[ASK] Standard Capacitor ....

[acbern]

So looks as if the 8904 alone, delivering only 10V will not do it on the extreme ends. I was execting this, but do not yet have an HV amplifier. A standard amplifier and a 100V transformer should do it, the only problem may be the capacitance to earth and between wingdings. I need to see. According to the 707 manual, 100V should be sufficient. Also, the lock-in amplifier is much more senstive than the ESI detector, that should compensate too. I will see.

Re. the qudarature bridge, if you want to determine the capacitance of your standard, either you change the resistance (diffcult for a coaxial resistor), the amplitude ratio (dekatran) or the frequency (changes the amplitude relationship too). So the best (and thats what the national labs do) is the ratio with a dekatran. 

[BravoV]

 

[amspire]

Should finish my tests of the Genrad Standard capacitor versus the eBay Russian Silver Mica capacitors. The results are looking really interesting and a little surprising. I should be able to post some results and some photos today.

[amspire]

I did some testing on the Russian SSG capacitors, and also a Genrad 1409-G Standard Mica capacitor. I saw the same capacitor that I am using was available on eBay at $10 for 4 capacitors. The Genrads often sell for over US$100.

My test setup was a ESI 701 Capacitance Measuring System, with  a small foam esky to act as my temperature chamber.

I put two 50nF Russian SSG capacitors in parallel is a small shielded box to make a 0.1uF capacitor (it measured 0.10001 uF).

The Genrad was a custom value type 1409-g. This is normally a 2nF +/- 0.1% part, but it was obviously customized by Genrad before sealing to be 1.9865 +/- 0.05%. This is probably important to the following tests. It measured at 1.9875 nF.

Here is the Capacitance variation versus frequency. The Russian capacitors were clearly better then my Genrad.

Now for the capacitance variation versus temperature at 1kHz test frequency.

The Russian capacitor is brilliant, but what is happening to the Genrad? The Genrad is meant to be +25 to +45 ppm/C between 13 and 33 degrees, but instead it is rubbish.

If you look at the dissipation factor, same story. The Russian capacitor dissipation factor was 0.00007 at 12 degrees to 0.00012 at 39 degrees. Again brilliant!

 The Genrad is meant to be better then 0.0003. The dissipation factor was 0.0003 at 10 degrees, about 0.006 at 23 degrees and 0.0013 at 33 degrees! Something is very wrong.

What I think has happened is that to get 1.9865 uF, they took an unsealed 2nF from the production line and added a 0.27uF cap in series. The trouble is that the 0.27 uF would not have been mica. It was built about 1959, and they could have even put a paper capacitor in. It would explain the poor curves now. Also, the Genrad capacitors are not hermetically sealed. They put in a packet of Silica Gel and seal it with something like Butyl Rubber (Mastic). After over 50 years, the silica gel may have lost the ability to absorb any more moisture. The Russian capacitor is hermetically sealed.

The new type 1409 caps from IET have a coefficient of +20ppm/C and this Russian cap is about +11ppm/C.

I have only tested one, but I think that the Russian SSG caps are about half the drift and dissipation of the IET Type 1409 Standard Capacitors. The SSG's look very stable on the bridge and I think with a more accurate temperature curve characterization, it would be very easy to work out the capacitance to 10ppm given the temperature.

The other message is: avoid custom value Genrad Standard caps.

[manganin]

The Genrad was a custom value type 1409-g. This is normally a 2nF +/- 0.1% part, but it was obviously customized by Genrad before sealing to be 1.9865 +/- 0.05%. This is probably important to the following tests. It measured at 1.9875 nF.

That is not a custom value (which would have been engraved on the name plate), but it has simply drifted out of the specs. That is very common. The 1409 is extremely sensitive to non-optimal storage conditions and many (if not most) of the used ones have problems.

The capacitor was factory adjusted within 500 ppm of the nominal value at 1 kHz according to the specifications, but within 100 ppm in practice. Except some of the early ones which had larger margins. If the capacitor is good it should still be within the one year specification.

What I think has happened is that to get 1.9865 uF, they took an unsealed 2nF from the production line and added a 0.27uF cap in series. The trouble is that the 0.27 uF would not have been mica. It was built about 1959, and they could have even put a paper capacitor in.

Every capacitor was hand made, there was no "production line". All you will find inside is a small factory made mica capacitor used for the final adjustment and sometimes not even that.

[amspire]

You could be right. The rear panel does have an old printed label that says: "Adjustment Accuracy +/- 0.05%" but I cannot be sure that the label came from Genrad, or that the back plate even originally came from this standard. The back plate is removable.

So is it possible that moisture or contamination can get into the standard enough to not only to cause a 0.7% error, but also totally change the temperature coefficient? If the sealant is that bad, it is really not much of a standard capacitor.

I might stick it in the oven at 65 degC (the maximum storage temperature) for a while and see if the capacitance or dissipation factor changes.

Doesn't look like it has any use as a standard capacitor anymore.

[manganin]

So is it possible that moisture or contamination can get into the standard enough to not only to cause a 0.7% error, but also totally change the temperature coefficient? If the sealant is that bad, it is really not much of a standard capacitor.

The moisture gets easily in but it is practically impossible to get it out. The first sign of the problem is the dissipation factor going up and when the damage progresses it will also change the capacitance.

Another but rare fault mechanism is with the clamp holding the stack together, which can change the capacitance a lot, but that usually doesn't much affect the dissipation factor.

I might stick it in the oven at 65 degC (the maximum storage temperature) for a while and see if the capacitance or dissipation factor changes.

Doesn't look like it has any use as a standard capacitor anymore.

The process is irreversible. A good candidate for teardown...

[BravoV]

Thanks amspire & manganin ! 

You guys inspired me, pulled trigger on eBay -> http://www.ebay.com/itm/321715093338 , 5 pairs. 

[manganin]

You guys inspired me, pulled trigger on eBay -> http://www.ebay.com/itm/321715093338 , 5 pairs. 

The long term stability is still the big question.

The SSG capacitors that I measured many years ago showed both systematic and random drift way too high for metrology use. Material purity issue?

[BravoV]

You guys inspired me, pulled trigger on eBay -> http://www.ebay.com/itm/321715093338 , 5 pairs. 

The long term stability is still the big question.

The SSG capacitors that I measured many years ago showed both systematic and random drift way too high for metrology use. Material purity issue?

Only for transfer standard, is it good enough ? Also does this type of cap sensitive to mechanical shock ?

[amspire]

I decided to find out a bit more about the SSG capacitors, and luckily I found all the details I could possible want from an Audio site:

These have such feathery, extended highs that almost make Mundorf gold/silver seem a bit less open up top. Not bad for a $5 capacitor, so what's the caveat? Unfortunately for the SSG-3, I've been listening to some of the best Teflon capacitors of late, and compared to the good Teflons, the silver mica's don't quite have the stop-dead-in-your tracks resolution and definition, especially in the midrange. Compared to Teflons, SSG-3 is a tad more laid-back and softer in the midrange, and the bass, while impactful, is not quite bounce-rain-off-the-drumskin tight. Still, I can see a lot of people being quite happy with these silver mica's, especially if used in equipment that leans in opposite sonic directions or if your tastes cotton to the open, airy, feathery, smooth sound.

So there you go - a good capacitor for a transfer standard, as long as you are happy with the feathery sound.

I can probably give some stability information. I have had my capacitors for about 10 years, and I think I measured the value for one of them. If I can dig up the details, I will measure it again.

[amspire]

Found the old measurement and repeated it. Over about 10 years, the SSG-1 50000pF capacitor drift against my ESI bridge is less the 2pF - that is less then 0.004%.

The original measurement was only a 2 wire measurement, so I cannot be precise about this error. It doesn't take much to cause a 2-wire error of a few pF when you do not have a proper testing jig. I just tried to measure it the same way I did before. Luckily, the room has the same temperature as the old measurement - 17 degrees C.

I have no idea of the long term stability of my bridge standard capacitor, but its calibration adjustment range is +/- 0.03%, so obviously ESI do not expect it to drift much during its life. 

A one-off sample, but again I am impressed. It is not enough to make conclusions about all SSG-1 capacitors, but it looks like there are some good ones out there.

Richard

[acbern]

It should be kept in mind that the genrad 1403 air capacitors may drift up to 0.05% per year, and even if they drift less, it is still the value to be taken when using them as a standard (unless one has made a statistical analysis showing it is better). Most silver mica capacitors are not worse. I will do some measurements once my gear is back from calibration in  2 weeks timeframe on 10pF to 100nF capacitors.

[amspire]

The new IET 1409 capacitors are rated at 0.01%/year maximum drift.

I don't think the original 1409 back in the 1950's had a yearly drift spec, but in-house tests over 3 years had shown less then 0.01% drift, and 0.01% was also the limit to Genrad's own in-house standards.

[amspire]

I might stick it in the oven at 65 degC (the maximum storage temperature) for a while and see if the capacitance or dissipation factor changes.

Doesn't look like it has any use as a standard capacitor anymore.

The process is irreversible. A good candidate for teardown...

Well, we currently have dry cool weather. I thought I would try baking my sick Genrad 2nF 0.1% standard cap.

Did it for 2 hours and the dissipation factor dropped from .00057 to .00040. Promising. The capacitance dropped from 1.9875nF to 1.9867nF that was closer to the value marked on the capacitor.

OK. I did another 8 hours of baking.

Dissipation factor is now 0.0003 which just meets the specifications. Capacitance has increased to 1.9988nf. It has gone from 0.7% below the engraved 2nF capacitance to 0.06% out. The original accuracy on the label was 0.1%, so the capacitor is now back in specification. Interesting thing is I have had this for about 8 years and it has always been stable and pretty close to 1.9865nF. It looks like once moisture gets in, it will not leave until the temperature is high enough. The fact it only took 8 hours is damning on the state of the sealant.

I am leaving it out in the sun all day today and keep baking it until there is no more change - I want to dry out the internal Silica Gel pack. Then I will add a new layer of sealant. We do not get too much humid weather, so I am pretty sure it will be stable after this.

Can it be trusted again as a Standard? I wouldn't use it commercially, but for use at home, you make the most out of what you have.

[manganin]

OK. I did another 8 hours of baking.

Dissipation factor is now 0.0003 which just meets the specifications. Capacitance has increased to 1.9988nf. It has gone from 0.7% below the engraved 2nF capacitance to 0.06% out. The original accuracy on the label was 0.1%, so the capacitor is now back in specification. Interesting thing is I have had this for about 8 years and it has always been stable and pretty close to 1.9865nF. It looks like once moisture gets in, it will not leave until the temperature is high enough. The fact it only took 8 hours is damning on the state of the sealant.

Very interesting findings!

One of the capacitors (0.5uF) that I tried to cure was baked several months in 60 C. It slowly affacted the dissipation factor but the capacitance stayed practically the same. The original error was smaller than yours though. My guess was that the humidity had corroded the thin silver layer

Because the capacitance changed so fast, could it be just a contact issue. One of the silver layers lost a contact because of the mechanical force by the expanded silica gel. Or the pressure of the expanded silica gel working against the stack spring. Hard to believe that the capacitor itself dries so fast.

Can you see any obvious slots or cracks in the sealant, when the bottom lid has been removed?

Can it be trusted again as a Standard? I wouldn't use it commercially, but for use at home, you make the most out of what you have.

It would be interesting to see how it reacts if exposed to humidity again. And if the change of the capacitance is smooth or sudden.

[BravoV]

Abit OT, about the precision mica caps, presumably closely matched, why they're packed in pair from factory ?

[woodchips]

On measuring capacitors, has anyone had a play with the Pacitor aircraft measuring systems?

These are used to measure the fuel remaining in the tanks and rely on the dielectric difference between air and fuel to vary the measured capacitance. The test sets have a clock type dial reading down to single pF. This is driven by an oscillator in some type of bridge arrangement.

Another manufacturer of precision capacitors is Muirhead, air spaced, quartz supports, loss angle 1 micro radian. Comes in a large wood box, looks really nice!

[Vtile]

When in old General Radio's era, apart from the proven mica, just wonder if the development of NPO material at that time already matured as now ?

NP0 is just plain ceramic, and yes they had them, but not in surface mount [obviously].

Mica is a very old material, and over the millions of years it has already done whatever it is going to do-- so, that's one of the main reasons that it is used in capacitors and resistors.  Mica *is* affected by humidity though.

A vacuum capacitor made of silver-plated Invar plates and Sapphire spacers would probably be quite stable with time and temperature, and it would have very low [almost zero] dissipation factor up into UHF.  There have been some experiments at NIST, but I can't remember the papers nor do I have any links.  I think everyone went with fused-silica capacitors, because they would be more robust during shipping than any capacitor made of interdigitated plates.

This is really old topic, I just ended up here by gugle.. When talking about Mica is it a nicname for special type of mica product or just general term referring to all types of mica sheets etc. (which some have resin like micanite and some are pure sheets)

On measuring capacitors, has anyone had a play with the Pacitor aircraft measuring systems?

These are used to measure the fuel remaining in the tanks and rely on the dielectric difference between air and fuel to vary the measured capacitance. The test sets have a clock type dial reading down to single pF. This is driven by an oscillator in some type of bridge arrangement.

Another manufacturer of precision capacitors is Muirhead, air spaced, quartz supports, loss angle 1 micro radian. Comes in a large wood box, looks really nice!

Capacitive level sensing is 'pretty common' in industrial settings.

[amspire]

This is really old topic, I just ended up here by gugle.. When talking about Mica is it a nicname for special type of mica product or just general term referring to all types of mica sheets etc. (which some have resin like micanite and some are pure sheets)

Mica is a rock that you mine. It cleaves perfectly in one plane and so that is how the sheets are made.

[Vtile]

This is really old topic, I just ended up here by gugle.. When talking about Mica is it a nicname for special type of mica product or just general term referring to all types of mica sheets etc. (which some have resin like micanite and some are pure sheets)

Mica is a rock that you mine. It cleaves perfectly in one plane and so that is how the sheets are made.

Yes, but there is also products that are some form of compressed mica laminates etc. composites.

PS. What I tried to ask is, when people are talking about mica capacitors are they referring to any Mica product or specific type of mica ie. natural 100% mica sheets.

[Conrad Hoffman]

I've only had one GR mica cap apart, but it had a desiccant pack, a cover plate and was potted in rosin. I think I replaced or baked the desiccant pack, probably a low temp bake of the whole thing, then potted the back with 2-part RTV. Sometimes over time those rosin jobs can pull away from the case and admit moisture. The rosin can be a nightmare to get out- I'm still trying to get inside a big GR inductor that went open.

[BravoV]

Yet another Russia silver mica capacitor, SSG-3 series 200nF 0.3% 350V rated.

I'm guessing this SSG-3 series is also as good as the SSG-1 as the one Amspire tested on ESI 701 above ?

Few shots, measured result and size comparison with the smaller SSG-1 series, a common thru hole Cornel Dubilier silver mica, and a TO220 as reference size.

[HighVoltage]

My first "standard" capacitor, bought as NOS and really looks like it was never used before.

But, what does "BELOW 750 KC" mean on the label?

[BravoV]

Olde calling ... KC = Kilo Cycle = Kilo Hertz