Date   

Re: How to troubleshoot a faulty 7B92A?

Dan G
 

On Sat, Mar 20, 2021 at 04:19 PM, Thierry Delaitre wrote:

Regarding pin 7 and the TRIG'D light, the light is off when the scope gets
switched-on (with AUTO, AC, INT). The TRIG'D light switches on if i change the
mode to NORM. The light stays on if I revert the mode back to INT.
This is definitely abnormal. When switching from AUTO to NORM, the TRIG'D
light should flash briefly, but then go out. This suggests that the sweep timing
capacitors did not charge, or that sweep end detection circuits are not working.
It also explains why the hold-off and auto ramps are stuck low.

As Mark suggested, if you have not already done so, make sure all mechanical
switches, especially S490, operate correctly.

I would also look at pins AV (Aux Gate signal) and AR (Main Gate signal)
to see whether the plug-in thinks it has started a sweep or not. These
are normally square pulses, but yours are probably stuck. Whether they
are stuck high or low may provide a clue as to what is wrong.

Pin 10 remains at a constant DC high voltage and Pin 8 remains at a constant
DC low voltage. both pin 10 and 8 do not show any visible waveform
This now makes sense given your TRIG'D explanation above.
While this is not what these signals should look like, I think it is the result
of problems earlier in the sequence of events. They may be ok once
the plug-in is able to complete the sweep ramp.

I use a scope probe on the same scope to troubleshoot.
Sounds good.


dan


Re: 1S1 on ebay

Sean Turner
 

I received it...seller did a great job packing it. It appears to be functional, certainly moreso than my prototype one. I think it will need some caps done as well, but having a mostly working one will be a massive help troubleshooting!

Sean

On Sun, Mar 14, 2021 at 06:35 PM, Sean Turner wrote:


Guilty as charged. LOL! I need another one so I can hopefully have a reference
to compare things I find in my prototype 1S1 to.

Sean
Hide quoted text ( #quoted-182801618 )

On Sun, Mar 14, 2021 at 04:17 PM, Charles wrote:



So who got the 1S1... there were 8 bids and it went for $104!


Re: PS 5010 troubleshooting (no negative, erratical display)

ditter2
 

Hello Martin,

On the PS5010, my reference to the voltage and current ramping is the programming that takes place in the current and voltage limit DACs on a user request through front panel ON/OFF switch, or through remote control. This action takes place in both ON > OFF and OFF > ON operations. If you are monitoring the DAC outputs with a scope when the Output is toggled, you will see this programming action. It is to protect the output relay. Because the supply will normally go into CC mode during this transition, the software does not monitor the loop status during this time, to prevent reporting of the mode change.

If you need to draw a small amount of current to get regulation, then the minimum load is not sufficient to keep the output regulated. This could be either in the supply itself, or possibly (??) C-E leakage around the pass transistor in the mainframe. That would explain different amounts of the effect in different mainframe slots.

For the PS5004 – I designed all of it. For some fun, get a schematic and quickly try to find the precision 16 bit DAC needed to support the output resolution. If you are looking for a large “brick”, you won’t find it. 16 bit DACs were available at the time, but cost about 2.5 times the proposed manufacturing cost target for the entire built and tested instrument. The DAC I designed is a gated charge pump, using simple digital counter logic to set the duty factor of the current gate clock. BTW, Sony independently came up with the same idea for the 16 bit DACs needed in CD players about the same time.

I believe the part of the cal procedure you are referring to asks you to set output V to max, then back down one count (If my memory is correct – that was designed 40 years ago.) What you are doing is setting the full scale of the fine span of the DAC. Rather than take the full 16 bit resolution in one span, there are two charge pumps, scaled – I believe 200 counts (Each is 8 bits, the fine is 1/200th the coarse. They don’t map 1:1 with the digital pot coarse-fine range). Since calibration requires use of the digital knobs to set ranges, and the display is a volt meter measuring the measured output (not the programmed value as in the PS5010), the only way to set known values with the knob is at the extremes. So “0” is easy to set – turn both knobs to the left several turns and you can calibrate out the offsets in the system. Turn either or both knobs to the right several turns and you are at full scale – it is now possible to set the full span gain of the coarse current pump. Turning only the fine control down one “click” reduces the coarse stage DAC count by 1, and sets the fine DAC to is max scale. Now the user can calibrate the full scale of the fine . The order might not be the same as what I wrote in the cal procedure 40 years ago, but the process is the same.
Steve


Re: 7A16A high frequency compensation.

unclebanjoman
 

Hi Albert,
thanks!
Yeah, it's that aberration "up" in the 10 ns point I can not eliminate. I didn't think it was the cable. I generally observe the "dribble up" phenomenon. Do you think is the cable ?
I also have a 7A16 (not A) but I don't think it has that reflection. Maybe because of the smaller bandwidth?

Anyway, now the response of my 7A16A is much better than before. Initially was horrible. I think the previous owner tried to calibrate it but then gave up.
Really a tour de force: more than 4 hours fiddling the pots!

What do you think about the trim pot not shown in the schematics? May it be the very latest revision? I did not find more recent schematics.

Max


Re: 7A16A high frequency compensation.

Albert Otten
 

Hi Max,

How much nicer do you want it? :-).
I wonder whether the slight aberration you see after 10 ns is due to reflection in a 5 ns cable.

Albert


Re: How to troubleshoot a faulty 7B92A?

Mark Vincent
 

Thierry,

Q410, Q412, etc are 0367. Check all that have this on them from C to E. Likely leaky.

Dan is right about U820 possibly being defective. I had one of these bad before. It is possible dirty contacts, if this one is in a socket.

If you are not getting any sweep on standard or delayed, I have seen Q812 open B to C. This transistor will make U820 look bad if the transistor is bad. A KSP10BU will replace this and the 0367 types. I have had that happen at least once. The Cherry switch on the right side of the plug-in at the front not physically mounted right. That will be open or grounded to P911. See if it is closed and open by a multimeter when you pull and push the time base selector. This is done with no power applied. The screws that mount to the readout board (A12) are oblong and allow movement of the switch. Make sure you have the pin that activates the microswitch there. You will see it with the right side panel removed. R479 and R939 should be 1W. These are stressed and could be high to open. I have heard of someone saying U856 can go bad. This is a 7474 flip-flop

Pulling the time base knob engages the delayed sweep. I do not know if you did or did not test this. See if you get sweep this way. The small intensity knob to the left of the time/div knob. That could be too low. Turn that CW to see of you get sweep. When working right, it will get very bright at full CW. The contrast to the right of this control may need adjusting. The trace sep. knob my be off.

If your S/N is below 60000, it is possible the tunnel diodes are bad on one or both triggering boards.

It is possible the internal dc bal. and/or trg. sens. pots were turned. They may need a slight adjustment. It may bring the triggering back.

Cleaning the switches might help if you have already done so.

I hope this helps.

Mark


Re: How to troubleshoot a faulty 7B92A?

Thierry Delaitre
 

thank you.

The 7B92A is still in the vertical slot.

Regarding pin 7 and the TRIG'D light, the light is off when the scope gets switched-on (with AUTO, AC, INT). The TRIG'D light switches on if i change the mode to NORM. The light stays on if I revert the mode back to INT.

Pin 10 remains at a constant DC high voltage and Pin 8 remains at a constant DC low voltage. both pin 10 and 8 do not show any visible waveform

I use a scope probe on the same scope to troubleshoot.

Thanks

Thierry


Re: 7A16A high frequency compensation.

unclebanjoman
 

Hi Albert,
thank you for the response.
After a further two hours, and comparing with the more detailed explanations given in the 7A16P manual, I was finally able to calibrate the step response in a decent manner.
I could not appreciate any variations by turning the last trim, C770. When I got the plugin the first time, it was almost completely unscrewed. I screwed in some turns to prevent the screw from falling out!

I posted some screenshots showing the final result.
Photos are in https://groups.io/g/TekScopes/album?id=262075
Album title:"A16A step response (with 7854)

What do you think about? All comments are welcome.

Furthermore, I noticed that my PCB is very recent: it shows a S/N 670-2323-08 (see photo). There's a trimmer that does not appear in the schematic.
I've marked in the photo with a red arrow. Green arrow points to the LF compensation trimmer R436 which is present in the schematic. The other one is marked LF (on the PCB) but is not present in the schematic. I presume was an additional trim point for the same purpose of the green one, so I adjusted it in the same manner as R436.

Max


4 photos uploaded #photo-notice

TekScopes@groups.io Notification <noreply@...>
 

The following photos have been uploaded to the 7A16A step response (with 7854) album of the TekScopes@groups.io group.

By: unclebanjoman <mmazza@...>


Re: OT: MC12080 prescaler chip weirdness

Ed Breya
 

Like Tom said, don't drive the control signals with direct logic. Some prescaler types have control interface that's directly compatible with TTL/CMOS, but I don't think the 12080 is. It appears to be pure ECL, so you never pull any inputs solidly down to Vee - they can go to Vcc for high, or left open for low, presuming they have built-in pulldown Rs. If the modulus is to be fixed, just hard wire according to the instructions. For logic-controlled modulus, you can pull high at low impedance, but only pull low via a high resistance, like tens of k ohms. Built in type pulldowns in ECL are typically 50-100 k. For the 12080, it looks like it would be simple to use PNP Qs to pull high to Vcc, and just put them in cutoff to let the ECL inputs pull low.

Ed


Re: 7L5 dot frequency reset bug - any clues or a fix?

Dan G
 

On Fri, Mar 19, 2021 at 12:52 PM, Richard Steedman wrote:

Dan, did you ever get round to trying your proposed fix of increasing the
C841/R841 time constant?
Hi Richard,

Good timing, and thank you for reminding me! I have not had occasion to
open the 7L5 since my last post on this topic, but, as it happens, I am just
now waiting for replacement parts for my L3 module to arrive from Europe.
I expect them to arrive in the next week or two. I will need to do a partial
re-calibration after the replacement, so it will be a perfect opportunity to
give the C841/R841 modification a try.

I will post my results here.

I can confirm that the 7934 mainframe also has the
issue. See two photos [here](https://groups.io/g/TekScopes/album?id=262040)
of the >1s delay after power-on before TS1 pulses appear. I suspect the issue
affects any mainframe using a readout system newer than the original analog
ROM board (e.g. microprocessor for the 7854 and EPROM-based board for 79xx).
I suppose these days one could also whip up a fix using a PIC to drive TS1
pulses into the serial clock input shortly after power-on and then get out the
way to allow option circuitry to take over.
Thank you for posting those photos. They provide a valuable data point, as I
had always assumed that this problem is specific to the 7854. By the way, it
is nice to see a 7623A in action. I have recently replaced a failed CRT in mine,
and it is now a real pleasure to use.


dan


Re: Changing out GR874 connectors +Printact relays

Stephen
 

On Sat, Mar 20, 2021 at 04:48 AM, Charles wrote:


I do too...
Yes, I think they look pretty cool.

Besides that, I have a working S-1 and S-2 sampling head with 874 connectors.
It looks like a delicate operation to put an N connector on the front of
either...
Indeed. At the end of the day, I decided I don’t want to have to mess around with them and leave them alone. I neither need the hassle, nor do ai want to have to deal with the possibility of ruining something for good.


Re: OT: MC12080 prescaler chip weirdness

Ed Breya
 

Is this a deal where you're just adding the 12080 to an existing board in the counter, or are you building the whole prescaler from scratch? Also, what frequencies and input levels are you using to test it with? The 12080 should work just fine if you apply it according to the datasheet, and have the right RF/HF environment around it. RF prescalers like this one must be AC coupled, and have a minimum cutoff frequency depending on the coupling cap size, and need sufficient amplitude and edge speeds to get proper operation. If the signal input isn't right, you can get strange results, and it can self-oscillate when the signal is not present or too small. Odd things can happen if you overdrive it too. Also, don't overlook the bypass caps - besides good supply bypassing, one is needed on the unused differential input.

Ed


Re: OT: MC12080 prescaler chip weirdness

Tom Lee
 

That is a very fussy chip. The most common error is to drive the control signals with TTL level signals. That won't work reliably. They have to be very close to the rails. I don't remember the specs any longer, but IIRC you have to be within a diode drop of the positive supply, and the signals have to be clean. And, of course, the input signal has to be clean as well.

HTH.

Tom

--
Prof. Thomas H. Lee
Allen Ctr., Rm. 205
350 Jane Stanford Way
Stanford University
Stanford, CA 94305-4070
http://www-smirc.stanford.edu

On 3/20/2021 10:46, Michael Dunn wrote:
A Fluke counter I’ve been fixing up and adding homebuilt options to is missing the prescaler option. The vast majority of current chips seem to be binary, so, short of also dividing the reference clock, I decided to use a decade divider, an NOS MC12080 ÷10/20/40/80 chip.

Funny thing is, instead of dividing by 10, it’s dividing by 2! The chip internals consist of a fixed ÷5, followed by a programmable ÷2/4/8/16 stage. It’s like the ÷5 isn’t there, or is bypassed.

After some searching, I found a few other people bemoaning this exact same behaviour. So…at least I’m not crazy. Anyone here experienced something similar?

Gotta wonder what’s going on. Is there some sort of flaw in the chip design that leads to this failure mode? Was it a custom version? Hmm…

md








Re: 7A16A high frequency compensation.

Albert Otten
 

Hi Max,

My 7A16 (not A) manual shows a picture. The 7A16A performance is the same or better I guess.
I did the same setup with PG506, 7854 and "as is" 7A16A, Nice smooth step with rise time 1.55 ns, followed by just one overshoot 6.5% and one undershoot 3.5%. After this negligible aberrations from flat. With a 7A19 rise time was 1.1 ns.

Albert

On Sat, Mar 20, 2021 at 01:01 AM, unclebanjoman wrote:

Moreover, in the manual there is not even a photograph of some typical
waveform that can be used as a reference. Does anyone know where to find a
screenshot of a typical 7A16A step response?


Re: How to troubleshoot a faulty 7B92A?

Dan G
 

On Sat, Mar 20, 2021 at 07:16 AM, Thierry Delaitre wrote:

U820 - Sweep control IC
- Pin 6 (auto delay) is 0.5V ; strange ?? other side of the R828 resistor
(180K) is 14.7V
- Pin 7 (trig D) is 0.5V
U820 contains comparator and discharge circuits on pins 6 and 8.

U820 allows C828 to start charging when it stops receiving trigger signals.
That is how it eventually reaches free-run in AUTO mode.

Your pin 7 being low suggests that your TRIG'D light is on. Is that correct?
It would explain why pin 6 is low, and why you never reach free-run mode.

If your 7B92A is still in the vertical slot, set for AUTO, AC, INT, then there
should not be any trigger signals, and the fact that U820 is in triggered mode
is suspect.

- Pin 8 (hold off timing) is 0V ; strange? according to the schematic there is
a pull-up resistor to the 15V line
Pin 6 and pin 8 cannot be reliably measured with a DMM, as they both
normally have positive-going ramps, and should never get much higher
than +5V. You will need to look at them with a 10X high-Z probe.
(Remember to use an IC probe tip adapter on your probe, or else
be extremely careful not to let the probe tip slip and short adjacent
IC pins. I find that it is usually safer to probe at some other location:
R828 for pin 6, R871 or R872 for pin 10, R834 for pin 8.)

U820 allows hold-off capacitors to start charging at the end of the
sweep, so if pin 8 is stuck at 0V DC, then U820 may think the sweep is
still in progress, or it may have never received a HOLDOFF START pulse
on pin 16.

- Pin 10 (hold off) shows a 4V DC voltage (no sweep)
- Pin 17 (trig disable) is 0V
- Pin 18 (lockout) is low
- Pin 19 (auto mode) is low as it should for auto triggered
Are pins 10 and 8 both constant DC, or is there a waveform visible with
a 10X probe?


dan


OT: MC12080 prescaler chip weirdness

Michael Dunn
 

A Fluke counter I’ve been fixing up and adding homebuilt options to is missing the prescaler option. The vast majority of current chips seem to be binary, so, short of also dividing the reference clock, I decided to use a decade divider, an NOS MC12080 ÷10/20/40/80 chip.

Funny thing is, instead of dividing by 10, it’s dividing by 2! The chip internals consist of a fixed ÷5, followed by a programmable ÷2/4/8/16 stage. It’s like the ÷5 isn’t there, or is bypassed.

After some searching, I found a few other people bemoaning this exact same behaviour. So…at least I’m not crazy. Anyone here experienced something similar?

Gotta wonder what’s going on. Is there some sort of flaw in the chip design that leads to this failure mode? Was it a custom version? Hmm…

md


Re: TDS544A with strange display #photo-notice

Siggi
 

On Fri, Mar 19, 2021 at 6:41 PM durechenew via groups.io <durechenew=
yahoo.com@groups.io> wrote:

Some news, but not what you may expect.
As there were some doubts (and I agree, for good reasons), I opened the
display area again; soldered two wires to shutter connections - totally
random, I have no idea for which colors and I believe it's not relevant.
Shutter connector in normal place, put everything back and started the
scope. Here is the moment to say, for Siggi, that drivers for the shutter
are indeed powerful, initially designed to drive (servo) motors at 20kHz,
max 0.5A DC.
OK, see pictures 5 to 8 in the album.
Synchro signal on Ch1, frame, as I said before, around 17 ms (by signal I
mean waveform).
Picture 5: on Ch2 (should be green but, at low light, camera doesn't make
much of a difference between colors) is signal from output of driver for
one of the shutter's colors.
Picture 6: same signals, different timing.
Picture 7: same as 6 but signal is displayed on Ch4, not Ch2. Some
vertical vectors are obviously of different color (naked eye sees clear
red). If Ch4 light is OFF (pressing Ch1, for example) signal is all blue
(no red vectors).
Picture 8: Same as Picture 6, different shutter channel ( different driver
output/color).
Conclusions:
1. Shutter works fine; period of signal, roughly 370uS or so (just to
have an idea how often the shutter refreshes the colors - definitely not
line frequency). As I don't have the pair signal for the same color I
cannot say if they are really opposite; it may be that some time they are
not (and that color is not ON), depends how the control signals come from
the memory.
The shutter plot thickens.
For the record, I agree with you that the artifact you see is not coming
from the shutter. Nonetheless, I've always been curious to understand how
it works, and sadly I don't think your data is conclusive.
Grousing through the LCD driver board yet again, I see the following:

- The LCS_CLK input to the shutter board is the horizontal line clock.
It's derived from the ~HSYNC signal from the CPU/Display board.
- The LCS_CLK signal then goes to a 74HC4040, a 12 bit ripple counter.
- The first output of the ripple counter used is Q2, which is then
LCS_CLK/4.
However, both edges of this signal are used to latch the output from the
EPROM, alternately into U505 and U510, so the fastest possible differential
output change from those two would be LCS_CLK/2, or every other line.
- The distribution of these signals to the output drivers is curious.
The top and bottom output drivers take signals from one or the other
octal latches, but the middle one takes alternate signals.
- The counter is reset on an ~FSYNC/VIDEO signal that comes from U2011
on the CPU/Display board (A11<24>, 5C),

... time passes ...

Looks like we're back to color-per-frame, rather than color-per-line.
Presumably the ~FSYNC signal is asserted every 3 or 4 frames, to reset the
LCD shutter counter to start again.
Is the line frequency of the CRT then really that high? It wouldn't serve
any purpose that I can see?

Presumably the complicated shutter driving machinery is about generating 0V
DC for each shutter on average. I don't understand why it needs to be
driving at line frequency/2, as it looks like it's really just switching
color shutters between frames. I also don't understand why it needs 12
lines - maybe each shutter is driven by 2 rows by 2 cols?

2. Some waveform information for Ch4 is incorrectly displayed when a
different color (driver output for red) is ON (and on condition that light
of Ch4 is also ON). This back the same conclusion it was suggested before:
Between the point(s) of split from VGA the signal path toward Video for CRT
is corrupted. Still to be discovered where and how...
Yups, I agree. On my TDS784D, the selected trace is intensified, supporting
the hypothesis that an intensity bit is one of the five bits going to the
TLC34075.
Happy hunting!


Re: How to troubleshoot a faulty 7B92A?

Thierry Delaitre
 

have also checked at the collector of Q412 and there is no sweep

T.


Re: Changing out GR874 connectors +Printact relays

 

I do too... and some years ago I bought a large bagful of GR adapters and cables. So I'll stick with them ;)
Besides that, I have a working S-1 and S-2 sampling head with 874 connectors. It looks like a delicate operation to put an N connector on the front of either...

6021 - 6040 of 186402