Noting much happens here (at the moment), so I nag (OT) a bit more about the amplifier.
Yesterday we (me and the amplifier) did have a 'sound battle' on foreign ground, it was between the 'Signature Transresistance Current Feedback BIBGT' http://www.diyaudio.com/forums/analogue-source/154210-mpp-880.html#post3446896 and a 'Sphinx Project 18' Welcome to Audiopaul.nl! - Passion for Audio and Music, with show reports, news, articles, and much more!, to be short, and it was a complete victory/walkover for my amp. The(my) amp sounds more controlled, more detailed, and projects the sound into the room, detaching it completely from the load speakers (in this case a pair of 'Apoge Duetta II's Apogee Duetta II loudspeaker | Stereophile.com) and by this creating a much wider and deeper soundstage.
I can tell you I AM HAPPY
Yesterday we (me and the amplifier) did have a 'sound battle' on foreign ground, it was between the 'Signature Transresistance Current Feedback BIBGT' http://www.diyaudio.com/forums/analogue-source/154210-mpp-880.html#post3446896 and a 'Sphinx Project 18' Welcome to Audiopaul.nl! - Passion for Audio and Music, with show reports, news, articles, and much more!, to be short, and it was a complete victory/walkover for my amp. The(my) amp sounds more controlled, more detailed, and projects the sound into the room, detaching it completely from the load speakers (in this case a pair of 'Apoge Duetta II's Apogee Duetta II loudspeaker | Stereophile.com) and by this creating a much wider and deeper soundstage.
I can tell you I AM HAPPY
Last edited:
We could drive it very, very loud (no sweat) way more loud than the Project 18 could. And even when driven into clipping, the new(my) amplifier does not sound nasty, anyway at that point the volume level was so high that no one could tolerate it for very long
According to the RMS-multimeter we could go to 50Vrms into the Apogee without any stress.
Last edited:
In the HP amp i'm toying with I'm using the BC550C/560C almost everywhere for simplicity. Cordell models are clearly very optimistic, but I adjusted the hfe to the one I have measured and they are very consistent (i believe I have either onsemi or fairchild production, since they were bought from USA).
The 2SA1015/2SC1815 both Y-grade that I have about a hundred pieces are much less consistent, probably GR grade would be more similar to the high BC"C" current gain. I didnt test all nor trace all, but I have the same Peak device (btw get the new firmware/sw is more detailed ^^) and I could do that.
I believe to not have very good models of them, or at least not optimistic as the ones I have for the BC. As drop-in spice simulation worsens the THD.
They run at around 10mA as drivers, so I think I should try as replacement for the BC550/560 as i thought
Unfortunately I dont have a spectrum analyzer nor a very low distortion oscillator yet.
The BC5xx have good quasi-sat for both polarities. The A1015/C1815 do not. So don't use the latter for current mirrors. Try changing back to BC5xx for just the current mirrors and see what happens.
Actually, BC550/A1015 may be a great complimentary combination where high Hfe and good quasi-sat behavior are the top priorities.
Actually, BC550/A1015 may be a great complimentary combination where high Hfe and good quasi-sat behavior are the top priorities.
The circuit is not mine, I just changed some values and the output stage, to use the readily available PCB.
I'm learning spice (among other things) btw.
Kean: I need to get BC550B and higher grade toshibas to match them otherwise 500 and 200 hfe doesnt sound that good
but it's an interesting idea. There are all BC mounted, exactly as in the schematic I posted in the linked thread.
Telstar, as I recall the greatest thing to improve THD was to make the current mirror gain as equal as possible, IE get R2/R11 as close as possible, increasing gain in other parts of the amplifier. The current mirrors should provide the LEAST amount of the total gain, rather than most of it. Unless you fix this it's unlikely that swapping transistors will change much.
Yeah, I agree. I'll leave it like it is, and just evaluate the sound.
I'll go work to the Marsh and my own design shortly.
Joachim, do you think you could do quasi-sat and linearity tests on some RF transistors like the BF240 and BF199? Or the BFG198/194? I don't know of a lot of transistors in this category so I couldn't pick out a complimentary pair.
One parameter that interests me for many transistors is log conformity, but that is even less useful than quasi-sat measurements, for most things.
One parameter that interests me for many transistors is log conformity, but that is even less useful than quasi-sat measurements, for most things.
In the HP amp i'm toying with I'm using the BC550C/560C almost everywhere for simplicity. Cordell models are clearly very optimistic, but I adjusted the hfe to the one I have measured and they are very consistent (i believe I have either onsemi or fairchild production, since they were bought from USA).
The 2SA1015/2SC1815 both Y-grade that I have about a hundred pieces are much less consistent, probably GR grade would be more similar to the high BC"C" current gain. I didnt test all nor trace all, but I have the same Peak device (btw get the new firmware/sw is more detailed ^^) and I could do that.
I believe to not have very good models of them, or at least not optimistic as the ones I have for the BC. As drop-in spice simulation worsens the THD.
They run at around 10mA as drivers, so I think I should try as replacement for the BC550/560 as i thought
The toshiba pair have higher early voltage so would have lower distortion but this only at lower frequencies than say 1 Khz. Above those frequencies their higher Cob will cause higher distortion than the BCs. What you should do with the Kumisa design is use baxandall super pair as the vas. Not only is loopgain increased but its more linear and vas temperature is more stable due to the vbe´s cancelling. I dont believe the current mirror plays any part in lowering THD and the same can be had without its use.
