Test Equipment Anonymous (TEA) group therapy thread

[Vince]

ROCHAR DMM REPAIR

OK been studying the thing some more, trying to trace grounds and all the wires coming from the centronics interface connector.
A good flash light and DMM in buzz mode, is essential. Patience too.

So here are my findings :

There are 3 different grounds :

#1 : chassis ground. It's not being used by the plugin at all. The binding post on the front panel is connected straight to a pin on the interface connector, via a black and green wire.

#2 : ground on the diff pair board : goes to both the Volts and Amps boards, as well as their respective black/negative terminal on the front panel.

#3 : found two different black ground wires on the interface connector.... which would help confirm my theory of an intermittent ground connection to the Amp board while the ground on the Volts board is fine and reliable. UNFORTUNATELY this hope soon vanished when I buzzed these two wires and found they were connected    So even if one would go intermittent, the other one should be able to take over. PLUS... when I traced the actual wires, I found that actually thes two ground wires go to the Volts board as well as the range switch, but they don't go to the Amps board.

So it's still a mystery... 

Then I tried to trace every single wire of the interface connector. There aren't that many so I though what the hell, let's give it a try.

- x2 black wires : ground for Volts board and range switch.

- yellow-green = goes straight to the Volts board. I think this must be the supply voltage because the red wire of the hour counter is soldered to the same pin on the connector...

- white -blue = again goes straight to the Volts board. Maybe another supply voltage, negative or god knows what. I guess I could measure that...

- a row of x7 solid colour wires, rainbow style : from orange to white. These all go straight to the range switch. They are most likely used to tell the main frame which neon bulb to light up for the decimal point.
There are not 7 DP though, so maybe some of these wires do something else. This should be detailed in the main frame manual I would think, can't wait to see that.


So not much going on on this interface connector really ! ground, V+, decimal points, that's about it.
Signal measure goes to the diff pair / 4mm connectors, not the centronics connector.


Hey, we are making progress aren't we ?! 

After some more wiggling of the plug-in, I noticed that the display goes nuts only with some forms of wiggling.... I need to be more accurate : I noticed that I can "trigger" the display if I apply a twisting action to it, i.e try to rotate it. However if I apply purely a linear motion, either going from left to right, or top to bottom, then it is wayyyyyyyyyyyy less sensitive. For what it's worth.


So I am learning stuff, but sadly it's going nowhere : I am at the same time finding evidence that are incompatible : stressing the interface connector by wiggling the plug(in definitely triggers the problem... but the wiring of that connector is such that even intermittent contact should not cause my symptoms.

Honestly  this thing is starting to drive me completely nuts, I am lacking ideas now...

I guess now that I have identified that very few pins of that interface connector are actually doing something useful... I could try to bridge them see if that helps. Somehow I bet it won't and I will be back to square one.  I think I will soon be in need of professional psychiatric help in order to retain  at least some bits sanity....

[TERRA Operative]

Like these from Teledyne.

I just replaced some of those in my SG5030 Levelled Sine Wave Generator last week.

[Vince]

Oh, you mean these things are still made these days ?! 

[TERRA Operative]

Yeah, don't bother checking the new price... I found some at RS Components, for around $200 a pop! 

I ended up getting a set of 5 NOS from Ebay for waaayyyyy less. I only needed 4 but a spare is nice to have.

[m k]

For Labor I ordered T727 something, from China also.
Not exactly sure how it fits since there seems to be different inserts.

Local and old service shop

Visiting revealed that wrong city.

It's model was mentioned to be T727.
Molded model is PA-16CX and actual connection jack seems to be bigger than in China model.

[Vince]

@Vince, that blue multi-contact connector looks like a standard Centronix/GPIB/etc connector. Could you make up an extender cable (or just unbolt the connector on one side and pull it out a bit) to make clearance so you can plug it in with the diff-pair banana jacks not connected to the main unit, and then measure the diff pair from the plug-in module directly with a multimeter?
That way you can narrow down if the gremlins are in the plug-in module (you'll see the weird readings on your multimeter) or in the mainframe somewhere.

At least those plugs should be easy to source. They are still pretty common.

Just tried it.... digged an old // port printer cable... bingo, same number of pins, 36, it fits in the main frame !
Now just need a female scket for the other end of the cable.... just looked, my local electronics shop has them for only 1.50 Euro ! 

https://www.e44.com/connectique/connecteurs/chassis/Centronics-36/Femelle/centronic-36-points-femelle-chassis-CA36C.html

Hmmm... for that price, OK it's worth trying to make a cable... not much to lose...

[RolandK]

...
Now the question is: Selling the AVM Fritzbox again or keeping it as a spare one? From experience Fritzboxes never die.

HW gets old, cheaper and outdated...

Murphy's law: if you keep you don't need it.
if you sell it, you get the first early fail ever - promissed this never happend before.

You can sell, so we'll see if you have a monday's model.

[TERRA Operative]

@Vince, that blue multi-contact connector looks like a standard Centronix/GPIB/etc connector. Could you make up an extender cable (or just unbolt the connector on one side and pull it out a bit) to make clearance so you can plug it in with the diff-pair banana jacks not connected to the main unit, and then measure the diff pair from the plug-in module directly with a multimeter?
That way you can narrow down if the gremlins are in the plug-in module (you'll see the weird readings on your multimeter) or in the mainframe somewhere.

At least those plugs should be easy to source. They are still pretty common.

Just tried it.... digged an old // port printer cable... bingo, same number of pins, 36, it fits in the main frame !
Now just need a female scket for the other end of the cable.... just looked, my local electronics shop has them for only 1.50 Euro ! 

https://www.e44.com/connectique/connecteurs/chassis/Centronics-36/Femelle/centronic-36-points-femelle-chassis-CA36C.html

Hmmm... for that price, OK it's worth trying to make a cable... not much to lose...

Told ya they were common!



As for me, I got an old almost worthless Trio (Kenwood) 35MHz CS-1577 scope on the bench, and my HP8160A 50MHz pulse gen boatanchor on the bench now. The amount of time spent on this thing is more of a stubborn refusal to quit on my part than any sense of value or return on investment. It won't beat me!

The Trio works perfectly in X-Y mode but no trace in any other mode, I suspect the horizontal sweep is to blame.

The 8160A has a faulty board which was narrowed down in about 30 seconds using the service manual.
Now I just need to make up some riser cards to let me poke at it in the usual way with old HP gear.

[tautech]

The Trio works perfectly in X-Y mode but no trace in any other mode, I suspect the horizontal sweep is to blame.

Pretty darn easy to diagnose with just a DMM.
Adjust to a really slow timebase and on any horizontal plate terminal you should see the voltage rise as the sweep completes.
Just make sure you set it up to default settings for any of the service manual DC voltages to be correct.
Might be something as simple as a dicky XY switch.

[Messtechniker]

Now the question is: Selling the AVM Fritzbox again or keeping it as a spare one? From experience Fritzboxes never die.

I use older wlan routers - fritz! boxes in particuar - as additional
access points around the house (and garden). Often with a timer
on their power supply. Saves money on hardware and power costs.

[Robert763]

<SNIP>


As for me, I got an old almost worthless Trio (Kenwood) 35MHz CS-1577 scope on the bench, and my HP8160A 50MHz pulse gen boatanchor on the bench now. The amount of time spent on this thing is more of a stubborn refusal to quit on my part than any sense of value or return on investment. It won't beat me!

The Trio works perfectly in X-Y mode but no trace in any other mode, I suspect the horizontal sweep is to blame.

The 8160A has a faulty board which was narrowed down in about 30 seconds using the service manual.
Now I just need to make up some riser cards to let me poke at it in the usual way with old HP gear.

Sounds like a trigger issue...

[TERRA Operative]

I thought it might be, but there's no free-running trace in Y-T mode.

[factory]

ROCHAR DMM REPAIR

OK been studying the thing some more, trying to trace grounds and all the wires coming from the centronics interface connector.
A good flash light and DMM in buzz mode, is essential. Patience too.

...snip...

- a row of x7 solid colour wires, rainbow style : from orange to white. These all go straight to the range switch. They are most likely used to tell the main frame which neon bulb to light up for the decimal point.
There are not 7 DP though, so maybe some of these wires do something else. This should be detailed in the main frame manual I would think, can't wait to see that.


So not much going on on this interface connector really ! ground, V+, decimal points, that's about it.
Signal measure goes to the diff pair / 4mm connectors, not the centronics connector.

...snip...

As well as the decimal points being selected by the plug-in, there are also the symbols for DC, AC, ohms, of course only certain ones will be used depending on the plug-in fitted, guessing the plug-in enables the appropriate symbol, the mainframe will select between +/- if DC is enabled.

David

[Neomys Sapiens]

What some people call a 'test lead kit'. Hmpfff.
Well,together with the 1754 kit already here it WILL be worth calling it one.
 

[Neomys Sapiens]

ROCHAR DMM REPAIR

I guess now that I have identified that very few pins of that interface connector are actually doing something useful... I could try to bridge them see if that helps. Somehow I bet it won't and I will be back to square one.  I think I will soon be in need of professional psychiatric help in order to retain  at least some bits sanity....

Hi Vince! The symptoms described (error by wriggling/nudging..) and some of the pictures gave me an idea: might there be a chassis connection on one of the PCBs in the signal path going bad (toothed washer missing, too much corrosion..)?

[Vince]

ROCHAR DMM REPAIR

OK been studying the thing some more, trying to trace grounds and all the wires coming from the centronics interface connector.
A good flash light and DMM in buzz mode, is essential. Patience too.

...snip...

- a row of x7 solid colour wires, rainbow style : from orange to white. These all go straight to the range switch. They are most likely used to tell the main frame which neon bulb to light up for the decimal point.
There are not 7 DP though, so maybe some of these wires do something else. This should be detailed in the main frame manual I would think, can't wait to see that.


So not much going on on this interface connector really ! ground, V+, decimal points, that's about it.
Signal measure goes to the diff pair / 4mm connectors, not the centronics connector.

...snip...

As well as the decimal points being selected by the plug-in, there are also the symbols for DC, AC, ohms, of course only certain ones will be used depending on the plug-in fitted, guessing the plug-in enables the appropriate symbol, the mainframe will select between +/- if DC is enabled.

David

No symbols nor AC on this plugin so can't be that. However I had a more plausible idea : these 7 rainbow wires can't be all for DP... there are only 4  DP on the display.. so that's 3 extra wires.

Then I realized hey... we do need those extra wires ! Look at the range switch... there are 4 positions used for calibration purposes, no less. 2 are related to the main fraime, to adjust Zero and 20,000 points.
So we need at least one wire for each feature. The mainfraime can't read in a crystal ball, unless the plugin tells it to go into calibation mode, mainframe can't guess. By default it will display the voltage from the diff pair. So out of these 3 extra wires, I guess one is there to tell the mainframe to adjust its zero, and the other tells is to display it's internal Vref, which can then adjust with the adequate trimmers located below the display.

I don't have any idea for the third and last wire though.... but we will soon find out once I get the manual 

[Vince]

ROCHAR DMM REPAIR

I guess now that I have identified that very few pins of that interface connector are actually doing something useful... I could try to bridge them see if that helps. Somehow I bet it won't and I will be back to square one.  I think I will soon be in need of professional psychiatric help in order to retain  at least some bits sanity....

Hi Vince! The symptoms described (error by wriggling/nudging..) and some of the pictures gave me an idea: might there be a chassis connection on one of the PCBs in the signal path going bad (toothed washer missing, too much corrosion..)?

It screams bad ground indeed I agree, but all my investigations so far say otherwise, so I am bit baffled.

There is no chassis ground used by the plug in. There is just the ground provided by the centronics connector and that must be good because the voltmeter part of the plugin works fine.
Then there is the ground used for the diff pair carrying the signal, but I already bridged these 3 banana terminals and that didn't change a thing.

It screams bad ground I agree, but I can't find any evidence of that anywhere I look 

Once I have made an extension cable I will be able to do more torture tests and hopefully that will give me new clues to elaborate a new theory....

At least one can say it's a bad connection problem, and that it's limited to the amp part of the plugin. It's already a win compared to previously where I was totally clueless.... need to narrow it down further now...

[Neomys Sapiens]

At least one can say it's a bad connection problem, and that it's limited to the amp part of the plugin. It's already a win compared to previously where I was totally clueless.... need to narrow it down further now...

And I wish you success wholeheartedly on this endeavour! While I'm not generally known as a Nixie aficionado or a collector of instruments from this period, the build quality of the thing is amazing!
BTW, should you need a TO-can relay, you can ask. I'm rather well stocked in this department.

[Neomys Sapiens]

There is no OBVIOUS chassis ground used by the plug in. There is just the ground provided by the centronics connector and that must be good because the voltmeter part of the plugin works fine.
Then there is the ground used for the diff pair carrying the signal, but I already bridge these 3 banana terminals and that didn't change a thing.

That, combined with your earlier observation of the error resembling a charging effect, led me onto this path of thought.
So while it might not be easily visible, maybe there IS a contact to one of the attachment points of a PCB. Something where a common-mode cap might cause such a phenomenon...
But I'm sure that you are giving it a thorough look, just wanted to add this thought.

[Vince]

Thanks Neomys ! 

I am not into any TE in particular...I have no love affair for tube stuff or Nixie stuff as a religion.. I like TE of all periods.
It's a case by case thing... I bought the Rochar TE I got because they happened to "turn me on".. but I am not going to buy any Rochat TE just because it's Rochar... There is Nixie Rochar TE I find ugly so.. .I won't buy it. Same for Tek scopes, I have many Tek scopes but it's just because there are many Tek scopes that I happen to like... but there is plenty of Tek scopes I also don't like and wil never buy, be it hollow-state or digital or analog...

I have zero dogma....

However as you say, one thing I like in general about old stuff, at least for decent brands, is the engineering (both electrical and chassis/mechanical) and build quality, quality of the materials, their durability.

The modern stuff is invariably bult cheap no matter many thousands of Euros it costs. That's why the cheap chinese manufacturers can grow I think.... the look and build quality of their scopes is no better or worse than the cheap stuff. It's all pretty much even. Cheap plastic box with overwhelming user interface full of quirks and bugs waiting for you to discover them.

[Vince]

There is no OBVIOUS chassis ground used by the plug in. There is just the ground provided by the centronics connector and that must be good because the voltmeter part of the plugin works fine.
Then there is the ground used for the diff pair carrying the signal, but I already bridge these 3 banana terminals and that didn't change a thing.

That, combined with your earlier observation of the error resembling a charging effect, led me onto this path of thought.
So while it might not be easily visible, maybe there IS a contact to one of the attachment points of a PCB. Something where a common-mode cap might cause such a phenomenon...
But I'm sure that you are giving it a thorough look, just wanted to add this thought.

I have some experience in fixing stuff now, so I try to never say never... because I got bitten once or twice.... I would say " it CANNOT "be that... then after wasting lots of time exploring other avenues, I eventually find the problem..... ah yes, finally it WAS that indeed ! 

That's why I am taking baby steps in my troubleshooting here, I don't want to jump to conclusions only to realise 6 months later that I was wrong from the start...

Anyway, need to order a centronics connector and receive the manual. That will take a few days, 10 maybe, especially in this holiday period... so in the meantime I will work on other things in the lab...plenty to choose from...

[factory]

ROCHAR DMM REPAIR

OK been studying the thing some more, trying to trace grounds and all the wires coming from the centronics interface connector.
A good flash light and DMM in buzz mode, is essential. Patience too.

...snip...

- a row of x7 solid colour wires, rainbow style : from orange to white. These all go straight to the range switch. They are most likely used to tell the main frame which neon bulb to light up for the decimal point.
There are not 7 DP though, so maybe some of these wires do something else. This should be detailed in the main frame manual I would think, can't wait to see that.


So not much going on on this interface connector really ! ground, V+, decimal points, that's about it.
Signal measure goes to the diff pair / 4mm connectors, not the centronics connector.

...snip...

As well as the decimal points being selected by the plug-in, there are also the symbols for DC, AC, ohms, of course only certain ones will be used depending on the plug-in fitted, guessing the plug-in enables the appropriate symbol, the mainframe will select between +/- if DC is enabled.

David

No symbols nor AC on this plugin so can't be that. However I had a more plausible idea : these 7 rainbow wires can't be all for DP... there are only 4  DP on the display.. so that's 3 extra wires.

Then I realized hey... we do need those extra wires ! Look at the range switch... there are 4 positions used for calibration purposes, no less. 2 are related to the main fraime, to adjust Zero and 20,000 points.
So we need at least one wire for each feature. The mainfraime can't read in a crystal ball, unless the plugin tells it to go into calibation mode, mainframe can't guess. By default it will display the voltage from the diff pair. So out of these 3 extra wires, I guess one is there to tell the mainframe to adjust its zero, and the other tells is to display it's internal Vref, which can then adjust with the adequate trimmers located below the display.

I don't have any idea for the third and last wire though.... but we will soon find out once I get the manual 

I thought you were trying to identify all the wires, I know I mentioned three that would be on the mainframe side of the connector, but one should still be part of the plug-in for the DC voltmeter/ammeter plug-in.
And yes you must have a wire to select which symbol type the plug-in requires, how else would the mainframe know to select between the DC (+/-), AC, or Ohms symbols?

This won't help with the intermittent fault of course.

David

[Vince]

OK OK.... nope I was not trying to identify all the wires on the mainframe side of the connector, of course... there are 36 of them, it's 100 populated !   

I was only doing it on the plugin side of course, where there are only a few wires.

Good point about the plugin having to tell the main frame what unit symbol to use, eh eh....  OK. 

You win, I lose !

Suddenly these 3 extra wires don't seem enough now, we should have more of them  ! 

Maybe the DP, one of 4, is coded on 2 "bits", to save wires..... can't wait to see all the details in the manual...

[TERRA Operative]

New few acquisitions are on the way to meeee!... (Photos from the auctions)

I finally got myself an AC volt/current calibrator.  This one is a Yokogawa 2558, good for 1000V@6mA and [email protected] up to 500Hz (or seemingly 800Hz with an external frequency reference). Now I'll be able to calibrate all my multimeters!

Also, I got a Trilithic DB-50 BNC DC block, good for 10MHz to 1GHz at 500V

And finally, two HP 8141A power sensors. I already have the power meter and cable, so fingers crossed at least one of them works..........


Merry Christmas to meeeeee..... 

[factory]

Interesting, it can be together used with the newer 2558A, both are good for 1200.0V, two can be used together for calibrating power meters.
Also has a sweep mode, not seen that on the old HP calibrators.
https://tmi.yokogawa.com/eu/solutions/products/generators-sources/standard/2558a-ac-voltage-current-standard/

Hope it works better than the HP 745A, still not got round to fixing it, doesn't do 1kV-ish without the 746A, which I don't have.

David