Test Equipment Anonymous (TEA) group therapy thread

[Vince]

Nixie DMM continuation :

OK so I replaced the two Yellow temporarily with something ballpark.

I went for  820 ohms for the lower branch, the 4V ref. That gave me 3.85V, almost there ! 
Will put a pot instead to let me  figure out what resistor I need exactly. 1k probably...

1k again, when I replaced the other yellow one, that feeds the comparator. Again based on what makes sense according to the schematic.

DMM didn't care about my efforts on this Ref board, symptoms still the same.

I scoped the comparator ref output on the working DMM, so I can compare it with the faulty DMM.
See below.  We get a waveform of about 400mV peak to peak. There is a positive "plateau" , the DC Voltage output by the ref I guess, followed by a negative "pulse" of similar amplitude.
The overall period of this "signal" is about 400ms, which seems correlated to the refresh rate of the display on the front panel.. which would make perfect sense.

Then I scope the faulty DMM. Get the exact same waveform and amplitude, except the period os longer, like 750ms... but that also seem sot match the refresh rate which is indeed slower than that of the working DMM. Now I don't think the Vref is involved in the determination of the refresh rate, doesn't make sense... it must be rather be the ramp generator, which at te very least, inherently set the absolute limit as to how fast it can sample the input signal.

Will study the schematic a bit more in the comparator board department....

Anyway. Now that I have fixed the power supply and got the Vref board in a "working" condition, short of being calibrated... and since this has not led to any progress what so ever... I guess now is time to start swapping boards !!! 

Tomorrow probably... 00H30 here, getting tired...

[TaylorD93]

Wow, that's pure genius, how to make do with what you have, fine tune your own 300mohms resistors, well done !! 

 

This time i just got some bigger 500W resistors on a giant heatsink from another lab... I guess if you operate anything bad enough it will let the smoke out. I think i could give it a good go and these bigger ones will still be alright haha

Silly thing is ive just repaired the instrument, its not a formal cal, just a rough yes its close enough so that it can be sent to our ISO Calibration Lab. They wont repair the instrument as its +20 years old, so i said id sort it. in the last 8 months 6 of the 8 have broken down, most often due to dried Electrolytic caps.

Not bad getting to do your hobby at work

[Vince]

Nixie DMM update

PROGRESS has just been made !!!

There is now measurization going on here  !!!

Could not help, could not go to sleep, I was too curious and impatient... I started swapping boards with the working DMM.

Remember I said it looked like the first / LSD digit was counting, but the second one next to it, looked like it was counting only between '0' and '1' and no more ?

Well that was it indeed ! I swapped that board and now all 4 digits are alive and counting !!! 

So I need to fix that board...

There is still a long way to go though, but I am happy that basic operation has now been restored, it basically "works".

So from what I see, remaining problems are :

I fed it with 2.5V from my Chinese Vref.

- On the 5V range (High impedance input, > 1,0000 Meg) = 2,65V, close enough for now...
- On the 5V range ( 5 Meg impedance) = 2,65V also.
- On the 50V range = 4.36V ohhh..... not good.
- On the 500V range = 22.6V !!! 

So since the reading is "good" (enough) on the 5V range which is the straight path to the A/D converter, that means the entire chain of measurement is basically fine.
It only reads non-sense on the 50V and 500V ranges, which means the problem must be limited to, and located, on the attenuator board. So I will need to look at that...

Also, new problem : the polarity Nixie quit working ! It doesn't light up at all anymore, there is no MINUS or PLUS sign anymore ! 

Plus of course I need to make a new Vref board of course. that won't happen overnight of course, so In the meantime I could swap board with the working DMM to let me move on with my troubleshooting.


So to sum it up for now :

1) A failed cap in the power supply, need to tidy up my repair now.
2) A Vref board to redesign/make
3) Failed polarity Nixie to troubleshoot
4) LSD+1 Nixie board to troubleshoot
5) Attenuator board to troubleshoot.
6) Power switch that's stuck open.. sealed unit it looks like, so not too hopeful
7) Cosmetic restoration / corrosion / pitting to attend to.

So still far from done, but at least now it basically "works", the rest has a clear path to it so not worried...

2H30 AM, I am about to lose consciousness so going to bed, see ya later....     

[mnementh]

Good sleuthing, V. 

You might wanna check the boards you swapped around; maybe there's something different on one of them to drive the +/- indicators. You know, when you're conscious again!

Now go get some  ya maniac! 

mnem
*toddles off to ded*

[Zucca]

 2022 RPN summary after a night of research.

Swissmicros

DM42:    fastest and most precise in the market
DM41X:  not as fast, but immense programs catalogue available

Community driven FW
Those guys are nuts and heros... what they achieved is remarkable.

W34S: two line display on HP-20b and HP-30b (hack) hardware, just tons² of functions and precision equal to DM42!

....and here the bomb I did not know...

So Swissmicros DM42 HW is/will be compatible with the WP34S successor, the ultimate RPN calculator all time....

WP43S

In fact Swissmicros is planning to produce the WP43S, a DM42 HW with WP43S FW and key overlay....

here some pics...





holy shit...

[psykok]

definitely need one

[med6753]

When engineers have wet dreams. 


Glad I'm not an engineer

[Robert763]

You need to slow down a bit, take a bit more time & be very careful next time. 
Could you have disconnected the wires instead? i.e. the yellow one at the top right (in picture below) and the red to the pot. Then measured the each resistor without disturbing the fragile parts with green spot corrosion.

Sorry mam ! 
Thing is, from the get go I expected to replace these things just based on how crusty they looked, and the green stuff leaking from the terminals... so I was not even trying to rescue these yellow buggers ! 
To me it was like looking at an old cap or battery...if you find one that leaks and is so crusty, it does not even cross your mind to even try to reuse them 

Now obviously rewriting history is easy and had I known what I would find inside these yellow jackets.. yes maybe I would have tried to open them up much more carefully..... next time ?     I will take that as a learning experience... maybe one day I will get another piece of TE that might make use of this experience...

P.S. the capacitor & diodes that supply the +200V for the tubes looks to be a voltage doubler circuit.

Wow, I don't only get moral support on TEA... I also get ACTUAL technical support as well !  Thanks for taking the time look at this thing, much appreciated !  :-+0

Good spot !    I had been trying to make sense of this part of the PSU... and of the fact that I measured only 100V or so at the input, half the expected voltage ! Makes perfect sense now !  Boy it didn't even cross my mind that they could be doubling the voltage...never seen it done before in such circumstances. Transformer is always custom made, so the secondaries are just wired to output whatever it is that you need usually ! 
No idea why they didn't make that winding produce 200V to begin with ! Save on copper ?!... at the detriment of adding more components and cost ?! 
I will keep that at the back of my mind next time I look at an antique power supply... I won't get fooled twice !

Another problem now I've had more time to look at the circuit I found last night, is that is differs a bit from what you have, also the reference feeds two ways, one via the "4000" switch to the attenuator and the other to the comparator board. Really could do with the correct diagram for this variant.

Yes it's been bugging me all day as well... I got carried away when I said I would make a replacement board... I can't do that until I figure out what this second output does, how it interacts with where it's going... 

My board has 2 extra resistors...
So I cracked the other metrer opened, since it is the same model that the manual is based on.
It's Vref board, see below, is 100% identical to the schematic. The resistors on it look a bit more "modern". The two "special" / yellow resistors have been replaced by small plastic cylindrical ones. They are clearly marked. Their value is exactly what the schematic says. The 1,050 tempco is also printed on them.
It's.. perfect.

So I put the two meters/boards side by side and compared them carefully.
I eventually figured the older board out.

To help with understanding, I will call "lower branch" the vertical path that start at the Zener diode anode, and goes downward. That's the branch that produces the 4V ref and goes to the attenuator. Then there is the "upper branch", starting rom the Zener anode also, and that buggers off to the comparator board. That's the mysterious / tricky branch to me 

So. basically, the two boards are 100% identical in topology, wiring, same interface to the outside world, they are interchangeable for all I can see.
All that changes, is the value f the resistors, slightly, and also how what they are made out of (single resistor or a // or series combination).

Both the upper and lower branches are affected :

- Upper branch:  schematic shows 6,7k total resistance, about 50/50 split between that 1% 3.7k resistor, and the 1050 tempco 3k one.
In the older board, the total resistance is still the same, but the split is different : 5,7k for the 1% resistance, and only 1k for the yellow one. Well I think at one point I did manage to measure it and it was about 1k IIRC... and since it would add up here, so let's say it was indeed 1k ! 

- Lower branch : The fixed  1% 3,9k at the very bottom, has been replaced by what you spotted, a // combination of 56k and 3,75k, which measures at 3,5k.
So still ballpark the same as the newer board.
Then the other resistance, the 1050 tempco one, has been replaced with a combination of two R in series : the yellow one in series with a 1,33k resistor. Which means I guess that yellow one which I could never measure, must have been in the order of 1,000 ohms or something (like the other one then), if we want everything to add up...


So.... I feel a bit less confused and scared than I did this morning.. things are clearer now. The mods are minor. They are basically the same boards, interchangeable, same pinout same everything really... just minors mods to how the resistors are "implemented".

What I think :

1) it's clear the lower branch is what generates the -4V ref, so if I make my own Vref circuitry, I need to connect it to this output.

2) The upper branch I still have no idea what it does in this comparator board, nor why it too needs a 1050 tempco resistor. However, what I think at the least, is that the Zenr diode provides a low output impedance 6,2V source to BOTH the upper and lower branches. Hence, to me, the upper branch is not concerned with what's going on in the lower branch, so I could separate the two : disconnect the lower branch and replace it with my own Vref, and leave the upper branch as it is, so it can keep doing whatever it's doing...

3) The lower branch, the ref, I can just replace the dead yellow resistor with a 1k resistor, or trimmer if I get fancy, to restore basic operation. Of course out the window with the 1050 special R, but for now, it should be enough to make the board operate again. It's quick and easy and free to do, so I think that's what I am going to do right now, it's only 10PM I still have a bit of juice and time left before I go snoring.

Hi Vince,
I would NOT replace the zener with "modern" reference. The well aged zenner is likely much more stable than most modern references. There are a number of high prformance meters that use this type of zener reference.
Keeping a constant current through the zenner is more important.  The current in the existing design is about 7mA
A LM317L and 180R (178R if you want to go to E48) resistor in place of the existing 820R resistor would do the trick. Replace the "special" resistors with standard metal film ones.
The 317L circuit is: output pin and one end of 180R to the -12V suply, other end of 180R to Adjust pin, input pin to the zener anode.

There is a significant flaw in the design of the "test" circuit of this meter. You can't use a voltage derived from the meter reference to test the meter accuracy. The voltages will be ratiometric. So if the reference goes down the so will the test voltage. This makes the test meaningless in terms of accuracy checking. 

[Ice-Tea]

Nah, cleaning can wait a bit longer.

[TERRA Operative]

Vectron OCXO mod for my Tek SG503, all working on the breadboard and PCB's ordered today.
Original crystal was flakey and intermittent, so I figured why not replace it with a bit of an upgrade?   

[Saskia]

@zucca

Don't you ever sleep ?

I need to check which HP calculators are currently residing here and if I can reflash them.
Unfortunately the 10b2+ cannot be reflashed with scientific firmware if I read the announcement correctly. Got mine for 10€ or so, it's still sitting here. Don't have time to fart around with it right now, anyway.

[mnementh]

   

Hi Vince,
I would NOT replace the zener with "modern" reference. The well aged zenner is likely much more stable than most modern references. There are a number of high prformance meters that use this type of zener reference.
Keeping a constant current through the zenner is more important.  The current in the existing design is about 7mA
A LM317L and 180R (178R if you want to go to E48) resistor in place of the existing 820R resistor would do the trick. Replace the "special" resistors with standard metal film ones.
The 317L circuit is: output pin and one end of 180R to the -12V suply, other end of 180R to Adjust pin, input pin to the zener anode.

There is a significant flaw in the design of the "test" circuit of this meter. You can't use a voltage derived from the meter reference to test the meter accuracy. The voltages will be ratiometric. So if the reference goes down the so will the test voltage. This makes the test meaningless in terms of accuracy checking.

Robert -

You evidently missed it in one of V's later posts... there is a completely isolated, dedicated Vref just for the test voltage. It just happens to be on the same PCB.

The one he was talking about replacing was this second Vref; I believe he fully intends to repair the circuit in the main Vref, as he has not damaged the temp compensating resistor for that ckt.

mnem
 

[Vince]

   

Hi Vince,
I would NOT replace the zener with "modern" reference. The well aged zenner is likely much more stable than most modern references. There are a number of high prformance meters that use this type of zener reference.
Keeping a constant current through the zenner is more important.  The current in the existing design is about 7mA
A LM317L and 180R (178R if you want to go to E48) resistor in place of the existing 820R resistor would do the trick. Replace the "special" resistors with standard metal film ones.
The 317L circuit is: output pin and one end of 180R to the -12V suply, other end of 180R to Adjust pin, input pin to the zener anode.

There is a significant flaw in the design of the "test" circuit of this meter. You can't use a voltage derived from the meter reference to test the meter accuracy. The voltages will be ratiometric. So if the reference goes down the so will the test voltage. This makes the test meaningless in terms of accuracy checking.

Robert -

You evidently missed it in one of V's later posts... there is a completely isolated, dedicated Vref just for the test voltage. It just happens to be on the same PCB.

The one he was talking about replacing was this second Vref; I believe he fully intends to repair the circuit in the main Vref, as he has not damaged the temp compensating resistor for that ckt.

mnem
 

Yes and no...

The second yellow resistor is fucked too, I can't seem to measure anything meanful from it any more... I though I had measured it at 1k early on, but lzast night was 6k which doesn't make sense.  Plius since it's corroded inside, it's bound to get worse and worse over time, so I would have wanted to get rid of it anyway...

[Zucca]

@zucca

Don't you ever sleep ?

When I was little my mom was responsible to turn the light off, now at night it is just me and my shit in my lab (wife and kid are in sleep mode).
Quoting one of my favorite writer, I am again the biggest problem in my life....

[Vince]

Hi Vince,
I would NOT replace the zener with "modern" reference. The well aged zenner is likely much more stable than most modern references. There are a number of high performance meters that use this type of zener reference.
Keeping a constant current through the zenner is more important.  The current in the existing design is about 7mA
A LM317L and 180R (178R if you want to go to E48) resistor in place of the existing 820R resistor would do the trick. Replace the "special" resistors with standard metal film ones.
The 317L circuit is: output pin and one end of 180R to the -12V suply, other end of 180R to Adjust pin, input pin to the zener anode.

OK OK... you are making my life easier and cheaper here, so I vote for that ! 

At the end of the day it's a mere 0.2% 2,500 counts meter, it's not like it needed rocket science to work properly ?! 

OK so let's move on to the next issues now...

[nixiefreqq]

2022 RPN summary after a night of research.

you spent a night on rpn research?

DUDE!   surely you could find a more productive way kill time?  (there are lots of stripper bars down near fayetteville.  i personally recommend "bottom's up" on bragg boulevard.)


edit  full disclosure.  have not been in that dive since 1982.  that was back when downtown was still called "fayettenam".  it may not exist today.

[tggzzz]

Hi Vince,
I would NOT replace the zener with "modern" reference. The well aged zenner is likely much more stable than most modern references. There are a number of high performance meters that use this type of zener reference.
Keeping a constant current through the zenner is more important.  The current in the existing design is about 7mA
A LM317L and 180R (178R if you want to go to E48) resistor in place of the existing 820R resistor would do the trick. Replace the "special" resistors with standard metal film ones.
The 317L circuit is: output pin and one end of 180R to the -12V suply, other end of 180R to Adjust pin, input pin to the zener anode.

OK OK... you are making my life easier and cheaper here, so I vote for that ! 

At the end of the day it's a mere 0.2% 2,500 counts meter, it's not like it needed rocket science to work properly ?! 

OK so let's move on to the next issues now...

Certainly not

Mouser sells 1000ppm/C thick film resistors. SMD, so you'd have to take account of heat dissipation.

If you do stick in a modern voltage ref, they can drift a lot in the first few weeks. IIRC I saw 1mV in 5V in one example (5.0005 -> 5.0015V)

[mnementh]

      

Hi Vince,
I would NOT replace the zener with "modern" reference. The well aged zenner is likely much more stable than most modern references. There are a number of high prformance meters that use this type of zener reference.
Keeping a constant current through the zenner is more important.  The current in the existing design is about 7mA
A LM317L and 180R (178R if you want to go to E48) resistor in place of the existing 820R resistor would do the trick. Replace the "special" resistors with standard metal film ones.
The 317L circuit is: output pin and one end of 180R to the -12V supply, other end of 180R to Adjust pin, input pin to the zener anode.

There is a significant flaw in the design of the "test" circuit of this meter. You can't use a voltage derived from the meter reference to test the meter accuracy. The voltages will be ratiometric. So if the reference goes down the so will the test voltage. This makes the test meaningless in terms of accuracy checking.

Robert -

You evidently missed it in one of V's later posts... there is a completely isolated, dedicated Vref just for the test voltage. It just happens to be on the same PCB.

The one he was talking about replacing was this second Vref; I believe he fully intends to repair the circuit in the main Vref, as he has not damaged the temp compensating resistor for that ckt.

mnem
 

Yes and no...

The second yellow resistor is fucked too, I can't seem to measure anything meanful from it any more... I though I had measured it at 1k early on, but lzast night was 6k which doesn't make sense.  Plius since it's corroded inside, it's bound to get worse and worse over time, so I would have wanted to get rid of it anyway...

No, no, no... you need it. As you surmised, that resistor is part of the temperature comp for the Vref. Take it apart carefully and repair it.

I'm 99% positive that "green stuff" on the legs and bobbin caps is just corrosion from environmental exposure; once you clean the bobbin and re-tin the legs, I have no doubt you can find the broken wire end and make it 100% serviceable again. Then, wrap the bobbin in Kapton tape and re-coat in epoxy. That way, if you ever have to fix it again, you can still dismantle it carefully.

As picky as you are about the cosmetic details, I have no doubt you can "restore" this special-purpose resistor to its former glory... if you wanna get really anal about it, print up a little label on your prehistoric laser printer and put it under the last layer of Kapton tape before you epoxy it all up.

Cheers!

mnem
 

[McBryce]

Anyone looking to add a "screenshot" option to their Tek 400 Series? https://www.ebay.de/itm/115332266476

McBryce.

[med6753]

Anyone looking to add a "screenshot" option to their Tek 400 Series? https://www.ebay.de/itm/115332266476

McBryce.

And surprisingly you can still get the B/W film for the camera. No, not interested.

[Vince]

Hi Vince,
I would NOT replace the zener with "modern" reference. The well aged zenner is likely much more stable than most modern references. There are a number of high performance meters that use this type of zener reference.
Keeping a constant current through the zenner is more important.  The current in the existing design is about 7mA
A LM317L and 180R (178R if you want to go to E48) resistor in place of the existing 820R resistor would do the trick. Replace the "special" resistors with standard metal film ones.
The 317L circuit is: output pin and one end of 180R to the -12V suply, other end of 180R to Adjust pin, input pin to the zener anode.

OK OK... you are making my life easier and cheaper here, so I vote for that ! 

At the end of the day it's a mere 0.2% 2,500 counts meter, it's not like it needed rocket science to work properly ?! 

OK so let's move on to the next issues now...

Certainly not

Mouser sells 1000ppm/C thick film resistors. SMD, so you'd have to take account of heat dissipation.

Wow... incredible ! I will look into that then, see I can find them locally and with the right value.

If you do stick in a modern voltage ref, they can drift a lot in the first few weeks. IIRC I saw 1mV in 5V in one example (5.0005 -> 5.0015V)

Change of plan, Robert prefers to keep the old Zener and its quantum patina, and instead use a current source to stabilze the Zener, then I don't need 1050 tempco resistors anymore, just regular precision resistors with low tempco.

[Vince]

No, no, no... you need it. As you surmised, that resistor is part of the temperature comp for the Vref. Take it apart carefully and repair it.

I'm 99% positive that "green stuff" on the legs and bobbin caps is just corrosion from environmental exposure; once you clean the bobbin and re-tin the legs, I have no doubt you can find the broken wire end and make it 100% serviceable again. Then, wrap the bobbin in Kapton tape and re-coat in epoxy. That way, if you ever have to fix it again, you can still dismantle it carefully.

As picky as you are about the cosmetic details, I have no doubt you can "restore" this special-purpose resistor to its former glory... if you wanna get really anal about it, print up a little label on your prehistoric laser printer and put it under the last layer of Kapton tape before you epoxy it all up.

Cheers!

mnem
 

No no no no... you have the wrong impression. I am not anal about cosmetic, or else I would not be able to leave with this meter or any old gear.. they inevitably have dings and dents and imperfections all over the place, wear marks... it's just a matter of what I can live with, and what can technically be improved for a reasonable effort and reasonable budget both time and money. It's all a compromise, there is no anality here, just pragmatism...

In the case at hand, mutilating that poor yellow resistor and make an ugly hack to try to close it back up, is way worse to me (and lots of work) than making a clean replacement solution, whatever said solution might be.

If you think otherwise I can understand, but then it's YOU being anal, not me any more ! 

[mnementh]

Hi Vince,
I would NOT replace the zener with "modern" reference. The well aged zenner is likely much more stable than most modern references. There are a number of high performance meters that use this type of zener reference.
Keeping a constant current through the zenner is more important.  The current in the existing design is about 7mA
A LM317L and 180R (178R if you want to go to E48) resistor in place of the existing 820R resistor would do the trick. Replace the "special" resistors with standard metal film ones.
The 317L circuit is: output pin and one end of 180R to the -12V suply, other end of 180R to Adjust pin, input pin to the zener anode.

OK OK... you are making my life easier and cheaper here, so I vote for that ! 

At the end of the day it's a mere 0.2% 2,500 counts meter, it's not like it needed rocket science to work properly ?! 

OK so let's move on to the next issues now...

Certainly not

Mouser sells 1000ppm/C thick film resistors. SMD, so you'd have to take account of heat dissipation.

If you do stick in a modern voltage ref, they can drift a lot in the first few weeks. IIRC I saw 1mV in 5V in one example (5.0005 -> 5.0015V)

I think as srb1954 suggested, it is not a matter of better or worse tempco... but that the circuit and this resistor are designed with a specific tempco response in mind. I know that 1000ppm/C is close to the spec in that hand-written schematic, but do we know if the original has a flat slope? Probably not even close.

As we don't have the original designer or correct documentation to draw from, I'd at least try to fix the existing wire-bobbin resistors. That will be easy enough to get quite close to original spec with just a little care. Myself, I wouldn't even be afraid to try fixing the one Vince cut, unless there's oodles of cut strands we can't see.

That path also has the benefit of not requiring Vince to search all over Hell and Creation in a greatly limited market to find some suitable substitute.

mnem
 

[mnementh]

No, no, no... you need it. As you surmised, that resistor is part of the temperature comp for the Vref. Take it apart carefully and repair it.

I'm 99% positive that "green stuff" on the legs and bobbin caps is just corrosion from environmental exposure; once you clean the bobbin and re-tin the legs, I have no doubt you can find the broken wire end and make it 100% serviceable again. Then, wrap the bobbin in Kapton tape and re-coat in epoxy. That way, if you ever have to fix it again, you can still dismantle it carefully.

As picky as you are about the cosmetic details, I have no doubt you can "restore" this special-purpose resistor to its former glory... if you wanna get really anal about it, print up a little label on your prehistoric laser printer and put it under the last layer of Kapton tape before you epoxy it all up.

Cheers!

mnem
 

No no no no... you have the wrong impression. I am not anal about cosmetic, or else I would not be able to leave with this meter or any old gear.. they inevitably have dings and dents and imperfections all over the place, wear marks... it's just a matter of what I can live with, and what can technically be improved for a reasonable effort and reasonable budget both time and money. It's all a compromise, there is no anality here, just pragmatism...

In the case at hand, mutilating that poor yellow resistor and make an ugly hack to try to close it back up, is way worse to me (and lots of work) than making a clean replacement solution, whatever said solution might be.

If you think otherwise I can understand, but then it's YOU being anal, not me any more ! 

I'm anal about function, not appearance. I've seen you be anal about both aspects of a project at different times. 

Take that as a compliment; in these circles, that's what it is.

Like I said... I'd at least try and fix it. I think you're smart enough to outwit a coil of wire. 

mnem
*toddles off to do some leather surgery*

[mnementh]

...So Swissmicros DM42 HW is/will be compatible with the WP34S successor, the ultimate RPN calculator all time....   WP43S

In fact Swissmicros is planning to produce the WP43S, a DM42 HW with WP43S FW and key overlay....      holy shit...

I think I hear the sound of somebody wailing on his piggy bank with a comically oversized wooden mallet...

mnem
There, there, Z... just think of it as an excuse to put off buying a DM42 for a little while longer...