Choosing the correct coupling capacitor

[john12ax7]

This is something that I would really like to study one day. Specifically, take a 10uF film and 10uF electrolytic and maybe a 10uF bipoloar electrolytic and null test them against each other. If the added distortion is really negligible as I claim, then it should really not be taboo to use electrolytics in filters.

Even if the electrolytic distortion is acceptable, their tolerance is usually rather bad and more prone to drift. Not qualities you want in something repeatable like a filter.

 

[abbey road d enfer]

This is the article:

http://www.aes.org/e-lib/download.cfm/20891.pdf?ID=20891

This article is so unfounded I had forgotten about it.

Looking at it again I see the paper is strictly about coupling and not using an electrolytic small enough to cut lows.

Indeed. 100uF with 100k (-3dB @0.015Hz!) is more than what I advocate. This guy thinks that the very presence of a capacitor is enough to create distortion. No. It's the voltage that's developed across it when driving a load that counts.

 

[Matador]

Note that there are caveats when talking about tube circuits, especially in guitar amps. Many inter-stage coupling capacitors are deliberately sized low to provide low-cut (high pass) between stages, which can help prevent blocking distortion as well.

As an example, the JCM800 (2203) utilizes 22nF into 150K for the output tubes, which is about 50Hz. Going 10x or 100x might have unwanted side effects.

 

[abbey road d enfer]

Guitar amps are quite different. Distortion is typically more than 1%, so a liitle more from the capacitors is negligible. And the coupling caps are never electrolytics. In addition, the signal from an electric guitar has a very large subsonic content (pick attack), so it's necessary to provide significant high pass filtering, before it reaches the output xfmr.

 

[Pief]

A coupling capacitor is an interstage capacitor for blocking DC.
Values for these types of caps never really exceed 200 nanos. So why people are talking uF's in the range of 47uf to 470uF is beyond me.

Grtz...

Pief

 

[Bo Deadly]

A coupling capacitor is an interstage capacitor for blocking DC.
Values for these types of caps never really exceed 200 nanos. So why people are talking uF's in the range of 47uf to 470uF is beyond me.

This is simply not true. Even for tube gear.

 

[Newmarket]

Huh...Sounds like mumbo jumbo.

JR

Indeed. The sort of stuff people come up with when showing no concept of the physics of how electricals signals propagate in a component or conductor.

 

[Newmarket]

Even if the electrolytic distortion is acceptable, their tolerance is usually rather bad and more prone to drift. Not qualities you want in something repeatable like a filter.

Yes - and lots of things many don't appreciate wrt electrolytics. eg the stated life spec is usually defined as having lost half of their initial value. Use them to block dc (and power decouplig) when required. Not for filters / EQ / time constants.

 

[abbey road d enfer]

A coupling capacitor is an interstage capacitor for blocking DC.
Values for these types of caps never really exceed 200 nanos.

This is true for tube circuits that operate with high impedances.

So why people are talking uF's in the range of 47uf to 470uF is beyond me.

Have you not heard? there are new devices that are called transistors and operate at low impedance.

 

[Pief]

I know. But please...if its such common knowledge: Post me 1 proven cirquit diagram that uses 47uf (or above) interstage coupling capacitors for AF amplification and i will offer you my sincere apologies.

 

[abbey road d enfer]

No need for apologies. But please be a little more open-minded. That is part of scientific approach.
Don't want to waste my time but I could probably find dozens of other examples.

 

[Bo Deadly]

I know. But please...if its such common knowledge: Post me 1 proven cirquit diagram that uses 47uf (or above) interstage coupling capacitors for AF amplification and i will offer you my sincere apologies.

This is part of the Mackie 1202VLZ3. There are actually 7 coupling electrolytics at 47uF and a 3300u in the gain control network which is just as much in the signal path as bypass caps in a tube amp.

 

[neil.johnson]

Doug Self, Small Signal Audio Design, Chapter 2 - investigates capacitor distortion, with measurements. And for further details look for Cyril Bateman's papers on capacitor behaviour ("Capacitor Sound").

Neil

 

[JohnRoberts]

Back in the 1980s when I was still writing my "Audio Mythology" magazine column, I dedicated one column to capacitor mythology. Again capacitors are a mature technology and well characterized. In other words learn how to read the spec sheets, not the marketing hype.

I don't have copies of my old columns (I hear they are around on the WWW) but the physics surrounding capacitors hasen't changed in the decades since.

JR

 

[Lee_M]

I prefer using a THAT1646, that uses small bipolar 'lytics.

I like those chips too, I tend to use the Wima "red box" 10uf film caps for the "output sense" capacitors.

They're a little bit more expensive than bipolar electrolytics (the Wima caps are around £2 each when buying small amounts), but their rated life expectancy is around 300,000 hours, so without getting into things like "mojo", they should be a significant improvement (in the long term, at the very least) over the 10,000 hours that most "long life" electrolytics are rated for.

 

[warpie]

I might be wrong but I have the impression that the hours of the lytics are rated according to their maximum temperature. So for an 85C cap the 10,000 hours lifetime would be in the case the cap runs close to 85C.

 

[abbey road d enfer]

I might be wrong but I have the impression that the hours of the lytics are rated according to their maximum temperature. So for an 85C cap the 10,000 hours lifetime would be in the case the cap runs close to 85C.

That's correct. There is a formula that anticipates the expected lifetime with lower ambient temperature, based on chemical activity. It says that lifetime doubles for every 10°C temperature drop.
A 2000h/85°C cap submitted to an ambient temperature of 40°C can be expected to live 5 years.
Which justifies using 105°C types, which would last 20 years. Actually, there maybe other causes of defect in a 20 years span. And who designs gear for 20 years of service, today?
Actually, many lytic caps that are supposed to be long past according to this law are still operating, probably not like the day they came out of the factory, but well enough for their owners to ignore it.

 

[Newmarket]

Actually, many lytic caps that are supposed to be long past according to this law are still operating, probably not like the day they came out of the factory, but well enough for their owners to ignore it.

Indeed. Other factors too eg standard 'wet' electrolytics are (arguably) best operated in a vertical position as this helps with the continual 'self-healing' process. But there are loads of cases in which they will be horizontally orientated eg a PSU pcb is mounted to a chassis vertically or they are on a PCB in a vertical rack type system (eg Schroff).

 

[warpie]

Indeed. Other factors too eg standard 'wet' electrolytics are (arguably) best operated in a vertical position as this helps with the continual 'self-healing' process. But there are loads of cases in which they will be horizontally orientated eg a PSU pcb is mounted to a chassis vertically or they are on a PCB in a vertical rack type system (eg Schroff).

I never heard this before. So that's pretty much the case for all 500 modules. Interesting.

 

[Tubetec]

I read somewhere that in tube stages much larger foil caps can sometimes be used to good effect with regards noise . The idea was that by making the cut off frequency of the filter much lower , the output impedence of the preceeding valve could help reduce low frequency noise at the following valve grid.
Anyone any thoughts on this ?