There are to find three versions of this power amp in the book "Power Amp Projects" (collection of released amp projects in Audio Amateur 1970 - 1989):
1) Citation 12 page 65/66, Fig. 5/7
2) MOSFET 12 (page 65/66 Fig. 6/8) the quasi complementary version
3) Revised MOSFET 12 (page 73, right side near "Correction"
The crazy clipping behavior is to observe by number 2, the the quasi complementary version. The reason is clearly only cause the output power buffer topology.
For only N-Channel in the output there are some other topologies - look about
http://www.diyaudio.com/forums/soli...better-audio-non-complements-audio-power.html
But only if I use a quasi complementary arrangement there are this clipping behavior with overshooting transients; even without front end voltage driver stages and even in pure class A mode.
Where I must introduce anti saturation diodes to avoid such overshooting by clipping?
1) Citation 12 page 65/66, Fig. 5/7
2) MOSFET 12 (page 65/66 Fig. 6/8) the quasi complementary version
3) Revised MOSFET 12 (page 73, right side near "Correction"
The crazy clipping behavior is to observe by number 2, the the quasi complementary version. The reason is clearly only cause the output power buffer topology.
For only N-Channel in the output there are some other topologies - look about
http://www.diyaudio.com/forums/soli...better-audio-non-complements-audio-power.html
But only if I use a quasi complementary arrangement there are this clipping behavior with overshooting transients; even without front end voltage driver stages and even in pure class A mode.
Where I must introduce anti saturation diodes to avoid such overshooting by clipping?
I found oscilation in my quasi project early on but only in the VAS and lower MOSFET driver transistor. Both sources of oscilation were killed stone dead by 220pF between the collector and the base.
What appears to happen is there is too much gain and so you get over shoot.
The capacitors slow the amp down a bit and stop the overshoot.
Does the amp clip on positive or negative phases or both ?
What appears to happen is there is too much gain and so you get over shoot.
The capacitors slow the amp down a bit and stop the overshoot.
Does the amp clip on positive or negative phases or both ?
Last edited:
Yes I had a similar experience playing around with a naim clone.
When changing TIP41/42 for faster drivers (eg 2sc4793/2sa1837) I got oscillation on the positive phase. I was able to kill the oscillation by putting 47pF across collector and base of the driver responsible.
My question about this is, do you guys think it is better to go with fast drivers (and low cob) and kill the oscillation with a polystyrene capacitor or would it be better to select drivers that don't cause the oscillation in the first place?
When changing TIP41/42 for faster drivers (eg 2sc4793/2sa1837) I got oscillation on the positive phase. I was able to kill the oscillation by putting 47pF across collector and base of the driver responsible.
My question about this is, do you guys think it is better to go with fast drivers (and low cob) and kill the oscillation with a polystyrene capacitor or would it be better to select drivers that don't cause the oscillation in the first place?
Last edited:
only on the negative phase, as you can see by second pdf at post#1
Please note, this is not the complete amplifier - only the output buffer stage. Please note additional, that primilliary the clipping behavior is the main problem, not an oscillation problem. The observed oscillation occurs only during clipping. If I avoid clipping, there is no oscillation effect and no advice by upper roll off of the frequency response.
But I have found a better solution to minimize this crazy clipping behavior, still unsatisfactory, but a step in the right direction see pdf files at botton.
By some old quasi complementary output stages with NPN BjTs from british power amplifier (A&R Cambridge e. g.) similar solutions are to find, mostly a parallel network in series to the emitter line of lower driver transistor, that consist a diode, a resistor and sometimes a capacitor.
Last edited:
Is R23 really 2000 ohms ?
I usually use 390R in series with the gates.
Quite a few designs protect the gates with zener diodes too.
There are no differences between 200 ohms and 2000 ohms (R23) regarded clipping behaviour.
The complete article and the modification advices from Audio Amateur now as PDF download available:
http://www.passdiy.com/pdf/citation.pdf
If I perform the modification (introduce of an P-Channel MOSFET, see schema last page of this pdf file) in the power stage, perfect clipping behavior is present.
But I don't understand the reason for the curious clipping behaviour in case of quasi complementary topology like the first mentioned MOSFET circuit. If I enhance the driver collector current of the lower driver transistor, no influence is to observe.
The complete article and the modification advices from Audio Amateur now as PDF download available:
http://www.passdiy.com/pdf/citation.pdf
If I perform the modification (introduce of an P-Channel MOSFET, see schema last page of this pdf file) in the power stage, perfect clipping behavior is present.
But I don't understand the reason for the curious clipping behaviour in case of quasi complementary topology like the first mentioned MOSFET circuit. If I enhance the driver collector current of the lower driver transistor, no influence is to observe.
Last edited:
Citation 12 "Citation Twelve" Photos from commercial Version
by post 89 about
http://www.diyaudio.com/forums/soli...-models-quasi-complementary-power-output.html
you will find some pics about the commercial version
by post 89 about
http://www.diyaudio.com/forums/soli...-models-quasi-complementary-power-output.html
you will find some pics about the commercial version
imho, such capacitor will be more linear than inherent capacitances of semiconductors, and hence the distortion for higher frequencies would in essence be lower.
About Ebay Item number: 280524141846 there is to find an old Citation 12, certainly of interest for evaluation.
- Status
- Not open for further replies.