Repair of Return LeCroy/Preamble 1855A differential amplifier

[rx8pilot]

Yesterday, I received a LeCroy DA1855A differential pre amplifier from eBay. Advertised as working and has 14 day return option. It looks great in pictures, was packed very well, and physically presents very nice. Not a single scuff mark anywhere on it. But it does not power up. 

I contacted the seller and he agreed to let me take a look inside for easy and obvious issues. I took the cover off and was greeted with the most convoluted mechanical arrangement I have ever seen.  I am nearly positive the mechanical design was done by one of the electrical engineers that has never once had to assemble or service anything. The unit upon application of power - simply has a low level chirp and small pulses of voltage at about 1hz. First thing to check is the power supply, however that requires (apparently) disassembling all of the hand built and very delicate looking pre-amplifier section. I looked at it for an hour trying to guess how it went together in the first place. There is a lot of point to point hand construction on the front end - looks super laborious to manufacture!

Anyway - has anyone every serviced one of these? Trying to decide if I should send it back or dive in. It's a bargain if I can get it to work, but obviously a waste if I damage it in the process of finding the power problem.

[TiN]

I'd pry it out of your cold hands, even dead, especially if it's a bargain. Don't mess with front end unless absolutely necessary, it's done that way due to difficult to achieve specs.
Keep it, so we can see moar photos and teardown, please. Let the community wish prevail   .
Maybe seller can give you some refund to sweeten the pill.

[rx8pilot]

I am highly motivated to make it work. The approximate cost was about $615 US - there were other things like the matching probe in the purchase so I am guessing this part was about that number. Not a bad deal if I can get it working. Most of them are listed at 4x the price on eBay and new ones are $5190 USD! I already have one other of the older Preamble versions - they are nearly the same except this one has a remote option that hooks up to LeCroy scopes. It is a very nice instrument to have for power supply design among other things. I would really like to have two working units.

Challenge number 1:
Disassemble without messing with the clearly delicate front end. So far, I have not figured out how to get the PCB on top removed without needing to de-solder the cables going directly into the hand wired cages that house the first stage of the amplifier. There are also 3 ribbon cables that are soldered direct to the front panel. The other ends are removable but on the bottom of the unit. the screws that hold the whole sandwich assembly inside the chassis are under the bottom PCB - meaning I have to pull all the boards  and connections to gain access to the power supply. I have repaired all kinds of very expensive and delicate commercial electronics, but this thing is nothing short of a real challenge.

If a single pot or trimmer gets bumped, I can see LeCroy charging an arm and a leg to get it back to calibration.

Check out the hand work - Nuts. The PCB has no silkscreen either. I have not found a service manual and probably will not. The owners manual states that calibration requires special equipment not commercially available. They are not going to give up any secrets from what I can tell.

[rx8pilot]

I dug into it. What an enormous pain in the ass. Stunningly poor mechanical design. I have never seen anything quite like it in all my decades of taking things apart.

Anyway - the power supply, as suspected, is history. The label indicates it should output +5v and +-29V. The main problem is that the +29v is pulsing out +45 to +65v! Crap, even if I fix the the power supply - the rest of the system may be toast after being pummeled with twice the expected voltage. Super crappy news. These types of supplies are hard to work on too - very difficult to get to any of the parts and I don't think this weird thing is going to be easy to source to replace it.

Bummed.

[H.O]

Stating the obvious here but can't you try powering it up with an (3) external supply - just to see if the rest of the circuit works? If it does it might be slightly more motivating trying to fix the existing supply or possibly build/buy a supply to fit in the old ones place or even use an external supply to retrofit the unit with.

[nctnico]

Stating the obvious: I'd start by checking the diodes and electrolytic capacitors in the PSU. The PSU sending pulses at least shows it has some signs of life. It wouldn't surprise me if the 5V rectifier diode is a short. The PSU starts up but can't regulate (hence the 65V pulse) and then goes into short circuit protection.

[rx8pilot]

For tonight, I have to give up. The disassembly took hours and I decided to put it back together before I forgot how to do it. I left a message for the seller to see what he thinks.

I do have enough power supplies to make the voltages happen, the problem is that I have to make a fixture of sorts to allow me to get it wired up and assembled to see if it works. I would not easily be able to do full testing because the shielding would not be in place. If there are some subtle problems, I think I would need to have it assembled correctly. On top of all that - the process of disassembly and assembly is rather risky to the front end and the BNC connectors. The solid signal wires take a lot of bending and they may fatigue after a few more passes.

 

Was hoping for better, but it seems the seller will likely just give me a refund and I can start shopping again.

[TiN]

Surprised to see SMPS in this LNA to be honest. 

[rx8pilot]

Surprised to see SMPS in this LNA to be honest. 

It is massively shielded from radiated emissions but a single pass through a ferrite is all the exiting wires get. There does not appear to be any protection on the main PCB's to deal with over voltage either, so I am not encouraged. I cannot find a block diagram or schematic and there is not even a silkscreen to help.

With that said - I powered it up externally based on the voltages printed on the PSU. The logic comes up just fine in the internal precision voltage generator works. The inputs, however, appear to be dead. The internal voltage source appears at the output, but no signals appear to pass. The current draw is right at the stated max current of the power supply:

Stated capacity vs Actual
+29v
300mA   300mA

-29v
300mA 260mA

+5v
1000mA    850mA

And the analog section with all the discrete transistors get hot - certainly hotter than my other one. I checked the various OP AMPS to make sure the supply voltages were as expected, one of the voltage references checked out ok. The 5v logic seems to be normally operating - that rail was just dead and I don't think it ever saw any high voltage. The only rail that went high voltage was the +29v. The negative rail is the one pulling more current.

This is not looking good. A lot of mystery and not much time to chase it down. The eBay seller has not yet responded to my notes. I got so far as to explain that the PSU was spewing out random high voltage and likely destroyed or at minimum seriously compromised the rest of the amplifier. This thing is rather high-end and well outside of my understanding. I believe it is past the point of practical repair - at least for me. I am sure someone with a ton of patience could go from transistor to transistor looking for bad things. Even if I could find something and get it working - I would have to send it to LeCroy or someone for calibration adding to the cost. If it failed calibration, that would suck even worse.

[nctnico]

It would be an interesting repair project but -based on your findings- likely needs replacing almost all of the semiconductors due to the overvoltage. However if it is well designed then changing the semiconductors shouldn't affect calibration. Because semiconductors are very sensitive to temperature variations in them should already be dealt with by the design itself.

[TheSteve]

We're all thinking it - it needs to end up in TiN's hands, lol.

[rx8pilot]

Part of me would love to skip work for a week and see if I can meet the challenge. It would be a sweet victory to find the issues and bring this gem back to life.

Sent from my horrible mobile....

[rx8pilot]

It's officially a dead project. On it's way back to the eBay seller. I hope the refund process is smooth and easy.

Parting notes: This instrument has the shittiest mechanical design I have ever experienced. The electronics on the other hand seem very well designed. I have another one of these already, and it works really well huge common mode range, CMRR and low noise even at 10x amplification. As long as I never have to attempt service ever again - I love these things. Looking for another one now.

[TiN]

Likely a wise move. Could be just one opamp fried, but more like whole cluster of issues. Easily one of those 200+hr projects, with steep learning curve. Thanks for teardown anyway!

P.s. I have plenty ToDo's for next year already. And much more to come soon

[steve_tech]

Too bad I didn't see this earlier. I was involved in calibration, repair, and sustaining engineering for the 1822 and the 1855.

It was difficult cramming all the guts in a PacTec box. The have been around for more than 15 years. Not bad for test equipment.

They were never intended to be serviced by the user. The original calibration procedure was 26 pages, single spaced, double sided.

Early on, production was one per day.

[rx8pilot]

Awesome!

I just purchased another one (my second) it should arrive in a few more days. Hopefully that one is in good shape.

The assembly was clearly very difficult - so many carefully hand placed parts. Do you have a suggestion for getting two calibrated? Do you know if any schematics or cal documents can be available (legally).
It's a unique piece of equipment and useful in power supply design. They still command a lot of money including the matched probe sets. 

Glad you chimed in.

[sbvr4]

Just picked up one of these up today. Does anyone know the calibration procedure for this? Specifically, the PVG, which is roughly +.8112v off. Not very precise.

Everything else seems within spec, but the PVG is way off and would like to correct the error. Any help would be greatly appreciated.

Also, @Steve_tech any chance the enclosure is the PacTec CF-225? It looks right, but PacTec oddly don't seem to offer dimension for that unit. I have a single rackmount unit that I would like to convert to the smaller bench model.
 
Thank you

[rx8pilot]

Perhaps if the PVG is that far off it is not a calibration challenge - but a repair of something broken.

If you need to take it apart - good luck. It is a weird one that is also rather delicate.

[ZGoode]

Too bad I didn't see this earlier. I was involved in calibration, repair, and sustaining engineering for the 1822 and the 1855.

It was difficult cramming all the guts in a PacTec box. The have been around for more than 15 years. Not bad for test equipment.

They were never intended to be serviced by the user. The original calibration procedure was 26 pages, single spaced, double sided.

Early on, production was one per day.

Total long shot, but having worked on the 1855 (hopefully non-a version), you don't happen to know anywhere I can find those manuals for it?  I have a 1855, but can only find the manuals for a 1855A.