Computing History from 1972 onwards

Re: TRS80 Model 4P CRT contrast fail

[Adrian Graham via cctalk]

> On 24 Nov 2017, at 08:41, Brent Hilpert via cctalk  
> wrote:
> 
> On 2017-Nov-23, at 5:07 PM, Adrian Graham via cctalk wrote:
>>> On 23 Nov 2017, at 21:18, Brent Hilpert via cctalk  
>>> wrote:
>>> ...
>>> You might check the contrast control.
>>> I can't find the actual schematic for the contrast control connections, but 
>>> I 'm guessing the ends are connected across 12V (video board pins 5 & 
>>> 1/4/10) and the wiper to video board pin 8.
>>> The contrast control can be expected to take pin 8 somewhere + to get 
>>> pixels fully on.
>> 
>> You’re right with pin 8, tracing the circuit I can see why someone 
>> recommended I check Q101 and Q102 which look OK in my tester.
>> The contrast control itself tests OK resistance wise, it goes from pretty 
>> much 0ohms to 610M.
> 
> 610M ?
> I hope that M is a typo.
> 

Morning!

It is, 610k or 0.6M :)

Cheers

A

Re: TRS80 Model 4P CRT contrast fail

[Adrian Graham via cctalk]

> On 24 Nov 2017, at 17:23, Tony Duell via cctalk  wrote:
> 
> On Fri, Nov 24, 2017 at 8:41 AM, Brent Hilpert via cctalk
>  wrote:
>> On 2017-Nov-23, at 5:07 PM, Adrian Graham via cctalk wrote:
 On 23 Nov 2017, at 21:18, Brent Hilpert via cctalk  
 wrote:
 ...
 You might check the contrast control.
 I can't find the actual schematic for the contrast control connections, 
 but I 'm guessing the ends are connected across 12V (video board pins 5 & 
 1/4/10) and the wiper to video board pin 8.
 The contrast control can be expected to take pin 8 somewhere + to get 
 pixels fully on.
>>> 
>>> You’re right with pin 8, tracing the circuit I can see why someone 
>>> recommended I check Q101 and Q102 which look OK in my tester.
>>> The contrast control itself tests OK resistance wise, it goes from pretty 
>>> much 0ohms to 610M.
>> 
>> 610M ?
>> I hope that M is a typo.
> 
> Indeed...
> 
> I am working from the Model 4P service manual which I downloaded from
> archive.org I
> think. I can't find a schematic for the control panel, but the parts
> list says the 2
> pots are 500 Ohms (contrast) and 500k brightness. One of diagrams at
> the start of
> the manual suggests the former has 3 wires (wired as a potentiometer),
> the latter
> 2 (wired as a variable resistor).

Yep, see my correction this morning. Just checked it again and it’s 0-641ohms.

> I am assuming the fault is correctly described, a lack of contrast and not
> brightness. Can you get a bright raster if you turn the brightness control up?

Yep, that’s how I can *just* see the boot logo. 

> OK... Firstly, measure the voltage on pin 8 of the monitor PCB. Can you get
> it to swing from 0 to around 12V?

I’ll have to partially reassemble out of the case to test, but I can get to the 
other end of pin 8 with the control panel removed.

> Scope the video output from the CPU board (on pin 2 of the monitor PCB). Is
> it a good TTL level signal?

Good point.

> 
> Measure the voltage on pin A of the CRT base PCB. Expect around 60V here.
> 
> Check that C101 is not shorted. You could try the machine with it removed
> to see what happens.

I removed it and tested it last night, looks OK.

> Have you checked the TR103 on the CRT base PCB?

Not yet, sleep was more important at that point :) 

I’ll report back tomorrow, cheers!

—
Adrian/Witchy
Binary Dinosaurs - Celebrating Computing History from 1972 onwards

Re: TRS80 Model 4P CRT contrast fail

[Brent Hilpert via cctalk]

On 2017-Nov-23, at 5:07 PM, Adrian Graham via cctalk wrote:
>> On 23 Nov 2017, at 21:18, Brent Hilpert via cctalk  
>> wrote:
>> ...
>> You might check the contrast control.
>> I can't find the actual schematic for the contrast control connections, but 
>> I 'm guessing the ends are connected across 12V (video board pins 5 & 
>> 1/4/10) and the wiper to video board pin 8.
>> The contrast control can be expected to take pin 8 somewhere + to get pixels 
>> fully on.
> 
> You’re right with pin 8, tracing the circuit I can see why someone 
> recommended I check Q101 and Q102 which look OK in my tester.
> The contrast control itself tests OK resistance wise, it goes from pretty 
> much 0ohms to 610M.

610M ?
I hope that M is a typo.

Re: TRS80 Model 4P CRT contrast fail

[Adrian Graham via cctalk]

> On 23 Nov 2017, at 20:43, Chuck Guzis via cctalk  
> wrote:
> 
> On 11/23/2017 11:34 AM, Adrian Graham via cctalk wrote:
> 
>> None of the pots on the video board deal with contrast so I’ve got it out on 
>> the bench to remove and test the capacitors. What else can I look at at the 
>> same time? Someone else has already mentioned the transistors Q101, Q102 and 
>> Q103 (on the yoke board).
> 
> Have you tried replacing C101?  What happens if you temporarily remove
> TR101?
> 

It tests ok at 4816nF with 1.4ohms ESR. I need to work out a way of getting 
everything connected together outside the case so I can do measurements. I’m 
hoping it’s something fairly easy to spot as high voltage even at low current 
is daunting.

Cheers

A

Re: TRS80 Model 4P CRT contrast fail

[Adrian Graham via cctalk]

> On 23 Nov 2017, at 21:18, Brent Hilpert via cctalk  
> wrote:
> 
> On 2017-Nov-23, at 11:34 AM, Adrian Graham via cctalk wrote:
>> Hi folks,
>> 
>> A random facebook post made me dig out my model 4P which has never worked as 
>> long as I’ve owned it, it’s always just sat on a shelf looking cute. 
>> 
>> Good old ASTEC PSU problem so I fixed that, reseated all the thankfully 
>> socketed chips and it burst into life. Sort of. I know it’s running because 
>> I can turn the brightness up to raster lines and see the pattern actually 
>> change when RESET is pressed. With contrast right up I can JUST see the boot 
>> logo appear while it reads the floppy.
>> 
>> None of the pots on the video board deal with contrast so I’ve got it out on 
>> the bench to remove and test the capacitors. What else can I look at at the 
>> same time? Someone else has already mentioned the transistors Q101, Q102 and 
>> Q103 (on the yoke board).
>> 
>> Screen pic: http://www.binarydinosaurs.co.uk/trs80Model4PScreen.jpg 
>>  - you can JUST see 
>> the boot logo, I know it’s not screen burn because it does disappear on 
>> RESET and comes back again.
>> 
>> Video board schematic: 
>> http://www.binarydinosaurs.co.uk/trs80Model4PVideoBoardSchematic.png 
>>  
> 
> 
> You might check the contrast control.
> I can't find the actual schematic for the contrast control connections, but I 
> 'm guessing the ends are connected across 12V (video board pins 5 & 1/4/10) 
> and the wiper to video board pin 8.
> The contrast control can be expected to take pin 8 somewhere + to get pixels 
> fully on.

You’re right with pin 8, tracing the circuit I can see why someone recommended 
I check Q101 and Q102 which look OK in my tester. The contrast control itself 
tests OK resistance wise, it goes from pretty much 0ohms to 610M.

> Or get back to us with some voltage measurements at the BCE of TR101, 2 and 
> 3, and at the external connections to the little board on the end of the CRT.
> Note that there will (should) be several hundred volts on some of those CRT 
> board connections.
> 
> Does the "internal brightness" control on the video board have some effect?

Yes, it works with the external control. It was the first thing I checked last 
night.

Cheers

A

Re: TRS80 Model 4P CRT contrast fail

[Brent Hilpert via cctalk]

On 2017-Nov-23, at 11:34 AM, Adrian Graham via cctalk wrote:
> Hi folks,
> 
> A random facebook post made me dig out my model 4P which has never worked as 
> long as I’ve owned it, it’s always just sat on a shelf looking cute. 
> 
> Good old ASTEC PSU problem so I fixed that, reseated all the thankfully 
> socketed chips and it burst into life. Sort of. I know it’s running because I 
> can turn the brightness up to raster lines and see the pattern actually 
> change when RESET is pressed. With contrast right up I can JUST see the boot 
> logo appear while it reads the floppy.
> 
> None of the pots on the video board deal with contrast so I’ve got it out on 
> the bench to remove and test the capacitors. What else can I look at at the 
> same time? Someone else has already mentioned the transistors Q101, Q102 and 
> Q103 (on the yoke board).
> 
> Screen pic: http://www.binarydinosaurs.co.uk/trs80Model4PScreen.jpg 
>  - you can JUST see 
> the boot logo, I know it’s not screen burn because it does disappear on RESET 
> and comes back again.
> 
> Video board schematic: 
> http://www.binarydinosaurs.co.uk/trs80Model4PVideoBoardSchematic.png 
>  


You might check the contrast control.
I can't find the actual schematic for the contrast control connections, but I 
'm guessing the ends are connected across 12V (video board pins 5 & 1/4/10) and 
the wiper to video board pin 8.
The contrast control can be expected to take pin 8 somewhere + to get pixels 
fully on.

Or get back to us with some voltage measurements at the BCE of TR101, 2 and 3, 
and at the external connections to the little board on the end of the CRT.
Note that there will (should) be several hundred volts on some of those CRT 
board connections.

Does the "internal brightness" control on the video board have some effect?

Re: TRS80 Model 4P CRT contrast fail

[Chuck Guzis via cctalk]

On 11/23/2017 11:34 AM, Adrian Graham via cctalk wrote:

> None of the pots on the video board deal with contrast so I’ve got it out on 
> the bench to remove and test the capacitors. What else can I look at at the 
> same time? Someone else has already mentioned the transistors Q101, Q102 and 
> Q103 (on the yoke board).

Have you tried replacing C101?  What happens if you temporarily remove
TR101?

--Chuck

TRS80 Model 4P CRT contrast fail

[Adrian Graham via cctalk]

Hi folks,

A random facebook post made me dig out my model 4P which has never worked as 
long as I’ve owned it, it’s always just sat on a shelf looking cute. 

Good old ASTEC PSU problem so I fixed that, reseated all the thankfully 
socketed chips and it burst into life. Sort of. I know it’s running because I 
can turn the brightness up to raster lines and see the pattern actually change 
when RESET is pressed. With contrast right up I can JUST see the boot logo 
appear while it reads the floppy.

None of the pots on the video board deal with contrast so I’ve got it out on 
the bench to remove and test the capacitors. What else can I look at at the 
same time? Someone else has already mentioned the transistors Q101, Q102 and 
Q103 (on the yoke board).

Screen pic: http://www.binarydinosaurs.co.uk/trs80Model4PScreen.jpg 
 - you can JUST see 
the boot logo, I know it’s not screen burn because it does disappear on RESET 
and comes back again.

Video board schematic: 
http://www.binarydinosaurs.co.uk/trs80Model4PVideoBoardSchematic.png 
 

Cheers!

—
Adrian/Witchy
Binary Dinosaurs - Celebrating Computing History from 1972 onwards