GW Instek GDS-2204 oscilloscope - repair log [fixed]

[sean0118]

Time to fix another scope. 

This one is a GW Instek GDS-2204, made in 2007 judging by the IC date codes.

It seems to have some intermittent noise issues, things like spurious spikes, that don't appear on other scopes. I tried to reproduce it today but couldn't, might be temperature dependent?


I opened it up and straight away saw a lot of corrosion on the power supply PCB. At first I thought the capacitors were leaking, but turns out the epoxy used to secure the components is the type that turns corrosive, and it's on everything! 

I'm not sure if the capacitors are actually compromised, I put a probe on its output and couldn't see much ripple. There was high frequency noise ~0.1V but that's probably due to how I was measuring it.


I need to investigate further, I suspect the power supply might not be causing the other issues. But here's some photos anyway:

[Grandchuck]

Wow; that is a lot of corrosion.  Was the scope in a high humidity environment?  Good luck with it.

[sean0118]

Wow; that is a lot of corrosion.  Was the scope in a high humidity environment?  Good luck with it.

It's even worse than I thought initially. I'm really surprised it put out the correct voltages still. 

I haven't had it for long, but it was stored in a normal lab environment before I had it (but not humidity controlled). I think the epoxy must just be really corrosive. The weird thing is there's a few places where the epoxy is still a light colour and hasn't caused corrosion yet, maybe it was a different batch?


I'm starting to suspect the power supply might be the cause of the odd behavior. Some pads on the bottom of the PCB are even corroded, one pin was completely cracked and loose.

I was hoping to not do a complete re-cap, but even some that looked okay initially are lumpy with corrosion under the plastic sleeve. I couldn;t even get a capacitance reading of one of the bad ones, so I suspect they are damaged internally too.









 

[Grandchuck]

I am wondering if there was some kind of fumes/corrosive atmosphere in the storage environment?  You could be right about the epoxy.  I wonder if other items from this manufacturer have exhibited this?  Anyone know?

Good chance you will wind up with a decent scope if you follow through.

[sean0118]

I am wondering if there was some kind of fumes/corrosive atmosphere in the storage environment?  You could be right about the epoxy.  I wonder if other items from this manufacturer have exhibited this? 

The brown glue is known to go corrosive, it's more common on 1980s audio amps etc though. But I think you are right, seems high humidity might be required as well, and I guess here in Sydney it is high at times.


Also, the PSU is a LYTEC LP5422, turns out there's already repair threads about it (and a partial schematic):

https://www.eevblog.com/forum/repair/help-with-a-smps-(lytec-lp5422/
https://www.eevblog.com/forum/repair/gw-instek-gds-2104-power-supply-failure/
https://www.eevblog.com/forum/beginners/lytec-lp5422/

[tautech]

You're very lucky it hasn't already gone BANG such is the conductivity of some of these adhesives when they absorb enough moisture and start breaking down.

Short of stripping the PCB to clean all that muck away first stop should be a cycle in the dishwasher and see what you have left.

[sean0118]

Okay, I think this one is a success. 

I ended up removing a lot of the components, they were just too corroded to recover. I have access to an ultrasonic cleaner so I ended up putting it through that for the max time. Thankfully the main transformer didn't have any glue on it and seems undamaged, the inverter transformer did and the fragile windings made it difficult to get the glue off the legs, but I got it pretty clean. Could always replace it with an LED backlight if it fails eventually.

I've attached a spreadsheet with all the components I've replaced in case it helps someone (lists capacitor pitch / diameter etc). Also just a warning for anyone working on one, the primary side heatsink is live, I know it's common, but just thought I'd point it out.

Surprisingly it still works with all the new components populated. I was a little worried I'd make a mistake due to the amount I was replacing. I've run the scope for a few hours now and haven't noticed any issues.

I also upgraded the firmware to V1.11 and ran the vertical calibration. GW still provides the firmware on their website, but I'll attach it here too so it's backed up.

It seems to be working properly now, and it's nice that it's a four channel and 200MHz, but it's a shame the screen is so low resolution. You can see individual pixels easily.
 














Also, I was curious and opened up the front end acquisition section:






I also took the front panel off. The Ch 3 LED (LED1703) was blinking occasionally and I wanted to check there was no conductive epoxy on the front panel PCBs (there's not luckily). The LED for Ch 3 had some black marks on the back of it, I think it might have just been sharpie though, it cleaned off with alcohol easily. I re-soldered it again and it seems to have fixed it, maybe just temporarily though, I guess it might have broken bond wires internally?