HP3458A Schaffner mains line filter failure

[atsthng]

Not unexpected actually, given mention of this filter failing in other equipment in the forum. Still, gave me a good scare (loud bang, horrendous stink) in the middle of making a measurement - thought I'd stuffed up a $$ meter.

Mess cleaned up with IPA - most on the chassis/cover, some on the digital board.

Replaced with the Schaffner modern equivalent, powered up and fortunately all seems OK. The old filter is fastened with two rivets; it looked difficult to drill out without leaving metal chips all over the place, so a sharp chisel wedged between chassis and the plastic housing of the filter, application of a hammer, and it eventually came out.

This HP3458A is nearly 20 years old. Filter has a HP part number on it, so the filter might also be that old.

So, for HP3458A owners who are have very old machines and are unaware of the tendency of this filter to fail, you now know :-)

[lowimpedance]

 Ahh yes it is amazing how fast you move to switch off, even though its to late. Had this happen on my Tek 2235, quite horrid smell !.
As you say a good clean and a new inlet and off you go again.

[amyk]

There's a thread here where someone had two failures, since he replaced it with another new-old-stock of the same model which failed shortly thereafter. The capacitors in them are what fails shorted, releasing all the smoke and tar potting.

[atsthng]

Yes I read that. Hopefully the modern equivalent, ie non NOS, are more reliable.

It was a short. Tripped the panel breaker, didn't activate the residual current breaker, so it is unlikely to be a short to earth but rather between line and neutral. If I'm motivated enough I should tear open the filter to confirm this. A short to earth would be quite dangerous if there wasn't a working earth connection.

[HighVoltage]

This happened to me with an old Phillips counter with a plastic housing that came on fire.
Luckily there was nothing above the counter at the time of the fire.
It just totally melted and burned the top cover.

I would not have expected this on a quality instrument.
Definitely not on a 3458A !

[c4757p]

These tend to last quite a long time in general - but if I've got some piece of old equipment and it's got one labeled Schaffner, that thing gets changed posthaste. Seems they used a large number of completely shit capacitors around that time period...

[saturation]

Interesting you report this, there is a similar coincidental? just posted thread on the HP equipment Yahoo group on the same topic, but not on the same instrument.

[David Hess]

I was going to mention the recent hp_agilent_equipment@yahoogroups.com discussion about these Shaffner filters failing.  There have been other discussions there and in the TekScopes@yahoogroups.com group in the past.

There are similar failures in some paper X and Y class capacitors that happen with age and the general recommendation is to replace them when convenient.

[mimmus78]

Same problem with old Solatron 7150 ... there is thread here about this too.

[David Hess]

The problem appears to be moisture infiltration into the paper dielectric in the capacitors which swells breaking the case and causing eventual failure.

[HighVoltage]

The one in my Phillips counter was filled with tar / bitumen / black stuff.
It appeared to be completely sealed.
How would the moisture get infiltrated in these sealed cans?

I think these capacitors increase their ESR over time and then get hotter and hotter until finally the
capacitors explode and in a worst case, the tar start to burn.
But then I did not take it apart to search for the real reasons of breakdown.

[David Hess]

The one in my Phillips counter was filled with tar / bitumen / black stuff.
It appeared to be completely sealed.
How would the moisture get infiltrated in these sealed cans?

They are not hermetically sealed.

I think these capacitors increase their ESR over time and then get hotter and hotter until finally the
capacitors explode and in a worst case, the tar start to burn.
But then I did not take it apart to search for the real reasons of breakdown.

These specific paper capacitors fail in this way even when not used.  Age enough enough.

[electron69]

I am renewing this old topic a little bit: I recently consulted it because I just bought a venerable 3458A from the years 86-87 according to the date codes written on the components.

After connecting it to the mains I felt after a few minutes an awful smell of tar or bitumen and so I stopped to look at the mains filter. Indeed, I consult this blog so often that I had already seen many warnings about this well-known Schaffner filter... And I can confirm that I had to clean the A5 board and the separator plate a little bit because black bitumen had sunk.

Having read the answers of my predecessors, I therefore decided to apply (on the filter only, disconnecting the connections to the transformer, and not forgetting to add a piece of solid adhesive tape on either side of the metal-plastic junction of this filter) a DC voltage of 100V then gradually rising to 320V, I left this DC voltage for at least two days, reversing the polarity from time to time.

Then I reconnected the filter alone to the mains, then after an hour or two I put everything back in the original state... and this 3458A has been running for two weeks now without any problem and there is no more smell and drips. Indeed, I didn't have the heart to change the filter... problems with rivets and dust that this kind of operation would cause. So I tried to find a simplistic method to avoid the chore of change.

If others want to try to see if this solution could save a lot of work...

[drussell]

...I left this DC voltage for at least two days, reversing the polarity from time to time.

Then I reconnected the filter alone to the mains, then after an hour or two I put everything back in the original state... and this 3458A has been running for two weeks now without any problem and there is no more smell and drips. Indeed, I didn't have the heart to change the filter... problems with rivets and dust that this kind of operation would cause. So I tried to find a simplistic method to avoid the chore of change.

That is a terrible idea.

One of those capacitors could short out at any moment and potentially cause a fire.

They are known-defective.  The only proper course of action is to replace it.

[floobydust]

Your method to apply high DC V for a few days- that is what you do to (wet) electrolytic capacitors to reform the dielectric.

But this is a paper dielectric film capacitor that fails, so I don't believe your method is sound  The failed capacitors have probably gone open-circuit after the original burning up.

Just bite the bullet and replace the mains filter

[Alex Eisenhut]

Put me in the "replace on sight" camp. Saw these Schaffner filters in some Tek gear I bought, removed them before even applying power. I bought Corcom equivalents (I needed the PCB mounted ones).
I paid more for the filters than for the gear ... 
Meh, what can you do.

[Alex Eisenhut]

And I can confirm that I had to clean the A5 board and the separator plate a little bit because black bitumen had sunk.

"leaked"  C'est vrai que "couler" se traduit comme "sunk" si on parle de couler un navire...

[Samogon]

What is the replacement part number, where did you bougt it?
I have my 3458A on its way to me, so want to replace it before
I even apply power to the unit.
Thank you.

[ArthurDent]

I learned that these filters were a ticking time bomb just recently so I checked and found my very nice HP3457A had one. I replaced it immediately with another brand. There are a number of companies that make these.

[TiN]

I am renewing this old topic a little bit: I recently consulted it because I just bought a venerable 3458A from the years 86-87 according to the date codes written on the components.

I would love to see photos of boards , as AFAIK this is the oldest unit that I've ever heard. Unit was introduced to marked around 1989, might be one of first prototypes or so even. Would be cool to see it's guts.

Having read the answers of my predecessors, I therefore decided to apply (on the filter only, disconnecting the connections to the transformer, and not forgetting to add a piece of solid adhesive tape on either side of the metal-plastic junction of this filter) a DC voltage of 100V then gradually rising to 320V, I left this DC voltage for at least two days, reversing the polarity from time to time.

Leaving time bomb and fire hazard in very special meter like 3458A, considering than 3458A has to be run continuously for long time to ensure accuracy, often unattended sounds wrong to me.  Saving 10$ to risk on XXXX$ meter?

[amyk]

Having read the answers of my predecessors, I therefore decided to apply (on the filter only, disconnecting the connections to the transformer, and not forgetting to add a piece of solid adhesive tape on either side of the metal-plastic junction of this filter) a DC voltage of 100V then gradually rising to 320V, I left this DC voltage for at least two days, reversing the polarity from time to time.

Then I reconnected the filter alone to the mains, then after an hour or two I put everything back in the original state... and this 3458A has been running for two weeks now without any problem and there is no more smell and drips. Indeed, I didn't have the heart to change the filter... problems with rivets and dust that this kind of operation would cause. So I tried to find a simplistic method to avoid the chore of change.

If others want to try to see if this solution could save a lot of work...

What you did probably just burnt the cap(s) open, the results just aren't as dramatic because you limited the voltage applied. If it didn't clear them all, the same thing could happen again to the other(s) in the module.

[ArthurDent]

I literally ripped the filter I showed in post #18 apart to see the capacitors. The 0.047 Mf is across the line at the input and is rated 250 V.A.C., the others didn't do well when I destroyed the filter so I couldn't read them. I wonder if 250 is cutting it close if your mains voltage is 240. It would be interesting to know if the failures are in countries where the mains are 240 and the filter is o.k. where the line is 120 V.A.C.. Either way, I'd play it safe and replace any of these filters you find to make sure the equipment is really protected and give you peace of mind.

[madires]

The problematic filters got Rifa caps. I'm not sure, but the filter's production year could give a hint, since Schaffner dropped Rifa at some point in time. WIMA caps are quite reliable.

[ArthurDent]

Although the complaints on Schaffner filters seem to be for older production, even into the mid 1970s, the photo of the failed one that started this thread was 1994 and the one I replaced that hadn't failed (yet) in my HP 3457A was 1996. I don't know if there's enough data to say if the cause is capacitor make, age, or transients on a capacitor that could use a higher voltage rating for 240 V.A.C. mains. I have a feeling that the one I had would have continued working fine on 120 V.A.C. but I'm happy I took the time to replace it.

Some of the pieces of equipment I have from the 1980s just had a ceramic capacitor from each line to ground and the caps were rated 1 KV.

[Samogon]

Got my 3458A today and error annunciation which was shown on the picture revealed only one error: unit requires calibration. So let me join 3458A club  
Also got mains inlet/filter from digikey it is 10EEJ1 from TEcorcom. This is first step i will do before any further attempts to switch unit on. Dont want spoil good deal i got.