Tale of two Tektronix AM503 Current Probe Amplifiers

[Hexley]

Recently I received a bad AM503 current probe amplifier from a buddy. He had a couple of good AM503s and did not need the bad one so he generously passed it on.

Initial evaluation showed that the output signal had a very large AC component and the overload LED was lit.



The immediate suspicion was that a dipped tantalum had shorted and caused some issue. But none of the three in the device was shorted.

So the supply rails were examined. Sure enough, the problem was obvious – the ripple on the negative 16 volt rail was 7 volts. Ouch.



A visual inspection of the PCB showed what had happened.

The 47uF electrolytic on the -16V output, C165, lost its negative lead right off the case:



I surmised this might be due to excessive ripple causing the capacitor to overheat. My buddy calculated that the capacitor should be rated for at least 0.5A ripple current and have low ESR in order to handle the transients that could be caused when the probe is measuring very large signals.
 
After the capacitor was replaced the AM503 worked properly.

But then I happened to look inside one of the good AM503s, and found the same problem with the same capacitor. In this case, the lead was still attached (by a thread, it seems -- it fell off when the capacitor was desoldered.)

Here is the second AM503, with C165 leaking badly just above the large blue electrolytic.



Interesting to see the same failure on two separate units, one with 1981 date codes and the other with 1991 date codes. Also, it is interesting that the corresponding capacitor on the +16 volt rail, C155, looked perfectly intact on both units.

It may be just a fluke that two units had problems with exactly the same part. But I thought it worth passing along in case other owners want to examine their AM503s.

[tautech]

Thirty and forty year old electrolytic caps failing, no surprises here......might be a similar problem in my type 134. 

[Hydron]

I also have 2 AM503s, the first (early 90s vintage) died on me due to those exact caps going bad, though visually they were fine, no eaten-up leg. A bodged radial cap (and another on the matching rail) fixed it.

The second unit I have is from over 10 years earlier with '81 date codes throughout, but all the electrolytics in it were 100% fine in both capacitance and ESR when I checked them (desoldered one leg to do it properly). They all went back in - I'll be keeping an eye on them for corrosion (very easy to do with these modules) but didn't seem much point swapping parts that tested and worked fine (especially when I don't have any matching axial parts).

[artag]

I checked mine and there seems to be no problem yet.

My pcb layout is slightly different though I don't see that affecting the corrosion. It's 670-4353-04 and the capacitor has a datecode of 8410 so it could be a little older than yours - or perhaps closer to the one you fixed first.

[srb1954]

Recently I received a bad AM503 current probe amplifier from a buddy. He had a couple of good AM503s and did not need the bad one so he generously passed it on.

Initial evaluation showed that the output signal had a very large AC component and the overload LED was lit.

(Attachment Link)

The immediate suspicion was that a dipped tantalum had shorted and caused some issue. But none of the three in the device was shorted.

So the supply rails were examined. Sure enough, the problem was obvious – the ripple on the negative 16 volt rail was 7 volts. Ouch.

(Attachment Link)

A visual inspection of the PCB showed what had happened.

The 47uF electrolytic on the -16V output, C165, lost its negative lead right off the case:

(Attachment Link)

I surmised this might be due to excessive ripple causing the capacitor to overheat. My buddy calculated that the capacitor should be rated for at least 0.5A ripple current and have low ESR in order to handle the transients that could be caused when the probe is measuring very large signals.
 
After the capacitor was replaced the AM503 worked properly.

But then I happened to look inside one of the good AM503s, and found the same problem with the same capacitor. In this case, the lead was still attached (by a thread, it seems -- it fell off when the capacitor was desoldered.)

Here is the second AM503, with C165 leaking badly just above the large blue electrolytic.

(Attachment Link)

Interesting to see the same failure on two separate units, one with 1981 date codes and the other with 1991 date codes. Also, it is interesting that the corresponding capacitor on the +16 volt rail, C155, looked perfectly intact on both units.

It may be just a fluke that two units had problems with exactly the same part. But I thought it worth passing along in case other owners want to examine their AM503s.

Capacitor C165 is a wet tantalum type. These use sulfuric acid as an electrolyte and it appears that the case seal has failed and the acid has corroded away the lead. Luckily you caught it in time before the acid migrated to the PCB and caused all sorts of damage.

These wet tantalum capacitors are extremely expensive and, if you were to replace this capacitor with the original type, it would cost more than the AM503 is worth. Fortunately, as this capacitor seems to be relatively non-critical in this application, it should be possible to replace it with a modern low-ESR aluminium electrolytic. I would replace C155 as well as it will also probably eventually fail in the same manner.

[RBBVNL9]

Please add me to the list of those who found these 47uF 50V caps to be broken in their current probe amp. My AM503 had an awful self-oscillation of 600kHz or so. Replacing both broken capacitors in the power amplifier (C155, C165) with some caps I had lying around solved the issue. Great to see my probe making clean measurements again!

Yet, I would like to replace these capacitors with decent ones, also to bring the ripple on the 16V power rail down to the level specified in the instruction manual (2mVpp max). It is now much higher.

There are a few points for which I appreciate views and advice:

• What is a good replacement for the broken 47uF 50V? I do not necessarily want the original types (like wet tantalum), good fitting alternatives (e.g., low ESR high ripple axial aluminium electrolytes) will do - unless people tell me this is a bad idea. The VISHAY "120 ATC" series 47uF 100V 20% seems one option that is in stock with at least some suppliers.
• I probably want to replace C402 and C406 in the primary PSU as well (these are large, axial 1250uF 50V electrolytes). Did not yet find good alternatives that are in stock at Mouser etc. It is key that they fit well in the available space.
• Any other caps that I better replace in the AM503?

Advice is appreciated (if possible with link to a supplier...) Thanks!