9 Feb
2004
9 Feb
'04
8:21 a.m.
On Sat, 2004-02-07 at 23:35, Jules Richardson wrote:
Oh, and if anyone's dishing out Perq wisdom
there's also a Perq 1 with
a
display fault too - on that one the image is centered on the screen,
but
squashed into about 80% of the normal width, and maybe 10% of normal
height. It's also very non-rectangular in nature.
Chatted with my good buddy who still makes a living repairing
terminals. His guess is that the horizontal yoke isolation cap is
leaking/bad. This will definitely make the horizontal squoosh. Many
terminals derive the drive voltage for the vertical from the horizontal
flyback - thereby causing the problem you are seeing.
His recommendations:
1. look for an oversized cap (electrolytic - yup one of those
beasties...) near the horizontal drive transistor (generally
heatsinked) that has a rating between 35 and 60 volts and a value
between 2.2 and 20 microfarads. This is the isolation cap. If it is in
the 2.2 to 4.7 volt range replace it with a film cap - you will never
have a problem again. You can't reform this one since they put AC
(albeit mostly DC) through the cap and once it goes it's gone.
2. Replace any electrolytic near anything that looks hot.
3. Inspect, clean, lubricate, and replace, if necessary, the yoke
connector socket. His experience is that when the cap goes, the current
through the yoke goes up and fries the connector. The bad connection
will cause early failures.
If the vertical problem persists, you probably have a similar problem
with a cap in the vertical drive or a power supply problem.
Good luck
Claude Ceccon
19 Feb
19 Feb
10:50 a.m.
On Sat, 2004-02-07 at 23:35, Jules Richardson wrote:
Ve _very_ careful. The VMI monitor used on the PERQ 1 is unconventional
in a lot of respects (I have an idea that the EHT doesn't come from the
horizontal output stage, but from a second transistor/transformer circuit
also driven from the horizontal drive signal). I will have to dig out the
schematics sometime (yes I have the full manuals...)
I would start, though, by checking the 55V line from the PERQ's PSU to
the monitor. If that's low or noisy, you'll have all sorts of problems...
Oh, and if anyone's dishing out Perq wisdom
there's also a Perq 1 with
a
display fault too - on that one the image is centered on the screen,
but
squashed into about 80% of the normal width, and maybe 10% of normal
height. It's also very non-rectangular in nature.
Chatted with my good buddy who still makes a living repairing
terminals. His guess is that the horizontal yoke isolation cap is
leaking/bad. This will definitely make the horizontal squoosh. Many
terminals derive the drive voltage for the vertical from the horizontal
flyback - thereby causing the problem you are seeing.
His recommendations:
1. look for an oversized cap (electrolytic - yup one of those
beasties...) near the horizontal drive transistor (generally
heatsinked) that has a rating between 35 and 60 volts and a value
Eh? The coupling capacitor is normally a lot higher voltage than that
(600V and above in my experiece!). It's also likely to be a non-polarised
electrolytic (this is similar to having 2 normal electrolytics 'back to
back'), and no, those can't easily be re-formed.
between 2.2 and 20 microfarads. This is the isolation
cap. If it is in
the 2.2 to 4.7 volt range replace it with a film cap - you will never
have a problem again. You can't reform this one since they put AC
(albeit mostly DC) through the cap and once it goes it's gone.
-tony
12:30 p.m.
On Thu, 2004-02-19 at 21:50, Tony Duell wrote:
I will have to dig out the
schematics sometime (yes I have the full manuals...)
I thought if anyone had details on them it'd be you :-)
(Aside - does PERQ stand for something, or was it originally all
capitals just as a trendy marketing thing?)
I would start, though, by checking the 55V line from
the PERQ's PSU to
the monitor. If that's low or noisy, you'll have all sorts of problems...
Rats, I did last weekend and it didn't seem *too* low but I haven't got
the exact figure with me at the mo. I believe it was sitting at 50V or
so, just checked on a meter though and not a scope. The LS logic chip in
it was getting a healthy 5V too. I'll give it a quick (and careful) poke
with a 'scope at the weekend.
Then I think I'll pull the fuses out of it and lock the HD so nobody
fiddles with it and breaks it before the display can be sorted :)
(does appropriate media exist out in the public domain to rebuild the
hard disk contents if needs be? Of course finding a replacement drive if
that breaks is another matter entirely)
cheers
Jules
4:28 p.m.
On Thu, 2004-02-19 at 21:50, Tony Duell wrote:
Alas at the momnet (just come out of hospital, hence the absnece from
this list for a couple of weeks), I don't really feel like climbing over
the VAX rack to get to them...
I will have to dig out the
schematics sometime (yes I have the full manuals...)
I thought if anyone had details on them it'd be you :-)
(Aside - does PERQ stand for something, or was it originally all
capitals just as a trendy marketing thing?)
No, it was claimed to be an acronym : Pascal Evaluation Real Quick (or
Really Quickly).
The machine, remember, was intended to run Q-code (a bit like UCSD
p-code, but with some graphics instructions) as its machine code...
I would start, though, by checking the 55V line
from the PERQ's PSU to
the monitor. If that's low or noisy, you'll have all sorts of problems...
Rats, I did last weekend and it didn't seem *too* low but I haven't got
the exact figure with me at the mo. I believe it was sitting at 50V or (does appropriate media exist out in the public domain
to rebuild the
hard disk contents if needs be? Of course finding a replacement drive if
that breaks is another matter entirely)
Public Domain, not as far as I know (I am not sure who owns the copyright
now). Exists in the hands of enthusiasts? Yes, most certainly.
-tony
20 Feb
20 Feb
5:10 a.m.
On Fri, 2004-02-20 at 03:28, Tony Duell wrote:
I would still check it with a 'scope. I'll see
if I can find the manual
in the next day or so...
No rush - I'll get a scope on it tomorrow and then after that I probably
won't even be back at the museum for a week. There's about a billion
other things that badly need doing there too, so not like I'm stuck for
things to sort out :)
Just to confirm, we're talking about a PERQ 1 here
-- it's confusing to
me wether this is a PERQ 1 (verically-mounted SA4000 hard disk at the
left side, monito connects on a single DC37 connector) or a PERQ 2T1
It's a 2T1, vertical drive mounted at the back.
(there is a PERQ 1 also with a video fault which I think I mentioned;
that's where the confusion probably arises. That one has a totally
mangled display, squashed horizontally and vertically, very non-linear,
and no recognisable character output on it at all)
cheers
Jules
12:18 p.m.
Just to
confirm, we're talking about a PERQ 1 here -- it's confusing to
me wether this is a PERQ 1 (verically-mounted SA4000 hard disk at the
left side, monito connects on a single DC37 connector) or a PERQ 2T1
It's a 2T1, vertical drive mounted at the back.
(there is a PERQ 1 also with a video fault which I think I mentioned;
that's where the confusion probably arises. That one has a totally
mangled display, squashed horizontally and vertically, very non-linear,
and no recognisable character output on it at all)
Do you _know_ if the PERQ is giving the right scan frequencies? (64kHz
horizontal, 60Hz vertical IIRC).
-tony
2:45 p.m.
On Fri, 2004-02-20 at 23:18, Tony Duell wrote:
Oh, right...
Can you please re-describe the 2 faults and the machines they occur on --
I was under the impression I was considering a PERQ 1 (VMI monitor) with
a small picture all round...
Well the 2T1 is the one I'll try and sort out first, as that one
certainly appears to have a healthy (or almost-healthy) system unit.
So, the 2T1 has a standard KME as you say - it looks to have correct
horizontal and vertical drive, with the exception that it skips every
few lines, leaving a blank horizontal gap. I can see what look to be
characters on the display, but they're not at all readable. Difficult to
tell if it's scrambled pixels (suggesting a system fault) or an analogue
problem (suggesting the monitor). The display has an unstable shimmering
right margin to it, which definitely looks analogue in nature.
Onto the PERQ 1. Standard monitor as you say - only that one is
*totally* stuffed. The picture (such as it is) is collapsed into maybe
2" vertically, and about 80% of the screen width horizontally. It's
extremely non-linear; left and right margins weave all over the place. I
can see beam traces all over the shop within the squashed picture too.
It's actually quite a work of art! Not sure if there's more than one
fault going on here, or whether some critical point isn't getting to the
right voltage and that's causing the whole thing to fail.
No idea on that one unfortunately - that why I thought I'd worry about
the 2T1 first as it at least looks like it's trying to do something
useful :)
cheers
Jules
Just to confirm, we're talking about a PERQ 1 here
-- it's confusing to
me wether this is a PERQ 1 (verically-mounted SA4000 hard disk at the
left side, monito connects on a single DC37 connector) or a PERQ 2T1
It's a 2T1, vertical drive mounted at the back. OK, back to the T1. I think it's the machine I
worked on some years ago
(don't worry, everything is 'stock', apart from the fact you may find the
odd IC in a socket if I replaced it).
Ahh, yes it will be. I checked all the connectors and reseated the
boards plus chips - I remember seeing a few socketed ones that didn't
look like they would have originally been socketed.
The only non-standard feature is
the monitor power cable -- I couldn't find the original lead, and the
connector at the monitor end is something odd. So I removed that
connector, fitted a length of cable with a 3 pin DIN plug (to fit the
PERQ's 55V output socket) and wired the other end inside the monitor,
using a strain-relief bush in the holw where the power connector used to
be. That shouldn't cause many problems.
That's definitely the machine. I remember thinking it odd that all the
cables from the monitor were socketed apart from the power lead :)
one), the vertical osciallator components, and the
vertical output stage
comopnents. The latter 2 connect to presets and a TDA1170 chip. AFAIK,
the circuitry is conventional.
I think I mentioned in the original post that, IIRC, those same chips
are used in Cub monitors and have a habit of cooking themselves. Pretty
sure they tend to result in just a single hoizontal line in the centre
of the screen when they go though, not 'missing' lines as I'm seeing.
Do you have an EHT voltmeter?
Personally, no. I'll hunt around in the museum for one tomorrow (I'm
sure some of the radio room or Colossus guys must have one)
(there is a
PERQ 1 also with a video fault which I think I mentioned;
that's where the confusion probably arises. That one has a totally
mangled display, squashed horizontally and vertically, very non-linear,
and no recognisable character output on it at all)
Do you _know_ if the PERQ is giving the right scan frequencies? (64kHz
horizontal, 60Hz vertical IIRC).
4:48 p.m.
On Fri, 2004-02-20 at 23:18, Tony Duell wrote:
Oh, right...
Can you please re-describe the 2 faults and the machines they occur on --
I was under the impression I was considering a PERQ 1 (VMI monitor) with
a small picture all round...
Well the 2T1 is the one I'll try and sort out first, as that one
certainly appears to have a healthy (or almost-healthy) system unit.
So, the 2T1 has a standard KME as you say - it looks to have correct
horizontal and vertical drive, with the exception that it skips every
few lines, leaving a blank horizontal gap. I can see what look to be
That does _not_ sound like a monitor problem. You're getting missing
video every few scan lines, right? In which case, either you've got some
Analogue signal that's blanking the display _and_ which is locked to the
vertcial sync signal, or you've got a problem with the memory board
(middle board in the cardcage) in the PERQ.
How far does it get through the self-tests? What's the final number on
the DDS display (front of the CPU box) when it tries to boot?
Just to confirm, we're talking about a PERQ 1 here
-- it's confusing to
me wether this is a PERQ 1 (verically-mounted SA4000 hard disk at the
left side, monito connects on a single DC37 connector) or a PERQ 2T1
It's a 2T1, vertical drive mounted at the back. characters on the display, but they're not at all
readable. Difficult to
tell if it's scrambled pixels (suggesting a system fault) or an analogue
problem (suggesting the monitor). The display has an unstable shimmering
right margin to it, which definitely looks analogue in nature.
That may be normal. There's some marginal timing in the vidro circuitry,
and often the right few pixels are a bit unstable :-( At least one of my
PERQs does that...
Onto the PERQ 1. Standard monitor as you say - only that one is
*totally* stuffed. The picture (such as it is) is collapsed into maybe
2" vertically, and about 80% of the screen width horizontally. It's
extremely non-linear; left and right margins weave all over the place. I
can see beam traces all over the shop within the squashed picture too.
It's actually quite a work of art! Not sure if there's more than one
fault going on here, or whether some critical point isn't getting to the
right voltage and that's causing the whole thing to fail.
I will ook at the VMI manuals... This one does sound to have some kind of
monitor problem...
OK, back to the T1. I think it's the machine
I worked on some years ago
(don't worry, everything is 'stock', apart from the fact you may find the
odd IC in a socket if I replaced it).
Ahh, yes it will be. I checked all the connectors and reseated the
boards plus chips - I remember seeing a few socketed ones that didn't
look like they would have originally been socketed.
10:36 p.m.
On Sat, 2004-02-21 at 03:48, Tony Duell wrote:
Ahh, fair call - I was wondering along the lines of bad cap or otherwise
in the vertical drive, such that it was the display itself which was
producing uneven deflection (because I can only see vague character
shapes I can't tell if I'm actually missing data, or if there are blank
lines effectively inserted into the display by the monitor itself)
I am not a display expect by any means! :)
Fair enough :) I love the dual-stacking of chips on the memory board!
cheers
Jules
So, the 2T1
has a standard KME as you say - it looks to have correct
horizontal and vertical drive, with the exception that it skips every
few lines, leaving a blank horizontal gap. I can see what look to be
That does _not_ sound like a monitor problem. You're getting missing
video every few scan lines, right? In which case, either you've got some
Analogue signal that's blanking the display _and_ which is locked to the
vertcial sync signal, or you've got a problem with the memory board
(middle board in the cardcage) in the PERQ. How far does it get through the self-tests? What's
the final number on
the DDS display (front of the CPU box) when it tries to boot?
It gets to 999, which I assumed was normal - but maybe not! Maybe the
character data on the screen is telling me there's a video problem, but
that isn't much use if there's a video problem ;)
The display has an unstable shimmering
Ahh, ok.
right margin to it, which definitely looks
analogue in nature.
That may be normal. There's some marginal timing in the vidro circuitry,
and often the right few pixels are a bit unstable :-( At least one of my
PERQs does that... Ahh, yes it
will be. I checked all the connectors and reseated the
boards plus chips - I remember seeing a few socketed ones that didn't
look like they would have originally been socketed.
You'd be suprised. The only chips I _remember_ socketting was an op-amp
on the hard disk analogue PCB.
21 Feb
21 Feb
11:38 a.m.
On Sat, 2004-02-21 at 03:48, Tony Duell wrote:
Ahh, fair call - I was wondering along the lines of bad cap or otherwise
in the vertical drive, such that it was the display itself which was
producing uneven deflection (because I can only see vague character
I'd love to know what fault could essentially put stair-steps in the
vertical deflection waveform. That waveform is generated in an analogue
manner, by chaeging a cap connected to the TDA1170 chip, BTW...
Anyway, what I'd do at this point is try to get a grey raster from the
monitor. Unplug the video BNC cable, leave the other cables connected (so
the monitor gets the right scan frequencies, etc. Turn up the brightness
(and if necessary, tweak the sub-brightness pot on the monitor PCB -- I
can find out which one that is if it's not obvious). Look at the screen.
You're aiming for an even, grey, rectangle. If you get that, the
deflection sside is fine. If you get black lines across it, then I am
going to have to think very carefully...
So, the
2T1 has a standard KME as you say - it looks to have correct
horizontal and vertical drive, with the exception that it skips every
few lines, leaving a blank horizontal gap. I can see what look to be
That does _not_ sound like a monitor problem. You're getting missing
video every few scan lines, right? In which case, either you've got some
Analogue signal that's blanking the display _and_ which is locked to the
vertcial sync signal, or you've got a problem with the memory board
(middle board in the cardcage) in the PERQ. shapes I can't tell if I'm actually missing
data, or if there are blank
lines effectively inserted into the display by the monitor itself)
I am not a display expect by any means! :)
That's right for POS (which is what I left on the machine when I last
fiddled with it). 255 is right for PNX, 400 for Accent (IIRC). Unless
there's a really strange problem, I think we can assume the PERQ is
booting properly.
OK, that means the PERQ's CPU can access memory, and that the memory has
passed the self-tests. So there's not a RAM problem. But there could be a
video controller problem.
Once we know whether the monitor can give a plain raster, I will dig out
the appropriate schematics (memory board (that's where all the video
stuff is ) or monitor and see what I can deduce
-tony
How far does it get through the self-tests?
What's the final number on
the DDS display (front of the CPU box) when it tries to boot?
It gets to 999, which I assumed was normal - but maybe not! Maybe the 
12:28 p.m.
I'd love to know what fault could essentially put
stair-steps in the
vertical deflection waveform. That waveform is generated in an
analogue manner, by chaeging a cap connected to the TDA1170 chip,
BTW...
Wouldn't coupling some AC of appropriate frequency into it do the
trick? Assuming the AC is phase-locked to the vertical scan, of
course, otherwise the bands wouldn't stay still....
/~\ The ASCII der Mouse
\ / Ribbon Campaign
X Against HTML mouse(a)rodents.montreal.qc.ca
/ \ Email! 7D C8 61 52 5D E7 2D 39 4E F1 31 3E E8 B3 27 4B
12:07 p.m.
Jules mentioned to me by e-mail that the blanking pattern is not stable,
and that there's a lot of ripple on the output of the 48V regulator in
the monitor. These 2 observations must be connected :-) -- it sounds like
excessive power ripple is upsetting the monitor's video circuity.
I would therefore check :
The 2 capactiors -- particularly the 0.1uF one -- on the regulator PCB
(on the back of the heatsink).
The following capacitors on the monitor main PCB :
C11 (47uF)
C10 (10uF)
C2 (1000uF)
C19 (47uF)
C3 (1000uF)
And
C3 (47uF) on the CRT base PCB
-tony
22 Feb
22 Feb
6:28 a.m.
On Sat, 2004-02-21 at 23:07, Tony Duell wrote:
Jules mentioned to me by e-mail that the blanking
pattern is not stable,
and that there's a lot of ripple on the output of the 48V regulator in
the monitor.
bleh - I meant to send that to the list and it went only to Tony. What
happens when posting whilst half asleep :)
These 2 observations must be connected :-) -- it
sounds like
excessive power ripple is upsetting the monitor's video circuity.
I would therefore check :
The 2 capactiors -- particularly the 0.1uF one -- on the regulator PCB
(on the back of the heatsink).
The following capacitors on the monitor main PCB :
C11 (47uF)
C10 (10uF)
C2 (1000uF)
C19 (47uF)
C3 (1000uF)
And
C3 (47uF) on the CRT base PCB
Nice one! Thanks, will take a look when I'm up there next (probably
Saturday, or possibly sooner if the enormous Marconi system we're
expecting delivery of turns up mid-week. Getting that through the door
is going to be a major challenge :-)
cheers
Jules
7438
days inactive
7451
days old
12 comments
4 participants
participants (4)
-
ard＠p850ug1.demon.co.uk -
julesrichardsonuk＠yahoo.co.uk -
mouse＠Rodents.Montreal.QC.CA -
technobug＠comcast.net