OTOH you could try to use these errr..varicaps to generate a lot of distortion like the "high end" tube guys love. H2,H4 would be nice but knowing my luck, the non-linearity will give H3,H5...etc out to VHF. Darn!
Its ironic, you know. The electros distort signals when they have no polarizing DC and ceramics distort themselves with it. You'd think they'd get together and cut a deal for linearity.
In my dreams you say?
There aren't many opportunities for most of us to test sound gear, let alone capacitors, with real instruments. You would seem to be uniquely placed there, though I imagine trying higher frequency percussion stuff like triangle, 'sandpaper', maybe harpsichord harmonics etc. that might be appropriate to focus on the upper registers, would be a bit weird to test. I can imagine the neighbours peeking to see what we're doing whacking a triangle at a mic. at 4 a.m. heh heh!
You're right about the effect of the 5uF being swamped by the 5,000 of electro - at frequencies below those where the impedance of the ceramic becomes substantially lower than the electro. At that point, most signal is now flowing via the ceramic and its capacitance is prone to modulation by the AC voltage across it. It's then a non-linear device at AF. In RF, correctly applied, they can be less of a problem. In amps and general AF work they do fine on PSU and rail bypass where such effects are no drama. BTW, if you crack open an FM or AM ceramic fiter you wont see a lot of difference. That should say something. I don't think resonance is an issue here though. Pehaps there is another view...
Its ironic, you know. The electros distort signals when they have no polarizing DC and ceramics distort themselves with it. You'd think they'd get together and cut a deal for linearity.
In my dreams you say?
There aren't many opportunities for most of us to test sound gear, let alone capacitors, with real instruments. You would seem to be uniquely placed there, though I imagine trying higher frequency percussion stuff like triangle, 'sandpaper', maybe harpsichord harmonics etc. that might be appropriate to focus on the upper registers, would be a bit weird to test. I can imagine the neighbours peeking to see what we're doing whacking a triangle at a mic. at 4 a.m. heh heh!
You're right about the effect of the 5uF being swamped by the 5,000 of electro - at frequencies below those where the impedance of the ceramic becomes substantially lower than the electro. At that point, most signal is now flowing via the ceramic and its capacitance is prone to modulation by the AC voltage across it. It's then a non-linear device at AF. In RF, correctly applied, they can be less of a problem. In amps and general AF work they do fine on PSU and rail bypass where such effects are no drama. BTW, if you crack open an FM or AM ceramic fiter you wont see a lot of difference. That should say something. I don't think resonance is an issue here though. Pehaps there is another view...
You have to Qualify that John. Not NPO - they have great temp stability and do not suffer from any of the dielectric and DF problems of the other ceramic types.
Have to smile at that
I have seen and worked on so many intermitent faults caused by small compressed disc ceramics over the years... they have something of a deserved reputation. I know they come in all different flavours but once bitten and all
They can be good, but when they are bad, they are very very bad.
Well thanks for the details on HF caps. I'm using vile Z5U dielectric Aerovoxs, at $7 each and since they are made in Mexico there are not likely to be quality problems, but as far as Audio Frequency modulation of capacitance I don't know. I have some left over from the organ recap project is the main reason, also as a former digital (TTL STTL) designer, I was taught to put ceramic caps everywhere to cover switching transients. I've designed hundreds of .01 uf ceramics to go in boards in the JSC communications room (1979). NPO 10 uf 50V Aerovox ceramics are $17 each and weren't available in stock when I recapped the organ last winter. The ceramic cap has the wonderful feature of fitting where electrolytics used to be. I have 1 each 10 uf 63 V metalized polyester cap left over, otherwise the next size is some .022 uf polypropylenes left over from the PAS2 recap project. Or a lot of 1960 vintage paper .1 ufs removed from the PAS2. Funny the ST120 sounds so good with the Z5U 5 uf ceramic at the input, but the imput impedance is so high with the hi-gain input transistor (BC108A original Dynaco) that maybe impedence modulation doesn't matter much. Experiment paused until I need to spend $6 shipping something from Newark in NC. Also I need to dissect this 555 containing pcb marked "parts only-defective" and figure out what I need to do to change it to an AF square wave oscillator with .5-1.2 VAC output. Also a second scope probe would be handy, this HH Smith probe I got for $30 twenty years doesn't seem to be made anymore.
I make a lot of weird noises at 1AM in the morning. I live across the street from the industrial park and 50' from the nearest house with 1" plaster walls, so I play anything as loud as I want anytime I want. Scott Joplin on the Steinway, Pictures at an Exhibition as hard as I can hit the piano, JS Bach with the go pedal all the way down and the bass pot in the preamp turned all the way up. The H182 has a 15" woofer and a 40 W 7591 woofer amp, I can drown out the cars driving by with the boom tubes. Joan Jett or ZZ Top on the SP2's, also Moussorski Bald Mountain last night. I got my training in a high school band that marched 12 tubas and 24 drummer with twin bass drums and 10 tenor drums, there is no such thing as too much bass.
I make a lot of weird noises at 1AM in the morning. I live across the street from the industrial park and 50' from the nearest house with 1" plaster walls, so I play anything as loud as I want anytime I want. Scott Joplin on the Steinway, Pictures at an Exhibition as hard as I can hit the piano, JS Bach with the go pedal all the way down and the bass pot in the preamp turned all the way up. The H182 has a 15" woofer and a 40 W 7591 woofer amp, I can drown out the cars driving by with the boom tubes. Joan Jett or ZZ Top on the SP2's, also Moussorski Bald Mountain last night. I got my training in a high school band that marched 12 tubas and 24 drummer with twin bass drums and 10 tenor drums, there is no such thing as too much bass.
Last edited:
High value Ceramic is crap. Please measure them for DISTORTION at low frequencies or IM, they have 'non return to zero, and very high DA. Indianjo, you would do better with an aluminum cap. I published this in 1978. For DA, look at Walt Jung's website. There is no worse cap than a high value ceramic.
the numbers suggest bypassing a 5000-10000 uF output coupling cap with a few uF of anything is not likely to have audible effect
if it does it just means you've suceeded in coupling in RF pollution to your feedback loop - because it has no meaningful effect on audio frequencies where the esr of the big electros is milliOhms and |Z| of 1 uF is 16 Ohm @ 10 KHz - remember you have to divide the "effect" again by the esr/R_speaker and you easily get -100 dB and that only at the highest audible frequencies
if it does it just means you've suceeded in coupling in RF pollution to your feedback loop - because it has no meaningful effect on audio frequencies where the esr of the big electros is milliOhms and |Z| of 1 uF is 16 Ohm @ 10 KHz - remember you have to divide the "effect" again by the esr/R_speaker and you easily get -100 dB and that only at the highest audible frequencies
Thanks J Curl, your LF distortion results for ceramic caps did not make wikipedia, which warned me of microphonics on ceramic caps beofre I bought them. On the loud Hammond organ recapped with 10 uf Z5U ceramics, microphonics so far not a problem. Distortion on a sine wave addition organ may be audible but not obvious; what is an electronic organ "supposed" to sound like. Everybody on the internet tells me a Hammond H100 is tinny and lacks bass, which means to me they haven't changed their electrolytic caps since 1968. JCX, this3300- 5000 uf output cap is outside the feedback loop, outside the terminals of the amp. (I was proposing 10 uf, not 1uf). JW Erath built speakers that had the optically sensed feedback from the woofers of speakers into the amp feedback loop, but including feedback for speaker response for treble would inolve microphones and bit bucket delay lines to align the phase with what is going out, not really a viable idea considering the slow speed of sound in air, and that music involves many transients, not just sine waves.
A short story for everyone else: Back 1971, I was trying to improve my Fisher FM 60 tuner with improved output caps. What did I choose? A 2 uF monolythic ceramic. Why not?. Yes, I got improved bass transmission. However, over time, I got listeners fatigue. Yes, that's what the cap gave me. Later, in 1974, I saw a graphic reproduction of what ceramics actually do. They actually have hysteresis. Non-return to zero. It was on the screen. No other cap does that in any significant way. IF you want hi fi, go for fidelity. If you want an effects generator, then maybe ceramic might be an interesting way to go.
Z5Us are really horrible if the voltage across them varies, they are only useful for logic supply decoupling and hf switched mode.
Any decent electrolytic has negligible inductance at audio frequencies. Poor track design and long flying leads can easily have orders of magnitude higher inductance.
Any decent electrolytic has negligible inductance at audio frequencies. Poor track design and long flying leads can easily have orders of magnitude higher inductance.
Hi Moly, your explanation is very good but i think it's is incomplete yet.
For your split supply and dc coupled amp figures, can you illustrate the complete AC signal Path (also in the power supply)?
Then you must redraw this AC Signal path, if you substitute the input symmetrical signal with a unsymmetrical signal.
Now the result is very different, i think.
and that way is more important because an audio amp not manages symmetrical signal or at least proper circuit distortion transform the signal to unsymmetrical.Francesco.
Last edited:
Hi mos57
Oh heck
So why not try it and see for yourself
Do you think the output signal would be distorted in some way if for example I switched the 50/50 squarewave to say 90/10 ratio ?
Hi Mooly!!!
I am going OT, i think, because my consideration was mainly for class A bias circuit and not for class B as you seems suppose it.
But analisys of signal paths for class A, class B, simmetrical and not symmetrical input signal was very interest.
The psu caps have different importance
Sorry but now i not have the time to prepare a real circuit.
Have a good day, Francesco
Read the first four pages and wanted to reply. Mooly has it right. If the PSU caps are in the signal path why dont the relatively huge charging spikes get in the output signal? These caps are buffered to the output by the output transistors. As stated earlier this is the PSRR, and its usually very high for well designed amps. The one thing that may be audible with PSU caps is increasing there value to increase the amps peak current output. (but this also depends on the size of the transformer).
The thread
Well folks, its been great to read the examination of most of Gordy's above posers. The contributions have, I'm sure, been appreciated. Would anyone care to step up to offer their findings on servo substitution or even dare to utter (qualified I suggest) subjective opinion on electrolytic types in the o/p of power amps?
I conclude from the thread that these are the only practical type of power o/p capacitor. There are many flavours of electro apart from the already treated bipolar construct. Low ESR, High ripple current, AC power etc. may be a consideration for optimum sound. Of course there are brand and graded series of audio capacitors but this may be asking for a bun fight on just that brand/undefined type preference.
But the views are......
Well folks, its been great to read the examination of most of Gordy's above posers. The contributions have, I'm sure, been appreciated. Would anyone care to step up to offer their findings on servo substitution or even dare to utter (qualified I suggest) subjective opinion on electrolytic types in the o/p of power amps?
I conclude from the thread that these are the only practical type of power o/p capacitor. There are many flavours of electro apart from the already treated bipolar construct. Low ESR, High ripple current, AC power etc. may be a consideration for optimum sound. Of course there are brand and graded series of audio capacitors but this may be asking for a bun fight on just that brand/undefined type preference.
But the views are......
Last edited:
Hi people, what i try to tell you is that after Mooly's explanation, we can think that psu capacitors, not being in the signal path, they not are important for sound reproducion: Its' false!!!! 
In the past years i have substitute the two Mallory 10.000 microF. psu caps in my Spectral DMA50 (50W class B complementary amp) with these others:
1)LCR
2)SPRAGUE
3)ROE
4)JENSEN 4 POLE.
The sound was very different for each type of caps, the influence was for me as strong as series output caps!!!
So we need a better and a deeper investigation about.
Francesco.
In the past years i have substitute the two Mallory 10.000 microF. psu caps in my Spectral DMA50 (50W class B complementary amp) with these others:
1)LCR
2)SPRAGUE
3)ROE
4)JENSEN 4 POLE.
The sound was very different for each type of caps, the influence was for me as strong as series output caps!!!
So we need a better and a deeper investigation about.
Francesco.
Ok, but my last comment was derived from fig.1 in post 111.
Indianajo, i think that ps supply caps in this tipical circuit aren't a simply "filter" element and my experience wrote in the post above, confirm this.
So it's difficult separate ps caps from output caps
Speaking rigorous, nobody answered to Gordy questions, seems to me!!!!
Indianajo, i think that ps supply caps in this tipical circuit aren't a simply "filter" element and my experience wrote in the post above, confirm this.
So it's difficult separate ps caps from output caps
Speaking rigorous, nobody answered to Gordy questions, seems to me!!!!
Gentlemen, it has been a very interesting thread. Overall the conclusion seems to be that there is no conclusion, and no definite agreement as to impact of output capacitors.
I will be experimenting with the output of a power follower shortly. As for sound quality of PSU caps I feel that it is a potentially complex subject as it is difficult to separate the cap from the system, as they work as one whole network. Indeed it would need to be assessed with and without regulation, with different types of rectifier, different configurations of active circuits, etc., etc.. This suggests a separate thread, but not one for me to start.
As for here, then let conversation continue as however you please. I think Ian's call for opinions on output servos (and their impact on the substitution for capacitors) is a sensible request...
... so any opinions?
I will be experimenting with the output of a power follower shortly. As for sound quality of PSU caps I feel that it is a potentially complex subject as it is difficult to separate the cap from the system, as they work as one whole network. Indeed it would need to be assessed with and without regulation, with different types of rectifier, different configurations of active circuits, etc., etc.. This suggests a separate thread, but not one for me to start.
As for here, then let conversation continue as however you please. I think Ian's call for opinions on output servos (and their impact on the substitution for capacitors) is a sensible request...
... so any opinions?
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.