TDS460 PSU repair success

[james_s]

I had picked up this TDS460 as well as a TDS420A several years ago, both had bad power supplies. This one had experienced a flashover on the primary side which did considerable damage including killing the strange hybrid these use and vaporizing several traces. The other power supply I got going but all outputs were high and not regulating and I didn't really understand the unusual design these use.

Some time later an eevblog member in Germany kindly offered to give me another defective power supply from one of these so I took them up on the offer. I was busy with other projects so the power supply sat on my shelf for some time until I finally took it out the other day. I reinstalled a few parts that had been removed and replaced some capacitors and then tried it out. It was cycling the over-voltage protection so I temporarily disabled that and found that the -15V output was working correctly but the other outputs were high. Someone pointed me to this http://www.ti.com/lit/ml/slup129/slup129.pdf which explains the strange (to me) magnetic amplifier topology used to regulate the separate outputs. After some hours of troubleshooting I found one of the outputs on the LM324 op amp was only going up to about 1.5V so I replaced that, output now looked good but three of the voltages were still high. I banged my head against the wall with this thing for several more hours, testing all the parts I could think of and finding nothing wrong.

Then it finally occurred to me that the working output is only rated to deliver 0.75A while the other outputs are anywhere from 3.5A to 15A so maybe a minimum load is required. Well I connected an automotive brake light bulb to the 5V output, powered it up and... bingo! The 5V output could now be dialed in exactly! Tried the same with the other outputs and all but one was good. After a bit of troubleshooting I found I had broken a trace going to the adjustment pot for the still misbehaving output while removing parts to test them, fixed that and all outputs are working 100%. I hate to think how much time I wasted troubleshooting when the problem was no load on the output.    It should have been obvious too given how it works, but I could swear I saw Mr Carlson test one of these PSUs on the bench without a load but perhaps I'm just remembering wrong.

Anyway after all this I dug out the scope, spent some time putting it all back together and... It's alive!!! For the first time since I've owned it I was able to power it up, calibration looks good, all four channels work, display looks good, I was amazed. I have better scopes but it was going to bug me too much to give up on this one.

A couple of issues remain though. The acquistion board needs a re-cap, the notorious SMT electrolytics have leaked and made a bit of a mess, I scrubbed it clean some years ago before getting stuck on the PSU and do not see any signs of further leakage but obviously those parts need to be changed. Also SPC fails, I'm going to replace the caps first but if that doesn't fix it I may have to replace all the attenuator relays as I did in my TDS784C.

Does anybody have the latest firmware for the TDS460? Any hacks or options I can unlock? There doesn't seem to be a whole lot of information out there on these older TDS scopes. They feel a bit sluggish and not nearly as nice to use as the newer TDS5/6/700 and TDS3000 series but they're still reasonably capable instruments and 4 channels, TekProbe support and 350MHz is not too shabby.

[Smoky]

James, I rebuilt two of those power supplies, part number 119-3939-00, for my TDS420 (one for backup).

Take a close look at the solder joints of the connector pins. Both of the units I worked on had cracks around the solder joints. Reflow them even if they look OK.

          

Also, it was recommended to me to replace the Bipolar TVS, part number P6KE400CA, if the power supply had seen any stress.

     

Get to those caps sooner than later. I've seen where the leaking electrolyte had shorted a voltage regulator. It burnt a hole in the board!

[madao]

Your Question about Firmware.

I have here a TDS430A Corpse with  TDS420 CPU (it is use also for TDS460) Version:  v2.5e

I am writing on   unified tektool for all  TDS500/600/700.  Almost, now it run, but not 100% relaible (TDS400 series is a bit difficult).
I have checked also him on this  corpse and  it functions. (but i must  replay flash programming)  I must improve code.

Have you NI GPIB-Card/Adapter ?

Greetings
matt

[james_s]

Yeah I resoldered those pins years ago, then I got stuck and set it aside. Now I need to figure out the next one of these PSUs, it was in a similar state, all voltages high but now it is dead. I can see a 32kHz signal on the primary of the drive transformer but there is nothing on the chopper transistor, gotta be something simple.

I don't have a GPIB card, only a USB GPIB interface that I built. I could get one easily enough though, just need to set up one of my old PCs that has an ISA slot.

[shakalnokturn]

Look for high value resistors gone open.

[Dacke]

Sounds like the same experience I had with a TDS420A power supply,  i was scratching my head until i realized that I had to load down the outputs to get the thing functioning.  Initially the main switching transistor was shorted (thankfully the hybrid was fine),  and I replaced it but I was trying to figure out why it shorted in the first place before I tried to start it up.  Turns out it was a small shorted electrolytic that had shit itself and leaked all over the board,  right next to the switcher.   The fluid had also started eating through the traces.  Even though I recapped the entire supply to prevent future issues,  I found out that this one little cap is a really common troublemaker.  After all that was done,  I also ended up disabling the OCP because I simply could not find any issue on the supply and something was causing it to mysteriously shut down.   Everything is fine now and the scope has been solid for almost a year,  but working on that power supply was a strange experience indeed.

[james_s]

Well I looked at the power supply for the TDS410A again and found a 75k power resistor on the primary side that had gone open, not sure how I missed that before. I replaced that and now it works too so I've got a working PSU for both of my TDS400 series scopes. I think I actually repaired this one a couple years ago before breaking it again, I just didn't know it was repaired since I hadn't tried loading it and didn't want to put it in the scope. Anyway mission accomplished there.

I cleaned up the bits of the TDS410A and put that all back together and was astonished to find that it works perfectly! Even SPC passes on that one no problem, only issue left is the fan has bad bearings.

I did some poking around with GPIB and succeeded in enabling options 05 and 2F which are video triggering and advanced math (FFT) however I have not managed to enable 1M despite the label on the back listing it. There are some empty RAM spots on the board but I believe those are for option 2M.

Back to the TDS460 I tried enabling options on that using the same process and had no luck. Does anyone know how to enable those? I found a thread on another forum from almost 10 years ago where someone mentioned they got some disks with option enable scripts but another post in 2014 from someone else asking got no replies. This information has GOT to be out there somewhere, these scopes are not rare.

[Smoky]

James, Capt Bullshot made this comment on another thread relating to the TDS420 and enabling the options. Maybe it works with the 460 too   


"From my own tests, TDS420 uses slightly different addresses. Didn't record each step, so I can't tell the exact values, but by setting each address from 327686 to 327695 to 1, FFT, Memory and Video trig. were enabled. Beware, either 327693 or 327694 enables RS232 / Centronics which leads to a POST message if the HW isn't installed."

[james_s]

Well I wasn't able to get anywhere enabling anything on the 460 so I went back to the 410A. I realized I had a scrap board with some 128k*8 SRAMs so I fired up the hot air and pulled those, then soldered them onto the DSP board from the scope. Now when I turned it on I got a failure error for the DSP, dang! Then I looked and realized that the 120,000 points capture is available now, ok so apparently despite the sticker the scope did not actually have sufficient memory. Either that or when it detects all the memory it enables that without needing anything set.

So now when I look in the error log I get this:
dsp68kD2MemTest. ** addr = 0xb40012 exp data = 0x5555(cut off)

Another time I looked and the address was 0x40010. 0x55555... is reasonable as that's 101010... in binary

It's entirely possible that one of the salvaged SRAMs is bad, however I certainly wish there was a memory map somewhere as I have no idea which chip corresponds to that address. I also haven't figured out a way to scroll over to the right in the error log to read the rest of the line that is cut off. Apparently it can be read via GPIB but that brings up another issue, I can send GPIB commands just fine but any time I try to read I get a timeout error. I don't know whether there's a fault with the GPIB port on the scope or an incompatibility with my DIY interface but the interface works with every other Tek scope I have. I suppose I could try writing to the addresses in question via GPIB and see what chip selects light up. It's just annoying because the scope was working perfectly and now it has the deep memory option which seems to capture fine but the error means something is wrong.

[james_s]

Well I'm not even sure quite what I did but I got GPIB working correctly so now I can read from it and after some experimentation I managed to pull the error log and read it.

:ERRLOG:FIRST "Sat 02-22-20 22:54:08  ERROR: diagnostic test failure, dsp68kD2MemTest, ** addr = 0xb40012  exp data = 0x5555  actual = 0x557f, ** addr = 0xb40014  exp data = 0x5555  actual = 0x55ff, ** addr = 0xb40016  exp data = 0x5555  actual = 0x55ff, ** addr = 0xb40018  exp data = 0x5555  actual = 0x55ff, ** addr = 0xb4001a  exp data = 0x5555  actual = 0x55ff, ** addr = 0xb4001c  exp data = 0x5555  actual = 0x55ff, ** addr = 0xb4001e  exp data = 0x5555  actual = 0x55ff, ** addr = 0xb40020  exp data = 0x5555  actual = 0x55ff";NEXT ""



So that boils down to this:

0xb40012  actual = 0x557f, 0b0101010101111111
0xb40014  actual = 0x55ff,  0b0101010111111111
0xb40016  actual = 0x55ff,  0b0101010111111111
0xb40018  actual = 0x55ff,  0b0101010111111111
0xb4001a  actual = 0x55ff,  0b0101010111111111
0xb4001c  actual = 0x55ff,  0b0101010111111111
0xb4001e  actual = 0x55ff,  0b0101010111111111
0xb40020  actual = 0x55ff   0b0101010111111111

These are each expecting 0x5555 which is 0b0101010101010101

This is looking a lot like a bad RAM or two. I'm starting to think it might be easiest to just buy a set of 6 new SRAM chips. I could just pull off the RAM I added and call it done but this has turned into more of a hacking project than meeting any particular need. Unfortunately the way the DSP board installs in the scope I can't really get in there to probe it while the thing is running.