Hi Mark,
I made a surface mount transistor adapter that plugs into my first two board designs. A block of steel keeps them both seated and near the same temperature. Not ideal, but it works better than anything else I have tried. I'd be worried about losing the darned smt packages with a stream of air, but I guess you have to reduce velocities for that.
Ever sneeze by surprise and find that you don't have very many surface mount parts left? Especially annoying after you've matched most of them!
-Chris
I made a surface mount transistor adapter that plugs into my first two board designs. A block of steel keeps them both seated and near the same temperature. Not ideal, but it works better than anything else I have tried. I'd be worried about losing the darned smt packages with a stream of air, but I guess you have to reduce velocities for that.
Ever sneeze by surprise and find that you don't have very many surface mount parts left? Especially annoying after you've matched most of them!
-Chris
Hi SGK,
I work in the basement away from the air duct openings. That helps. Any variation that occurs will cause an equal shift in both DUTs, so that cancels out. No problems there. The outer box really helps a lot with air current changes like the ones you are talking about.
No, I still use the capacitor. It isn't nearly as detrimental to sound quality as you might think. Perhaps in stuff I design, I'll do away with that capacitor and correct for DC offset in the Vas or pre-drive stages. That's the only way to allow a servo to work where it doesn't have an effect on sound quality. Sometimes you are lucky doing it the standard way, sometimes not so lucky.
-Chris
I work in the basement away from the air duct openings. That helps. Any variation that occurs will cause an equal shift in both DUTs, so that cancels out. No problems there. The outer box really helps a lot with air current changes like the ones you are talking about.
Yes, in a high quality, well designed amplifier, I do use my tightest devices and also match the degeneration resistors. The better the circuit design, the more doing this matters.
No, I still use the capacitor. It isn't nearly as detrimental to sound quality as you might think. Perhaps in stuff I design, I'll do away with that capacitor and correct for DC offset in the Vas or pre-drive stages. That's the only way to allow a servo to work where it doesn't have an effect on sound quality. Sometimes you are lucky doing it the standard way, sometimes not so lucky.
I've never had that luck no matter how carefully I tried or what instrument or jig I use. In your case the jig would give you very high yields for very tight matched parts. My experiences doing as you are lead me to design the jig. One thing I can say is that the jig gives me very consistent results. That's why I stick with it when matching is important.
-Chris
As I described in my book, the use of a quality NP capacitor can add very little distortion. A key is that it have a voltage rating of greater than 50V. Capacitors intended for loudspeaker crossover networks often work well. Nevertheless, I usually prefer a DC servo. When properly designed, they certainly do not compromise the sound.
Most of the time, correcting for dc offset in the VAS or pre-driver is the wrong thing to do. This is especially true where input stages properly use a current mirror load, which tends to enforce equal currents in the two transistors of the IPS LTP. Amplifiers that just use an IPS load resistor and single-ended drive to the VAS (no current mirror) can be another story, and I never recommend that type of topology.
Cheers,
Bob
I've built fully complimentary amplifiers (see the e-Amp and the Ovation 250) and CFA designs and I've never had output drift problems. I tend AC couple the feedback network on VFA designs with a 470 or 1000uF cap that puts the LF pole at 1 or 2 Hz. Offset adjustment is with a pot. On the VFA designs the drift is below 10 mV across temperature. The fully DC coupled CFA designs are about 50 mV at switch on, but settle to within 1-2 mV within a few minutes.
I am just finishing up a new CFA design that is using a DC servo. So far so good, but I still have to do the LF high power testing.
I am just finishing up a new CFA design that is using a DC servo. So far so good, but I still have to do the LF high power testing.
Hi Bob,
I agree, my suggestions do not work well with amplifier designs as you've described. However, the experience I have is working with stuff that's out there already. Just adding a CCS for tail current and using a tight matched pair will improve the sound. One real mark of success is a DC offset that doesn't change as the amplifier warms up.
I have your book. I even read it and do refer to both it and the ones written by Doug Self for reference. Excellent material. As for capacitors, noticed that higher voltage ratings seem to sound better. They also fit the PCB locations better than the ones made today with the same ratings in a smaller case. One thing that should be touched on is to make sure that when bending the leads to fit, make certain the areas where the leads exit the rubber plug are not stressed as the leads are bent. Allowing the leads to stretch the rubber will lead to early capacitor death.
-Chris
I agree, my suggestions do not work well with amplifier designs as you've described. However, the experience I have is working with stuff that's out there already. Just adding a CCS for tail current and using a tight matched pair will improve the sound. One real mark of success is a DC offset that doesn't change as the amplifier warms up.
I have your book. I even read it and do refer to both it and the ones written by Doug Self for reference. Excellent material. As for capacitors, noticed that higher voltage ratings seem to sound better. They also fit the PCB locations better than the ones made today with the same ratings in a smaller case. One thing that should be touched on is to make sure that when bending the leads to fit, make certain the areas where the leads exit the rubber plug are not stressed as the leads are bent. Allowing the leads to stretch the rubber will lead to early capacitor death.
-Chris
I'm looking good forward to your new edition coming out. Hopefully it is soon. I'm interested in your chipamp section for car audio. I noticed since your previous edition 2010, the national line of chipamps have been obsoleted. I'm interested in what recommendations you have for replacement for the lm48930 and LT1166 bis spread controller is elegant .
Sincerely,
David P.
Sincerely,
David P.
Hi David,
Thank you for your interest in my second edition. It is coming along well, and there is so much new material I want to put in there. Once again, as happened with the first edition, I am probably up against a page-count limitation (this time 750 pages as compared to the previous 600 pages). The book will include 5 new chapters.
Unfortunately, my progress of late has been somewhat stalled by the distraction of some significant family matters. My target for publication remains this coming Fall, but for that to happen the publisher would have to move significantly faster than usual to go from completed manuscript to publication.
It is deeply disappointing that TI has obsoleted so much of the fantastic National audio ICs. One debate I am having with myself is whether to ditch those sections in the book covering those IC, or to keep much of that material as an example of what was possible using those devices when they were available.
Cheers,
Bob
you don't need my advice, but if I were faced with page cropping to fit the publisher's limit, then removing the old material applicable to obsoleted devices seems the first candidate to be trimmed.
That creates a secondhand market for the first edition and makes those who now want the second edition more likely to spend some more money.
That creates a secondhand market for the first edition and makes those who now want the second edition more likely to spend some more money.
Or make the obsolete chapters from the first edition available for download in a
nicely formatted .pdf so people can just download and print if they want it.
I find it a bit annoying that there is so much repeat when I buy a first and
second edition. I also find it annoying when some material is deleted in the
second edition since then I hesitate to give away the first edition. Being able
to buy the second and print what I need from the first is a solution. Not sure
if there is anyway to work in a volume 1 and 2 or introductory/advanced.
nicely formatted .pdf so people can just download and print if they want it.
I find it a bit annoying that there is so much repeat when I buy a first and
second edition. I also find it annoying when some material is deleted in the
second edition since then I hesitate to give away the first edition. Being able
to buy the second and print what I need from the first is a solution. Not sure
if there is anyway to work in a volume 1 and 2 or introductory/advanced.
Andrew and Pete,
These are all good thoughts and ideas. There are interesting tradeoffs in going to a Vol 1 & 2 in a second edition, even assuming the publisher would go for it. It would not necessarily be as easy as a slice down the middle of the chapters. Indeed, given the nature of the material, I'm guessing it might be a challenge to segregate it into introductory/advanced. A lot of people who already own the first edition would likely want to buy both volumes anyway, and end up spending more money. Just thinking out loud here.
I think the idea of dumping the parts dependent on the obsolete devices and putting it up on the web page is a good one.
Cheers,
Bob
Hi Bob,
IMHO, there is always a market for an advanced book. It's the nature of DIY. A lot of us are trying to push our limits. I, for one, would be very interested in a book with your version of "advanced". Kolinummi's book holds some interesting ideas. But I'm sure there's other avenues that can be explored.
Paul
IMHO, there is always a market for an advanced book. It's the nature of DIY. A lot of us are trying to push our limits. I, for one, would be very interested in a book with your version of "advanced". Kolinummi's book holds some interesting ideas. But I'm sure there's other avenues that can be explored.
Paul