Output decoupling Caps on TeddyRegs

Agent_Cooper79 · #1 28-10-08, 04:08 PM

Hi everyone,

I just want to pass on a few things me and Gunnar have noticed from trying different types and sizes of capacitors on the outputs of TeddyRegs feeding Naim preamps.

All tests were performed in the context of Gunnars setup where we used TeddyRegs in his tw HiCaps feeding an 82 (modded with MMK feedback, Avondale Current source decoupling in the Gain stages and Elna Starget decoupling Caps), fed by CdSII and 135s into Totem Mani2.
One of the HiCaps was modded at a time and after given at least a few days to burn in it was tried against the earlier best sounding capacitor HiCap. This gave us the possibility to try two combinations against each other back an forth reasonably fast.

It started of by trying to cure a slight sibilance in the treble heard mostly on less than stellar recordings, and since so much really good music is on crap recordings this actually posed a problem.

The TeddyRegs are implemented almost right according to spec with Fetlington output vbe, the only deviation from Teddy's spec was the output capacitor where we originally used 47uF Rubycon ZA.
We started our experiments with Elna Silmics. Right from the start they were extremely good, treble very easy on the ear with tons of deep bass. However, while burning in everything got progressively worse and after proper burning in they only gave a slight improvement over the Rubycons. Disappointing at first but showed us there was more to be had with the proper choice of parts!
Second up for test was a 0.1uF ceramic (Kemet multilayer, X7R dielectric) or 0.1uF film (Wima metallized polyester). Given that these caps was of equal size and often are used in similar applications I expected very similar results from these. How very wrong I was! The ceramics resulted in an almost explosive performance with bass like thunder and enormous dynamic capabilities, the treble was the culprit with no organic qualities and rather harsh. With the film it was quite the opposite, treble was very sweet, mids very organic but slam and bass extension were really lacking.

By this time we were really scratching our heads to figure out how to get the dynamics of the ceramics and the voices and treble of the films. We figured that since new Silmics=good, burned in Silmics=bad, burned in Rubycon ZA=bad that the combination of big (as in 47uF) and low ESR caps was not a good combination. Also, small sizes didn't seem to be able to cut it, at least not on their own.
So we tried something reasonably big with a bit more ESR, a 10uF tant. This proved to be a really good allrounder, with deep firm bass and ok mids and highs. I could have lived with the tants. But the thought of the magic-pixie-dust mids and highs of the films haunted us. The tants really gave what I feel is a house Naim sound, ie good pace and bass slam but lacking in conveying detail, nuances and organic qualities in the music.
So we tried one more thing: the 10uF tants bypassed with the 0.1uF films. This proved to be the answer (for us in our setups), but only after at least a week of burn in. The entire range from lowest lows to highest highs are in my ears just right. Very organic, realistic and gives you a real sense for the music. Pace, timing and temporal changes are very palpable. The best of everything tried and described above.

This tant/film combination also gave excellent results in my 102 where I have individual TeddyRegs fed by Tracos ala Mr Tibbs, other mods are identical to the ones in Gunnars preamp described above (Avondale A260Z power and Guru qm10 equivalent speakers). So they are not too setup specific.
Only one word of caution; don't judge them too early , they really do need at least a week to really start performing on their optimum. Since they have such an impact on the sound the process of burn in is very evident.

Cheers to Teddy for making these Regs public!

/Magnus

Ps The purpose of this ramble is not to proclaim the ultimate output capacitor of TeddyRegs, but encourage you to try different things and also combinations of caps. It would be really nice to hear from others trying this and other combinations and maybe discuss it further. For us, this combination gave sufficiently excellent results to not venture further (for now.....

)

sq225917 · #2 28-10-08, 04:17 PM

http://media.bestkiteboarding.com/mm/321-729-mods.jpg

i don't think there's anything that doesn't benefit from a bit of poly bypassing ;-)

bivalve · #3 28-10-08, 05:57 PM

Gunnar & Magnus,

A nice piece of research. I'm not surprised ZAs and Silmic IIs don't sound as good as alternatives. ZAs generate lots of distortion and Silmics become inductive at frequencies below 100kHz. Stargets are much better behaved but not as low ESR at low frequencies.

I would guess a 10uF film in series with a 1-3 ohm resistor would perform well.

David

Agent_Cooper79 · #4 29-10-08, 01:26 AM

Hi bivalve,

Yes, I was thinking about trying a group buy Wima in series with a small resistor but when bypassed tant worked so well I didn't bother. If you try it please report back, these things are always interesting!

sq229517, nice, are they yours?

/Magnus

Gunnar · #5 29-10-08, 01:34 AM

Thanks David.

As Magnus points out there are two hicaps in my set up and therefore comparison between the two were very easy. The winner stays and we could then try another contender.

What we also found is that there is a significant difference between film and cheramic in this context. So just plonking in a film or a cheramic thinking, ´they sound the same´is not what we found.

If not Magnus had the good idea with this combo I would have used the film.
Film appealed more to my right brainhalf than the cheramic.

The tant/film combo though wins in this survey over the strenghts of the film and the cheramic.

This survey only proves that in our set ups, the best choice is the tant/film combo by far.

Gunnar

sq225917 · #6 29-10-08, 09:00 AM

Magnus, thanks, yes they are my current cards. I have a feeling I will be playing with them some more in the near future. These things seldom stay stood still for long..

teddy_pardo · #7 29-10-08, 11:35 AM

Hi Magnus and Gunnar,

Very nice report. I was looking forward to read it as I spent much time as well doing the same kind of tests. My results are in-line with yours in most aspects.

There were some minor differences. I didn't hear a degradation with the Silmics, but maybe we are not using the same value (I used 100uF 50V).

Tantalum give a warm Naimie sound, no doubt about it. Some like it some don't.

Both Ceramic and Polyester give a very fast and dynamic sound, very transparent. Sometimes too much, but that depends on the rest of your system.

The combination of 10uF Tantalum and a small Polyester or Ceramic is a good mix. Actually I use this combination in many other places too (e.g. for decoupling of op-amps). This combination is also suggested in several application notes. The reason is that Tantalum capacitors have high ESR and relatively flat curve (unlike film that become a coil at high frequencies). This combination allows having large capacitance at low frequencies and no ringing at high frequencies. My friend Avi (AR_sound) suggests using 10nF film instead of 0.1uF, and I agree with him. If you are using Ceramic capacitors, the 10nF will probably sound much better than the 0.1uF (with Ceramic it's more critical than with Film).

Cheers, Teddy

flatpopely · #8 29-10-08, 01:53 PM

Magnus.

And I had just got to love the TRd 72, now I'm going to have to try some different caps.

Andrew

neiljadman · #9 29-10-08, 02:05 PM

Quote:

Originally Posted by flatpopely

... try some different caps...

How about a BC128?

Agent_Cooper79 · #10 30-10-08, 06:13 AM

Hi Teddy,

Thanks for your reply, nice to get more scientific details of why it is working. Bugger though, now I will have to try the 10nF films/ceramics....

Andrew, please try som different caps, it will be very interesting to see what results you get since you are not using the 27R of the input RC-filter to the stages. If you try a BC128 as suggested by Neil, please report back.

/Magnus

teddy_pardo · #11 30-10-08, 09:16 AM

I know that Avi tried the BC128 (for lower voltage circuit) and say they are the best. The problem is that they are 25V, so it's risky. I already had problems with BC128 running near their limits, so be careful...

Agent_Cooper79 · #12 30-10-08, 10:27 AM

Same old stupid question... but what would you consider too close to the limit, would ca 21V be ok or is that too high?

/Magnus

neiljadman · #13 30-10-08, 10:35 AM

BC128s run on a 24v circuit in the 321 quite happily - though what the effective voltage is who knows - time to get the DMM out...

hacker · #14 30-10-08, 10:50 AM

Tants don't fair so well on the edge. I'm currently building Starfish v2 and I just had 3 tants fail short-circuit in quick succession. They were 16V rated parts running at 15V.... No explosions, but short circuits are not good! I moved to 25V parts to be safe.

Gunnar · #15 01-11-08, 02:35 PM

Quote:

Originally Posted by teddy_pardo

Hi Magnus and Gunnar,

Very nice report. I was looking forward to read it as I spent much time as well doing the same kind of tests. My results are in-line with yours in most aspects.

There were some minor differences. I didn't hear a degradation with the Silmics, but maybe we are not using the same value (I used 100uF 50V).

Tantalum give a warm Naimie sound, no doubt about it. Some like it some don't.

Both Ceramic and Polyester give a very fast and dynamic sound, very transparent. Sometimes too much, but that depends on the rest of your system.

The combination of 10uF Tantalum and a small Polyester or Ceramic is a good mix. Actually I use this combination in many other places too (e.g. for decoupling of op-amps). This combination is also suggested in several application notes. The reason is that Tantalum capacitors have high ESR and relatively flat curve (unlike film that become a coil at high frequencies). This combination allows having large capacitance at low frequencies and no ringing at high frequencies. My friend Avi (AR_sound) suggests using 10nF film instead of 0.1uF, and I agree with him. If you are using Ceramic capacitors, the 10nF will probably sound much better than the 0.1uF (with Ceramic it's more critical than with Film).

Cheers, Teddy

Hi Teddy!

Thanks for your input, much appreciated.

The tant alone gave an all to much of a grey sound presentation.
I though it was life less.

I must once more point out the difference between alternative choices in this test.

As there is probably no such thing as a musical sounding pre amp (It´s the musicians who are musical, not the amps

)
I would in my humble opinion state that the alone tant is not as transparent as the hybrid, the one we came find was best one. I would have chosen the film but with a loss of low end grip and authority.

The hybrid might simply let the music through in the best way of the configurations we have tried.

I´m fully aware that this is not to be considered as complete survey as it must be dependent on many variables, but Magnus and my pre, psu and amps are absolutely on par before we started this. So somekind of consensus might be drawn.

One consensus though is that a Hi Cap can be vastly improved upon.

What I would really like is a consensus that optimises this amazing regulator.

Best regards