16 Apr
2020
16 Apr
'20
12:23 a.m.
Wonderful: A few weeks ago I forgot to turn off my VT52 and left it
running for a day or two. Now the screen is filled with snow and it
looks like the text is all over the place horozontally.
Any tips or thoughts on where to start looking to fix? The keyboard
seems to be working as does the RS232 input (the snow on the screen
changes when the pdp11 talks to it)
Thanks!
Chris
16 Apr
16 Apr
11:18 p.m.
On 04/15/2020 07:23 PM, Chris Zach via cctalk wrote:
Wonderful: A few weeks ago I forgot to turn off my
VT52
and left it running for a day or two. Now the screen is
filled with snow and it looks like the text is all over
the place horozontally.
Any tips or thoughts on where to start looking to fix? The
keyboard seems to be working as does the RS232 input (the
snow on the screen changes when the pdp11 talks to it)
Unseat and replug all connectors for a first start. Your
description sounds like the monitor may not be getting sync
from the display generator board.
Jon
17 Apr
17 Apr
6:21 a.m.
On 2020-Apr-15, at 5:23 PM, Chris Zach via cctalk wrote:
Wonderful: A few weeks ago I forgot to turn off my
VT52 and left it running for a day or two. Now the screen is filled with snow and it looks
like the text is all over the place horozontally.
Any tips or thoughts on where to start looking to fix? The keyboard seems to be working
as does the RS232 input (the snow on the screen changes when the pdp11 talks to it)
Try turning the contrast or brightness up till the black level starts becoming
illuminated. Is the raster (the screen character field / framing rectangle) stable?
If no rectangle or stable display field becomes apparent then as Jon suggests it may be
loss of sync, probably horizontal, and likely to be in the monitor circuitry.
(Any affect from twiddling the H or V sync controls?)
If the raster is stable then it might be something in the character generation pipeline
(the character generator or the row pixel serializer), or in the monitor video amp.
("snow" as a description leaves for a range of possible interpretations.)
There's always checking power supply levels to start with. In addition to the +5,
there's probably a +/-12 or -5 for the character generator.
4:35 p.m.
On Fri, Apr 17, 2020 at 2:21 AM Brent Hilpert via cctalk <
cctalk at classiccmp.org> wrote:
On 2020-Apr-15, at 5:23 PM, Chris Zach via cctalk wrote:
Ian Primas knows a lot about these. (snark).....but def check the power
first, the large filter caps. I have a working VT50, which is similar but
not exactly. If you leave a terminal on for a long time and it dies,
that's a power issue or power related (caps resistor fried diode, etc.)
IMHO.
Wonderful: A few weeks ago I forgot to turn off
my VT52 and left it
running for a day or two. Now the screen is filled with snow
and it looks
like the text is all over the place horozontally.
Any tips or thoughts on where to start looking to fix? The keyboard
seems to be
working as does the RS232 input (the snow on the screen changes
when the pdp11 talks to it)
Try turning the contrast or brightness up till the black level starts
becoming illuminated. Is the raster (the screen character field / framing
rectangle) stable?
If no rectangle or stable display field becomes apparent then as Jon
suggests it may be loss of sync, probably horizontal, and likely to be in
the monitor circuitry.
(Any affect from twiddling the H or V sync controls?)
If the raster is stable then it might be something in the character
generation pipeline (the character generator or the row pixel serializer),
or in the monitor video amp.
("snow" as a description leaves for a range of possible interpretations.)
There's always checking power supply levels to start with. In addition to
the +5, there's probably a +/-12 or -5 for the character generator.
18 Apr
18 Apr
8:45 a.m.
Hello Chris!
Sorry that your trusty VT52 failed. But it shouldn't be too hard to repair
it!
I have successfully restored a VT52 (actually a VT50 with a brain
transplant): http://www.datormuseum.se/home/dec-vt5x. My experience is that
the capacitors were in good shape. Especially the bigger filter capacitor.
I think that DEC used great quality capacitors.
It is always a good idea to check the voltages so that they are within the
spec. But it should not be necessary to replace all the capacitors in my
opinion.
The VT52 is a quite interesting design with a very simplistic CPU built
from TTL components and a microprogram that has 1024 by
8 bits. Most of the
operations inside it is controlled from this program. In mine
the sockets
for the mikroprogram were bad which meant that it failed to do anything.
Then I had a couple of gates that were bad.
Den tors 16 apr. 2020 kl 23:07 skrev Chris Zach via cctalk <
cctalk at classiccmp.org>:
Wonderful: A few weeks ago I forgot to turn off my
VT52 and left it
running for a day or two. Now the screen is filled with snow and it
looks like the text is all over the place horozontally.
It would be great to have a picture since it could give some hints on what
is wrong with it. My understanding is that the scan is working both
horizontally and vertically. Otherwise you would only have a straight line.
Unlike many common CRTs this CRT circuitry has no oscillator in it that
creates a picture without input. And since the horisontal
deflection circuit also creates the high voltage there would be no picture
at all unless the horisontal signal is there.
This means that the two control boards in the bottom creates some kind of
horisontal and vertical signals. These signals comes from a simple divider
chain. There could be some kind of problem in the divider chain. But then
it would be more likely to not work at all if one of the counters are bad.
But it is a good idea to check the H and V signal som they are inline with
the spec. Around 15kHz and 60 or 50 Hz respectively.
Since I haven't seen how the output looks like I speculating a bit. It
could be the case that the divider chain is correct but the video signal is
not generated in sync with the divider chain. Then the characters would end
up all over the place on screen.
There is a flip-flip, made out of a 7400 (E16) and 74H10 (E14) that is the
video flip-flop. It controls when the screen is rendered. It might be
the case it could cause your problem. I think it is a good idea to check
the signals around this flip-flop.
I am always a bit skeptical towards 74Hxx. My impression is that they fail
more often than standard 74xx.
When I repaired my VT52 I made a quick circuit that created a composite
video signal from the control board. The signal was fed into a standard CRT
monitor which I had around. I could then run the control boards flat on
the bench to take measurements rather than assembling prior to each test. I
thought that it was quite handy.
Any tips or thoughts on where to start looking to fix?
The keyboard
seems to be working as does the RS232 input (the snow on the screen
changes when the pdp11 talks to it)
This indicate that since you are using RS232 you have at least some kind of
voltage level on +12 V and -12 V since otherwise there would be no
communication. And +5V is probably not that bad either. Since not much
would operate in that case. Likewise the -5V is probably ok since the char
rom would not work very well without the bias.
Good luck with your repair!
/Mattis
Thanks!
Chris
4:46 p.m.
Here's a video of it with sound on.
https://i.imgur.com/X1qVYGP.mp4
C
On 4/18/2020 4:45 AM, Mattis Lind wrote:
Hello Chris!
Sorry that your trusty VT52 failed. But it shouldn't be too hard to
repair it!
I have successfully restored a VT52 (actually a VT50 with a brain
transplant): http://www.datormuseum.se/home/dec-vt5x. My experience is
that the capacitors were in good shape. Especially the bigger filter
capacitor. I think that DEC used great quality capacitors.
It is always a good idea to check the voltages so that they are within
the spec. But it should not be necessary to replace all the?capacitors
in my opinion.
The VT52 is a quite interesting design with a very simplistic CPU built
from TTL components and a microprogram that has 1024 by 8 bits. Most of
the operations inside it is controlled from this program. In mine the
sockets for the mikroprogram?were bad which meant that it failed to do
anything. Then I had a couple of gates that were bad.
Den tors 16 apr. 2020 kl 23:07 skrev Chris Zach via cctalk
<cctalk at classiccmp.org <mailto:cctalk at classiccmp.org>>:
Wonderful: A few weeks ago I forgot to turn off my VT52 and left it
running for a day or two. Now the screen is filled with snow and it
looks like the text is all over the place horozontally.
It would be great to have a picture since it could give some hints on
what is wrong with it. My understanding is that the scan is working both
horizontally and vertically. Otherwise you would only have a straight
line. Unlike many common CRTs this CRT circuitry has no oscillator?in it
that creates a picture without input. And since the horisontal
deflection?circuit also creates?the high voltage there would be no
picture at all unless the horisontal signal is there.
This means that the two control boards in the bottom creates some kind
of horisontal and vertical signals. These signals comes from a simple
divider chain. There could be some kind of problem in the divider chain.
But then it would be more likely to not work at all if one of the
counters are bad. But it is a good idea to check the H and V signal som
they are inline with the spec. Around 15kHz and 60 or 50 Hz respectively.
Since I haven't seen how the output?looks like I speculating a bit. It
could be the case that the divider chain is correct but the video signal
is not?generated in sync with the divider chain. Then the characters
would end up all over the place on screen.
There is a flip-flip, made out of a 7400 (E16) and 74H10 (E14) that is
the video flip-flop. It controls when the screen is rendered. It might
be the?case it could?cause your problem. I think it is a? good idea to
check the signals around this flip-flop.
I am always a bit skeptical towards 74Hxx. My impression is that they
fail more often than standard 74xx.
When I repaired my VT52 I made a quick circuit that created a composite
video signal from the control board. The signal was fed into a standard
CRT monitor which I had around.? I could then run the control boards
flat on the bench to take measurements rather than assembling prior to
each test. I thought that it was quite handy.
Any tips or thoughts on where to start looking to fix? The keyboard
seems to be working as does the RS232 input (the snow on the screen
changes when the pdp11 talks to it)
This indicate that since you are using RS232 you have at least some kind
of voltage level on?+12 V and -12 V since otherwise there would be no
communication. And?+5V is probably not that bad either. Since not much
would operate in that case. Likewise the -5V is probably ok since the
char rom would not work very well without the bias.
Good luck with your repair!
/Mattis
Thanks!
Chris
5:07 p.m.
On Sat, Apr 18, 2020 at 10:47 AM Chris Zach via cctalk <
cctalk at classiccmp.org> wrote:
Here's a video of it with sound on.
https://i.imgur.com/X1qVYGP.mp4
We had a VT52 do something akin to that once back in the day... The tech we
had that worked on this stuff said he had to fix something in the vertical
refresh circuit to achieve vertical hold... My vague memories likely won't
help you much, but this is quite distinctive and others of the group will
be able to suggest things :)
Warner
C
On 4/18/2020 4:45 AM, Mattis Lind wrote:
Hello Chris!
Sorry that your trusty VT52 failed. But it shouldn't be too hard to
repair it!
I have successfully restored a VT52 (actually a VT50 with a brain
transplant): http://www.datormuseum.se/home/dec-vt5x. My experience is
that the capacitors were in good shape. Especially the bigger filter
capacitor. I think that DEC used great quality capacitors.
It is always a good idea to check the voltages so that they are within
the spec. But it should not be necessary to replace all the capacitors
in my opinion.
The VT52 is a quite interesting design with a very simplistic CPU built
from TTL components and a microprogram that has 1024 by 8 bits. Most of
the operations inside it is controlled from this program. In mine the
sockets for the mikroprogram were bad which meant that it failed to do
anything. Then I had a couple of gates that were bad.
Den tors 16 apr. 2020 kl 23:07 skrev Chris Zach via cctalk
<cctalk at classiccmp.org <mailto:cctalk at classiccmp.org>>:
Wonderful: A few weeks ago I forgot to turn off my VT52 and left it
running for a day or two. Now the screen is filled with snow and it
looks like the text is all over the place horozontally.
It would be great to have a picture since it could give some hints on
what is wrong with it. My understanding is that the scan is working both
horizontally and vertically. Otherwise you would only have a straight
line. Unlike many common CRTs this CRT circuitry has no oscillator in it
that creates a picture without input. And since the horisontal
deflection circuit also creates the high voltage there would be no
picture at all unless the horisontal signal is there.
This means that the two control boards in the bottom creates some kind
of horisontal and vertical signals. These signals comes from a simple
divider chain. There could be some kind of problem in the divider chain.
But then it would be more likely to not work at all if one of the
counters are bad. But it is a good idea to check the H and V signal som
they are inline with the spec. Around 15kHz and 60 or 50 Hz respectively.
Since I haven't seen how the output looks like I speculating a bit. It
could be the case that the divider chain is correct but the video signal
is not generated in sync with the divider chain. Then the characters
would end up all over the place on screen.
There is a flip-flip, made out of a 7400 (E16) and 74H10 (E14) that is
the video flip-flop. It controls when the screen is rendered. It might
be the case it could cause your problem. I think it is a good idea to
check the signals around this flip-flop.
I am always a bit skeptical towards 74Hxx. My impression is that they
fail more often than standard 74xx.
When I repaired my VT52 I made a quick circuit that created a composite
video signal from the control board. The signal was fed into a standard
CRT monitor which I had around. I could then run the control boards
flat on the bench to take measurements rather than assembling prior to
each test. I thought that it was quite handy.
Any tips or thoughts on where to start looking to fix? The keyboard
seems to be working as does the RS232 input (the snow on the screen
changes when the pdp11 talks to it)
This indicate that since you are using RS232 you have at least some kind
of voltage level on +12 V and -12 V since otherwise there would be no
communication. And +5V is probably not that bad either. Since not much
would operate in that case. Likewise the -5V is probably ok since the
char rom would not work very well without the bias.
Good luck with your repair!
/Mattis
Thanks!
Chris
7 p.m.
We had a VT52 do something akin to that once back in
the day... The tech
we had that worked on this stuff said he had to fix something in the
vertical refresh circuit to achieve vertical hold... My vague memories
likely won't help you much, but this is quite distinctive and others of the
group will be able to suggest things :)
Vertical hold implies that there is an oscillator that need to get into
hold or lock with an incoming signal. There is none in a VT52 so that is
not possible. There is simply no way to adjust vertical hold in a VT52.
/Mattis
Warner
C
On 4/18/2020 4:45 AM, Mattis Lind wrote:
Hello Chris!
Sorry that your trusty VT52 failed. But it shouldn't be too hard to
repair it!
I have successfully restored a VT52 (actually a VT50 with a brain
transplant): http://www.datormuseum.se/home/dec-vt5x. My experience is
that the capacitors were in good shape. Especially the bigger filter
capacitor. I think that DEC used great quality capacitors.
It is always a good idea to check the voltages so that they are within
the spec. But it should not be necessary to replace all the capacitors
in my opinion.
The VT52 is a quite interesting design with a very simplistic CPU built
from TTL components and a microprogram that has 1024 by 8 bits. Most of
the operations inside it is controlled from this program. In mine the
sockets for the mikroprogram were bad which meant that it failed to do
anything. Then I had a couple of gates that were bad.
Den tors 16 apr. 2020 kl 23:07 skrev Chris Zach via cctalk
<cctalk at classiccmp.org <mailto:cctalk at classiccmp.org>>:
Wonderful: A few weeks ago I forgot to turn off my VT52 and left it
running for a day or two. Now the screen is filled with snow and it
looks like the text is all over the place horozontally.
It would be great to have a picture since it could give some hints on
what is wrong with it. My understanding is that the scan is working
both
horizontally and vertically. Otherwise you would
only have a straight
line. Unlike many common CRTs this CRT circuitry has no oscillator in
it
that creates a picture without input. And since
the horisontal
deflection circuit also creates the high voltage there would be no
picture at all unless the horisontal signal is there.
This means that the two control boards in the bottom creates some kind
of horisontal and vertical signals. These signals comes from a simple
divider chain. There could be some kind of problem in the divider
chain.
But then it would be more likely to not work at
all if one of the
counters are bad. But it is a good idea to check the H and V signal som
they are inline with the spec. Around 15kHz and 60 or 50 Hz
respectively.
Since I haven't seen how the output looks like I speculating a bit. It
could be the case that the divider chain is correct but the video
signal
is not generated in sync with the divider chain.
Then the characters
would end up all over the place on screen.
There is a flip-flip, made out of a 7400 (E16) and 74H10 (E14) that is
the video flip-flop. It controls when the screen is rendered. It might
be the case it could cause your problem. I think it is a good idea to
check the signals around this flip-flop.
I am always a bit skeptical towards 74Hxx. My impression is that they
fail more often than standard 74xx.
When I repaired my VT52 I made a quick circuit that created a composite
video signal from the control board. The signal was fed into a standard
CRT monitor which I had around. I could then run the control boards
flat on the bench to take measurements rather than assembling prior to
each test. I thought that it was quite handy.
Any tips or thoughts on where to start looking to fix? The keyboard
seems to be working as does the RS232 input (the snow on the screen
changes when the pdp11 talks to it)
This indicate that since you are using RS232 you have at least some
kind
of voltage level on +12 V and -12 V since
otherwise there would be no
communication. And +5V is probably not that bad either. Since not much
would operate in that case. Likewise the -5V is probably ok since the
char rom would not work very well without the bias.
Good luck with your repair!
/Mattis
Thanks!
Chris
8:35 p.m.
On Sat, Apr 18, 2020, 1:01 PM Mattis Lind <mattislind at gmail.com> wrote:
Told you my memory sucked...
Warner
/Mattis
We had a VT52 do something akin to that once back
in the day... The tech
we had that worked on this stuff said he had to fix something in the
vertical refresh circuit to achieve vertical hold... My vague memories
likely won't help you much, but this is quite distinctive and others of the
group will be able to suggest things :)
Vertical hold implies that there is an oscillator that need to get into
hold or lock with an incoming signal. There is none in a VT52 so that is
not possible. There is simply no way to adjust vertical hold in a VT52.
> Warner
>
>
>>
>> C
>>
>> On 4/18/2020 4:45 AM, Mattis Lind wrote:
>> > Hello Chris!
>> >
>> > Sorry that your trusty VT52 failed. But it shouldn't be too hard to
>> > repair it!
>> >
>> > I have successfully restored a VT52 (actually a VT50 with a brain
>> > transplant): http://www.datormuseum.se/home/dec-vt5x. My experience
>> is
>> > that the capacitors were in good shape. Especially the bigger filter
>> > capacitor. I think that DEC used great quality capacitors.
>> >
>> > It is always a good idea to check the voltages so that they are within
>> > the spec. But it should not be necessary to replace all the capacitors
>> > in my opinion.
>> >
>> > The VT52 is a quite interesting design with a very simplistic CPU
>> built
>> > from TTL components and a microprogram that has 1024 by 8 bits. Most
>> of
>> > the operations inside it is controlled from this program. In mine the
>> > sockets for the mikroprogram were bad which meant that it failed to do
>> > anything. Then I had a couple of gates that were bad.
>> >
>> > Den tors 16 apr. 2020 kl 23:07 skrev Chris Zach via cctalk
>> > <cctalk at classiccmp.org <mailto:cctalk at classiccmp.org>>:
>> >
>> > Wonderful: A few weeks ago I forgot to turn off my VT52 and left it
>> > running for a day or two. Now the screen is filled with snow and it
>> > looks like the text is all over the place horozontally.
>> >
>> >
>> > It would be great to have a picture since it could give some hints on
>> > what is wrong with it. My understanding is that the scan is working
>> both
>> > horizontally and vertically. Otherwise you would only have a straight
>> > line. Unlike many common CRTs this CRT circuitry has no oscillator in
>> it
>> > that creates a picture without input. And since the horisontal
>> > deflection circuit also creates the high voltage there would be no
>> > picture at all unless the horisontal signal is there.
>> >
>> > This means that the two control boards in the bottom creates some kind
>> > of horisontal and vertical signals. These signals comes from a simple
>> > divider chain. There could be some kind of problem in the divider
>> chain.
>> > But then it would be more likely to not work at all if one of the
>> > counters are bad. But it is a good idea to check the H and V signal
>> som
>> > they are inline with the spec. Around 15kHz and 60 or 50 Hz
>> respectively.
>> >
>> > Since I haven't seen how the output looks like I speculating a bit. It
>> > could be the case that the divider chain is correct but the video
>> signal
>> > is not generated in sync with the divider chain. Then the characters
>> > would end up all over the place on screen.
>> >
>> > There is a flip-flip, made out of a 7400 (E16) and 74H10 (E14) that is
>> > the video flip-flop. It controls when the screen is rendered. It might
>> > be the case it could cause your problem. I think it is a good idea to
>> > check the signals around this flip-flop.
>> >
>> > I am always a bit skeptical towards 74Hxx. My impression is that they
>> > fail more often than standard 74xx.
>> >
>> > When I repaired my VT52 I made a quick circuit that created a
>> composite
>> > video signal from the control board. The signal was fed into a
>> standard
>> > CRT monitor which I had around. I could then run the control boards
>> > flat on the bench to take measurements rather than assembling prior to
>> > each test. I thought that it was quite handy.
>> >
>> >
>> > Any tips or thoughts on where to start looking to fix? The keyboard
>> > seems to be working as does the RS232 input (the snow on the screen
>> > changes when the pdp11 talks to it)
>> >
>> >
>> > This indicate that since you are using RS232 you have at least some
>> kind
>> > of voltage level on +12 V and -12 V since otherwise there would be no
>> > communication. And +5V is probably not that bad either. Since not much
>> > would operate in that case. Likewise the -5V is probably ok since the
>> > char rom would not work very well without the bias.
>> >
>> > Good luck with your repair!
>> >
>> > /Mattis
>> >
>> >
>> > Thanks!
>> > Chris
>> >
>>
>
10:56 p.m.
Well, let's see: I took the bottom off the VT52 and checked the
connections to the high voltage board on the back. Cleaned up both
connectors, plugged in, same problem.
I do hear a medium to high pitched squeal coming from the back board.
Can't quite triangulate it (I need a hose and some time) but it might be
coming from those center transistors. Any idea what these boards are
doing, and if there is a schematic for these circuits?
Given that it went weird after being powered on for awhile I would
suspect it has something to do with a component that generates heat. I
could also take an IR picture of the back board and see if anything is
unusually hot/not hot.
Audio of the noise: https://i.imgur.com/ps4Ovkj.mp4
Pic of the back: https://i.imgur.com/HgNoTYK.jpg
C
11:33 p.m.
Here's a link to the vt52 maintenance manual. Has testing and service info. Hope it
helps.
http://bitsavers.informatik.uni-stuttgart.de/pdf/dec/terminal/vt52/EK-VT52-…
Wayne
-----Original Message-----
From: cctalk [mailto:cctalk-bounces at classiccmp.org] On Behalf Of Chris Zach via cctalk
Sent: Saturday, April 18, 2020 3:56 PM
Cc: General Discussion: On-Topic and Off-Topic Posts
Subject: Re: Great, my VT52 is shot.
Well, let's see: I took the bottom off the VT52 and checked the connections to the
high voltage board on the back. Cleaned up both connectors, plugged in, same problem.
I do hear a medium to high pitched squeal coming from the back board.
Can't quite triangulate it (I need a hose and some time) but it might be coming from
those center transistors. Any idea what these boards are doing, and if there is a
schematic for these circuits?
Given that it went weird after being powered on for awhile I would suspect it has
something to do with a component that generates heat. I could also take an IR picture of
the back board and see if anything is unusually hot/not hot.
Audio of the noise: https://i.imgur.com/ps4Ovkj.mp4 Pic of the back:
https://i.imgur.com/HgNoTYK.jpg
C
19 Apr
19 Apr
5:54 a.m.
s?ndag 19 april 2020 skrev Chris Zach via cctalk <cctalk at classiccmp.org>:
Well, let's see: I took the bottom off the VT52
and checked the
connections to the high voltage board on the back. Cleaned up both
connectors, plugged in, same problem.
I do hear a medium to high pitched squeal coming from the back board.
Can't quite triangulate it (I need a hose and some time) but it might be
coming from those center transistors. Any idea what these boards are doing,
and if there is a schematic for these circuits?
The left is the video /CRT part. The right is the power supply part. The
transistors are mostly regulator transistors.
I have linked to both maintenance manual and schematic from the page I
referred to in one of my previous replies to this thread.
http://www.datormuseum.se/home/dec-vt5x
Take a look there!
/Mattis
Given that it went weird after being powered on for
awhile I would suspect
it has something to do with a component that generates heat. I could also
take an IR picture of the back board and see if anything is unusually
hot/not hot.
Audio of the noise: https://i.imgur.com/ps4Ovkj.mp4
Pic of the back: https://i.imgur.com/HgNoTYK.jpg
C
3:22 p.m.
After looking at the video, and Mattis' description of the
sweep method, I'm guessing a counter
in the chain has gone bad. Assuming this uses a 5 x 7
character matrix, then there would be probably
a divide by 6 for the scan across each character plus space
between, then a counter for the number
of characters per line, a counter for the 8 or 9 raster scan
lines for each row, and then a counter
for the number of rows. At least one of those counters
might be counting too fast. Does the vertical just give one
reset pulse, or does it give a pulse to bump the trace down
the screen one scan line at a
time? Also, a cap in the vertical sweep circuit may have
gone bad, causing the vertical to move too
fast, as evidenced by sloping lines on the screen.
Just guessing from a distance.
Jon
6:58 p.m.
Den s?n 19 apr. 2020 kl 17:22 skrev Jon Elson <elson at pico-systems.com>:
After looking at the video, and Mattis'
description of the
sweep method, I'm guessing a counter
in the chain has gone bad. Assuming this uses a 5 x 7
character matrix, then there would be probably
a divide by 6 for the scan across each character plus space
between, then a counter for the number
of characters per line, a counter for the 8 or 9 raster scan
lines for each row, and then a counter
for the number of rows. At least one of those counters
might be counting too fast. Does the vertical just give one
reset pulse, or does it give a pulse to bump the trace down
the screen one scan line at a
time? Also, a cap in the vertical sweep circuit may have
gone bad, causing the vertical to move too
fast, as evidenced by sloping lines on the screen.
As I wrote earlier in the thread I think it is a good idea to check the V
sync and H sync signals to check that they are right in pulse length and
shape.
There are indeed electrolytic capacitors in the vertical deflection circuit
but I am not sure if that would make the beam move much faster since for a
faster move of the beam also requires a higher voltage over the deflection
coil to create a faster ramp up of the current trough it. But it is
definitely worth checking. Especially since the scan lines are sloping a
bit and not straight as one would expect.
The interesting thing is that the counter chain that develops the V and H
sync signals are different from the on the ones that generate addresses to
the picture memory. There are separate X and Y counters that addresses the
memory and then for each scan line the lower bits of the accumulator are
used for addressing the char ROM together with the information from screen
memory.
The VT52 is a quite complex juggle where counters interact with the program
which runs in sync with all operations.
So, yes, a counter might have gone bad. Myself I have never seen a
counter that has counted too fast unless the input was too fast. But I
probably still have peculiarities to experience.
Just guessing from a distance.
So do I. And now we need Chris to do some measurements on it to get further.
/Mattis
Jon
10:09 p.m.
As I wrote earlier in the thread I think it is a good
idea to check the V
sync and H sync signals to check that they are right in pulse length and
shape.
Good starting point. From the schematics would that be scope probe to J1
pin 5 for horizontal and j1 pin 9 for vertical? Also which is J1, and
how are they numbered :-) Also is there a good ground reference point?
Also is there a document that links the part numbers on the boards with
what's on the schematics?
There are indeed electrolytic capacitors in the
vertical deflection circuit
but I am not sure if that would make the beam move much faster since for a
faster move of the beam also requires a higher voltage over the deflection
coil to create a faster ramp up of the current trough it. But it is
definitely worth checking. Especially since the scan lines are sloping a
bit and not straight as one would expect.
I would believe a capacitor failure due to heat, any ones in particular
I should check?
So do I. And now we need Chris to do some measurements
on it to get further.
More than happy, just haven't worked on TV sets before and would rather
not blow off my hand. I'm guessing the really dangerous voltage is the
one on the bottom left that comes off the step up transformer over to
the odd plug going into the VT52 rear bulkhead (HV to display)
C
11:27 p.m.
On 2020-Apr-19, at 3:09 PM, Chris Zach via cctalk wrote:
Using the schematic and maintenance manual from bitsavers:
http://bitsavers.org/pdf/dec/terminal/vt52/MP00035_VT52schem.pdf
http://bitsavers.org/pdf/dec/terminal/vt52/EK-VT52-MM-002_maint_Jul78.pdf
Monitor schematic: MP00035_VT52schem pdf-page 10
Monitor board component layout: MP00035_VT52schem pdf-page 8
There are *two* horizontal signals from the logic to the monitor (plus the vertical &
video).
There are some waveform diagrams for what to expect around the monitor board presented in
the maint manual:
EK-VT52-MM-002_maint_Jul78 pdf-page 92
Something else that might help with diagnosis is taking pictures of the screen with known,
simple elements on the screen.
e.g.:
- clear the screen
- take pic if it doesn't clear
- type a simple character like a "-"or "1"
- pic
- type a half line or full line of characters
- pic
- type different characters on two different lines
- pic
The idea being to find out, from simple known patterns, where the pixels end up being
displayed,
so it may be possible to figure out what's happening with the scan, whether it's
getting stretched, folding over, etc.
Yes, you want to watch out for HV around the flyback transformer.
The very HV is reasonably insulated, but there's also a few-hundred volts around the
open componentry around the focus and intensity controls, the stuff feeding the neck end
of the CRT.
As I wrote
earlier in the thread I think it is a good idea to check the V
sync and H sync signals to check that they are right in pulse length and
shape.
Good starting point. From the schematics would that be scope probe to J1 pin 5 for
horizontal and j1 pin 9 for vertical? Also which is J1, and how are they numbered :-) Also
is there a good ground reference point?
Also is there a document that links the part numbers on the boards with what's on the
schematics?
There are indeed electrolytic capacitors in the
vertical deflection circuit
but I am not sure if that would make the beam move much faster since for a
faster move of the beam also requires a higher voltage over the deflection
coil to create a faster ramp up of the current trough it. But it is
definitely worth checking. Especially since the scan lines are sloping a
bit and not straight as one would expect.
I would believe a capacitor failure due to heat, any ones in particular I should check?
So do I. And now we need Chris to do some
measurements on it to get further.
More than happy, just haven't worked on TV sets before and would rather not blow off
my hand. I'm guessing the really dangerous voltage is the one on the bottom left that
comes off the step up transformer over to the odd plug going into the VT52 rear bulkhead
(HV to display)
20 Apr
20 Apr
6:40 a.m.
Something else that might help with diagnosis is taking pictures of the
screen with known, simple elements on the screen.
e.g.:
- clear the screen
- take pic if it doesn't clear
- type a simple character like a "-"or "1"
- pic
- type a half line or full line of characters
- pic
- type different characters on two different lines
- pic
The idea being to find out, from simple known patterns, where the pixels
end up being displayed,
so it may be possible to figure out what's happening with the scan,
whether it's getting stretched, folding over, etc.
That is a good idea. Loop the terminal and create a bunch of "known
references".
Might add that it would be interesting to see one single character on a
lets say 4 different positions. Column 0, Column 20, Column 40 and end of
line. Approximately since it is hard to be exact.
/Mattis
5:40 p.m.
I looked at the video once more and paused around 0:21. It appears to be
some text there on the uppermost row.
Maybe the vertical sweep is highly non-linear. Check all capacitors in the
vertical stage. C43, C41,C40, C42, C35 and C45. Would probably be good to
measure the voltage over the current sensing resistor R73 and see if it is
a proper ramp as it should.
What about the R92 linearity potentiometer. Is it ok? Diode D31? Actually
check all semiconductors.
/Mattis
m?ndag 20 april 2020 skrev Mattis Lind <mattislind at gmail.com>:
Something else that might help with diagnosis is taking pictures of the
screen with known, simple elements on the screen.
e.g.:
- clear the screen
- take pic if it doesn't clear
- type a simple character like a "-"or "1"
- pic
- type a half line or full line of characters
- pic
- type different characters on two different lines
- pic
The idea being to find out, from simple known patterns, where the pixels
end up being displayed,
so it may be possible to figure out what's happening with the scan,
whether it's getting stretched, folding over, etc.
That is a good idea. Loop the terminal and create a bunch of "known
references".
Might add that it would be interesting to see one single character on a
lets say 4 different positions. Column 0, Column 20, Column 40 and end of
line. Approximately since it is hard to be exact.
/Mattis
21 Apr
21 Apr
2:57 p.m.
Well I'm starting to walk through this. First I took an IR picture of
the boards in operation, then started troubleshooting.
First I started checking voltages. J2 is easily accessible so I put a
ground probe on pin 10, and checked voltages as follows:
J2-3 Should be -5v, reading -3.8v
J2-2 Should be -12v reading -15v
J2-4 Should be 5v, reading 5.04v
J1-10 Should be 15v reading 14.84
Hm. That's odd. Looks like the negative voltages are a bit off.
According to the schematic the key transistor on the 5v rail is Q6,
which is a 2n3055. The key transistor on +15 is Q10 and that circuit
looks ok. However whatever is running the -5 and -12 volts is not
working right.... Looks like Q12 is the key power transistor there with
E2 serving as the control.
Hm....
C
On 4/20/2020 1:40 PM, Mattis Lind via cctalk wrote:
I looked at the video once more and paused around
0:21. It appears to be
some text there on the uppermost row.
Maybe the vertical sweep is highly non-linear. Check all capacitors in the
vertical stage. C43, C41,C40, C42, C35 and C45. Would probably be good to
measure the voltage over the current sensing resistor R73 and see if it is
a proper ramp as it should.
What about the R92 linearity potentiometer. Is it ok? Diode D31? Actually
check all semiconductors.
/Mattis
m?ndag 20 april 2020 skrev Mattis Lind <mattislind at gmail.com>:
>
>
>
>>
>>
>> Something else that might help with diagnosis is taking pictures of the
>> screen with known, simple elements on the screen.
>> e.g.:
>> - clear the screen
>> - take pic if it doesn't clear
>> - type a simple character like a "-"or "1"
>> - pic
>> - type a half line or full line of characters
>> - pic
>> - type different characters on two different lines
>> - pic
>>
>> The idea being to find out, from simple known patterns, where the pixels
>> end up being displayed,
>> so it may be possible to figure out what's happening with the scan,
>> whether it's getting stretched, folding over, etc.
>>
>>
>>
> That is a good idea. Loop the terminal and create a bunch of "known
> references".
>
> Might add that it would be interesting to see one single character on a
> lets say 4 different positions. Column 0, Column 20, Column 40 and end of
> line. Approximately since it is hard to be exact.
>
> /Mattis
>
5:23 p.m.
On 2020-Apr-21, at 7:57 AM, Chris Zach via cctalk wrote:
Well I'm starting to walk through this. First I
took an IR picture of the boards in operation, then started troubleshooting.
First I started checking voltages. J2 is easily accessible so I put a ground probe on pin
10, and checked voltages as follows:
J2-3 Should be -5v, reading -3.8v
J2-2 Should be -12v reading -15v
J2-4 Should be 5v, reading 5.04v
J1-10 Should be 15v reading 14.84
Hm. That's odd. Looks like the negative voltages are a bit off.
According to the schematic the key transistor on the 5v rail is Q6, which is a 2n3055.
The key transistor on +15 is Q10 and that circuit looks ok. However whatever is running
the -5 and -12 volts is not working right.... Looks like Q12 is the key power transistor
there with E2 serving as the control.
Hm....
It's conceivable there was a revision change where the -12V was changed to -15V, but
it does seem unlikely.
The -12 is used by the vertical deflection circuit, the CRT filament, (and the RS-232).
I wouldn't immediately anticipate -12 going to -15 alone would produce the upset
you're seeing on the screen but there could well be interrelated failures between the
vertical deflection and the over-voltage -12, so the excessive -12 should probably be
addressed.
In the power supply, you could measure some voltages around the -12 regulator.
E2.2 and E2.3 should be around 0V.
E2.6 I expect would normally be within +/- 2V of GND.
If E2.6 is way-positive (>Q4.E, > ~ +2.6V), then E2 is attempting to throttle down
the output but failing due to some failure around Q4/Q12.
Measure the voltages on Q4 and Q12.
They could be leaky or low-gain.
It's conceivable Q12 has a CE short, but I'd expect the input to the reg to be
more than -15.
The low -5 is odd, zener D14 looks to be a 5.1V type.
R45 is probably burning up.
I haven't spotted anything in the schematic that actually uses the -5,
it looks to be provided for some alternative character generator (schematic pdf.42),
so that may depend on what character generator chip / daughter board is used in your
unit.
8 p.m.
It's conceivable there was a revision change where
the -12V was changed to -15V, but it does seem unlikely.
The -12 is used by the vertical deflection circuit, the CRT filament, (and the RS-232).
Well, vertical deflection seems to be my problem so this may make sense.
Until I figure this out I'm going to pull that jumper to remove power
from the HV circuits. I'm guessing the circuit can
regulate with no load.
E2.2 and E2.3 should be around 0V.
E2.6 I expect would normally be within +/- 2V of GND.
If E2.6 is way-positive (>Q4.E, > ~ +2.6V), then E2 is attempting to throttle down
the output but failing due to some failure around Q4/Q12.
Measure the voltages on Q4 and Q12.
They could be leaky or low-gain.
It's conceivable Q12 has a CE short, but I'd expect the input to the reg to be
more than -15.
*nod* I just took a look at Q12 and it's not shorted which is good. It
looks to be the regulator power transistor, the other components vary
the base to keep the voltage at the output constant under load. The
input voltage to the circuit is around -24v so it looks like the
rectifier diodes are ok enough.
The low -5 is odd, zener D14 looks to be a 5.1V type.
Yeah, that is really odd, especially if -5v isn't really used. I wonder
if the zener diode is shorted, that could cause all sorts of weirdness.
I'll check.
C
9:36 p.m.
Ok, some test results below.
1) Are we sure J2-2 should be -12? It really seems to like being -15.
2) The voltage across the zener diode there is a nice solid -3.8-3.9v.
It appears to be a germanium diode (.1v drop across, both directions) I
could disconnect it and see if that makes things better.
3)
E2.2 and E2.3 should be around 0V.E2.2 is -.8 volts
referenced to ground.
E2.3 is -22mv referenced to ground
E2.6 is around .6v. Ramps up from about .58 at startup to .61 then down
to .60something. Note that is a positive voltage wrt ground.
Q12 was reading
-24.4v emitter
-15v collector
-23.8v base
Q12 does not have a CE short
What is this using as a reference to get the -12v from the transistor?
Maybe the zener is just out of spec. Can someone with a VT52 cross check
my voltages to see if they are accurate?
Thanks all!
E2.6 I expect would normally be within +/- 2V of GND.
If E2.6 is way-positive (>Q4.E, > ~ +2.6V), then E2 is attempting to throttle down
the output but failing due to some failure around Q4/Q12.
Measure the voltages on Q4 and Q12.
They could be leaky or low-gain.
It's conceivable Q12 has a CE short, but I'd expect the input to the reg to be
more than -15.
The low -5 is odd, zener D14 looks to be a 5.1V type.
R45 is probably burning up.
I haven't spotted anything in the schematic that actually uses the -5,
it looks to be provided for some alternative character generator (schematic pdf.42),
so that may depend on what character generator chip / daughter board is used in your
unit.
9:47 p.m.
I wonder if I would go straight to hell if I just pulled the Q12
transistor and wired in a 12 volt wall wart between ground (J2-2) and
+12 (ground) to get -12....
C
On 4/21/2020 5:36 PM, Chris Zach via cctalk wrote:
Ok, some test results below.
1) Are we sure J2-2 should be -12? It really seems to like being -15.
2) The voltage across the zener diode there is a nice solid -3.8-3.9v.
It appears to be a germanium diode (.1v drop across, both directions) I
could disconnect it and see if that makes things better.
3)
E2.2 and E2.3 should be around 0V.E2.2 is -.8
volts referenced to ground.
E2.3 is -22mv referenced to ground
E2.6 is around .6v. Ramps up from about .58 at startup to .61 then down
to .60something. Note that is a positive voltage wrt ground.
Q12 was reading
-24.4v emitter
-15v collector
-23.8v base
Q12 does not have a CE short
What is this using as a reference to get the -12v from the transistor?
Maybe the zener is just out of spec. Can someone with a VT52 cross check
my voltages to see if they are accurate?
Thanks all!
> E2.6 I expect would normally be within +/- 2V of GND.
> If E2.6 is way-positive (>Q4.E, > ~ +2.6V), then E2 is attempting to
> throttle down the output but failing due to some failure around Q4/Q12.
>
> Measure the voltages on Q4 and Q12.
> They could be leaky or low-gain.
> It's conceivable Q12 has a CE short, but I'd expect the input to the
> reg to be more than -15.
>
> The low -5 is odd, zener D14 looks to be a 5.1V type.
> R45 is probably burning up.
> I haven't spotted anything in the schematic that actually uses the -5,
> it looks to be provided for some alternative character generator
> (schematic pdf.42),
> so that may depend on what character generator chip / daughter board
> is used in your unit.
>
22 Apr
22 Apr
12:27 a.m.
Meantime reading the manual I found an interesting test: If you short
emitter to base on Q4 (easiest way is to jumper diode D10) the voltage
on the -12v supply goes to .4 volts. They're saying it's E2, R15,R17,R14.
Is there a way I can test the op-amp in circuit? Maybe it's dead.
C
On 4/21/2020 5:36 PM, Chris Zach via cctalk wrote:
Ok, some test results below.
1) Are we sure J2-2 should be -12? It really seems to like being -15.
2) The voltage across the zener diode there is a nice solid -3.8-3.9v.
It appears to be a germanium diode (.1v drop across, both directions) I
could disconnect it and see if that makes things better.
3)
E2.2 and E2.3 should be around 0V.E2.2 is -.8
volts referenced to ground.
E2.3 is -22mv referenced to ground
E2.6 is around .6v. Ramps up from about .58 at startup to .61 then down
to .60something. Note that is a positive voltage wrt ground.
Q12 was reading
-24.4v emitter
-15v collector
-23.8v base
Q12 does not have a CE short
What is this using as a reference to get the -12v from the transistor?
Maybe the zener is just out of spec. Can someone with a VT52 cross check
my voltages to see if they are accurate?
Thanks all!
> E2.6 I expect would normally be within +/- 2V of GND.
> If E2.6 is way-positive (>Q4.E, > ~ +2.6V), then E2 is attempting to
> throttle down the output but failing due to some failure around Q4/Q12.
>
> Measure the voltages on Q4 and Q12.
> They could be leaky or low-gain.
> It's conceivable Q12 has a CE short, but I'd expect the input to the
> reg to be more than -15.
>
> The low -5 is odd, zener D14 looks to be a 5.1V type.
> R45 is probably burning up.
> I haven't spotted anything in the schematic that actually uses the -5,
> it looks to be provided for some alternative character generator
> (schematic pdf.42),
> so that may depend on what character generator chip / daughter board
> is used in your unit.
>
3:09 a.m.
On 2020-Apr-21, at 5:27 PM, Chris Zach via cctalk wrote:
Meantime reading the manual I found an interesting
test: If you short emitter to base on Q4 (easiest way is to jumper diode D10) the voltage
on the -12v supply goes to .4 volts. They're saying it's E2, R15,R17,R14.
Is there a way I can test the op-amp in circuit? Maybe it's dead.
The three regulators are all referenced via D8 (~ +5V).
Circuit-wise, for the -12, 0V/GND is the reference input at E2.3.
The D8 5V ref and the targetted output V are divided via R15 and R17,R18 to provide the
sense input at E2.2.
If one does the R ratio of the three resistors, it comes out, as would be expected, to ~
0V.
I'm a little surprised there's that much difference between E2.2 & E2.3 (
(-0.022) - (-0.8) ) without sending the E2 output off to +V, but not sure how much device
variability to expect normally.
You might look for the on-board values of R15,17,18. If they have been changed from those
values specified in the schematic, then the -12 may have been changed to -15 (could do the
ratio calc).
Also what is the V at Q4.E (should be ~ +2.6V), also Q4.B & C.
On 4/21/2020 5:36 PM, Chris Zach via cctalk wrote:
> Ok, some test results below.
> 1) Are we sure J2-2 should be -12? It really seems to like being -15.
> 2) The voltage across the zener diode there is a nice solid -3.8-3.9v. It appears to
be a germanium diode (.1v drop across, both directions) I could disconnect it and see if
that makes things better.
> 3)
> E2.2 is -.8 volts referenced to ground.
> E2.3 is -22mv referenced to ground
> E2.6 is around .6v. Ramps up from about .58 at startup to .61 then down to
.60something. Note that is a positive voltage wrt ground.
> Q12 was reading
> -24.4v emitter
> -15v collector
> -23.8v base
> Q12 does not have a CE short
> What is this using as a reference to get the -12v from the transistor?
3:35 a.m.
The D8 5V ref and the targetted output V are divided
via R15 and R17,R18 to provide the sense input at E2.2.
If one does the R ratio of the three resistors, it comes out, as would be expected, to ~
0V.
Ah hah, that is clever. I wonder if the .8 volts means the output is
higher than what should be expected and the op amp isn't amping down or
something. However if the op amp was blown it should just allow full
voltage through. Maybe (is there a crowbar circuit in there).
I'm a little surprised there's that much
difference between E2.2 & E2.3 ( (-0.022) - (-0.8) ) without sending the E2 output off
to +V, but not sure how much device variability to expect normally.
You might look for the on-board values of R15,17,18.
If they have been changed from those values specified in the schematic, then the -12 may
have been changed to -15 (could do the ratio calc).
I tried measuring the resistance in circuit, but that never works. It's
possible they are measuring out properly and that .6 volts is the
representation of too much voltage at the output side (-15 instead of
the expected -12).
Also what is the V at Q4.E (should be ~ +2.6V), also
Q4.B & C.
I'll check that tomorrow. Connecting Q4.E to B did bring the output
voltage down to pretty much zero so that does seem to work.
If I can't figure this out I might just pull Q10 and put a 7912 in its
place. One chip does the whole job of regulating the output, end of
story. Bad me of course, but what the heck and if the display came up I
would know where the problem was.
If it turns out the op amp is dead, would a 741 work as a replacement?
4:15 a.m.
On 2020-Apr-21, at 8:35 PM, Chris Zach wrote:
Look at the marked (as opposed to measured) values of the (3) installed Rs, to see if they
have been intentionally changed in a revision to alter the output V.
When Q4.BE shorted, E2.6 should swing well-negative.
The D8 5V ref
and the targetted output V are divided via R15 and R17,R18 to provide the sense input at
E2.2.
If one does the R ratio of the three resistors, it comes out, as would be expected, to ~
0V.
Ah hah, that is clever. I wonder if the .8 volts means the output is higher than what
should be expected and the op amp isn't amping down or something. However if the op
amp was blown it should just allow full voltage through. Maybe (is there a crowbar circuit
in there).
I'm a little surprised there's that much
difference between E2.2 & E2.3 ( (-0.022) - (-0.8) ) without sending the E2 output off
to +V, but not sure how much device variability to expect normally.
You might look for the on-board values of
R15,17,18. If they have been changed from those values specified in the schematic, then
the -12 may have been changed to -15 (could do the ratio calc).
I tried measuring the resistance in circuit, but that never works. It's possible they
are measuring out properly and that .6 volts is the representation of too much voltage at
the output side (-15 instead of the expected -12). Also what is
the V at Q4.E (should be ~ +2.6V), also Q4.B & C.
I'll check that tomorrow. Connecting Q4.E to B did bring the output voltage down to
pretty much zero so that does seem to work. If I can't figure this out I might just pull Q10
and put a 7912 in its place. One chip does the whole job of regulating the output, end of
story. Bad me of course, but what the heck and if the display came up I would know where
the problem was.
If it turns out the op amp is dead, would a 741 work as a replacement?
Note comment lower left corner of schematic page.
3:52 p.m.
On 04/21/2020 10:09 PM, Brent Hilpert via cctalk wrote:
On 2020-Apr-21, at 5:27 PM, Chris Zach via cctalk
wrote:
Well, if the circuit **IS** regulating, then the voltage on
the two inputs will be identical.
But, since it might not be regulating, then these voltages
would not be equal.
But, if you can see that the + input is more positive than
the - input, yet the output
is pegged negative, for instance, then you know either the
op-amp is bad, or another circuit is overloading
the output and forcing it that way.
Jon
Meantime reading the manual I found an
interesting test: If you short emitter to base on Q4 (easiest way is to jumper diode D10)
the voltage on the -12v supply goes to .4 volts. They're saying it's E2,
R15,R17,R14.
Is there a way I can test the op-amp in circuit? Maybe it's dead.
7 May
7 May
1:53 a.m.
Well, I pulled the E2 op-amp and replaced it with a NOS one of the same
model. Put the supply together and now I am getting -17 volts on pins E2
to ground (E10).
I'm thinking of just replacing the power transistor Q12 with a 7912 -12v
regulator that I have here and bypassing the whole op amp/transistor
mess. That should give me a solid -12v on the E2 line and provide power
for the -5 volt divider circuit.
Thoughts?
C
On 4/22/2020 11:52 AM, Jon Elson via cctalk wrote:
On 04/21/2020 10:09 PM, Brent Hilpert via cctalk
wrote:
On 2020-Apr-21, at 5:27 PM, Chris Zach via cctalk
wrote:
Well, if the circuit **IS** regulating, then the voltage on the two
inputs will be identical.
But, since it might not be regulating, then these voltages would not be
equal.
But, if you can see that the + input is more positive than the - input,
yet the output
is pegged negative, for instance, then you know either the op-amp is
bad, or another circuit is overloading
the output and forcing it that way.
Jon Meantime reading the manual I found an
interesting test: If you short
emitter to base on Q4 (easiest way is to jumper diode D10) the
voltage on the -12v supply goes to .4 volts. They're saying it's E2,
R15,R17,R14.
Is there a way I can test the op-amp in circuit? Maybe it's dead.
13 May
13 May
6:56 p.m.
New subject: Great, my VT52 is shot.--Progress!
So with the op-amp replaced and the unit still whining and putting out
-15+ volts on the -12 rail I decided to just punt and pull the Q12
transistor. Removed it, will test but doesn't look happy and replaced it
with a 7912 -12 volt regulator mounted to the big heat sink with the
three wires going to both sides of capacitor C17 for ground and input
voltages and the -12 volt output going off to the top side of capacitor
C54. Kind of a hack, but I just want to see if this thing works.
Tested the power supply outputs with a bench voltage tester at -20 volts
and sure enough E2 is reading -11.98 and E3 is reading a solid -5 volts.
Put the power supply board in the unit, hooked it up, fired up the VT52.
Silence.
Dead silent. No whine, no high pitched sounds, no nothing. Quiet. Then I
saw the little blinking cursor underline on the screen and found that
the keys make the clicking and I realized not only is it working, but
it's better than before.
Turned it off, will wait a few hours for the HV to dissipate, then will
screw everything down. I'm thinking this -12 volt supply has been a
problem for some time as it used to make a high pitched TV whine but now
it's nice and quiet. Possible the root cause is something in the Q12
power transistor, but to be honest more modern technology in the power
supply can only be a good thing :-)
Should have it all back together tonight, I'll see if it talks to the
computer. If so we have a working VT52 and now people know that if the
screen looks like mine did the problem is most likely in the -12v circuit.
C
On 5/6/2020 9:53 PM, Chris Zach via cctalk wrote:
Well, I pulled the E2 op-amp and replaced it with a
NOS one of the same
model. Put the supply together and now I am getting -17 volts on pins E2
to ground (E10).
I'm thinking of just replacing the power transistor Q12 with a 7912 -12v
regulator that I have here and bypassing the whole op amp/transistor
mess. That should give me a solid -12v on the E2 line and provide power
for the -5 volt divider circuit.
Thoughts?
C
On 4/22/2020 11:52 AM, Jon Elson via cctalk wrote:
> On 04/21/2020 10:09 PM, Brent Hilpert via cctalk wrote:
>> On 2020-Apr-21, at 5:27 PM, Chris Zach via cctalk wrote:
>>> Meantime reading the manual I found an interesting test: If you
>>> short emitter to base on Q4 (easiest way is to jumper diode D10) the
>>> voltage on the -12v supply goes to .4 volts. They're saying it's E2,
>>> R15,R17,R14.
>>>
>>> Is there a way I can test the op-amp in circuit? Maybe it's dead.
>>
>>
> Well, if the circuit **IS** regulating, then the voltage on the two
> inputs will be identical.
> But, since it might not be regulating, then these voltages would not
> be equal.
> But, if you can see that the + input is more positive than the -
> input, yet the output
> is pegged negative, for instance, then you know either the op-amp is
> bad, or another circuit is overloading
> the output and forcing it that way.
>
> Jon
21 Apr
21 Apr
5:32 p.m.
The raster seems folded at the left and at the top, which indicates a
problem with both the vertical and horizontal sweep.
As Mattis Lind already mentioned, the VT52 uses direct drive of the
vertical and horizontal circuitry (no oscillators to synchronize) yet
the character rows in the video are running diagonally totally out of
sync with the sweep, so I'd check the timing generator (pages 4-17 and
4-18 of the Maintenance Manual, pages 19-21 of the Schematics PDF). But
that's after checking that the power voltages are solid and ripple-free.
I'd also check a few components around the horizontal output stage, e.g.
if capacitor C44 goes bad (page 10) then the pulses from the horizontal
output will get into other parts of the circuit via the power supply
lines and/or the driver stages, affecting the logic's work.
A final note: don't keep the terminal running like that for a long time,
since the horizontal frequency is out of specs the fly-back can get
damaged (the core can saturate, increasing the current drawn; that may
also explain the noise you're hearing).
Hector.
On 4/20/2020 7:40 PM, Mattis Lind via cctalk wrote:
I looked at the video once more and paused around
0:21. It appears to be
some text there on the uppermost row.
Maybe the vertical sweep is highly non-linear. Check all capacitors in the
vertical stage. C43, C41,C40, C42, C35 and C45. Would probably be good to
measure the voltage over the current sensing resistor R73 and see if it is
a proper ramp as it should.
What about the R92 linearity potentiometer. Is it ok? Diode D31? Actually
check all semiconductors.
/Mattis
m?ndag 20 april 2020 skrev Mattis Lind <mattislind at gmail.com>:
>
>
>>
>> Something else that might help with diagnosis is taking pictures of the
>> screen with known, simple elements on the screen.
>> e.g.:
>> - clear the screen
>> - take pic if it doesn't clear
>> - type a simple character like a "-"or "1"
>> - pic
>> - type a half line or full line of characters
>> - pic
>> - type different characters on two different lines
>> - pic
>>
>> The idea being to find out, from simple known patterns, where the pixels
>> end up being displayed,
>> so it may be possible to figure out what's happening with the scan,
>> whether it's getting stretched, folding over, etc.
>>
>>
>>
> That is a good idea. Loop the terminal and create a bunch of "known
> references".
>
> Might add that it would be interesting to see one single character on a
> lets say 4 different positions. Column 0, Column 20, Column 40 and end of
> line. Approximately since it is hard to be exact.
>
> /Mattis
>
--
Hector Peraza
BrightSpec NV/SA
Waterfront Researchpark
Galileilaan 15 (Darwin)
2845 Niel
BTW/VAT: BE 0846.102.393
http://www.brightspec.be
20 Apr
20 Apr
4:22 a.m.
m?ndag 20 april 2020 skrev Chris Zach via cctalk <cctalk at classiccmp.org>:
As I wrote earlier in the thread I think it is a good
idea to check the V
I have a picture of the connector on my page. Have a look there. The
connector have the standard color codes to map to numbers. Black for zero,
Brown for one. Etc.
sync and H sync signals to check that they are
right in pulse length and
shape.
Good starting point. From the schematics would that be scope probe to J1
pin 5 for horizontal and j1 pin 9 for vertical? Also which is J1, and how
are they numbered :-) Also is there a good ground reference point?
Also is there a document that links the part numbers on the boards with
what's on the schematics?
No DEC document. When I did my repair I mapped all ICs to the board but not
the connectors, sorry. You can find all here. Including links to schematis
and maintenance manual: http://www.datormuseum.se/home/dec-vt5x
There are indeed electrolytic capacitors in the
vertical deflection circuit
C45 and C42 could be useful to check.
but I am not sure if that would make the beam
move much faster since for a
faster move of the beam also requires a higher voltage over the deflection
coil to create a faster ramp up of the current trough it. But it is
definitely worth checking. Especially since the scan lines are sloping a
bit and not straight as one would expect.
I would believe a capacitor failure due to heat, any ones in particular I
should check?
So do I. And now we need Chris to do some measurements on it to get
What you could do is what I did. To remove all control board on the bottom
and lay it out flat on the table. Make a small circuit that create a
composite video signal and feed it into a standard composite video monitor
or TV set. If you remove a jumper to th CRT logic there will be no HV at
all generated.
/Mattis
further.
More than happy, just haven't worked on TV sets before and would rather
not blow off my hand. I'm guessing the really dangerous voltage is the one
on the bottom left that comes off the step up transformer over to the odd
plug going into the VT52 rear bulkhead (HV to display)
C
18 Apr
18 Apr
6:52 p.m.
l?rdag 18 april 2020 skrev Chris Zach <cz at alembic.crystel.com>:
Here's a video of it with sound on.
https://i.imgur.com/X1qVYGP.mp4
You have video even in the horisontal retrace. I still think you have a
problem somewhere around the video flip flop. Check that part of the
circuit.
/Mattis
C
On 4/18/2020 4:45 AM, Mattis Lind wrote:
> Hello Chris!
>
> Sorry that your trusty VT52 failed. But it shouldn't be too hard to
> repair it!
>
> I have successfully restored a VT52 (actually a VT50 with a brain
> transplant): http://www.datormuseum.se/home/dec-vt5x. My experience is
> that the capacitors were in good shape. Especially the bigger filter
> capacitor. I think that DEC used great quality capacitors.
>
> It is always a good idea to check the voltages so that they are within
> the spec. But it should not be necessary to replace all the capacitors in
> my opinion.
>
> The VT52 is a quite interesting design with a very simplistic CPU built
> from TTL components and a microprogram that has 1024 by 8 bits. Most of the
> operations inside it is controlled from this program. In mine the sockets
> for the mikroprogram were bad which meant that it failed to do anything.
> Then I had a couple of gates that were bad.
>
> Den tors 16 apr. 2020 kl 23:07 skrev Chris Zach via cctalk <
> cctalk at classiccmp.org <mailto:cctalk at classiccmp.org>>:
>
> Wonderful: A few weeks ago I forgot to turn off my VT52 and left it
> running for a day or two. Now the screen is filled with snow and it
> looks like the text is all over the place horozontally.
>
>
> It would be great to have a picture since it could give some hints on
> what is wrong with it. My understanding is that the scan is working both
> horizontally and vertically. Otherwise you would only have a straight line.
> Unlike many common CRTs this CRT circuitry has no oscillator in it that
> creates a picture without input. And since the horisontal
> deflection circuit also creates the high voltage there would be no picture
> at all unless the horisontal signal is there.
>
> This means that the two control boards in the bottom creates some kind of
> horisontal and vertical signals. These signals comes from a simple divider
> chain. There could be some kind of problem in the divider chain. But then
> it would be more likely to not work at all if one of the counters are bad.
> But it is a good idea to check the H and V signal som they are inline with
> the spec. Around 15kHz and 60 or 50 Hz respectively.
>
> Since I haven't seen how the output looks like I speculating a bit. It
> could be the case that the divider chain is correct but the video signal is
> not generated in sync with the divider chain. Then the characters would end
> up all over the place on screen.
>
> There is a flip-flip, made out of a 7400 (E16) and 74H10 (E14) that is
> the video flip-flop. It controls when the screen is rendered. It might be
> the case it could cause your problem. I think it is a good idea to check
> the signals around this flip-flop.
>
> I am always a bit skeptical towards 74Hxx. My impression is that they
> fail more often than standard 74xx.
>
> When I repaired my VT52 I made a quick circuit that created a composite
> video signal from the control board. The signal was fed into a standard CRT
> monitor which I had around. I could then run the control boards flat on
> the bench to take measurements rather than assembling prior to each test. I
> thought that it was quite handy.
>
>
> Any tips or thoughts on where to start looking to fix? The keyboard
> seems to be working as does the RS232 input (the snow on the screen
> changes when the pdp11 talks to it)
>
>
> This indicate that since you are using RS232 you have at least some kind
> of voltage level on +12 V and -12 V since otherwise there would be no
> communication. And +5V is probably not that bad either. Since not much
> would operate in that case. Likewise the -5V is probably ok since the char
> rom would not work very well without the bias.
>
> Good luck with your repair!
>
> /Mattis
>
>
> Thanks!
> Chris
>
>
1692
days inactive
1719
days old
33 comments
8 participants
participants (8)
-
bhilpert＠shaw.ca -
billdegnan＠gmail.com -
cz＠alembic.crystel.com -
elson＠pico-systems.com -
imp＠bsdimp.com -
mattislind＠gmail.com -
peraza＠newsguy.com -
Wayne.Sudol＠hotmail.com