Wavebourn,
A very interesting article you posted. It seems there are a lot of considerations that can also be linked to frequencies outside of the hearing range linked to harmonics and perception of sound. It will be interesting to see what you find!
If anyone is having difficulty opening the doc just right click the link and save as!
Regards
M. Gregg
A very interesting article you posted. It seems there are a lot of considerations that can also be linked to frequencies outside of the hearing range linked to harmonics and perception of sound. It will be interesting to see what you find!
If anyone is having difficulty opening the doc just right click the link and save as!
Regards
M. Gregg
Thanks for re-summarizing.
I think this is a fascinating topic and an explanation is eminently desireable. Yes, we could accept it on faith and just bypass everything as a matter of policy, but I for one am reluctant to either accept the end results OR a casual explanation (no offense Anatoliy) at face value. I want to know why it happens or at least know the circumstances of the effect being noticeable (for now).
Something like this, at least the currently offered explanation, should be testable.
Think of this; if there is something to it, an AES paper is clearly warranted. In the context of scholarly work, what we have now is a strawman.
Cheers,
Michael
Thank you Michael; but I believe AES members and guests already many times heard about non-linear capacitances, phase inter-modulations, and binaural perceptibility thresholds. And, I don't like to write complex documents by requirements, and to read lectures. Let's people who don't know how to do real things continue teaching others. 
By the way, couple of rows from your stand on AES convention guys demonstrated some rack mount cabinet with glass window on it's door. It was deadly silent, despite of extremely noisy computer inside. Have you seen it?
By the way, couple of rows from your stand on AES convention guys demonstrated some rack mount cabinet with glass window on it's door. It was deadly silent, despite of extremely noisy computer inside. Have you seen it?
I think there's certainly an opportunity for the audio current vector of the electrolytic what with it's dielectric absorption, ESR, ESL etc
to be such that it leaves room for the smaller film cap to become significant. A measurement would be simple. Someone want to borrow my tektronix clamp-on current probe?Meteor was a monoblock. Come to think of it, all my finished amps are monoblocks. There's an interesting question right up front whether this effect only occurs in a stereo amp with common B+, or also in amps with separately powered channels e.g. mono or "dual mono" or even amps with separate final LC filters. If the effect is specific to amps with common B+ rail, then I think that takes it in a somewhat different direction.
If putting the cap behind the inductor still makes a difference, I'd start looking at noise...
Michael
Just for interest,
If you use mono blocks would the effect be greater with each channel using different supply caps?
Regards
M. Gregg
Thank you Michael; but I believe AES members and guests already many times heard about non-linear capacitances, phase inter-modulations, and binaural perceptibility thresholds. And, I don't like to write complex documents by requirements, and to read lectures. Let's people who don't know how to do real things continue teaching others.
By the way, couple of rows from your stand on AES convention guys demonstrated some rack mount cabinet with glass window on it's door. It was deadly silent, despite of extremely noisy computer inside. Have you seen it?
Yes, we've all heard about these in isolation. But connecting them all together in this way is still a logic strawman until some corroborating evidence is produced.
Also, I guess I need my hand held through the logic:
In what way is the capacitance non-linear? AC current vs. voltage nonlinearity is all I can think of, except for maybe how capacitance varies with applied DC voltage (?)
How does non-linear capacitance lead to phase inter-modulation?
How does phase inter-modulation create delays that spatially de-localize the stereo image?
But if all this is patently obvious to everyone and I'm being especially dense, please forgive me.
Anyway, I think this is what I would expect to see explained along with evidentiary measurements, if there were a paper to be written.
If there is something there, it would amount to new knowledge and would impact the way (some) amplifiers are designed.
---
I spoke with the iso cabinet folks and got a follow-up phone call already. I diy'ed my own some years ago with good success, and they are doing basically the same. The walls of the cab are heavy MDF with possible extra layers, absolutely sealed joints and bracing as with a speaker cab, labyrinthine air path and internal low speed fans. Air inlet and exhaust underneath. The USB/FW drives need to go in the cab also; they're noisy! I supplemented the DAW with a networked fanless PC with CD burner.
It's a good formula and in my experience the easiest way to set up a one-room DAW based recording studio. (sorry for the OT but he asked...)
Last edited:
Member
Joined 2009
Paid Member
Yes, post 186 is the current as-built version:
http://www.diyaudio.com/forums/tubes-valves/167872-my-cellini-triode-amp.html
You see, no bypass caps used anywhere. I have some caps I can use, I think they are high voltage 1uF MKP.
Wait...
Actually; I have some more non-linear capacitances...
800V source is stacked 2 x 400V sources.
Screen grids are powered by +270V stabilized from 400V rectifier.
Drivers are powered from 400V stabilized from 800V rectifier.
Stabilizers are parametric, with source followers.
Drain of 400V stabilizer is connected to 800V, drain of 270V stabilizer is connected to 400V.
That means, already non-linear dynamic capacitances of MOSFETs are connected in parallel with filter electrolytics, so if electrolytics have plain ESR and ESL only, MOSFETs may impact...
Actually; I have some more non-linear capacitances...
800V source is stacked 2 x 400V sources.
Screen grids are powered by +270V stabilized from 400V rectifier.
Drivers are powered from 400V stabilized from 800V rectifier.
Stabilizers are parametric, with source followers.
Drain of 400V stabilizer is connected to 800V, drain of 270V stabilizer is connected to 400V.
That means, already non-linear dynamic capacitances of MOSFETs are connected in parallel with filter electrolytics, so if electrolytics have plain ESR and ESL only, MOSFETs may impact...
If adding 1% makes a difference, then presumably the effect is at the upper end of the frequency range where the electrolytic impedance begins to rise due to some combination of inductance and dielectric losses. 2uF would not have much effect below 10-20kHz unless electrolytics are much worse than I/we think.
If it is RF, then 0.1uF might work just as well. Does the improvement by adding a bypass vary according to how far away is the nearest LW/MW broadcast transmitter?
If it is RF, then 0.1uF might work just as well. Does the improvement by adding a bypass vary according to how far away is the nearest LW/MW broadcast transmitter?
As I see it, its not about the percentage, but about the dynamic impedance at any given moment. If the impedance of the film cap at a certain moment for a certain frequency is 10% of the impedance of the electrolitic, then 90% of that signal is flowing through the filmcap.
I really wonder what current flows through the filmcap.
Ever did that measurement, wavebourn? Fft on that signal with pink noise would be interesting...
Also, is the single powersupply directly connected to the centertaps of the xformers, without an own capacitor close to it?
Member
Joined 2009
Paid Member
I tried the MKP audiophool caps. I put one across each of the last electrolytics on the B+ supply (I have LCRC) and across each of the cathode bypass electrolytics on the output tube. I didn't bypass the input tube.
No obvious change in sound. It must be too subtle for me to hear it if there has been an improvement.
No obvious change in sound. It must be too subtle for me to hear it if there has been an improvement.
It is cold outside, and I feel not so well to carry the amp to my barn where a pile of measurement equipment is, ... I will do it later.
Yes, CTs are connected directly. Also, as I said before, drains of MOSFETs for +400 and +270 voltage stabilizers as well. They may be guilty! They should have on their drains dynamic capacitances equal to their amplification factors multiplied by drain to gate capacitances.
Ok. Well, Ummmm... Since this thread is kind of related to your problem I'll do this reply here. You've already taken a lot of heat over on your other thread about the 6as7 I see. I didn't go through all those posts to see if anyone touched on the OPT. I have deep sysmpathy for you and your amp. But Hammond explicitly says that OPT is not Hi-Fi material..... Ummmm...So... I doubt a bypass cap is going to undo that sitchiation...
On a lighter note, I lost a canoe out on James Bay at the mouth of the Moose 20 miles from Moosonee and could have been eaten by bears and wolves on my 2 day hike back to town!
Touche! There's no substitute for good iron, as I have found out (the hard way).
I haven't read the "Cellini" thread personally, but this link is pretty illustrative of 20to20's point. Not only does the manufacturer "gird its loins" through its "they're not exactly hi-fi" disclaimer, but it also seems to have trouble with consistency from unit to unit.
I personally own a pair of 125ESE's, but I use them for SE experimentation purposes only. I figure if a circuit sounds pretty good through a pair o' these then it'll be absolutely stunning through a pair of James or Lundahls.
At least that's how it's supposed to work...
I haven't read the "Cellini" thread personally, but this link is pretty illustrative of 20to20's point. Not only does the manufacturer "gird its loins" through its "they're not exactly hi-fi" disclaimer, but it also seems to have trouble with consistency from unit to unit.
I personally own a pair of 125ESE's, but I use them for SE experimentation purposes only. I figure if a circuit sounds pretty good through a pair o' these then it'll be absolutely stunning through a pair of James or Lundahls.
At least that's how it's supposed to work...
Looking at the impedance of a KMH series 68uF 450V al cap, it has only 3.03 Ohms 120Hz, and Impedance vs freq shows that it will be about .1 that at 10KHz and flatten out to beyond 100KHz.
The B32923E X2 2.2uF 305VAC cap is rated at 10 Ohms at 10Khz and drops to 1 Ohm at 100KHz. Extrapolating back on the graph it would have 100 hms impedance at 1KHz.
So the film cap has higher Z out to 100KHz but drops to less than the AL until inductance starts the plot back up and is above there at 1MHz.
Unfortunately the specs aren't given the same way and it takes some digging to try to figure out what the plots mean.
It looks to me like the low impedance and low ESR caps for switching supplies are actually better than the X2 film caps (MKT and MKP) as far as impedance vs frequency out to 100KHz or possibly further.
That said, I still use a 1uF, 1.5uF, or 2.2uf X2 cap across my main Al filter caps, just because I have them.
The B32923E X2 2.2uF 305VAC cap is rated at 10 Ohms at 10Khz and drops to 1 Ohm at 100KHz. Extrapolating back on the graph it would have 100 hms impedance at 1KHz.
So the film cap has higher Z out to 100KHz but drops to less than the AL until inductance starts the plot back up and is above there at 1MHz.
Unfortunately the specs aren't given the same way and it takes some digging to try to figure out what the plots mean.
It looks to me like the low impedance and low ESR caps for switching supplies are actually better than the X2 film caps (MKT and MKP) as far as impedance vs frequency out to 100KHz or possibly further.
That said, I still use a 1uF, 1.5uF, or 2.2uf X2 cap across my main Al filter caps, just because I have them.
Member
Joined 2009
Paid Member
Ok. Well, Ummmm... Since this thread is kind of related to your problem I'll do this reply here. You've already taken a lot of heat over on your other thread about the 6as7 I see. I didn't go through all those posts to see if anyone touched on the OPT. I have deep sysmpathy for you and your amp. But Hammond explicitly says that OPT is not Hi-Fi material..... Ummmm...So... I doubt a bypass cap is going to undo that sitchiation...
On a lighter note, I lost a canoe out on James Bay at the mouth of the Moose 20 miles from Moosonee and could have been eaten by bears and wolves on my 2 day hike back to town!
You mean, it's possible to make a tube amp that sounds even better ? - well this isn't so much bad news as great news - as I'm already enthralled with the sound I have !
I have noticed something has changed though, now that the amps warmed up. It ain't the stereo imaging that's changed, but something else. It seems that the background noise has been reduced. I'm wondering if these extra caps have removed a bit more noise from the supply rail.
As for the Hammond iron - well I'm certainly a novice with regards these things but I did take some interest in these measurements: Square waves for Hammond 125ESE
p.s you definitely have to be careful of the bears !
Last edited:
Hmmm. I put known series resistance with elko caps, why?
The evil you know and trust, vs the one you don't and can't.
Paralleling different cap values, flavors, and colors helps too.
Layered in polywanncrackoline and/or Saskatchewan sealskin.
1uF appears a drop in the bucket to the overall power supply,
but might have low Z at some frequency that is troublesome
for bigger rolls of wet paper and foil.
The evil you know and trust, vs the one you don't and can't.
Paralleling different cap values, flavors, and colors helps too.
Layered in polywanncrackoline and/or Saskatchewan sealskin.
1uF appears a drop in the bucket to the overall power supply,
but might have low Z at some frequency that is troublesome
for bigger rolls of wet paper and foil.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.