|
Post by leonski on Oct 5, 2018 17:22:20 GMT -5
might be a way to run cooler, save energy and still have the benefits of 'A'. input level is monitored and feeds forward to the BIAS power supply. Level goes up? Bias does too. With low or no signal the amp is still in 'A', but not drawing lots of current and generating so much heat.
Krell calls it 'sliding bias'. Neat idea which they have tried to optimize.
|
|
|
Post by AudioHTIT on Oct 5, 2018 18:18:49 GMT -5
... Pass indicates in the linked article that 'A' has the most bias and 'B' has NONE. the A/B amp has some bias ... I think that’s kinda what I said.
|
|
|
Post by AudioHTIT on Oct 5, 2018 18:21:40 GMT -5
might be a way to run cooler, save energy and still have the benefits of 'A'. input level is monitored and feeds forward to the BIAS power supply. Level goes up? Bias does too. With low or no signal the amp is still in 'A', but not drawing lots of current and generating so much heat. Krell calls it 'sliding bias'. Neat idea which they have tried to optimize. Another of Pass’ ideas, he called it dynamic bias and it’s what put him and Threshold on the map, well, that the amps sounded great didn’t hurt either.
|
|
|
Post by leonski on Oct 5, 2018 20:33:53 GMT -5
I'm thinking not TOO many high fidelity amps use pure 'B' amplification. This in a push-pull configuration where you have a 'P' type device and and 'N' device dealing with 1/2 of the waveform each, respectively. But Transistors do nothing below the threshold voltage. They need to be over the 'threshold' voltage which is (by semiconductor physics) about 0.6 to 0.7 volts. If you apply that much voltage to a transistor (both of 'em in a PP arrangement) you are at the edge of A/B. In a zero bias condition, you will have no output between about +0.7 volts and -0.7 volts, where the transistors won't turn 'on'. Higher gate voltages will allow more current, so there is apparently an advantage to higher gate voltages. I don't offhand know what the 'limit' of gate voltage would be, but it SHOULD be over 10 volts and probably higher.
Yes indeed. Pass had 'dynamic bias' and the Threshold amps are STILL sought after. He also had some patents on the 'Stasis' technology which was taken up by Nakamichi, too. The difference between the Pass system and Krell might be where they sense the signal level. If you feed forward from the input to the PS is different than say.....sensing the demand from the OutPut and linking back to the PS. Might be a couple other ways to do it, too.
A transistors performance curve is a long, 'Lazy S'. The tail up and down are bad places. The long flat part in the middle is the sweet spot. The tail at one end will be a pretty sharp upswing as gate voltage rises to and than past the threshold. I guess you get into somekind of breakdown situation at the other end of the 'S'.
|
|
|
Post by AudioHTIT on Oct 5, 2018 23:48:53 GMT -5
I'm thinking not TOO many high fidelity amps use pure 'B' amplification. Above I speculated none ... Yes indeed. Pass had 'dynamic bias' and the Threshold amps are STILL sought after. He also had some patents on the 'Stasis' technology which was taken up by Nakamichi, too. The difference between the Pass system and Krell might be where they sense the signal level. If you feed forward from the input to the PS is different than say.....sensing the demand from the OutPut and linking back to the PS. Might be a couple other ways to do it, too. I think Nelson described his as feed forward, but I wouldn’t want to characterize anything beyond that. ... but we digress
|
|
|
Post by leonski on Oct 6, 2018 1:00:27 GMT -5
Actually, Audio, it's worse than either of us would suspect. I'm thinking Class 'B' amps would almost exclusively be confined to PA or sound reinforcement stuff. Nothing anybody I know would want in their living room. I suspect that at normal listening levels with normal program music, this is an audible defect. Even if you were to use Very $$$ 'matched' transistors. Lots of hand labor involved. And I LOVE to Digress. Lots of cool stuff out there. It would be interesting to compare the Pass system to whatever Krell came up with. Agreed. The Pass article pretty much speaks for itself and anyone interested in a 'Class Act' should read it. It's not that complicated and Pass writes very clearly for the Non-Engineering types. Another type of 'A' topology is what is called 'Single Ended' This is another 'flavor' of 'A' and is an interesting problem. Most or at least many 'A' amps are a push-pull arrangement. EACH transistor is 'on' all the time, but only one at a time handles the Plus part of the waveform while the other handles only the Minus. Their is NO 'crossover' problems since both transistors are always in the 'sweet spot' and don't shut off. Not to forget the on-off state of the transistor is based on Bias. Schematic of the Pass First Watt F5, a push-pull class 'A' amp. Not that hard to see if you look at Figure #5 which shows Q3 and Q4 as the 'push' and 'pull' transistors. The output to the speaker is in the MIDDLE and goes thru the speaker BACK to ground, not shown. You'll see if you look closely that the arrow on Q3 points IN while the arrow on Q4 points out. These are N channel and P channel, respectively. Push and Pull, as it were. www.firstwatt.com/pdf/prod_f5_man.pdfThe Amp Camp Amp, while about 8 watts per channel and over the horizon from most Emotiva fans, is a very interesting amp and Single Ended. Push and Pull are both handled thru a single transistor. www.firstwatt.com/pdf/art_amp_camp_1.pdfThe schematic immediatly after 'simple 3' looks to be one channel of the amp. Q1 is the output device. Look thru the other drawings and see how it 'builds' from a simple idea to illustrate the principle, adding parts for fidelity, stability and the inclusion of a feedback loop. I've looked over the documentation of this amp and think I could build one from the kit. It's COMPLETE including case and a small switching power supply. In another thread someone owns some Klipsch La Scala speakers which I'd LOVE to try with this amp. If anybody in SoCal wants to build one of these, let me know. We can set up shop in my Living Room for a weekend as long as we promise not to burn any Holes in my good Oak Dining Room Table! I'll light the grill, later in the afternoon.
|
|
|
Post by Casey Leedom on Oct 6, 2018 2:09:41 GMT -5
Well, another key difference between the Krell and Pass technologies is that, the Krell uses Feedback in a big way and Nelson Pass eschews Feedback. Interestingly, the Krell does Current Feedback instead of Voltage Feedback. Apparently this is somehow supposed to work better with loads which dynamically change their reactance.
Casey
|
|
|
Post by leonski on Oct 6, 2018 13:10:05 GMT -5
Pass has written a lot of stuff. In several articles he talks about different ways to configure transistors. For an FET? Common Source. Common Drain. or Common Gate, I suppose. Each way has different characteristics. Some of which are useful for hi-fi, some not so much. Some Pass amps are only 2 gain stages. Other amps are higher numbers with various 'nested' feedback loops and all sorts of stuff. Pass will try to make the simpliest amp which will actually amplify. Not the best measuring or sounding, but but it works.
Doubt many would argue with the ReSULT of whatever Krell is doing, without ragard to how. I'm not going to complain, that's for sure.
|
|
|
Post by AudioHTIT on Oct 6, 2018 15:05:35 GMT -5
|
|
|
Post by leonski on Oct 6, 2018 22:54:54 GMT -5
Had I known about this even 2 months ago, I'd have been in attendance. As it turns out, I'll defer until next year and than report back. DIY is alive and well. Thanks for the heads-up, AudioHTIT. See 'ya there!
|
|