I would expect a bigger issue (non-issue for audio) is line and load regulation. Both thermals (possibly balanced out if thought through) and current variations through the transistors as the load changes to cause a change in output voltage. The protection diode should ensure the transistor is fine unless C57 and C59 are big enough to fry it.
When using just two secondary's, one for + and one for - voltage, I have found it very important to make sure that the two secondary voltages are out of phase with each other just like you would have with a single CT secondary and one bridge for a bipolar supply.
It's easy to connect them up to two bridges and have them both in phase and the sound will not be as good, especially the bass, compared to having the two secondary's out of phase.
I wonder how many power supplies are unknowingly built with the secondary's connected to the bridges in phase and not out of phase. I have done it.
This power supply feeds the final regulators in the preamp. I actually do not want tracking or even matching rail voltages. Depending on the stage and load the final regulator may vary. If a stage drives an inductor or high currents I prefer a shunt.
Re :
http://www.diyaudio.com/forums/loun...ch-preamplifier-part-ii-1052.html#post3930667
Two silly questions :
1) Gain can be changed by changing R1 / R4 ?
2) Why not mirror image U1b at the diff pair to get balanced output ?
Patrick
http://www.diyaudio.com/forums/loun...ch-preamplifier-part-ii-1052.html#post3930667
Two silly questions :
1) Gain can be changed by changing R1 / R4 ?
2) Why not mirror image U1b at the diff pair to get balanced output ?
Patrick
I guess that speaks for itself.
(As an ex service tech it was the kind of arrangement that drew my eye)
"... since the sixties ..."
So what?
First, I've never seen anything ever by anybody which could not be improved in absolute terms. It's not always done because the improvment might cost a lot and bring only very modest results, but it CAN be done.
Second, I believe at least 95% of the power amp topologies we use today have been around from the 70ies. We use some variant or subvariant of them and never think twice. So why be picky here and now?
If something sounded great in the 60ies, today it will still sound at least good, although it is unlikely to hold the crown since its inception. So try to improve on it and don't bother with thoughts of its longevity.
So what?
First, I've never seen anything ever by anybody which could not be improved in absolute terms. It's not always done because the improvment might cost a lot and bring only very modest results, but it CAN be done.
Second, I believe at least 95% of the power amp topologies we use today have been around from the 70ies. We use some variant or subvariant of them and never think twice. So why be picky here and now?
If something sounded great in the 60ies, today it will still sound at least good, although it is unlikely to hold the crown since its inception. So try to improve on it and don't bother with thoughts of its longevity.
Second question means output stage?
I needed balanced output and in my experience for a good link output stage I need an opamp with high output drive capability.
I want fast opamp with high output current and high slew rate, and low noise and low distortion at the same time. So my choice, confirmed by listening and measuring, is ADA4898.
That AD4898 is a nice chip!
They don't spec the OL BW but from the PSRR and CMRR I guess it would be a few 100 Hz?
PMA do you know what causes the large drop in large-signal BW? Normally I would susp[ect slew rate limiting but at 55V/uS probably not.
Any idea what causes this, circuit wise?
Jan
They don't spec the OL BW but from the PSRR and CMRR I guess it would be a few 100 Hz?
PMA do you know what causes the large drop in large-signal BW? Normally I would susp[ect slew rate limiting but at 55V/uS probably not.
Any idea what causes this, circuit wise?
Jan
Hi Rick,
It's early in the morning and I'm scratching my head over what you have said here. I'm trying to see the difference in whether the two secondaries are out of phase in the topology you are talking about. I am assuming you are talking about two, non-center-tapped secondaries, each of which is connected to a bridge rectifier to form a floating full-wave power supply. Each of those supplies has one output end connected to ground. The other side of one provides the positive rail, the other side of the other provides the negative rail. Since full-wave bridges are used, both secondaries are conducting on both peaks of the AC waveform. So I'm missing the difference in why the reversal of one of these two secondary windings would matter.
Cheers,
Bob
1) Gain resistor goes between R1/R4 at JFET sources. This way gain can be very high but there is a lower limit to the gain. R1-R4 want to stay a matched quad for common mode circuit.
2) The idea was for single ended out, the originals were diff-in/diff-out.
Because one winding handles the positive (output peaks)
while the other handles the negative. If they are not phased
correctly they are in a sense fighting each other ...
"Because one winding handles the positive (output peaks)
while the other handles the negative. If they are not phased
correctly they are in a sense fighting each other ..."
But if both supplies use full wave bridges, there can be no imbalance between the mains half cycles.
Please explain yourself further, with a diagram.
Or is this just something you heard/read?
while the other handles the negative. If they are not phased
correctly they are in a sense fighting each other ..."
But if both supplies use full wave bridges, there can be no imbalance between the mains half cycles.
Please explain yourself further, with a diagram.
Or is this just something you heard/read?
Yes, and even if the two secondaries are each a CT secondaries, each secondary conducts on each sine peak, pos and neg.
Jan
Bob,
If one winding is used for the positive rail and the other negative, consider the AC analysis when the "in phase" winding is passing current with noise to the positive rail and the "out of phase" winding is passing current and the same noise out of phase to the negative rail. Now connect a capacitor between the rails! (Of course this assumes both supplies are commoned at ground.) Now consider the case where there are capacitors from each power rail to ground. In the rail to rail capacitor version if I used two capacitors in series what would the noise voltage be if I looked between ground (common) and the capacitor capacitor connection? (Zero for those who hate my Socratic method.)
If one winding is used for the positive rail and the other negative, consider the AC analysis when the "in phase" winding is passing current with noise to the positive rail and the "out of phase" winding is passing current and the same noise out of phase to the negative rail. Now connect a capacitor between the rails! (Of course this assumes both supplies are commoned at ground.) Now consider the case where there are capacitors from each power rail to ground. In the rail to rail capacitor version if I used two capacitors in series what would the noise voltage be if I looked between ground (common) and the capacitor capacitor connection? (Zero for those who hate my Socratic method.)
- Status
- Not open for further replies.