Tektronix TDS420/TDS460 SMPS Green Hybrid

[Smoky]

On my last TDS420 oscilloscope repair, I was given a parts scope with a bad power supply.

Today I decided to de-solder the green hybrid from it.

It had some discoloration on it from a nearby resistor going up in smoke while the rest of the power supply looks like it went through WWIII with wires clipped and obvious signs of a repair attempt.

Anyway, since I have it out, is there a sure-fire way to determine if the hybrid is bad as it sits on the bench?

I'm expecting a TDS460 parts machine/restoration project on Friday with a good power supply, so it's possible that I can do some comparing while I do some re-capping 

[shakalnokturn]

If it looks good, it can't be bad 
Most of them fail with a burn mark in the middle because the auto-voltage selector got the better of the switching transistor that fused primary current sense resistor that's tied to pin 8 of the green thing.

Do you have a scope with a component tester function or a Huntron tracker at hand? I have some reference curves from a working hybrid.

Edit: The clipped wires are the auto-voltage selector, hardwiring these PSU's to the correct line voltage is a wise thing to do.

[Smoky]

Yes, I have a Huntron Tracker 2000.

Thank you Shakalnokturn!

Btw, the part number on the power supply is 119-3939-00.

[james_s]

Nothing is ever sure fire, but there is a reverse engineered schematic of what is in the green hybrid, it's really nothing exotic. Someone did lay out a replacement PCB for it, I have a board somewhere but never actually needed it because it turned out my hybrids were both good.

[shakalnokturn]

Nothing is ever sure fire, but there is a reverse engineered schematic of what is in the green hybrid, it's really nothing exotic. Someone did lay out a replacement PCB for it, I have a board somewhere but never actually needed it because it turned out my hybrids were both good.

It was team work on the yahoo LecroyOwners group at the time:
I put a lot of time into that project although I had no immediate need for it other than curiosity. De-potting those modules is hell!
A guy called Martin did the schematic capture I don't know if he's on EEVblog forum.
I made a first single-sided layout that Sam ("Scopeman" on EEVblog) reworked into a double sided layout eliminating all the zero ohm resistors.
Sam kindly sent me a few milled PCB's and components, one of the assembled boards is in use in one of my LeCroy scopes.

For Smoky:
http://eninz.free.fr/PSI100004/WORKING_100004_PWM_curves.zip
(May not stay online forever, there should be a copy on groups.io anyway...)

[CaptDon]

The three clipped wires are for the voltage doubler circuit.
I hard wired the doubler for operation on 120vac only!!
If you apply 120vac you will see the required 300+ volts
across the pair of capacitors. Be aware the Lecroy schematic
is close but not exact and it has several value errors as well
as wrong designators. The switch transistor shorted EBC and
took out the current sense resistor (.15 ohm I think) as well
as all the stuff that normally burns out during this failure.
The FR4 material also burned, but I was able to scrape away
the carbonized areas. The unit never tried to oscillate and
produce drive into the primary, but the upside is it also didn't
blow apart on attempted power up, it simply didn't do anything
and therefore my assumption that the green hybrid was killed
during the original power surge. We took one hell of a hit when
the utility pole was struck.

[Smoky]

Alright guys, I was able to compare the signatures that Shakalnokturn provided.

All of the pins of my hybrid appear to match.

One note, I needed to adjust the "range" setting on the Huntron 2000 to make some of the images look identical. There was no mention as to how much voltage/current was injected to create the reference signatures. FYI, the "Med 1" range of the Huntron is 15 VP @ 9.0 mA and the "Med 2" is 20 VP @ 0.6 mA.

Here are a few of the results. Pin 7, 8, and 10:











So by the end of the week, an additional broken PS with unknown problems will arrive. Hopefully, I can get one power supply working out of the two


*Update* The mysterious brown box with a broken 119-3939-00 power supply arrived 

[shakalnokturn]

So the green thing should be ok, never 100% sure until it's running the host equipment though.

Although I didn't evaluate the current scale I did state in the "Notes.txt" in the .zip file that voltage scale is about 2.5V/div.
ISTR that the scope used has a peak test current of approx 10mA which looks consistent with what you get, assuming your pictures were taken on the "Med." Range.

[CaptDon]

I couldn't get the thing to begin oscillating. It had good
input voltage with the doubler bypass mod but drew very
little current from the line input. The SMPS power transistor
I put in was basically a H.O.T. used in T.V. sets which really
was nearly the same as the O.E.M. part which also found
use in T.V. sets. Because the current sense resistor went
open circuit I just assumed the green weenie was damaged
and gave up on it. If the hybrid is actually good you have
a real gem on your hands because they have been on the
unobtainium list for a least a decade!! I was just happy that
when I powered it back up after replacing the flamed out
components it didn't try a repeat performance. The replacement
current sense resistors were a bit of a kludge, but they were all
I had in the drawer that would allow testing. Anyway, glad you
are having fun with all the left over odds and ends!!! I check
the metal scrap dumpster every day for more test gear. They
have two bays of gear with black X's on them and slated for
disposal. I have sent several messages asking if I could have
some of the gear (most of it actually works) for our lab (we
are in a different building) but no replies. Big industry is so
wasteful. The worst waste stream I ever saw was when I worked
at Harley Davidson in York!!! Insanity at its finest!!!
They would fire you for dumpster diving!!!

[Smoky]

OK, I now have two broken power supplies on the bench. One from CaptDon and another off of Ebay. I'm now going to combine the best parts from the two so to have a spare working SMPS on hand.

We know that CaptDon's SMPS was struck with a voltage spike, and I know nothing of the other, so I'm going to replace the three IC's that I see. The main PC board has one LM339N and the small board has one LM339N and one LM324N:





I'm going to pull the second green hybrid too and do a comparative test using my Huntron to see if they both react the same. I'll take pictures and post the results in small thumbnails using the "med 1" range on the Huntron.

CaptDon's SMPS's main PC board has deep burn holes from sizzling resistors yet has a perfect secondary PC board. The Ebay SMPS is missing the large transistor that is bolted to the heat sink near the 680uf bulk capacitors (2SC3688) and its secondary PC board was attacked by leaking electrolyte turning the small pins black.

It looks to me like we'll just squeak through this

[james_s]

When you do test it, make sure you put a load on the output you're trying to test. I burned a whole bunch of hours troubleshooting before I realized the reason both of mine were running way high and wouldn't regulate is that I didn't have an adequate load on them. On the bright side, it forced me to understand how the unusual circuit works.

[Smoky]

Thank you James!

...and as usual, the connector pin's solder joints are cracked at the PC board, and in addition, the pins were bent from bad handling. Here's how I went about fixing it. The cracking of the solder joints is probably partially due to the warping of the PC board over time too.

I force on the ribbon cable connector to the SMPS pins. I then flip the SMPS over and re-flow the solder joints. What this does is it takes the stress off of the pins and the pressure of the connector aligns the pins to the socket when the solder is softened no matter how warped the board is.

The ribbon cable slips on and off easy now

[james_s]

Certainly looks like it worked. All I did was straighten out the pins with pliers and then resolder them.

[Smoky]

I realize now which transistor CaptDon said was taken out by the voltage spike. It's the one bolted to the independent heat sink near the 680uf bulk capacitors. This 2SC3688 transistor is what I found installed:





Since the Ebay SMPS transistor was missing in this spot, I needed to open one of the repaired scopes to see what was the original part number used in this location when the 119-3939-00 was assembled at the factory.

I turned out to be part number MJF16006A.

[shakalnokturn]

I've run the same PSU's you're working on from both BU508A and 2SD1887 transistors in place of the MJF16006A.
The BU508A was slightly better for transition times.

[Smoky]

Thank you Shakalnokturn!

I'm going to buy the BU508A


Here's what else I discovered. No one on the internet sells this Motorola MJF16006A but I did find the MJH16006A.

I did some research and it appears that the only difference is its case size designation.

But here's the best part, I decided to open up all three repaired TDS scopes to check the power supplies. I found a MJF16006A used in one power supply and MJH16006A's used in the other two


*I ordered a few of the BU508AW's from Digi-Key along with some extra interface pads. I checked the data sheet and the thermal resistance value of the BU508AW is the same as the BU508A:

https://www.digikey.com/product-detail/en/BU508AW/497-7196-5-ND/1852124/?itemSeq=330447269


The secondary board is now renewed with fresh IC's and a 330uf 6.3V capacitor:





In this picture, the wrapping material is brittle and it looks to be fiberglass reinforced, and what appears to be thermal paste is bone-dry, is it best to just add more paste and Super-glue the wrapper back around it? All I have is silver duct-tape laying around as an alternative 





*Update* I was able to compare both Green Hybrids. I believe CaptDon's Hybrid is indeed faulty. Here are the signatures of the good Hybrid from the Ebay SMPS. Every pin matched the signatures provided by Shakalnokturn without requiring any voltage adjustments to be made to the Huntron Tracker.

 HuntronHybrid.zip (2152.86 kB - downloaded 119 times.)

[shakalnokturn]

I suppose you're doing a full recap, just in case you aren't don't forget to check the two small electrolytics on the aux. supply near the green module and trimmer. They're usually bad.
I'm not sure that is thermal paste, I would have guessed just glue, I doubt there is much thermal consideration on the varistor. Maybe rather blast-proofing.

On these designs I've got into the habit of adding a 15V TVS diode between hybrid pin 8 and primary ground because of the chain of destruction previously described.
The point (I haven't blown a PSU to check yet) is that the TVS should save the hybrid if the current sense emitter resistor fails.

[Smoky]

Yes, the power supply is going to be fully re-capped.

Shakalnokturn, would you mind elaborating how the TVS diode would be installed

I'm not very knowledgeable about this stuff yet.

Here is a picture of the underside of the SMPS. The Hybrid is removed so the holes are easily visible. I put a red dot next to Pin 8.





And would I be using a 15V Bi-directional or Uni-directional type of TVS diode?

*Btw, the factory installed TVS diode is installed between the two yellow dots.

[shakalnokturn]

I shove them as on picture. (Unidirectional)
The idea is that if the fuse is too slow (anyway it's before the bulk capacitors so there can still be plenty of energy left once it has blown) and the current sense resistor R7 fuses that can leave over 300V on hybrid's pin 8 through R10 - 470 ohm resistor.
Internally pin 8 is a 1k ohm resistor (the available schematic that states 10k needs correcting, you can confirm 1k ohm value with your Huntron between pins 1-8) then a NPN base. Emitter is to ground. Total resistance is low enough to wreck the crap out of the hybrid in this scenario.
Adding the TVS should limit the voltage on pin 8 to something safe allowing for R10 to fuse while leaving the green hybrid green.

The TVS you marked with yellow dots is part of the active clamp.
Consider the one I add a unproven improvement 

[Smoky]

Here is a picture from above where R10 is located. The white cement current sense resistor R7 is just to the left:





To confirm Shakalnokturn's statement regarding Pin 8 as being a 1K Ohm resistor, here is Pin 8 on the Huntron set to the "Med 1" range:





In the Huntron handbook, when the range setting is on "Med 1", a 1K resistor would appear as a 45 degree angle:





*Some of the parts to complete the SMPS arrived today 

I took a gamble with the 680uf bulk capacitors. My last batch of Nichicon LGX 680uf capacitors came in low when measured (~580uf). So this time around, I bought six of the LGX 820uf caps. They measure around 725uf! I'd rather have the caps be 50uf higher than specified than 100uf low:





The usual 2200uf Nichicon capacitors that I use were out of stock too. I went through the Digi-Key catalog and sprung for these Panasonic FR caps. They had the second highest capacity for ripple current, ~3820mA. You're also limited to what size of capacitor can squeeze in between the secondary PC board, which goes back in next:





Saturday delivery 

[shakalnokturn]

For the sake of covering the Power Systems Inc. PSU's subject: One of the more common failures after bad electrolytics is the opto going leaky and causing intermittent failures to start or shutdowns.(Opto isn't used in the regulation loop.)

[CaptDon]

That advice on the opto seems wise. When plugged in and powered on
for the first time the scope would run for 10 minutes and then shut
itself down. Simply pressing the soft touch power button would turn
the scope back on and it would run the rest of the day without failure.

[Smoky]

Chalk-up another win for the EEVBlog Community 

After an hour of idling, the power supply remains stable as can be. The DMM is clipped to Pin 1, the -15V rail:





I installed the 15V TVS diode where Shakalnokturn suggested. I also checked the clearance between the SMPS and its tray and there is plenty of area and height for the diode. No interference whatsoever:








I also replaced the original white Motorola CNY17-2 Optocoupler with a new one made by Vishay:





In the upcoming weeks, I plan on building a variable resistive load to test the outputs for stability and ripple.

If anyone has tips on how to tweak the SMPS's DC voltage rails, please chime-in. I see four variable resistors mounted on the secondary board but I'm not sure if it's just as easy as turning them. If so, we're done!

[shakalnokturn]

Congratulations on the resurrection.

I have a personal note on the setting of the trimmer near the hybrid, measurements taken from a working PSU as a reference for readjustment after replacing the hybrid.

As for the mag-amp regulation trimmers, yes it is as easy as turning them, preferably warm and definitely loaded.
I doubt it has much importance but I'd start with the highest powered output and work down.

[james_s]

All four voltages are independently adjustable, it's pretty cool. It was a pain to figure out how it worked but with some help from other members here I finally wrapped my head around it. I used a 12V automotive brake light bulb as a test load for all four outputs and tweaked them one at a time, the order makes no difference. Once you put it back in the scope it's probably worth re-checking the voltages and tweaking as needed but in my case it was not necessary to mess with it further.