Just as a statement of personal policy...I try to answer all queries directed to me personally. That presumes that I am aware of them. If someone asks a question of me in a Digital thread, for instance, it's highly unlikely that I will ever see it, as I rarely visit that Forum. As is, I suppose, obvious, I spend most of my time in the threads in the Pass Forum, some in Solid State, with an occasional foray into Speakers and Everything Else. Once in a blue moon I look in on Tubes. That does not mean that I read each and every thread, even in Pass. Sorry, but I no longer have anything like the time required. The amount of time I do have varies widely from day to day. Some days I read a couple of dozen threads. Some days none.
We're also assuming that a hypothetical query post doesn't vanish into thin air.
With the arguable exception of the times when I start a new thread (e.g. Aleph-X), I don't know that I'm all that necessary as an individual, anyway. Things have progressed to the point where Mark Finnis and I are no longer the only ones who have built Alephs, etc. so there are plenty of people who can answer the usual MOSFET substitution questions and so forth.
And, of course, there's always Nelson. He's been known to drop a hint here and there as to how to do this or that. (...ahem...) Just don't wear the man out with simplistic questions that others can answer just as easily. Save him for the difficult stuff like why red wine is better than white or whether to drink then listen or vice versa.
Grey
(aka Just Me)
P.S.: For those who want to know how the various things I'm working on are progressing, I've had to wear my Author hat for the last two days. Stan (my editor) asked for some changes to the most recent story, and that's kept me busy. I squirted them up the wire to NY today and hope to get back to electronics tomorrow.
We're also assuming that a hypothetical query post doesn't vanish into thin air.
With the arguable exception of the times when I start a new thread (e.g. Aleph-X), I don't know that I'm all that necessary as an individual, anyway. Things have progressed to the point where Mark Finnis and I are no longer the only ones who have built Alephs, etc. so there are plenty of people who can answer the usual MOSFET substitution questions and so forth.
And, of course, there's always Nelson. He's been known to drop a hint here and there as to how to do this or that. (...ahem...) Just don't wear the man out with simplistic questions that others can answer just as easily. Save him for the difficult stuff like why red wine is better than white or whether to drink then listen or vice versa.
Grey
(aka Just Me)
P.S.: For those who want to know how the various things I'm working on are progressing, I've had to wear my Author hat for the last two days. Stan (my editor) asked for some changes to the most recent story, and that's kept me busy. I squirted them up the wire to NY today and hope to get back to electronics tomorrow.
HI Nelson,
We're all abuzz about the new CSS FR125 driver that is coming out. (the WR 125 which is a wide range rather than full range has been out for awhile) It is designed by Adire for CSS, and the word is that they have lots of top end and a lot of low distortion bass and in the mids are competition for the Jordens that you are using in the J-lo. I guess they would become lo-C horns if you put this new driver in them. One appealing thing about them is that they are about $75 each- what a deal! They have a lot more bass so the J-lo cabinet might need a bit of tuning. Heck, with a bit of tuning they might work in the Beeeeg horns!
And what would happen with Current drive? We have to know 'cause you are about to publish the F1 schematics!
http://www.diyaudio.com/forums/showthread.php?s=&threadid=59762&perpage=10&pagenumber=1
Sadly, they are waiting to be unloaded in Canada due to a dock strike so are available "any day now" at : http://www.creativesound.ca/
I have no connection to any companies listed....
We're all abuzz about the new CSS FR125 driver that is coming out. (the WR 125 which is a wide range rather than full range has been out for awhile) It is designed by Adire for CSS, and the word is that they have lots of top end and a lot of low distortion bass and in the mids are competition for the Jordens that you are using in the J-lo. I guess they would become lo-C horns if you put this new driver in them. One appealing thing about them is that they are about $75 each- what a deal! They have a lot more bass so the J-lo cabinet might need a bit of tuning. Heck, with a bit of tuning they might work in the Beeeeg horns!
And what would happen with Current drive? We have to know 'cause you are about to publish the F1 schematics!
http://www.diyaudio.com/forums/showthread.php?s=&threadid=59762&perpage=10&pagenumber=1
Sadly, they are waiting to be unloaded in Canada due to a dock strike so are available "any day now" at : http://www.creativesound.ca/
I have no connection to any companies listed....
CSS FR125
Variac, et al:
while I do have a loose affiliation with Bob Reimer of CSS, I think I can say I'm objective enought to state the following:
these drivers are definitely worth your close consideration as to whether their TS parameters would work in the J-lo design.
I've been fortunate enough to gain access to a couple of pair from an advance shipment, and in fact built the enclosures noted in the above thread.
Bottom line is these sound very good indeed, and are stunning matched with the mini-A per BrianGT, and pretty decent on the P/P EL34 triode amp as well.
The same XBL^2 motor technology* by Adire is also utilized in an absolutely astounding 6" midbass driver, the Extremis 6.8.
*http://www.adireaudio.com/Files/XBL2TechPaper.pdf
Variac said:
Variac, et al:
while I do have a loose affiliation with Bob Reimer of CSS, I think I can say I'm objective enought to state the following:
these drivers are definitely worth your close consideration as to whether their TS parameters would work in the J-lo design.
I've been fortunate enough to gain access to a couple of pair from an advance shipment, and in fact built the enclosures noted in the above thread.
Bottom line is these sound very good indeed, and are stunning matched with the mini-A per BrianGT, and pretty decent on the P/P EL34 triode amp as well.
The same XBL^2 motor technology* by Adire is also utilized in an absolutely astounding 6" midbass driver, the Extremis 6.8.
*http://www.adireaudio.com/Files/XBL2TechPaper.pdf
Mr. Pass,
All the recent discussion caused by the F1 service manual and F2 shipments brings to mind the "in-process" installment of the Zen series on a transconductance Zen. Are you still planning on publishing this installment?
I for one am still very interested in a transconductance Zen article for the reason that it would provide a forum for a more detailed discussion of the more basic design issues.
Two things I am particularly interested in seeing included are:
1) A discussion on input impedance vs. input transistor gate capacitance vs. distortion. I know that you have posted a brief comment on this regarding the F1 or F2 but it would be interesting to me at least to have a a more detailed discussion with perhaps a graph or two.
2) A discussion on "common gate" operation in MOSFET's as a complement to the detailed discussion on the CD and CS connections in the Zen articles and in the paper on MOSFET testing. For instance, you have stated that the output impedance of the upper CCS in the F1 is roughly 1000 ohms. I would like to know how this was calculated. This info available in a number of text books but you have a way of cutting through the clutter and presenting the essence of a subject.
Anyway, I hope the transconductance Zen article is still simmering and that you can find a way to include some discussion the above. Thank you again for all the great stuff.
Cheers,
GL
All the recent discussion caused by the F1 service manual and F2 shipments brings to mind the "in-process" installment of the Zen series on a transconductance Zen. Are you still planning on publishing this installment?
I for one am still very interested in a transconductance Zen article for the reason that it would provide a forum for a more detailed discussion of the more basic design issues.
Two things I am particularly interested in seeing included are:
1) A discussion on input impedance vs. input transistor gate capacitance vs. distortion. I know that you have posted a brief comment on this regarding the F1 or F2 but it would be interesting to me at least to have a a more detailed discussion with perhaps a graph or two.
2) A discussion on "common gate" operation in MOSFET's as a complement to the detailed discussion on the CD and CS connections in the Zen articles and in the paper on MOSFET testing. For instance, you have stated that the output impedance of the upper CCS in the F1 is roughly 1000 ohms. I would like to know how this was calculated. This info available in a number of text books but you have a way of cutting through the clutter and presenting the essence of a subject.
Anyway, I hope the transconductance Zen article is still simmering and that you can find a way to include some discussion the above. Thank you again for all the great stuff.
Cheers,
GL
gl said:
Actually, the F1 service manual has been accepted for publication
by AudioXpress, and it serves the purpose you mention. It will
be followed by the F2 service manual, presumably.
1) The capacitance of Mosfets is their biggest fault, particularly
when we are using them in single-stage systems where they
aren't buffered by an input gain device. Not only do we see
capacitance value at the gate (practically speaking, the Cgs
and Cgd) but these values also alter with the Drain to Source
voltage Vds, which makes it a nonlinear component.
So we see a high frequency rolloff and a distortion curve which
climbs at high frequencies. If we take the Cgd at 100 pF and
and Cgs at 1000 pF, we see in a typical (Zen type) amplifier that
the Cgd is multiplied by the gain, so that 12 dB means something
like 400 pF, and the Cgs is divided down, and becomes something
like maybe 200 pF. The result is a small signal model in the
region of 600 pF or more. However, the small signal model
thinks in terms of having maybe 15 volts across the device. As
the voltage varies, the voltage across the device experiences
lower values, and the capacitance increases, giving a decrease
in gain at the same point that the device is ordinarily seeing
a transconductance increase. It slightly resembles the load-line
cancellation you see in triodes, but it only happens at high
frequencies, and the overall net effect is distortion increasing
as you go above the mid-band.
You can minimize these effects in various ways. You can:
a) use low capacitance Mosfets with smaller dies.
b) use lateral Mosfets whose capacitance is often higher
but more more linear
c) bias the transistor at a higher Vds to stay out of the
higher capacitance region
d) run lower voltage gain - often a fine option in smaller amplifiers
e) use lower impedance sources to drive these stages
f) cascode these stages to minimize the variation in
Vgs and Vds
g) use any combination of the previous 6
Now to part 2:
I refer you to my paper from Audio Magazine on cascoding, which
can be seen at www.passdiy.com and elsewhere
As to the impedance of a CCS, I do not calculate, I measure.
There are numerous ways of coming up with such figures, but
the easiest is to perform a precision measurement of the AC
current through a CCS while it is experiencing an AC voltage
at it's output. Best to use isolated measurement apparatus
for this, but it is otherwise easy enough.
The (intrinsic output impedance of the F1 and F2 are in the
1 Kohm region, but I load them arbitrarily with some resistance
for practical reasons, the first of which is that no loudspeaker
actually cares about whether the source impedance is infinitely
high and usually wants something less than 50 ohms anyway.
In measuring the intrinsic output impedance, you simply operate
the circuit (without the loading resistors) into two different loads
and measure the difference in output. The output will not be
exactly proportional to the load, allowing you to calculate the
output impedance.
Let's take a real example. I measure the output into 100 ohms
and into 47 ohms. If I measure .91V with 100 ohms, and .45
with 47 ohms, I will solve for Rs in the two equations
.91 = 100 (Rs+100) and .45 = 47/(Rs+47)
a little algebra later, we see that Rs is about 1000 ohms.
Tadaaa....
.
Mr. Pass,
Thank you very much indeed for the detailed response to my questions. I too felt like Terry on the CCS output impedance measurement.
IMHO the publication of the F1 manual in AudioXpress plus the on-going discussion of the F1 and F2 on this forum truly equal or exceed the expectations for the transconductance Zen article.
Regards,
GL
Thank you very much indeed for the detailed response to my questions. I too felt like Terry on the CCS output impedance measurement.
IMHO the publication of the F1 manual in AudioXpress plus the on-going discussion of the F1 and F2 on this forum truly equal or exceed the expectations for the transconductance Zen article.
Regards,
GL
lumanauw said:
I did spend the first 20 or so years using bipolars for most
everything.
I still use bipolars, and I have no quarrel with them, however
it is nearly impossible to make a single-stage power amplifier
with them, and difficult to make a 2 stage amplifier.
Currently, I am still exploring minimalism, and there I find that
Mosfets are the most appropriate power devices.
Blues said:
Nelson Pass said:
I think there was a little ambiguity in an earlier post that said if you measure .91 volts etc.... Where in absolute terms .91V = 910 Millivolts. I did however understand the formulae to represent an X vs Y ratio relationship, or being a "Mad" Scientist did you mean Spv vs Spy?
Regards
Anyhony
Mr. Pass,
I'm sure many DIYers only knows you as a man who builds amps with 1 or 2 mosfets, and never build more complicated things.
You have build many wonderfull cct's, but only some (like threshold or your patents) are available for us to see.
Do you have archive of your cct's or experiment cct's (from you start) that we can see (or buy if it is a book? I will consider that book as a "must read")
Person like you must have 1000s of cct's. From 1 transistor amp only, you mentioned you have almost 100 variation.
I'm sure many DIYers only knows you as a man who builds amps with 1 or 2 mosfets, and never build more complicated things.
You have build many wonderfull cct's, but only some (like threshold or your patents) are available for us to see.
Do you have archive of your cct's or experiment cct's (from you start) that we can see (or buy if it is a book? I will consider that book as a "must read")
Person like you must have 1000s of cct's. From 1 transistor amp only, you mentioned you have almost 100 variation.
I have less than a thousand circuits, but more than a hundred.
Unfortunately I don't tend to properly document the crappier
ones. In the last couple of years, I have been on a program
of scanning all the old folders and stuff. One of these days I'm
just going to put it on a DVD and toss it out, probably after I
lose my mind.
Unfortunately I don't tend to properly document the crappier
ones. In the last couple of years, I have been on a program
of scanning all the old folders and stuff. One of these days I'm
just going to put it on a DVD and toss it out, probably after I
lose my mind.
lumanauw said:
Lumanauw,
suppose a bloke like Douglas Self would spend time on a regular basis on the web to explain stuff to guys like you and me. Suppose he would write articles and post them on his homepage for you to download.
Suppose he would design circuits for a living and hand you the specifics when there is no commercial need for keeping them under lock and key.
Would you still want to buy his books ?
If you downloaded, copied and printed as much Pass stuff as i did you should have your book by now.
On top of that you get extras like talk on wine and shoes.
I believe the CCS impedance approach is a common way for establishing the damping factor in outputs, and quite elegant.
Simplicity is elegance.
Mr Pass, wait till i've lost my marbles and think i am a dog, you can throw me a frisbee.
