Thanks for that ZM,
I will proceed to provide the putative 2.8kz items. For those who might not understand what Zen and I are poking at here let me post an email response that may give you a better view, or just a headache.
When you have a given load, let's use 1kZ ohm, you must divide that load by 2 x Pi and the frequency you want to extend to. That resultant number is the amount of inductance that the transformer must develop, at that frequency, to provide a -3db roll off from 1kHz.
Then you have to decide where in the power band you want to achieve that frequency roll off. This because core material, at low frequencies, below 100 Hz, gains in it's ability to produce inductance, from rest to full inductance (this is known as permeability, by a factor of 8 for commercial iron manganese core, the M series, to 10 for 48% nickle. The more expensive 80% nickle does not provide more permeability, it just has a steeper slope of onset and less tolerance, by 1/3. for voltage/amperage increase. The most inductance/permeability, for all core material I know of, is found at 27 Hz. Almost all permeability charts are designed to show permeability for 50/60 Hz. There are very few in existence.
Designing these things is either a science, read Jensen, Edcor, Cinemag and many others here, or an art, read Tribute, Intact Audio, Sowter, Lundahl and O-Netics here. The difference in audio between the two groups is much greater than the difference in measurement would lead you to believe, with the nod for music reproduction going to the art group and the nod for specification justification going to the science group. Both approaches are valid.
To obtain a -0.5 db roll off at the frequency and power point you choose you must increase the inductance by 2.76 times the -3db number.
When I say that my Dac Buffer transformers are designed for a 1kZ ohm load in the primary, I am saying that at just off rest voltage, down in the micro volt/amp range, at 20 Hz, I am able to support a -0.5 db roll off. Other designers will choose otherdesign points and criteria when meeting the very same specs.
So, except for Langford Smith showing, in Radiotron Designers Handbook volume three, that the ideal amount of inductance, for lowest distortion is to be limited to twice that needed for -3 db roll off, at frequencies below 100 Hz, exact impedance match is a complete waste of time. So long as you can support the frequency of interest at the power level desired, all other concerns are moot. Unless you go for gaped core of course and apply an appropriate DC bias current. At which point almost everyone throws their hands up in the air and run's away....
My question came about because I do not design or build amplifiers and so am not always certain what is the limiting specification that would disqualify one design from another.
Bud
I will proceed to provide the putative 2.8kz items. For those who might not understand what Zen and I are poking at here let me post an email response that may give you a better view, or just a headache.
When you have a given load, let's use 1kZ ohm, you must divide that load by 2 x Pi and the frequency you want to extend to. That resultant number is the amount of inductance that the transformer must develop, at that frequency, to provide a -3db roll off from 1kHz.
Then you have to decide where in the power band you want to achieve that frequency roll off. This because core material, at low frequencies, below 100 Hz, gains in it's ability to produce inductance, from rest to full inductance (this is known as permeability, by a factor of 8 for commercial iron manganese core, the M series, to 10 for 48% nickle. The more expensive 80% nickle does not provide more permeability, it just has a steeper slope of onset and less tolerance, by 1/3. for voltage/amperage increase. The most inductance/permeability, for all core material I know of, is found at 27 Hz. Almost all permeability charts are designed to show permeability for 50/60 Hz. There are very few in existence.
Designing these things is either a science, read Jensen, Edcor, Cinemag and many others here, or an art, read Tribute, Intact Audio, Sowter, Lundahl and O-Netics here. The difference in audio between the two groups is much greater than the difference in measurement would lead you to believe, with the nod for music reproduction going to the art group and the nod for specification justification going to the science group. Both approaches are valid.
To obtain a -0.5 db roll off at the frequency and power point you choose you must increase the inductance by 2.76 times the -3db number.
When I say that my Dac Buffer transformers are designed for a 1kZ ohm load in the primary, I am saying that at just off rest voltage, down in the micro volt/amp range, at 20 Hz, I am able to support a -0.5 db roll off. Other designers will choose otherdesign points and criteria when meeting the very same specs.
So, except for Langford Smith showing, in Radiotron Designers Handbook volume three, that the ideal amount of inductance, for lowest distortion is to be limited to twice that needed for -3 db roll off, at frequencies below 100 Hz, exact impedance match is a complete waste of time. So long as you can support the frequency of interest at the power level desired, all other concerns are moot. Unless you go for gaped core of course and apply an appropriate DC bias current. At which point almost everyone throws their hands up in the air and run's away....
My question came about because I do not design or build amplifiers and so am not always certain what is the limiting specification that would disqualify one design from another.
Bud
Is the glass with the ice there to put out the fire?
Safety first!.
Thanks Rush. My wife is jealous
I will be trying one of Buds transformers and will happily share the results. Needless to say, i am feeiling good about the results as he has played with them off and om gor just a couple of years. Sort of like Nelson with amps.
Last edited:
Buzzforb: Neat, and a neat build. Did you move away from the Beast of a Thousand Jfets front end?. Are you using CCS's on the main power board? What is your conceptual enclosure going to look like?
This is just a throw together. Nothing like the original. I am working on a board that will allow for all types of variations on both FE and output, feedback, no feedback. I will not go too far until i have actually heard the thing and see what it sounds like. These are ADC transformers similar to the Jensens, but capable of little more current. BudP's transformers will present an easier load to the FE,being 2K vs 600R, and will just require tweaking on the output. I have CCS on bias section of output feeding LED's. I will play with Salas shunt feeding input pairs 10V vs. Full rail voltage into Joachim's FE vs. full rail voltage and 10R degeneration. Dropping the degeneration resistor lowers the Zout of FE, as well as lowers the noise, but Bud's transformers relieve that problem. Doubling up FE jfets takes care of it as well. Who knows. I just try to avoid making the magic smoke.
I hope that there isn't anything flammable in the glass.
Safety first!.
Thanks Rush. My wife is jealous
I will be trying one of Buds transformers and will happily share the results. Needless to say, i am feeiling good about the results as he has played with them off and om gor just a couple of years. Sort of like Nelson with amps.
Since the F6 has (unspecified) negative feedback, and more open-loop voltage gain than the M2, I would expect even less harmonic distortion from the F6. But, the high and low frequency roll-off is present, at levels depending on the amount of negative feedback. Thus I would expect the F6 to sound even less like a tube amp than the M2.
What sound has Nelson achieved with his F6. Only he and his buying customers know and experience. Flip side question: What will your Teaser6, buzzforb's Franken6 , triode_al's, flg's other DIYers' sound in Mr. Pass's experimental setup? You'd love to know! Maybe for a modest fee, Passlabs [or an entrepreuneur DIYer amongst us] test the subjective and objective performance of your Teaser6 compared to an authentic [investment] F6, and issue a Certificate? For all I know, you are well-equipped to do the testing of amps by other DIYers! It shows through your posts.
The fragment you abstracted from the M2 User Manual belongs exclusively to all of the M2 amplifiers assembled [and sold to users] by First Watt. It is also a guiding light to the DIYer to expect the subjective and objective performance that can be attained in a diy build of M2.
Ilquam. Take the time to play with P3 on the F5, paying attention to the harmonic structure of the distortion and trying to match it what you hear. It will not tell you all, but will give you an idea of what can be done. Of course a non feedback design like the Ba3/BA2 output stage will have the same ability, but perhaps a slightly different sound due to no feedback. Nelson's brilliance is his ability to decide what he wants and make it happen on his terms. Takes knowledge, and perhaps more important, experience. Its easy for me to tell the difference, cause I got the "Golden Ears".