Conrad, one further point for your research, you might want to glance at this website. I found it to be an excellent resource.
Tom- good site! I compiled my values from a bunch of data books, web sites, and a few measurements. I'm sure manufacturers tend to specify maximums- my measurements are invariably lower. Mica is a unique case, since the raw material varies so much. I suspect the films vary in composition, especially the ubiquitous "plastic film" sold by Panasonic. No Teflon caps here to test, as I don't know where I'd use the available values to any great advantage. They'd seem to be ideal for an RIAA phono preamp, but I don't know if the necessary values could be easily had. Just curious how you measure those low DFs. My 1615 goes down to 0.000001, but I'd only attempt that with the cap installed in a completely shielded enclosure, and at a pretty high voltage. Even then I wouldn't trust the measurement to better than plus or minus a few counts. I also have a phase comparator that could do it in theory, but I've never tried it. Not sure what's happening with the background- here the text shows up with very high contrast on medium gray with a 17" LCD, and is plenty readable, but monitors and peoples adjustments do vary.
Conrad, I have a Quadtech 1920 LCR meter. For best results, I measure caps using a 2V bias voltage in a copper can I've grounded to the 1920. Even then, DF measures bounce around as much as a magnitude between successive measurements.
I've posted some of Bateman's findings here. I've also posted DA measurements here using an admittedly compromised test rig (a DMM), though the results still provide useful information.
I've posted some of Bateman's findings here. I've also posted DA measurements here using an admittedly compromised test rig (a DMM), though the results still provide useful information.
Conrad Hoffman: (you da man!!) " ... I have a somewhat incomplete table I'm in the process of fleshing out, and will put up a link to that soon. ... ... If you've got more into to fill in, or see some obvious errrors, let me know. ..." Link: http://members.rpa.net/~choffman/cap_losses.htm
serengetiplains: (... what i'm talkin 'bout!!) " ... some of Bateman's findings here. [Link: http://www.diyhifi.org/forums/viewtopic.php?f=13&t=388 ] ... I've also posted DA measurements here. [Link: http://www.diyhifi.org/forums/viewtopic.php?f=15&t=550 ] ... "
(Appologies to all for the redundancies, but some folks out in the hinterlands live behind firewalls that may block hidden web link references or and some others who may wish to run the links directly through their favorite language translators.)
Now ... suggestions for the wording of the DIYAudio Wiki definations: should we just edit, amend or append the Wikipedia entry or make our own from scratch ?? [Link: http://en.wikipedia.org/wiki/Capacitor ] ...
serengetiplains: (... what i'm talkin 'bout!!) " ... some of Bateman's findings here. [Link: http://www.diyhifi.org/forums/viewtopic.php?f=13&t=388 ] ... I've also posted DA measurements here. [Link: http://www.diyhifi.org/forums/viewtopic.php?f=15&t=550 ] ... "
(Appologies to all for the redundancies, but some folks out in the hinterlands live behind firewalls that may block hidden web link references or and some others who may wish to run the links directly through their favorite language translators.)
Now ... suggestions for the wording of the DIYAudio Wiki definations: should we just edit, amend or append the Wikipedia entry or make our own from scratch ?? [Link: http://en.wikipedia.org/wiki/Capacitor ] ...
FastEddy- I like Wikipedia. I use it quite often and am impressed with the quality of most of the content. Therefore I want it written by people way better than me! Maybe we can come up with something as a group effort, but I foresee a big problem the minute we try to say anything definitive about the way anything sounds. Far out at one end (close to where I stand) is the view that the cap distortions (other than electrolytics) are inaudible. At the other end are the people who claim to hear the difference between Teflon and polystyrene. My guess is the circuit the cap lives in has a huge amount to do with what you hear.
OTOH, in another few minutes I could change my whole science based view of things. I just rebuilt an amp with much higher quality parts, improved the stability and lowered the distortion. I'm about 80% sure it now sounds worse than it did before. Lifeless is about the best description I can offer. I'll spend tonight at the bench trying to figure out why. My initial guess is that I added a miller cap to better stabilize the thing, and that may have been a mistake. Fortunately the only other cap of real significance is the 220uF in the feedback divider (no couplers), and that's too large to do much about. Maybe I'll make it larger still. Just think, if it were Teflon, it would be bigger than the whole rest of the amp!
OTOH, in another few minutes I could change my whole science based view of things. I just rebuilt an amp with much higher quality parts, improved the stability and lowered the distortion. I'm about 80% sure it now sounds worse than it did before. Lifeless is about the best description I can offer. I'll spend tonight at the bench trying to figure out why. My initial guess is that I added a miller cap to better stabilize the thing, and that may have been a mistake. Fortunately the only other cap of real significance is the 220uF in the feedback divider (no couplers), and that's too large to do much about. Maybe I'll make it larger still. Just think, if it were Teflon, it would be bigger than the whole rest of the amp!
Conrad: " ... Therefore I want it written by people way better than me! Maybe we can come up with something as a group effort ..."
Likewise ...
" ... I foresee a big problem the minute we try to say anything definitive about the way anything sounds. ..."
Is thus and ever shall be ... Opinions can be put in quotes or labled as a "subjective opinion". I have no problem statements regarding quality. Example of a subjective opinion:"Because of the higher frequency response and recovery time of polystyrene snubbing capacitors (in parallel with larger electrolytic capacitors), the results are generally better than when left out of the power supply circuit." - FastEddy.
" ... I just rebuilt an amp with much higher quality parts, improved the stability and lowered the distortion. I'm about 80% sure it now sounds worse than it did before. Lifeless is about the best description I can offer. ..."
I have a tube pre-amp between my DVD-A player and my MOSFET power amp for this and other reasons ... Years ago methods to "color" solid state equipment included the introduction wire wound carbon resistors in the audio signal path to make them sound more like the "mellow" sound of a tube amp. Sony did it. ...
Likewise ...
" ... I foresee a big problem the minute we try to say anything definitive about the way anything sounds. ..."
Is thus and ever shall be ... Opinions can be put in quotes or labled as a "subjective opinion". I have no problem statements regarding quality. Example of a subjective opinion:"Because of the higher frequency response and recovery time of polystyrene snubbing capacitors (in parallel with larger electrolytic capacitors), the results are generally better than when left out of the power supply circuit." - FastEddy.
" ... I just rebuilt an amp with much higher quality parts, improved the stability and lowered the distortion. I'm about 80% sure it now sounds worse than it did before. Lifeless is about the best description I can offer. ..."
I have a tube pre-amp between my DVD-A player and my MOSFET power amp for this and other reasons ... Years ago methods to "color" solid state equipment included the introduction wire wound carbon resistors in the audio signal path to make them sound more like the "mellow" sound of a tube amp. Sony did it. ...
Whew! Almost two in the morning because I couldn't stop listening. Fixed the amp. As you know, I'm into differential comparisons. The amp was simply producing more difference between input and output than it should have. Distortion was reasonable, though not terrible low. Response was ok. Not sure what name to put on the error yet. The culprit was the miller cap that I added because I couldn't get the thing completely stable. Finally I changed the whole ground strategy that the amp has had for the last 20 years; that made it stable and I removed the miller cap. In/out difference went down and the life came back. Couldn't hear a significant difference when I shorted the output inductors. Doug Self has a good point about the electrolytic in the typical feedback network- select it for THD, not for the low frequency rolloff. I need to make mine bigger, if only to reduce the in/out difference due to phase shift. IMO, he doesn't have a good point about miller caps- they aren't an elegant solution to several problems, but a problem waiting to happen. Maybe that's why people have been less than enthusiastic about blameless amps? Not that this amp is anywhere near blameless 
FastEddy, I assume you mean of the grounding? No diagram, but it's simple. The original scheme had the common heavy copper wire between the filter caps. Everything was single pointed to the center of that. I ran a stub off the center point to get away from the cross currents, but that didn't help a bit. The real problem was that the caps are pretty tall, and the PCBs are quite a ways away. Bringing back grounds for the PCBs (which have split grounds), the chassis, the bypass caps, all separately was no good. I'm sure the wire inductance was brutal, and the system essentially had no high frequency ground. What I did was establish a new star point right between the PCBs, near the amp inputs. Now the local bypassing works properly and the thing is stable as a stone. I made one other important change based on early and later schematics by SWTP (this is a copy of their old "Tiger" amp). The early amps had a 220pF cap from the base of the diff amp input to ground. The later amps were identical, but inserted a 1K resistor in the base lead. This seems to help stability quite a bit. I've seen the same thing with op-amps, where a cap to ground on the input will give rise to intractable oscillation.
Conrad: " ... The real problem was that the caps are pretty tall, and the PCBs are quite a ways away ... [I established] a new star point right between the PCBs, near the amp inputs. Now the local bypassing works properly and the thing is stable as a stone. ..."
Effectively balancing distance and ground path impedence between the PCBs ... You seem to be well in advance of my previous rant(s): " ... Distance and timing (speed of electron flow and signal), conductor capacitance and resistance variations ... all resulting in an added contribution to resistance, inductance and capacitance variations ...", your clear thinking, bench work and intuition making up for and being better than "book learnin'".
" ... a 1K resistor in the [cap] base lead ... the same thing with op-amps, where a cap to ground on the input will give rise to intractable oscillation. ..." I believe that I have also seen this in some of the older op-amp cookbooks ... the theory being that there was "reflection" or "a bounce back" from certain types / sizes of caps, re-enforcing feedback ... or some such (I might be wrong here).
(Speaking of books, you might take a peek at "The Black Swan: The Impact of the Highly Improbable" by Nassim Nicholas Taleb, new at Amazon, et al ... definately good reading for DIY types.)
Effectively balancing distance and ground path impedence between the PCBs ... You seem to be well in advance of my previous rant(s): " ... Distance and timing (speed of electron flow and signal), conductor capacitance and resistance variations ... all resulting in an added contribution to resistance, inductance and capacitance variations ...", your clear thinking, bench work and intuition making up for and being better than "book learnin'".
" ... a 1K resistor in the [cap] base lead ... the same thing with op-amps, where a cap to ground on the input will give rise to intractable oscillation. ..." I believe that I have also seen this in some of the older op-amp cookbooks ... the theory being that there was "reflection" or "a bounce back" from certain types / sizes of caps, re-enforcing feedback ... or some such (I might be wrong here).
(Speaking of books, you might take a peek at "The Black Swan: The Impact of the Highly Improbable" by Nassim Nicholas Taleb, new at Amazon, et al ... definately good reading for DIY types.)
Resistor & Cap on input of amp to keep from oscilating: http://www.national.com/images/pf/LM3886/01183301.pdf ... no?
Well, near as I can tell, there is an "RC time constant" relationship between resistor and cap at the (+) input ... the resistor / cap acting as a band pass or low pass filter to the positive feedback loop = allowing only the frequency band from ~0 Htz up to the intended upper limit of the amp or op-amp before out of control oscillations can set in ... and the resistor / cap does take care of that pesky "bounce" or reflection off of the cap, "squaring up" the leading edge of the signal ... If you have a 'scope you might be able to see this comparing a square wave signal trace with & without the resistor ... 
cliff: " ... That CR is a classic method of keeping the DC loop gain at unity ... Nothing to do with oscillations ..."
mmm ... Yes, approximately correct. (I'm wrong again
) My hydraulic model suggests that the resistor (1K Ri) "pinches off" the flow of electrons into and out of the cap (22 uF Ci) = acting as a high pass filter at the (-) negative feedback input (rather than a low pass type at the (+) input as stated). That would, I believe, do approximately the same (required) thing = allow the higher frequencies into the (-) negative input (pin 9) and thus, via the (-) negative feedback, keep the amp from oscillating ...
I guess I should breadboard these thing up myself before shooting from the hip ... or at least pay closer attention ... other than that, I had a great weekend ... You?
mmm ... Yes, approximately correct. (I'm wrong again
) My hydraulic model suggests that the resistor (1K Ri) "pinches off" the flow of electrons into and out of the cap (22 uF Ci) = acting as a high pass filter at the (-) negative feedback input (rather than a low pass type at the (+) input as stated). That would, I believe, do approximately the same (required) thing = allow the higher frequencies into the (-) negative input (pin 9) and thus, via the (-) negative feedback, keep the amp from oscillating ...I guess I should breadboard these thing up myself before shooting from the hip ... or at least pay closer attention ... other than that, I had a great weekend ... You?
"My hydraulic model suggests that the resistor (1K Ri) "pinches off" the flow of electrons into and out of the cap (22 uF Ci) = acting as a high pass filter at the (-) negative feedback input (rather than a low pass type at the (+) input as stated). That would, I believe, do approximately the same (required) thing = allow the higher frequencies into the (-) negative input (pin 9) and thus, via the (-) negative feedback, keep the amp from oscillating ..."
You are on a hiding to nowhere trying to analyse an audio amp feedback circuit using a "hydraulic" model of current flow.
There is no direct current in Ri because of the series cap.
Only using the frequency domain model makes sense:
1) Dc gain is 1
2) At high frequencies XCi is very low so gain is 20
3) As XCi gets to 1K at low frequencies the gain will be 3dB down on 20. This will occur at 7Hz
Whether the amp oscillates or not depends only on its internal frequency response and phase margin. It is impossible to tell from this block diagram what those values are. However, since this is a commercial audio amp IC, that will be detailed elsewhere in the datasheet.
You are on a hiding to nowhere trying to analyse an audio amp feedback circuit using a "hydraulic" model of current flow.
There is no direct current in Ri because of the series cap.
Only using the frequency domain model makes sense:
1) Dc gain is 1
2) At high frequencies XCi is very low so gain is 20
3) As XCi gets to 1K at low frequencies the gain will be 3dB down on 20. This will occur at 7Hz
Whether the amp oscillates or not depends only on its internal frequency response and phase margin. It is impossible to tell from this block diagram what those values are. However, since this is a commercial audio amp IC, that will be detailed elsewhere in the datasheet.
I'm not sure we're talking about the same resistor here. My interest is in Rb, the series resistor on the input. It prevents grounding the input when the pot is turned all the way down. If there were a RFI cap to ground (the situation in my amp), the resistor would limit the impedance to ground at high frequencies. No doubt it's something to do with the small but real input current, but I don't see how the signal gets reinforced to cause oscillation if the input is grounded directly.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.