Given a decent build, imo the weighted sum of sound quality , appearance, reliability and connectivity is
better then Audio Lab 8200MB,
less then Cambridge Audio AZUR 840W
equal to Slewmaster/Wolf 2014 (DIY)
Question on a too me rather odd capacitor choice...
So we have C1 being this honking great PP film thing that is about an inch on a side, but for C9 we make use of an NP electrolytic?
Why?? Both are in the 'signal path' (I hate that term, it disguises a critical failure to understand the circuit) and distortion produced in either will appear at the output because both C1 and C9 are outside the feedback loop but inject distortion into it. The currents in both parts are even equal more or less.
If an BP Electrolytic is good enough for C9 it is surely good enough for C1 as both connect thru 1k resistors to the bases of the Q1,Q2 LTP? Alternatively it is not good enough for C9 (The distortion measurements would tend to say it is just fine), and both C1 and C9 should be those honking great film parts!
What have I missed?
So we have C1 being this honking great PP film thing that is about an inch on a side, but for C9 we make use of an NP electrolytic?
Why?? Both are in the 'signal path' (I hate that term, it disguises a critical failure to understand the circuit) and distortion produced in either will appear at the output because both C1 and C9 are outside the feedback loop but inject distortion into it. The currents in both parts are even equal more or less.
If an BP Electrolytic is good enough for C9 it is surely good enough for C1 as both connect thru 1k resistors to the bases of the Q1,Q2 LTP? Alternatively it is not good enough for C9 (The distortion measurements would tend to say it is just fine), and both C1 and C9 should be those honking great film parts!
What have I missed?
My post testing different capacitors for C1, showing that a NP lytic can be used, and maybe a mouser search of 220uF + Film caps
C1 has a high pass function while C9 has a low pass function, the later should eat its own distortion. The alternative is a servo.
dmills - A film cap for C1 and an electrolytic for C9 is a compromise. C9 is too large for a film cap.
Ed
Ed
You sure about C9 for that, I figure it is in effect a low shelf in that it plus R9 forces the gain to unity at DC but allows it to rise to 1+R21/R9 at high frequency.
I am not convinced it 'eats its own distortion' any more then C1 does, for all that NP electrolytic caps large enough to have nearly no voltage across them at any frequency of interest are generally pretty harmless anyway (Yea, I know, sacrilege!).
C10 amuses me, I mean do we really have meaningful open loop gain up where a 1pF cap is relevant? If we do, we do, but I expect that sort of thing in high VHF and UHF power amps, first time I have seen it in an audio amp.
Agree that big film is silly, motor start caps have no place in audio amps.
It would be interesting to see (And I might do the experiment) how this thing does as a DC coupled amp, I expect the offset to drift a bit, but it might settle to something that can reasonably be trimmed, no way to really know but to do the test. As my application has all sorts of in box doings at both input and output of the amp, servoing the DC out would not be the worst thing my motional feedback board has to do (I actually need to dial in a specific amount of DC offset so a voice coil temperature measurement bridge can do its thing).
I agree, C10 seems to me a bit of excess and I do not think anyone would would be able to hear the difference between C10R26 in and out but as designers of this amp aimed at the best quality then the difference most likely can be measured. So I ordered four 1pF silver mica caps.
When it comes to well designed amps C9 vs servo I do not think most would hear any difference in domestic environment - that is a listening room not larger than say 30-35m2. From my experience when my ears were young (for some time I was involved in music industry) servo advantage would be heard with large high quality speakers. Very low frequencies are really felt more than heard but with servo everything seems to sound cleaner, not only low frequencies but mid and high frequencies as well. Of course high damping factor amp is required to control low frequencies. Also high slew rate amps sound cleaner at higher listening levels in large rooms or in the open. Making scissors snap sound naturally is a good test for any amp (and speakers).
When it comes to well designed amps C9 vs servo I do not think most would hear any difference in domestic environment - that is a listening room not larger than say 30-35m2. From my experience when my ears were young (for some time I was involved in music industry) servo advantage would be heard with large high quality speakers. Very low frequencies are really felt more than heard but with servo everything seems to sound cleaner, not only low frequencies but mid and high frequencies as well. Of course high damping factor amp is required to control low frequencies. Also high slew rate amps sound cleaner at higher listening levels in large rooms or in the open. Making scissors snap sound naturally is a good test for any amp (and speakers).
I always like jangling a bunch of keys in front of a good mic. A very sensitive test of intermod well out into the ultrasonic, only thing was, sometimes it turned out to be the mic that was the embarrassment.
Incidentally that is also a nightmare test when cutting vinyl, gives both the velocity limiters and the cutting head thermal trip a really good workout.
Incidentally that is also a nightmare test when cutting vinyl, gives both the velocity limiters and the cutting head thermal trip a really good workout.
Its actually very relevant, in our extensive loop gain testing both in simulation and real life testing it has shown to increase the phase and gain margins about 3db and about 4-5°respectively.
This in turn gives you more loopgain which ultimately reduces distortion.
The value of C10 was specifically selected to optimise the slope of the loopgain throughout the entire frequency band right up to the unity gain crossover frequency and the point were the phase shift is 180°. In fact loopgain was one of our main focuses when we were designing the wolverine and selecting the final components. 100's of hours have gone in to optimisation of the simulation and the simulation environment.
Anyone tried 2sc4382 / 2sa1668 as drivers yet? I got some from digikey and the 1668's all have an hfe of around 150+ but the 4382's have hfes around 53-60 will this be a problem ?
That's a better choice. You will be happy with those
The MKP2 fits C112. I'm asking to make sure that a MKP2 (Metallized Polypropylene) can be used or do I need to stay with the BOM recommended MKS2 (Metallized Polyester). I don't want to mess up the sound of the amp.
No blown fuses or smoke 😀 however Im testing without output transistors in, and I'm getting the following issues, identical on both channels - testing with a Variac and a small 30VA 0.63A 15-0-15VAC tx / 20VDC power supply.
A. TP107 and TP108 I get only ~400mV initially, guide states this should be 750-800
G. Im using TTC/TTA for Q101 and only getting 1.2V for both of these measurements, guide states 1.5V
H. My DC offset is 255mV both channels. R25 cannot get this down enough.
I. bias circuit does still function with R109, but due to being so low initially (point A) the highest I can get it to is 1.39V.
Is this all due to a low voltage/bad power supply for testing perhaps? I have a larger transformer here I can try..
A. TP107 and TP108 I get only ~400mV initially, guide states this should be 750-800
G. Im using TTC/TTA for Q101 and only getting 1.2V for both of these measurements, guide states 1.5V
H. My DC offset is 255mV both channels. R25 cannot get this down enough.
I. bias circuit does still function with R109, but due to being so low initially (point A) the highest I can get it to is 1.39V.
Is this all due to a low voltage/bad power supply for testing perhaps? I have a larger transformer here I can try..