Tektronix TM503 and PG506 attempt repair recondition

[Tantratron]

Hello, there are two topics related to TM503 power module and specific voltage calibrator PG506 section.

In the past, I've been calibrating legacy 2445 and 2465 tektronix oscilloscope where the key sections are PG506, SG503-SG504 and TG501 installed in TM50x power supply set. I was quite happy with this modular approach so acquired through times total of four TM503 (no fan) and many other sections (DC503A, PS50A, DM501A, FG504, AA501...) used in lab testing plus another PG506 from Israel, the first one from Greece. Lately I've acquired TDS540C along with some repairs so was looking on specific testing gear, it seems PG506 is quite useful.

Last week I've decided to run again both PG506 and suddenly they failed each with different symptoms. I'm pretty certain a possible root cause could be one TM503 because these PG506 used to run fine since 2014. The 1st PG506 is severely damaged but I was able to repair partially the 2nd PG506.

After spending lot of time reviewing the service manual of PG506 and TM503, it seems the PG506 uses two transistors found inside the TM503. The schematics of the PG506 shows a complex relay switching with different modes involving many local power supplies plus an earth ground and a floating ground.

My first question about TM503: do you recommend for this legacy units to recondition capacitor or inspect to make sure no strange transient could transmitted to the tektronix section ?

My second question about PG506:after looking the schematics, I've discovered in fact different models through time with PCB variants where I guess tektronix improved or solved known failures. The 1st generation with S/N below B040000, the 2nd generation from S/N above B040000 then the PG506A. Unfortunately I did purchase on eBay B03xxx so probably the one with most failures.

What is your return of experience on 1st generation PG506, are they reliable and worth repair or better look for PG506A ?

My 3rd question about PG506: I found finally one tantalum 16.5Vdc filtering capacitor C170 to be shorted so I've replaced both C170 and C180 from the other unit. However I notice both units have very old capacitors so wonder if it is recommended to recap plus what could be the root cause that the Standard Amplitude output to have lot of noise (see attached picture).

I've probed down to Q255 output, the signal is clean so there must be an internal power supply filter somewhere but there are so many power supplies (+/-16.5V, 5V, 18V, 120V, -72V, 14V...).

Another way to ask or frame my questions:

• is is there any known recurrent specific failures of the PG506's versus the TM503 because most of the sections in my four TM503 always worked fine
• what check or recondition is suggested on old TM503's
  

The key point being if these legendary PG506 are still worth to be used in labs and worth being repaired.

Thank you, Albert

[anotherlin]

Regarding the TM503, have you tried using other know working modules to check each of its 3 slots?
Each slot has 2 pass transistors (PNP and NPN) that are used for voltage regulation purpose.
The TM503 supplies both AC and DC to each slot. The DC is filtered but not regulated.
Each plug-in module is "responsible" to regulate/filter-clean the power rails it uses.

[CaptDon]

Don't use 'other' modules to test the slots!!!!! If one of the TO3 pass
trannys is shorted you will blow up every module you plug in!!!!
Take the TM (Transport Module) apart and check the transistors
for shorts!!! I am almost positive you will find one or more bad.
Also the filter caps are dying these days with tons of ripple on the
D.C. supplies. Some of the TO3 devices may be darlingtons. I don't
think the TM503 has a 'high power' compartment, but the TM504
definately does. Most power supply plug in modules will be limited
to 400ma. if they are not plugger in to a high power compartment.
The raw voltage on the input side of the pass trannys is usually
around 33 to 35 volts and often the regulated output side (determined
by what module you plug in) is around 20 volts. Modules don't
like 33 volts where 20 is supposed to go.

[David Hess]

My first question about TM503: do you recommend for this legacy units to recondition capacitor or inspect to make sure no strange transient could transmitted to the tektronix section ?

There is not much to go wrong in the mainframe.  The transistors would only be damaged if a bad plug-in was used.  The capacitors could be worn out causing excessive ripple on the DC supplies which was the case with my TM501.  I also have a TM504, and two TM503s.

My second question about PG506:after looking the schematics, I've discovered in fact different models through time with PCB variants where I guess tektronix improved or solved known failures. The 1st generation with S/N below B040000, the 2nd generation from S/N above B040000 then the PG506A. Unfortunately I did purchase on eBay B03xxx so probably the one with most failures.

What is your return of experience on 1st generation PG506, are they reliable and worth repair or better look for PG506A ?

The later PG506 has some design improvements but I think both early and late are about equally reliable.

My 3rd question about PG506: I found finally one tantalum 16.5Vdc filtering capacitor C170 to be shorted so I've replaced both C170 and C180 from the other unit. However I notice both units have very old capacitors so wonder if it is recommended to recap plus what could be the root cause that the Standard Amplitude output to have lot of noise (see attached picture).

One of the 16.5 volt solid tantalum capacitors in my PG506 was shorted out and after I changed it, one of the others shorted while I was watching.  So I recommend changing all of the solid tantalum capacitors which were not properly voltage derated in the PG506 whether they are currently bad or not.

[texaspyro]

Bad tants are par for the course in all Tek equipment of that era...  also ba LM741's in the TMxxx module regulator circuits.  Also note that the power transistors in the  TMxxx mainframes have non-standard pinouts.

[Neomys Sapiens]

Don't use 'other' modules to test the slots!!!!!

On TekWiki you should be able to find a document with 'AX-3557' in it's name; this is a description of a build-it-yourself TM500 mainframe/slot tester and utility power supply. It allows for a variation of materials and all you need is an empty TM500 cassette (or 5xxx series oscilloscope module) and a TM500 breadboard PCB.

I don't think the TM503 has a 'high power' compartment, but the TM504 definately does.

This is correct.

[David Hess]

Also note that the power transistors in the  TMxxx mainframes have non-standard pinouts.

I think that just reflects a difference in the standard TO-126 pinout and the standard TO-127 (TO-225) pinout which replaced it.

[CaptDon]

I forget which transport module TM502/503 I had that needed the phase
switched. A FG504 would not run until I swapped the phases (factory
wiring error and there was a recall/update notice sent out). The FG504
puts a pair of phases in series to obtain a higher voltage (around 60
after regulation I think) and if your phases are wrong from the factory
the module puts out some really strange waveforms with 120hz crawling
all over them.

[Tantratron]

On TekWiki you should be able to find a document with 'AX-3557' in it's name; this is a description of a build-it-yourself TM500 mainframe/slot tester and utility power supply. It allows for a variation of materials and all you need is an empty TM500 cassette (or 5xxx series oscilloscope module) and a TM500 breadboard PCB.

Many thanks, I did not know of his specific plug-in which could be nice to build or purchase if available.
While searching on interner for AX-3557, I could not find really details except this page http://w140.com/tekwiki/wiki/067-1201-99

To go back on my four TM500, one of them must have a problem as explained before but not sure to proceed. For the moment, I've tested each TM500's transistors, they all seem OK (no short, around 0.6V between their Base-Collector and Base-Emitter).

If there are indeed different wiring options on these TM503, I might need to check the voltages output individually but without any plug-in, this will take time. What I notice for sure, there seems to be two PCB variants of the TM500 PCB (see attached picture).

I guess the question is about the return of experience or global consensus, do you think or experience these TM503 designed, constructed in 70's and 80's to be still failure free, no need to recondition their basic parts ?

The TM504 and higher is not the topic, their design is different wit active parts to carry more power, the fan which is why I've always preferred the TM503 even though takes more room in my lab.

If we go back to the 2nd topic of the PG503 first generation (S/N below B040000), the main A1 board seems a complex beast in charge of generating different power supplies rails, end amplifier for the BNC outputs, switching. There seems to be a consensus some power supply tantalum cap to short (C170 and C175) but would it be advised to replaced them with modern filtering caps ?

As for one board, both power transistors Q120 and Q85 have failed, do you know any modern compatible transistor ?

After scavenging some parts from one PG506 into the other one, as mentioned before it works functionally except there is a big filtering issue where the 21KHz switching power supply appears with ringings on the output. When I choose 20mV standard output, it adds ringing of the same magnitude (see picture) so if choosing 2mV or 200uV it is a nightmare. Any idea where to focus my effort with minimum cap reconditioning to solve that part ?

[anotherlin]

I guess the question is about the return of experience or global consensus, do you think or experience these TM503 designed,
constructed in 70's and 80's to be still failure free, no need to recondition their basic parts ?

The TM503 is basically passive, transformer for AC voltages, plus rectifier diodes and electrolytics caps for filtered DC.
It doesn't have tantalums which are less reliable. So there's not much for it inside to fail.
And of course, it's built with high quality components.  I've never had any problem.
So I would recommend to change parts only if tested as not working.

As for the different revisions of the TM503, just check the plug-in output voltages and only then have a look at individual components.

You may also want to check:

http://philippe.demerliac.free.fr/Main.htm
It has a lot of Tektronix TM500/TM5000 resources.

https://groups.io/g/TekScopes
Which is a dedicated forum for Tektronix stuff.

[David Hess]

I guess the question is about the return of experience or global consensus, do you think or experience these TM503 designed, constructed in 70's and 80's to be still failure free, no need to recondition their basic parts ?

Only my TM501 needed replacement capacitors.  My two TM503s and one TM504 have never needed any maintenance other than cleaning and replacement hardware.

The TM504 and higher is not the topic, their design is different wit active parts to carry more power, the fan which is why I've always preferred the TM503 even though takes more room in my lab.

The original TM504 has no fan.  There is a later TM504 design (TM504A?) which includes a fan but they are pretty rare.

After scavenging some parts from one PG506 into the other one, as mentioned before it works functionally except there is a big filtering issue where the 21KHz switching power supply appears with ringings on the output. When I choose 20mV standard output, it adds ringing of the same magnitude (see picture) so if choosing 2mV or 200uV it is a nightmare. Any idea where to focus my effort with minimum cap reconditioning to solve that part ?

Amplitude calibration is done at lower bandwidths so much of the noise can be removed with the 20 MHz bandwidth limit function on the oscilloscope.

I do not remember the details, but Tektronix was aware of the noise issue and took steps in the design to minimize it.  I think this included improvements in the late model PG506.

[Tantratron]

Amplitude calibration is done at lower bandwidths so much of the noise can be removed with the 20 MHz bandwidth limit function on the oscilloscope.

I do not remember the details, but Tektronix was aware of the noise issue and took steps in the design to minimize it.  I think this included improvements in the late model PG506.

Please check again my previous post with scope view picture showing the issue when selecting 20mV standard amplitude output of the PG506. The noise I'm seeing here is with fundamental harmonic of 42 KHz so double of the 21 KHz switching of the main inverter. For instance, I took the picture with my tek 2465 set to 20 MHz filter, you can see the repetitive ringing spikes.

I suspect the main issue of this design concerns the A1 board which hosts all different power supplies generation then the precision output resistor in charge to receive the current source to provide the standard output. Only problem there are so many caps, so many power supplies and not easy to probe or to access for removal.

Again if some of you have a similar PG506 with SN below B040000, it would be nice to see the standard output say at 20 mV then 2mV and finally 200 uV to see if it is noise free.

Thank you, Albert

[David Hess]

Again if some of you have a similar PG506 with SN below B040000, it would be nice to see the standard output say at 20 mV then 2mV and finally 200 uV to see if it is noise free.

The output you show looks normal for a PG506 at low output amplitudes to me.

[EHT]

I replaced the electrolytics in my TM503 and would recommend doing the same. Not much to the thing. Most annoying is the latches tend to get damaged on the old modules and can get jammed in the rack!

I have a couple of PG506s which I need to get around to fixing. A problem is I'm missing some parts of the main Allen & Bradley modpot which have been cannibalized...

[Tantratron]

Again if some of you have a similar PG506 with SN below B040000, it would be nice to see the standard output say at 20 mV then 2mV and finally 200 uV to see if it is noise free.

The output you show looks normal for a PG506 at low output amplitudes to me.

I've done now two tests 5mV and 20mV standard output shown at the same time with my good old tek2465 and my recent acquired TDS540C. As you can see, there is the main inverter power supply noise which does not really disappear so it gets worse with lower stand output. Can you imagine if I choose 2mV or 200uV so there must a way to solve this but there are so many old capacitors.

I replaced the electrolytics in my TM503 and would recommend doing the same. Not much to the thing. Most annoying is the latches tend to get damaged on the old modules and can get jammed in the rack!

I have a couple of PG506s which I need to get around to fixing. A problem is I'm missing some parts of the main Allen & Bradley modpot which have been cannibalized...

How many PG506's do you have, what are their generation model, their Serial Number and does some of them work correctly ?

[texaspyro]

I've had a couple of PG506's with problems with the relay contacts.  The relay is sealed in a plastic case that can't be opened.   I drilled a small hole in the plastic using a drill bit in a pin vise  (drill it upside down so the drill shavings don't fall into the case) and sprayed contact cleaner into case.

[Kwakerman]

Hi,  I'm just going through exactly the same fault with a 506.  Luckily I have another that works (although it is a different pcb layout the components are 99.9% the same). 
After a lot of trial and error and side-by-side comparisons I think it is caused by Q80 / Q125 which are the PSU primary side driver transistors.  There is a lot of ringing on the collectors of those transistors and the square wave is a right mess  with an additional short duration pulse in the square wave when Standard Amplitude (SA) is selected (the ringing also affects most other power lines), however it cleans up if fast rise or high voltage is selected.  The signal remained clean with SA selected and P6 disconnected (120V feed to the SA board).  I measured the SA board 120 V supply current draw and it was about 27mA so I used a 4.7K resistor as a dummy load on the 120v supply (to eliminate the SA board) and the ringing returned so that appears to eliminate the SA (I am aware that the SA does provide feedback to the PSU and that would be affected with the SA disconnected).  Next stage was to remove fuse F30 that provides power to the chopper transformer and just let the multivibrator section of the psu run on its 5v supply.  This showed that Q90 and Q100 were providing a clean square wave but Q85/120 and Q80/125 were showing distorted waveforms   After checking everything I could I removed Q80  & Q125 and the signal at the emitters of Q85/120 cleaned up into a nice square wave.  Both Q80 and 125 look ok on a Huntron tracker, the junction voltage drops look ok but one of them has about an 18 Mohm reverse leakage on the base/emitter so hopefully that leakage is causing the issue.  Currently waiting for some D44C9 replacement transistors to arrive and will report back.

[mansaxel]

I replaced the electrolytics in my TM503 and would recommend doing the same. Not much to the thing. Most annoying is the latches tend to get damaged on the old modules and can get jammed in the rack!

On that note, I've got a TM503 with an AA501 distortion analyser in. The AA501 has had its pull tab broken off. What is the best way to get a module out when the tab is broken?

[Kwakerman]

If you remove the bottom cover from the 503 (4 screws) then that will give you access to the underside of the plugin and also to the screw that holds the lower guide rail in place.  That may be sufficient to allow you to get something thin in there to push the plugin's latch to disengage it?

[TERRA Operative]

If you want to make your own TM500 mainframe tester, here's an updated version I designed.

https://drive.google.com/drive/folders/1yj52Z_BtGZ7Q9BB2jS9Lf2Uz39turS4-?usp=sharing




I second the notion of testing the pass transistors before using a newly acquired mainframe. You can just use a multimeter and one of those ebay transistor testers.
A basic voltage check will make sure things are looking ok from the transformer too.

[Zenwizard]

I have moved my SG503-504, PG 506, and TG501 To a 5000 mainframe. Running the modules in the TM503 frame they got a little warm for my liking. This also caused some thermal related issues with the 503. They are much happier in a TM5006 housing. One thing to note about the pass elements in the 500 series frames. Check the serial number then be sure which pass elements you need. The pin outs on the pass elements vary widely including in the same frame number. If I remember correctly the TM503 frame used no less then 5 different pass transistors, and several different pass elements.

For what it is worth I have a SG504 frame that is due for a capacitor refresh it is causing issues with a sg505 and AA501A with low distortion measurements. Current troubleshooting has the finger pointing at noisy power rails.

[Neomys Sapiens]

I replaced the electrolytics in my TM503 and would recommend doing the same. Not much to the thing. Most annoying is the latches tend to get damaged on the old modules and can get jammed in the rack!

On that note, I've got a TM503 with an AA501 distortion analyser in. The AA501 has had its pull tab broken off. What is the best way to get a module out when the tab is broken?

When that happened to me, I used a pin vise with a 0.8mm drill to gently drill a hole into the broken-off surface. then I screwed a small self-tapping screw into this hole. Pulled on the screw - latch released.

[David Hess]

Running the modules in the TM503 frame they got a little warm for my liking. This also caused some thermal related issues with the 503. They are much happier in a TM5006 housing.

I have a pair of TM503 mainframes now to refurbish and have been thinking of adding a small fan to aid cooling.  Even a silent one would make a big difference.

The late TM504A had a fan added but they are very rare and expensive.

[Zenwizard]

The only real issue with The 503 frame is there is not a lot of room back there. But something from the computer world should be able to be adapted. Mabey something low profile on the bottom of the chassis and extend the feet up a bit. Would give direct cooling to the plugins bottom forcing air out the top. Now the wheels are turning.

[TERRA Operative]

On that note, I've got a TM503 with an AA501 distortion analyser in. The AA501 has had its pull tab broken off. What is the best way to get a module out when the tab is broken?

Take the bottom panel off and you'll see a small screw for each rail at the front end of the mainframe.
Remove the relevant screw and you'll be able to remove the plugin along with the rail it's locked to. Then you can just reinstall the rail and bottom panel.