♫♪ My little cheap Circlophone© ♫♪

I dont mind designing a board.
It would be double sided with the output + driver xistors to one side for heatsinking. All components will be TH. Gerber files will be posted.
Elvee: I will need wattage values of Rs that are higher than 1/4W and #s of other components that should be on the heatsink.
Q: Is it worthwhile to epoxy Q3+4 and Q5+6 to each other to improve temp. tracking?
Cheers, E
 
I dont mind designing a board.
It would be double sided with the output + driver xistors to one side for heatsinking. All components will be TH. Gerber files will be posted.
Elvee: I will need wattage values of Rs that are higher than 1/4W and #s of other components that should be on the heatsink.
Q: Is it worthwhile to epoxy Q3+4 and Q5+6 to each other to improve temp. tracking?
Cheers, E
Thanks for your involvement.

-All resistors can be 1/4W
-The output transistors, Q8 and Q10 obviously need a heatsink, as do the drivers Q9 and Q11. All four can be a common heatsink, or be cooled individually, or grouped in any convenient way.
-Q5 and Q6 do not normally require a heatsink unless operation at supply voltages above 45V is desired.
-No direct thermal contact is required between transistors, but the following pairs should preferably be mounted side by side: Q3/Q4, Q2/Q7, Q5/Q6, Q12/Q13.
-For D7, any small TH schottky will do, BAT81 or BAT85 f.e.
-D4 and D5 can be any 3A or 5A schottky, 1N5820 or something larger in TO220.
-Splitting D8/D9 is not required, but gives more flexibility.

-I didn't include an inductive Zobel network, because it would only be necessary in very extreme cases, but provision for it could be made on the PCB (and strapped in most instances).


What is the purpose of R14 (27 Ohms, between base and collector of Q12) ?
It provides a first order cancellation of the dynamic resistance of Q12, making the quiescent current largely independent of the supply voltages.

And i suggest using 20kHz for THD testing as this is where an amplifier shows its true colors.

IF its above 0.08% at 20kHz, its not worth building because it will not sound good at all.
Is 0.08% a magic number?
Here is the sim at 20KHz, the THD is doubled compared to 1KHz, but as I said earlier, the sim is not reliable at higher frequencies, and I'll need to make an actual measurement to confirm the value.
 

Attachments

  • THD20K.gif
    THD20K.gif
    35.6 KB · Views: 4,067
Well i may just be rambling here, but in my world, a amp that does 0.1% THD at 20kHz is a bad amp.

Though i realize i contradict myself here since the stereo npn amp simulates 0.1% or worse at 20kHz and kinda wonky phase, yet its one of my best sounding amplifiers.
 
Elvee: It is your design. Show me where you would implement a Zobel so I can include it on the pcb.

Wavebourne: Good point, but wouldn't your ears/brain take more time to recover than the electronics? The 16V squarewave output into the 3 Ohm load shows that it does the job (power suply permitting).
 
Last edited:
Wavebourn: But for C10 there are no caps in the signal path. Nor a transformer or tube in sight. I do not believe this design was to be used s a "classic" guitar amp.

Tekko: I understand what was ment by a Zobel, but it is Elvees design and the designer should make that addition.

To the design of a board: I was thinking of using the DIN 100x160 mm board. Anybody would like to comment on a pin-out for the 64/96 pin connector or any other aspects of the pcb layout?

Cheers, E
 
Wavebourn: Enlighten us. I did not find your question about C1, or an answer to it, on this thread. Also, I did not state a belief (relevant, or not), I believe. E

I believe the author knows what I was asking about. I did not ask you about classic guitar amps, and I did not ask silly questions about what is in the signal path.

Here is again what I asked:

Very nice pictures, but what about dynamic characteristics, like small signal distortions right after huge long impulses?

According to my own experience, servo in audio power amplifiers works best when it is used to correct small deviations from design center values, only and only when needed, and it have to be with as minimal as possible 2'nd order non-linearity. Otherwise envelope rectification causes modulation of working point that causes dynamic distortions that are tricky, because they cause distortions of music, but hard to spot on static measurements.
 
Servos

@ Wavebourn

Otherwise envelope rectification causes modulation of working point that causes dynamic distortions that are tricky, because they cause distortions of music, but hard to spot on static measurements.

I was told some time back by someone "Skilled in the art" that servos are no good for audio. In particular, it can mess up the bass. What you said seems inline with his investigations & testing.
 
Elvee: It is your design. Show me where you would implement a Zobel so I can include it on the pcb.
Here is a modified version of the schematic.

Can we see more pics of your board?
Here they are, plus pics of another prototype.
Note that with the ~10cm of wiring between the heatsink and the board, the transient waveforms are a shade less immaculate than the ones shown for the other prototype (they remain excellent though).

Originally Posted by Wavebourn
Very nice pictures, but what about dynamic characteristics, like small signal distortions right after huge long impulses?

Actually, I've asked the author. Your belief is irrelevant, because the question is regarding C1.
This is a highly valid point, and when designing the bias processor, I had in mind the dismal behaviour of the Visch's topology when subjected to wide dynamic variations. See an example here:
http://www.diyaudio.com/forums/soli...rossover-distortion-1975-a-9.html#post2078842
The Circlophone shows nothing of the sort, but there might indeed be low level issues hidden somewhere.
I will think about ways of testing this aspect.
And if you have ideas about tests that could be made, both in sim and reality, they are welcome.
 

Attachments

  • TheCirclophone11.GIF
    TheCirclophone11.GIF
    32.3 KB · Views: 3,856
  • DSCF0010.JPG
    DSCF0010.JPG
    300.3 KB · Views: 3,714
  • DSCF0017.JPG
    DSCF0017.JPG
    308.4 KB · Views: 3,367
  • CircloBis2.JPG
    CircloBis2.JPG
    317.5 KB · Views: 3,219
  • CircloBis1.JPG
    CircloBis1.JPG
    333.1 KB · Views: 15,553
  • CircloBis3.JPG
    CircloBis3.JPG
    294.1 KB · Views: 1,477
  • CircloBis4.JPG
    CircloBis4.JPG
    343.8 KB · Views: 1,499
Here is a first test along the lines suggested by Wavebourn.

The signal consists of 9 1KHz cycles at full power and 1 cycle with the amplitude reduced to ~1% of the maximum.

First pic shows the general outlook and the distortion: there is a marginal increase in the THD, but it remains quite gentle (LTspice calculates the THD over the last cycle of the waveform).

The second pic shows the instant of the transition magnified

The third pic is the voltage across the loop filter cap, C3 in this schematic, C1in the first schematic.

This tends to indicate (and is confirmed by the actual tests), that the servo is very clean and well behaved, and cannot be "derailed" by transient conditions the way the Visch is.
 

Attachments

  • CircloDyn1.GIF
    CircloDyn1.GIF
    60.5 KB · Views: 1,567
  • CircloDyn2.GIF
    CircloDyn2.GIF
    56.2 KB · Views: 1,135
  • CircloDyn3.GIF
    CircloDyn3.GIF
    60.9 KB · Views: 1,017