I’ve been looking at the A75 driver stage, if split in half the topology bares a very strong resemblance to the A0, A0s, and Aleph1 (not to be confused with the Aleph1.2). It follows that the Aleph1 should be able to drive a complementary follower output stage in class AB. I have drawn it out in a pdf document. The driver stage is absolutely unchanged and the area in the red dotted line is the new output stage. It can be visualized of as half of an A75 or as an Aleph1 whit a complementary follower output stage.
I don’t know what purpose the 1k resistors from the gate drivers to the output serve, perhaps feedback. But I put them in because they where in the original A75 article. It should be able to drive 24 mosfets (12N and 12P). Thermal stability running class AB will need some fiddling whit to get right I would imagine.
After reading the A75 article I still don’t understand the benefits of making the amp symmetric. With out symmetric operation it cuts out 4 mosfets and one trim pot.
Dose anyone have any opinions on if and how you think the amp may perform in class AB and what may be done to it to improve it or make it simpler.
Well, it’s 1:48am here so it’s time for me to get to bed.
I don’t know what purpose the 1k resistors from the gate drivers to the output serve, perhaps feedback. But I put them in because they where in the original A75 article. It should be able to drive 24 mosfets (12N and 12P). Thermal stability running class AB will need some fiddling whit to get right I would imagine.
After reading the A75 article I still don’t understand the benefits of making the amp symmetric. With out symmetric operation it cuts out 4 mosfets and one trim pot.
Dose anyone have any opinions on if and how you think the amp may perform in class AB and what may be done to it to improve it or make it simpler.
Well, it’s 1:48am here so it’s time for me to get to bed.
I built a simplified A75 a few years ago and tried running the front end with one half silenced by a capacitor. The amp worked fine and, if anything, sounded better to my ears. I did not measure this setup, but you are cutting the drive current in half, so you can expect some nominal increase in HF distortion.
I got a harmless but audible turn-on "pop" with this setup, but that may have been due to the cap time constant rather than to asymmetry. DC stability was unchanged, which it should have been as the amp still had symmetrical DC gain. If you build as shown in your PDF, you may see a little more DC drift, but probably nothing prohibitive.
I got a harmless but audible turn-on "pop" with this setup, but that may have been due to the cap time constant rather than to asymmetry. DC stability was unchanged, which it should have been as the amp still had symmetrical DC gain. If you build as shown in your PDF, you may see a little more DC drift, but probably nothing prohibitive.
Would increasing the constant current source provide more drive current or would it just raise the dissipation of that stage? The A1 has a current of about ~33mA it looks to me, and the A75 has a current of about 20mA, which may compensate a little.
I guess a cap of a trivial value placed across the Vgs multiplier would be a good idea to.
I guess a cap of a trivial value placed across the Vgs multiplier would be a good idea to.
LBHajdu said:
It will improve the drive in two ways. First, it will increase the negative current available to drive the complementary follower gates. In the circuit you propose, this is equal to the CCS current minus whatever current flows in the 10K loading resistors.
Second, it will cause the active P-MOSFET to operate in a region of higher transconductance, showing a greater change in output current for a given drive voltage. This increases open-loop gain and helps compensate for the gain lost due to asymmetrical drive.
LBHajdu said:
Of course the A75 was a collaborative piece with Norm Thagard,
and I recall that was interested in the symmetric aspect of
the design, and not afraid of a couple extra parts. Me, I over-
built the output stage, so you see I'm not afraid of a couple
extra parts either.
I looked at your schematic, and it looks fine to me. Certainly it
will work well in a high-bias AB.
LBHajdu said:“Not afraid of a couple extra parts” I think that’s the Mark Levinson philosophy to.
Back when my amplifier company was a lot bigger than his,
Dan D'Agostino and I used to hang out. He told a lot of
hilarious Michael (Dayton) Wright stories, and one of my
favorite comments of his was his description of DW design:
"Never use one part when seven will do."
I confess to having a slightly schizoid approach to circuits. I try to keep the audio portion as simple as possible, but do not mind odd, complex power supplies. It is not unusual for me to have independent, regulated rails for each gain stage. Differentials' current draw sums to DC, which makes things easy. So do a few other topologies. However, take a basic opamp, wherein you start with a differential, then go to a single-ended voltage gain stage, then a follower (aka current gain) output. The second and third stages wreak havoc on the supply rails. The obvious first line of defense is to put in lots of power supply capacitance. It works pretty well, it's cheap, and it's conceptually simple. However, I prefer to separate the three stages and give each its own regulated supply so as to keep one stage from influencing any other. This also allows you to play tricks with asymmetrical rail voltages and such, but that's another story for another day.
No, I don't generally regulate the output stage of power amps. It's an awful lot of trouble.
But that doesn't stop me from thinking about it...
Another way to look at it is the old computer programmer's chestnut about loops: You're not solving a thousand problems--you're solving one problem a thousand times.
Design one regulator, then do-until-done.
Complexity, obviously, is in the eye of the beholder.
The old thing about Einstein saying that everything should be as simple as possible, but not simpler has been used to justify the most horridly complicated circuits...just by declaring that that was the simplest circuit that would do the job.
The A-75 is actually a very simple amp. The symmetry of the circuit is a reflection of a given set of design goals. Bear in mind that a push-pull circuit will cancel distortion without the application of feedback. If your intention is to lower distortion while minimizing feedback, then a push-pull circuit is a valid option. There are hidden benefits, too. There is a universal problem in that more gain=more distortion. Demanding that a single gain stage supply all the voltage swing, for instance, will lead to higher distortion than if the gain is generated by two similar stages working in opposition. Each device will stay in a more linear operating region.
It all depends on where you're trying to go. Sometimes more parts, sometimes less, and sometimes you manage to create magic by stumbling on a synergistic combination that goes beyond the sum of its components.
Grey
No, I don't generally regulate the output stage of power amps. It's an awful lot of trouble.
But that doesn't stop me from thinking about it...
Another way to look at it is the old computer programmer's chestnut about loops: You're not solving a thousand problems--you're solving one problem a thousand times.
Design one regulator, then do-until-done.
Complexity, obviously, is in the eye of the beholder.
The old thing about Einstein saying that everything should be as simple as possible, but not simpler has been used to justify the most horridly complicated circuits...just by declaring that that was the simplest circuit that would do the job.
The A-75 is actually a very simple amp. The symmetry of the circuit is a reflection of a given set of design goals. Bear in mind that a push-pull circuit will cancel distortion without the application of feedback. If your intention is to lower distortion while minimizing feedback, then a push-pull circuit is a valid option. There are hidden benefits, too. There is a universal problem in that more gain=more distortion. Demanding that a single gain stage supply all the voltage swing, for instance, will lead to higher distortion than if the gain is generated by two similar stages working in opposition. Each device will stay in a more linear operating region.
It all depends on where you're trying to go. Sometimes more parts, sometimes less, and sometimes you manage to create magic by stumbling on a synergistic combination that goes beyond the sum of its components.
Grey
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.