diyAudio (
-   Solid State (
-   -   ThermalTrak+TMC amp (

dadod 5th February 2011 03:31 PM

ThermalTrak+TMC amp
10 Attachment(s)
For some time I am trying an amp with TermalTrak ON semiconductor. I included TMC too and here is what I came up to now.
I simulated it with LTSpice and the results are here.
THP at 1kHz is 0.000011% and at 20kHz is 0.000186%.
I did PCT with Sprint Layout and tested it with my test bench power supply +-30V up to 2A.
I did not test it with +-50V for what it was designed.
The bias seams to be stable, it did not change from 20deg C to 50deg C.
This is just proto board, main board and small input board. I intend to mak all in one board, it was for testing porpoises. It was not tested with a musik jet, no power supply redy jet.

OITCP(Output Inclusive Two Pole Compensation) used in this TT amp instead of TMC. page 60, post #594

Album with the amp construction created:

dadod 6th February 2011 07:10 AM

No comments at all??
It was nice to find how is easy to go so low in THD below ppm with this quite simple schematic.
Input is cascoded LTP with LSK389 in metal TO71 from Linear Systems. First CCS gives cca

dadod 6th February 2011 07:29 AM

Sorry for previous post, I sent it accidentally.
The first CCS is cca 9mA, and this is 4mA per each side of the LTP. 1ma is for cascode to keep cca 10V on JFETs.
The bias control is ddivided between ThermalTrak diodes and Vbe multiplier.
Vbe multiplier is thermally connected with the drivers only as suggested by Bob Cordel.
I was thinking to use a triple as suggested by Cordel for ThermalTrak in one of the ThermalTrak treads. I will do that probably later, never use triple before. Symulation does not show important improvement.
PCT in first post should be PCB off course.

Fourier components of V(vout) 1kHz
DC component:-0.0296359

Harmonic Frequency Fourier Normalized Phase Normalized
Number [Hz] Component Component [degree] Phase [deg]
1 1.000e+03 2.665e+01 1.000e+00 -0.13 0.00
2 2.000e+03 1.973e-06 7.402e-08 87.49 87.62
3 3.000e+03 2.029e-06 7.615e-08 11.10 11.23
4 4.000e+03 2.899e-07 1.088e-08 -41.62 -41.49
5 5.000e+03 1.746e-07 6.550e-09 -15.60 -15.47
6 6.000e+03 2.994e-07 1.124e-08 -7.33 -7.19
7 7.000e+03 3.302e-07 1.239e-08 0.49 0.62
8 8.000e+03 4.099e-07 1.538e-08 -6.78 -6.64
9 9.000e+03 4.130e-07 1.550e-08 -0.83 -0.70
Total Harmonic Distortion: 0.000011%

Fourier components of V(vout) 20kHz
DC component:-0.029628

Harmonic Frequency Fourier Normalized Phase Normalized
Number [Hz] Component Component [degree] Phase [deg]
1 2.000e+04 2.663e+01 1.000e+00 -2.61 0.00
2 4.000e+04 2.881e-05 1.082e-06 96.30 98.91
3 6.000e+04 1.920e-05 7.208e-07 143.84 146.45
4 8.000e+04 1.628e-05 6.112e-07 -143.04 -140.44
5 1.000e+05 2.368e-06 8.890e-08 -34.05 -31.44
6 1.200e+05 1.431e-05 5.375e-07 -150.17 -147.56
7 1.400e+05 1.741e-05 6.536e-07 145.16 147.76
8 1.600e+05 1.129e-05 4.241e-07 -148.10 -145.49
9 1.800e+05 1.883e-05 7.072e-07 137.23 139.84
Total Harmonic Distortion: 0.000186%


Workhorse 6th February 2011 08:40 AM


Originally Posted by dadod (
I simulated it with LTSpice and the results are here.
THP at 1kHz is 0.000011% and at 20kHz is 0.000186%.


Simulated THD figures are far from reality. Better show real world prototype measurements.:)

dadod 6th February 2011 09:05 AM

I've got a tone generator (sinus no better then 0,5%) an oscilloscope and 8/4 ohm load.
With that I can't get any valide THD measurements. I will show square wave measurements on 8/4 ohm//1nF-1uF load to show stability when I finish +-50V 500VA power supply.

dadod 6th February 2011 09:54 AM

TT+TMC Slew Rate improvement as suggested by D. Self
2 Attachment(s)
To improve Slew Rate I change CCS-s according to D. Self.
That I did in LTSpice only.Intereting is when change was made that controling CCS to be VAS CCS instead LTP CCS as in build protype the THD at 20kHz increased threefold.
Then, when I added small capacitor of 6pF(D. Self suggestion) between CCS collector and base of the controlling transistor, it incresed dramatically. Is it worth to increment slew rate and in the same time the THD???

D. Self CCS
Fourier components of V(vout)
DC component:-0.0310529

Harmonic Frequency Fourier Normalized Phase Normalized
Number [Hz] Component Component [degree] Phase [deg]
1 2.000e+04 2.663e+01 1.000e+00 -2.63 0.00
2 4.000e+04 9.475e-05 3.558e-06 136.72 139.34
3 6.000e+04 1.843e-05 6.918e-07 148.38 151.00
4 8.000e+04 1.602e-05 6.014e-07 -143.55 -140.93
5 1.000e+05 1.573e-06 5.907e-08 -38.79 -36.17
6 1.200e+05 1.443e-05 5.419e-07 -150.26 -147.63
7 1.400e+05 1.813e-05 6.807e-07 145.10 147.73
8 1.600e+05 1.126e-05 4.229e-07 -148.17 -145.54
9 1.800e+05 1.934e-05 7.261e-07 137.05 139.67
Total Harmonic Distortion: 0.000387%

D. Self CCS with 6pF
Fourier components of V(vout)
DC component:-0.0322834

Harmonic Frequency Fourier Normalized Phase Normalized
Number [Hz] Component Component [degree] Phase [deg]
1 2.000e+04 2.663e+01 1.000e+00 -2.63 0.00
2 4.000e+04 1.196e-03 4.492e-05 -108.97 -106.34
3 6.000e+04 2.318e-05 8.705e-07 -178.36 -175.74
4 8.000e+04 1.662e-05 6.242e-07 -144.68 -142.06
5 1.000e+05 1.330e-06 4.995e-08 -34.05 -31.42
6 1.200e+05 1.484e-05 5.573e-07 -151.23 -148.60
7 1.400e+05 1.890e-05 7.095e-07 144.09 146.72
8 1.600e+05 1.144e-05 4.294e-07 -149.40 -146.77
9 1.800e+05 2.003e-05 7.521e-07 136.16 138.78
Total Harmonic Distortion: 0.004495%

Here you can use my LTSpice simulation.
DADO-TT-TMC-new is D. Self CCS

AndrewT 6th February 2011 09:59 AM

By the time Self published edition 2 of his power amplifier book, his writings on slew rate had become confused and he reached no conclusion other than he could not solve the asymmetrical slew rate, problem.
Read Cordell and many others for info on slew rate.

dadod 7th February 2011 08:13 AM

In the 5th edition D. Self suggests to use (figure 8.49) small Cs=6pF with result +42 -43 V/usec. I simulated it in LTSpice attached in previous post, but I have got to much distortion as result. Schematic is DADO-TT-TMC-new when unzip the file.

dadod 21st February 2011 01:33 PM

Amp powered up and some listening
7 Attachment(s)
I have powered up this amp from +-50V temporary power supply. You can see it on the photo.
I was listening it for short time, only one chanell, in mono. Sound was good and clear. I need stereo set up to evaluate it more accurately.
With 0.1 ohm as emiter resistor the bias increased for 17% from 25 degree C to 50 degree C.
With 0.22 ohm it increase for 15% for the same change in temperature.
I run it without input signal and no heathsink, only alu bracket. Output transistor temerature slowly increased from 25 to 50 degree C just run with bias current and at that point I switched off the power.
For 0.1 ohm I adjusted the voltage to 13mV, means bias current to 130mA, and it increased to 17mV at 50 degree.
For 0.22 ohm it was 20mV and it increased to 23mV (from 90mA to 105mA.
I think it is qute good result.
After that I connected amp to signal generator (quite high disstortion about 1%)
On the scope screen are two sinus signals input(lower amplitude) and output on 8 ohm.
Here is also sinus klipping where is visible TMC type of klipping.
With the square signall there are the photos at 1kHz and 20kHz without and with capacitor parallel to he load. First with 10nF and then with 0.1uF.

homemodder 21st February 2011 01:52 PM

I think your LTP degeneration resistors are too large for Jfets, the noise contribution from there will probably be higher than the THD. You shoulnt go higher than 22 ohms and that is pushing it.

All times are GMT. The time now is 12:17 PM.

vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2017 DragonByte Technologies Ltd.
Copyright 1999-2017 diyAudio

Content Relevant URLs by vBSEO 3.3.2