ampage Tube Amps / Music Electronics	For current discussions, please visit Music Electronics Forum.

ampage archive

Vintage threads from the first ten years

Re: SS tube power amp emulation

10/6/1997 9:25 AM
R.G.	Re: SS tube power amp emulation There are a number of difficulties with emulating tubes, not the least of which is that different tube circuits behave differently, as witness the large number of mods to amps to make them sound "better" somehow. Your list of things left out third harmonic distortion in the output transformer itself, an unavoidable consequence of driving a ferromagnetic material with a (relatively) high impedance source. I do have a question. You note --- >>- Class B power amp crossover distortion, this >>results in even order harmonics. Does crossover give even order harmonics?? I thought it was primarily very high order and non-harmonic content, not even order. Crossover is very disagreeable in even small amounts, and even is fairly sweet sounding, even in noticeable amounts. On bias shift - One of my pet projects is to generate an envelope and use it to modulate the bias point of the inverters in a Tube Sound Fuzz circuit. There is a problem there, and it's the same one that tubes run into. If you allow too much signal level, the bias shift runs you into discontinuous conduction, analogous to grid blocking in a triode, so you have to bound the amount of bias shift that you allow, either by dividing down the signal or by some kind of limiter. To my knowlege, signal width modulation is not the same as phase shift - or did I not understand your point correctly. There are some real vibrato circuits in tube amps, notably Magnatone. They use a phase splitter/RC combiner circuit to give a variable phase shift when the R changes. They work like phase shifters that do not mix the dry signal back in.

10/6/1997 9:25 AM

R.G.

Re: SS tube power amp emulation

There are a number of difficulties with emulating tubes, not the least of which is that different tube circuits behave differently, as witness the large number of mods to amps to make them sound "better" somehow.

Your list of things left out third harmonic distortion in the output transformer itself, an unavoidable consequence of driving a ferromagnetic material with a (relatively) high impedance source.

I do have a question. You note ---
>>- Class B power amp crossover distortion, this
>>results in even order harmonics.
Does crossover give even order harmonics?? I thought it was primarily very high order and non-harmonic content, not even order. Crossover is very disagreeable in even small amounts, and even is fairly sweet sounding, even in noticeable amounts.

On bias shift - One of my pet projects is to generate an envelope and use it to modulate the bias point of the inverters in a Tube Sound Fuzz circuit. There is a problem there, and it's the same one that tubes run into. If you allow too much signal level, the bias shift runs you into discontinuous conduction, analogous to grid blocking in a triode, so you have to bound the amount of bias shift that you allow, either by dividing down the signal or by some kind of limiter.

To my knowlege, signal width modulation is not the same as phase shift - or did I not understand your point correctly.

There are some real vibrato circuits in tube amps, notably Magnatone. They use a phase splitter/RC combiner circuit to give a variable phase shift when the R changes. They work like phase shifters that do not mix the dry signal back in.

10/6/1997 11:41 AM
Steve Morrison	Does crossover give even order harmonics?? I thought it was primarily very high order and non-harmonic content, not even order. Crossover is very disagreeable in even small amounts, and even is fairly sweet sounding, even in noticeable amounts. I've wondered about this too. I know I've read in more than one textbook that crossover distortion is predominantly second harmonic, but I've never found it at all pleasant to listen to.

10/6/1997 11:41 AM

Steve Morrison

Does crossover give even order harmonics?? I thought it was primarily very high order and non-harmonic content, not even order. Crossover is very disagreeable in even small amounts, and even is fairly sweet sounding, even in noticeable amounts.

I've wondered about this too. I know I've read in more than one textbook that crossover distortion is predominantly second harmonic, but I've never found it at all pleasant to listen to.

10/7/1997 1:13 AM
Mikko H.	I do have a question. You note --- >>- Class B power amp crossover distortion, this >>results in even order harmonics. Does crossover give even order harmonics?? I thought it was primarily very high order and non-harmonic content, not even order. Crossover is very disagreeable in even small amounts, and even is fairly sweet sounding, even in noticeable amounts. I guess it depends how it's created, I haven't ever heard crossover distortion, as the only tube amp I use is SE. I just found a text (it was actually from a patent related to tube amp emulation) where it was said crossover distortion generates even order harmonics, but now when I think back, the article somehow mixed the meaning of crossover distortion and the bias shift, propably the writer thought they were the same thing (different phases had different levels, when combined, the result was even order harmonics because of the "crossover distortion"). On bias shift - One of my pet projects is to generate an envelope and use it to modulate the bias point of the inverters in a Tube Sound Fuzz circuit. There is a problem there, and it's the same one that tubes run into. If you allow too much signal level, the bias shift runs you into discontinuous conduction, analogous to grid blocking in a triode, so you have to bound the amount of bias shift that you allow, either by dividing down the signal or by some kind of limiter. I haven't started my similar project yet, but I think you could use 4007's complementary pairs, use +/- 15V dual PSU, run complementary pair from lower voltage (+/- 5-10V) supply and limit the signal with diodes to +/- 15V supply (if it helps). If complementary pairs doesn't work, why don't just use two n-type JFETS (upper FET buffer in source follower manner, lower FET gate connected to source, drain to upper FET's source, which is the output, lower FET's source to negative supply, input only to upper FET's gate). This is what I'm going to try, anyway. To my knowledge, signal width modulation is not the same as phase shift - or did I not understand your point correctly. I haven't read the text Jack mentioned about, but I just thought signal width modulation as a phase shift in the middle of the phase. The problem is how to implement it, as if the phase is splitted and put through a RC filter, there's the possible crossover distortion danger when the signal is combined. One way could be to modulate R with the signal itself, this could be done with a CA3080, if the signal if first amplified and then fed through resistor (simple voltage to current converter) to 3080's current control input. CA3080 preceded by voltage divider, followed by Miller integrator, with feedback from integrator to (+) input of 3080. This is like a usual phase shifter done with 3080's except there's no LFO, but control signal is the input signal. I'm not sure how many stages are needed, no phase splitter is needed. It is kind of ring modulator, which modulates signal by itself, and generates octave higher signal (second harmonic). -Mikko

10/7/1997 1:13 AM

Mikko H.

I do have a question. You note ---
>>- Class B power amp crossover distortion, this
>>results in even order harmonics.
Does crossover give even order harmonics?? I
thought it was primarily very high order and
non-harmonic content, not even order. Crossover is
very disagreeable in even small amounts, and even
is fairly sweet sounding, even in noticeable
amounts.

I guess it depends how it's created, I haven't ever heard crossover distortion, as the only tube amp I use is SE. I just found a text (it was actually from a patent related to tube amp emulation) where it was said crossover distortion generates even order harmonics, but now when I think back, the article somehow mixed the meaning of crossover distortion and the bias shift, propably the writer thought they were the same thing (different phases had different levels, when combined, the result was even order harmonics because of the "crossover distortion").

On bias shift - One of my pet projects is to
generate an envelope and use it to modulate the
bias point of the inverters in a Tube Sound Fuzz
circuit. There is a problem there, and it's the
same one that tubes run into. If you allow too much
signal level, the bias shift runs you into
discontinuous conduction, analogous to grid
blocking in a triode, so you have to bound the
amount of bias shift that you allow, either by
dividing down the signal or by some kind of
limiter.

I haven't started my similar project yet, but I think you could use 4007's complementary pairs, use +/- 15V dual PSU, run complementary pair from lower voltage (+/- 5-10V) supply and limit the signal with diodes to +/- 15V supply (if it helps). If complementary pairs doesn't work, why don't just use two n-type JFETS (upper FET buffer in source follower manner, lower FET gate connected to source, drain to upper FET's source, which is the output, lower FET's source to negative supply, input only to upper FET's gate). This is what I'm going to try, anyway.

To my knowledge, signal width modulation is not the
same as phase shift - or did I not understand your
point correctly.

I haven't read the text Jack mentioned about, but I just thought signal width modulation as a phase shift in the middle of the phase. The problem is how to implement it, as if the phase is splitted and put through a RC filter, there's the possible crossover distortion danger when the signal is combined.

One way could be to modulate R with the signal itself, this could be done with a CA3080, if the signal if first amplified and then fed through resistor (simple voltage to current converter) to 3080's current control input. CA3080 preceded by voltage divider, followed by Miller integrator, with feedback from integrator to (+) input of 3080. This is like a usual phase shifter done with 3080's except there's no LFO, but control signal is the input signal. I'm not sure how many stages are needed, no phase splitter is needed. It is kind of ring modulator, which modulates signal by itself, and generates octave higher signal (second harmonic).

-Mikko

10/15/1997 3:50 PM
Bengt Palmkvist	I saw a shematic in an electronics magasin (from -75 ?) where they used a CMOS Logic (!) gate to emulate the tube sound ! Take a 4012U 4-input NAND (it must be unbuffered or otherwise it will oscillate), connect all four inputs together and put a resistor from output to inputs (feedback) and a resistor between input signal and input. Now its an analog circuit !! An inverter will also do it. This was used in a guitar amplifier and I remeber when I was laborating whith this circuit I didn't use unbuffered gates so it oscillated an disturbed the radio I was listening to. B.P.

10/15/1997 3:50 PM

Bengt Palmkvist

I saw a shematic in an electronics magasin (from -75 ?) where they used a CMOS Logic (!) gate to emulate the tube sound !
Take a 4012U 4-input NAND (it must be unbuffered or otherwise it will oscillate), connect all four inputs together and put a resistor from output to inputs (feedback) and a resistor between input signal and input. Now its an analog circuit !! An inverter will also do it.
This was used in a guitar amplifier and I remeber when I was laborating whith this circuit I didn't use unbuffered gates so it oscillated an disturbed the radio I was listening to.

B.P.

10/15/1997 10:25 PM
R.G.	That's the basis for several soft distorting effects, notably the ElectroHarmonix Hot Tubes and Craig Anderton's Tube Sound Fuzz in it's several variants.

<<First Page

<Prev

Page 2 of 2