Component Swapping
I am waiting on a set of boards from Kristijan to arrive. Looking over the circuit, I want to try IRF044 Mosfets and something else in the front end.
Can the ZVP3310 be used instead of the 9610's here? The lower input capacitance looks like a small improvement. I also see the BC250 has a very low input capacitance.
Does the circuit as shown on Kristijan's diagram have active current gain? Excuse the DA questions, the inner details of how the Aleph circuits work still eludes me.
Looking it appears that a higher than normal output impedance is unavoidable due to the single output devse used. The 044 should help compared to a 240.
George
I am waiting on a set of boards from Kristijan to arrive. Looking over the circuit, I want to try IRF044 Mosfets and something else in the front end.
Can the ZVP3310 be used instead of the 9610's here? The lower input capacitance looks like a small improvement. I also see the BC250 has a very low input capacitance.
Does the circuit as shown on Kristijan's diagram have active current gain? Excuse the DA questions, the inner details of how the Aleph circuits work still eludes me.
Looking it appears that a higher than normal output impedance is unavoidable due to the single output devse used. The 044 should help compared to a 240.
George
Re: Component Swapping
Don't know about the ZVP3310 but the 9610's are exceptionnally linear devices.
I took IRFP140 also in part of better transconductance (and as by Nelson Pass himself, the lower voltage transistors should sound better).
The circuit does have active current gain, yes, through the 1.3k / 220uF feedback network. Since I changed resistor values and IRF's, I had to readjust it.
BTW - My output Z is a riddle because I measured it into a resistor and got Z=0.16 Ohms. Then after some changes including raising the gain to 26 dB, I measured again, this time using real speaker driver, and got (now) Z=0.75 Ohms. So it's not that the devices can't go lower, but that either a resistive R gives a different result from a real speaker, or that (more likely) my daring gain change made the circuit suboptimal in that respect...
Panelhead said:
Don't know about the ZVP3310 but the 9610's are exceptionnally linear devices.
I took IRFP140 also in part of better transconductance (and as by Nelson Pass himself, the lower voltage transistors should sound better).
The circuit does have active current gain, yes, through the 1.3k / 220uF feedback network. Since I changed resistor values and IRF's, I had to readjust it.
BTW - My output Z is a riddle because I measured it into a resistor and got Z=0.16 Ohms. Then after some changes including raising the gain to 26 dB, I measured again, this time using real speaker driver, and got (now) Z=0.75 Ohms. So it's not that the devices can't go lower, but that either a resistive R gives a different result from a real speaker, or that (more likely) my daring gain change made the circuit suboptimal in that respect...
MikeW said:
Hot Damn, a question I can answer!
The 221 value of the gate resistors is not critical. They are there to damp any ocillations in the circuit.
The 221 value is used a lot here, but anything from 100 - 300 ohms should work as well. These are used the same as grid resistors in tubed circuits. In tubed circuits the resistor of choice is a carbon comp to minimize high frequency occilations. The grid resistors are normally around 1K ohms or higher. The high impedances used in most tube circuits is what creates these problems.
George
Listening, and component choices
Well... after sorting out my self inflicted weird problems, I built the second channel, using the Aleph30 front end this time. Now that I have both channels operating, I should say something about my impressions about the sound. Works like a charm.
I had my mid driver on chip amps before, now Mini-A. It surprised me that there seemed little gross differences. But there seems to be a lot of small differences which make all the difference... if that makes sense. The largest difference seems much better PRAT - everything is more "normal", in an unspectacular way. Cool, really. You don't even know it's there.
I don't have any bass problems (bi amped, the Mini-A's operate >310 Hz), in fact the bass improved in definition and clarity as well - because now I can use one chip amp per driver. Before, it was one chip amp for 2 drivers. No hum problems in spite of one single 375 VA transformer for 2 Mini-A's plus 4 chip amps, and of course I will fix that configuration too once the transformers arrive for the Mini-A.
Mid neutrality and clarity are exemplary, so are the highs. Again, I don't know how to verbalize it - it's like saying, "no difference, except that now there is nothing wrong anymore".... Before, I was quite happy with the chip amps - except, except... that here and there, there was some hitches, mostly, on certain passages, some low level distortion, something like "jittery highs". The highs on the Mini-A are just straight. Hmm, eventually I should make a proper AB test.
Now for the final part: rebuilding both channels to same topology, values and components, proper PSU and connectors, and final tests.
Here I have a resistor question:
I will so far stick with non-exotic parts. In stores here I found 2 types of power resistors: snow-white-coffin types, presumably inductive, but tested non magnetic, and long, thin ones, presumably carbon, but with end caps, and tested magnetic at those end caps.
What trade off do I choose: non inductive but magnetic - and slightly capacitive (virtue of end caps), or non magnetic but inductive?
My gut feeling is that a little inductivity shouldn't hurt because we want to get rid of that HF hash anway... BUT since these resistors are in the feedback loop, that could actually work the other way around (increasing feedback for HF hash) ... opinions?
Add-on question: I used BC549 instead of BC546. The Vceo is dangerously low (30V) and I have +-24V rails - wonder how come they haven't blown yet... then again, don't touch a running system... I get a Vbe drop of 0.64 V, fine with my choice of source R=0.39. Is it worth changing to BC546 for safety, and have to recalibrate everything from bias to AC current gain feedback?
Well... after sorting out my self inflicted weird problems, I built the second channel, using the Aleph30 front end this time. Now that I have both channels operating, I should say something about my impressions about the sound. Works like a charm.
I had my mid driver on chip amps before, now Mini-A. It surprised me that there seemed little gross differences. But there seems to be a lot of small differences which make all the difference... if that makes sense. The largest difference seems much better PRAT - everything is more "normal", in an unspectacular way. Cool, really. You don't even know it's there.
I don't have any bass problems (bi amped, the Mini-A's operate >310 Hz), in fact the bass improved in definition and clarity as well - because now I can use one chip amp per driver. Before, it was one chip amp for 2 drivers. No hum problems in spite of one single 375 VA transformer for 2 Mini-A's plus 4 chip amps, and of course I will fix that configuration too once the transformers arrive for the Mini-A.
Mid neutrality and clarity are exemplary, so are the highs. Again, I don't know how to verbalize it - it's like saying, "no difference, except that now there is nothing wrong anymore".... Before, I was quite happy with the chip amps - except, except... that here and there, there was some hitches, mostly, on certain passages, some low level distortion, something like "jittery highs". The highs on the Mini-A are just straight. Hmm, eventually I should make a proper AB test.
Now for the final part: rebuilding both channels to same topology, values and components, proper PSU and connectors, and final tests.
Here I have a resistor question:
I will so far stick with non-exotic parts. In stores here I found 2 types of power resistors: snow-white-coffin types, presumably inductive, but tested non magnetic, and long, thin ones, presumably carbon, but with end caps, and tested magnetic at those end caps.
What trade off do I choose: non inductive but magnetic - and slightly capacitive (virtue of end caps), or non magnetic but inductive?
My gut feeling is that a little inductivity shouldn't hurt because we want to get rid of that HF hash anway... BUT since these resistors are in the feedback loop, that could actually work the other way around (increasing feedback for HF hash) ... opinions?
Add-on question: I used BC549 instead of BC546. The Vceo is dangerously low (30V) and I have +-24V rails - wonder how come they haven't blown yet... then again, don't touch a running system... I get a Vbe drop of 0.64 V, fine with my choice of source R=0.39. Is it worth changing to BC546 for safety, and have to recalibrate everything from bias to AC current gain feedback?
Re: Listening, and component choices
Here I have a resistor question:
I will so far stick with non-exotic parts. In stores here I found 2 types of power resistors: snow-white-coffin types, presumably inductive, but tested non magnetic, and long, thin ones, presumably carbon, but with end caps, and tested magnetic at those end caps.
What trade off do I choose: non inductive but magnetic - and slightly capacitive (virtue of end caps), or non magnetic but inductive?
My gut feeling is that a little inductivity shouldn't hurt because we want to get rid of that HF hash anway... BUT since these resistors are in the feedback loop, that could actually work the other way around (increasing feedback for HF hash) ... opinions?
My opinions on this subject are very biased, but I never use magnetic components if avoidable. To my ears, a big fat Allen Bradley 2 watter is the way to go.
For higher power applications I check with my trusty pocket magnet. I sometimes find some plastic cased 5W 1% Dales. And some ceramic cased WW are non-magnetic too.
The long ones you saw may be a version of wirewound. If banded, and the bands are wide these are wirewound in a conventional shape. These are almost always magnetic.
Having the power resistor in the feedback loop should help correct for the non-linearities. Just like having a large value coupling cap. JMHO
George
Add-on question: I used BC549 instead of BC546. The Vceo is dangerously low (30V) and I have +-24V rails - wonder how come they haven't blown yet... then again, don't touch a running system... I get a Vbe drop of 0.64 V, fine with my choice of source R=0.39. Is it worth changing to BC546 for safety, and have to recalibrate everything from bias to AC current gain feedback? [/B][/QUOTE]
Here I have a resistor question:
I will so far stick with non-exotic parts. In stores here I found 2 types of power resistors: snow-white-coffin types, presumably inductive, but tested non magnetic, and long, thin ones, presumably carbon, but with end caps, and tested magnetic at those end caps.
What trade off do I choose: non inductive but magnetic - and slightly capacitive (virtue of end caps), or non magnetic but inductive?
My gut feeling is that a little inductivity shouldn't hurt because we want to get rid of that HF hash anway... BUT since these resistors are in the feedback loop, that could actually work the other way around (increasing feedback for HF hash) ... opinions?
My opinions on this subject are very biased, but I never use magnetic components if avoidable. To my ears, a big fat Allen Bradley 2 watter is the way to go.
For higher power applications I check with my trusty pocket magnet. I sometimes find some plastic cased 5W 1% Dales. And some ceramic cased WW are non-magnetic too.
The long ones you saw may be a version of wirewound. If banded, and the bands are wide these are wirewound in a conventional shape. These are almost always magnetic.
Having the power resistor in the feedback loop should help correct for the non-linearities. Just like having a large value coupling cap. JMHO
George
Add-on question: I used BC549 instead of BC546. The Vceo is dangerously low (30V) and I have +-24V rails - wonder how come they haven't blown yet... then again, don't touch a running system... I get a Vbe drop of 0.64 V, fine with my choice of source R=0.39. Is it worth changing to BC546 for safety, and have to recalibrate everything from bias to AC current gain feedback? [/B][/QUOTE]
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.