Electrolytic vs Tantalum for Supply ByPass?

[abbey road d enfer]

No. He means a particular series called "x2y" that has nothing to do with mains protection. They're SMD MLCCs that have two caps in one package so that inductance cancels and so on. We talked about them in one of the CMRR threads. They're good for common mode shunt on entry for signal.

Ah OK, thanks for the explanation. I was not aware of them. I wonder how in practice they offer a significant improvement over standard ceramic caps for audio circuitry.

 

[ccaudle]

how in practice they offer a significant improvement over standard ceramic caps

You could start here, it has a lot of their brochure information in one place:
Johanson X2Y caps brochure

The high points are small size and low inductance for bypass use, and good cap-to-cap matching for EMI filter use (less CM to diff-mode conversion from cap mismatch).

 

[abbey road d enfer]

You could start here, it has a lot of their brochure information in one place:
Johanson X2Y caps brochure

Yes, I read it. I still don't see how it's a game-changer in audio circuitry when used in bypass.

good cap-to-cap matching for EMI filter use (less CM to diff-mode conversion from cap mismatch).

The example given in this document leaves somewhat perplex. When they compare with a filter with 3 elements, what is it? It I had two inductors and one cap, I would do better than a simple 1st-order 8kHz LPF.
They insist on the superlative matching of capacitors, but don't publish any figure.

 

[ccaudle]

I still don't see how it's a game-changer in audio circuitry when used in bypass.

Agree, for bypass I don't think it is a big deal for audio circuits. Even for digital portions of a design I'm not sure it is that much better than long side plating (e.g. 0508 cap instead of 0805).

When they compare with a filter with 3 elements, what is it?

The resistors shown on the schematic, a cap from each line to GND, and a cap from line to line for additional differential mode filtering. The 3-to-1 reduction they mention is going from the three separate capacitors to a single X2Y device. Probably a bigger deal when trying to fit circuitry into a cell phone than when putting some devices behind an XLR connector.

They insist on the superlative matching of capacitors, but don't publish any figure.

Yeah, I always found that annoying. I tried to pin them down one time, the only thing they would claim was that the matching was "typically better than 5%" for 20% rated caps. I don't know what that means in practice, I don't think I have a way to measure capacitance more accurately than that to check a few samples to see how close they really are. Now that I double check I might, my Fluke meter claims +/- 1.2% accuracy for capacitance measurements, so maybe I could check to within a couple of percent. I'll have to get some and try.

 

[Bo Deadly]

Agree, for bypass I don't think it is a big deal for audio circuits.

Yeah, I always found that annoying. I tried to pin them down one time, the only thing they would claim was that the matching was "typically better than 5%" for 20% rated caps. I don't know what that means in practice, I don't think I have a way to measure capacitance more accurately than that to check a few samples to see how close they really are.

I think the marquee feature of these x2y caps is that the terminal inductance is "cancelled" to give better common mode attenuation at high frequencies. That's it's trick. Period. So that's not really applicable to audio bypass. For GHz devices sure. But not for audio.

I think the brochure might be a little misleading though. It advertises these x2y caps as an alternative to a common mode choke. But it's not immediately clear to me if that's a fair comparison because I think the inductance cancellation effect removes inductance so that the capacitors are more effective whereas with a common mode choke the inductance is in series with the signal and blocking RFI. But whatever works I guess. And it is one tiny little cheap part which is hard to argue with.

I have circular PCBs specifically designed to fit on XLR connectors but small enough to pass through the panel cutout with circuit protection / RFI filtering on them. I could see x2y being good for that purpose being one tiny part to shunt RFI to the chassis right at the XLR.

 

[Newmarket]

Lead inductance? Come on now.

HaHa - yes. Maybe too much time in the EMC lab But there is a view that all circuits exist in a 'rf' environment now - whether wanted or not ! But agree that it's unlikely to make your OPA 2134 sing like a canary if you use a leaded part on a double sided pcb.

 

[Newmarket]

It's a concern for RF and high-speed logic. I've never could note a difference between film and ceramic by-pass caps in audio applications.

Do you mean using these for rail decoupling? I use them for mains inlet; they're quite enormous.

Agreed re Audio. But "These Days" ( copyright acknowledged to comedian Stewart Lee we have high speed digital/DSP etc everywhere.

Apologies for the X2Y confusion - I was replying quickly before having to go and do some actual work The naming is a bit unfortunate. It relates to the topology but easily confused with mains safety rated caps'.

 

[neil.johnson]

It looks like the tests did use 0.5V pulses which is a lot better than the 70 VRMS in the last link. Also the articles are kinda old. His "Ultra low distortion test oscillator" THD is "under 0.01%" which is just not that great these days and easily beaten with a mediocre USB audio interface. So lets just appreciate that he did that work almost 20 years ago.

Mmm... according to the writeup in the first Capacitor Sound article at
https://linearaudio.nl/cyril-batemans-capacitor-sound-articlesHis oscillator was able to achieve quite respectable results (page 6):

Output distortion of the complete
oscillator design shown in Figure 7.
Outputting 3 volts into a 600 Ω load,
distortion of this prototype measured using
my pre-notch filter/amplifier is buried in
the measurement noise floor at -128 dB, or
less than 0.5 PPM.

So that's less than 0.00005%. Not bad.

 

[Bo Deadly]

So that's less than 0.00005%. Not bad.

Yeah, I think you must be right. Looking at it again, I'm not sure where I got the 0.01% number from. Maybe I was looking at something without the passive notch filter.

 

[neil.johnson]

Yeah, I think you must be right. Looking at it again, I'm not sure where I got the 0.01% number from. Maybe I was looking at something without the passive notch filter.

It's the very first number on the first page:

Most properly designed power amplifiers measure less than 0.01%, or 100 PPM distortion when sinewave tested at 1 kHz

So I think he did pretty good for a Wien oscillator, and about on par with an APx555 for significantly less outlay!