I was tried to designing blameless amplifier (Lin topology). First, I want highest slew rate as possible. Then I tried to make lowest THD at 20 kHz. And I want high PSRR, too. Like you said, everything must be compromise and I did not want to go too complex. I like the way it sound, although at high frequency CFA topology is my favorite.
If you mean adding the usual RC lowpass at the amplifier input, this is one of the worst things you can do. Power amplifiers with BJT inputs draw non-linear currents from the source. This is of no consequence if they are driven from low impedances like 50 Ohm, but the series resistance of the RC (usually 1k or more) turns the distorted current into distorted voltage, and this can easily double the distortion from the amplifier. With unregulated supply rails you will also get a side-order of hum. This dreadful business is fully covered in APAD5 and APAD6, (p142 onwards) but the message does not seem to have got out there.
I blame myself for not publicising this more. There must be thousands of amplifiers out there distorting and humming much more than they need to because of one apparently harmless resistor.
This is what I specifically tried to avoid in the Sixth Edition. You will note that a whole new and lengthy chapter was devoted to some detailed research on various balanced or push-pull VAS designs. This greatly extends the design options available, but unfortunately the only conclusion that anyone could draw from the investigations is that the Blameless configuration with its simple single-ended VAS works much better and more dependably.
Not for the first time, I have found that clever configurations turn out to be not so clever, which is a bitter result for someone like me who takes joy in ingenious circuitry. Instead there is the Blameless configuration, which looks conventional at a first glance, but actually contains a number of rather subtle features that enable its good performance. Maybe there's some sort of philosophical lesson there.
I am more than happy to discuss the balanced/push-pull VAS issue in detail.
I'm going to point out (rather belatedly) that I don't think I have ever used Zeners or LEDs in a power amplifier biasing scheme. I have used either ordinary silicon diodes or BJT base-emitter junctions.
Jan, I was just about to say that!
It's the preferred way to generate white noise if you don't want to get into digital PRBS generators. I've used it many, many times.
I don't know it is misleading name or not. I do not care
. It is current on demand input like VSSA (Lazy Cat) or NX-Amp (Bonsai).I have not build ABX switch, although I can do it with AVR microcontroller. I just an amateur, may be in the future I will do it.
Why on earth do you want to make SR as high as possible? That's exactly the kind of misplaced single-issue focus that will not get you to the best amp, all around.
Jan
Oops... was buying from Kindle without paying enough attention and got the wrong edition
If you contact Amazon customer service, I'm sure that they will allow you to change the order to the new edition for the price difference. They have always been very helpful to me.
When I build VSSA variant and compare to another amp, like Lin topology, symmetry topology (like Leach amp) it sound different at mid and high frequency. Even when compare with Marantz SR6200, VSSA variant is better.
Then I learn how to design amplifier with current on demand input.
I design the blameless with high slew rate (around 90V/uS with +-44V DC power supply), I want compare it with amplifier with current on demand input. Is slew rate the most importance specification?
My blameless THD at 71W/8Ohm, 20kHz -> 0.000287% and PSRR at 100Hz -> 118 dB on simulation. I will verify it if I get better sound card with ARTA software.
What do you think?
Last edited:
Yet another of my (probably unwelcome) subjective comments. Recently I had a chance to listen both Sansui AU217 and AU317 (35+ years) old amps in mint condition. The only unusual thing about the topology of both amps is push-pull VAS. I am surprised how consistently well received the sound of this topology is among users, for almost 40 years. Probably just because of subtle coloration from the VAS stage. My friend was tempted to dump his brand new expensive Lavardin when he heard old Sansui(s). And I have heard that some big manufacturers recently revived this VAS topology.
Last edited:
The point is that in normal use it is horrifically less than Class-B.
There is no inevitable link between distortion and efficiency. But an amplifier that is needlessly inefficient is not an elegant solution to any problem, unless the problem is room heating.
It certainly looks that way!
Let us be quite clear. The Blameless amplifier is not a "standard Lin configuration". The Lin amplifier was 2-stage, The Blameless amplifier is 3-stage.
You might have most/all the IC industry in disagreement. OTOH I do have an interesting data point that I might run by Samuel if I can dig out the schematics.
Doug,
I always keep the input R below 1k, 100-470r and set the corner above 1mhz. As waly stated I know this is to avoid rf hash from entering or at least attenuate it somewhat. Without measurements I have heard the difference of setting this filter corner too low and using too big of a resistor in series with the input.
Colin
I always keep the input R below 1k, 100-470r and set the corner above 1mhz. As waly stated I know this is to avoid rf hash from entering or at least attenuate it somewhat. Without measurements I have heard the difference of setting this filter corner too low and using too big of a resistor in series with the input.
Colin
They've got it easy. Matched devices, all at the same temperature.
As described in APAD6, one of the major problems is temperature drift, leading to unbalanced Ic's in the input pair, and a one-way ticket to Distortion City.
Member
Joined 2009
Paid Member
Doug, with all respect, I am confused. On your website you say:
Distortion In Power Amplifiers
"Fig 1a shows the generic Lin power amplifier circuit, with the now universal differential input stage, representing something like 98% of the amplifiers ever built. It is the obvious starting point for amplifier investigation. [1] Fig 3 shows its distortion plot; there are two distortion regimes. Below 1 kHz THD is low at 0.002% but not zero, the noise floor being 0.0006% approx. Above 1 kHz, THD quadruples with each octave and reaches 0.5% before 20 kHz.
The basic topology is a transconductance amplifier (voltage- difference input, current output) driving a transimpedance (current-to- voltage converter) Voltage Amplifier Stage, followed by a unity-gain power buffer. The voltage at the VAS transistor base is typically only a couple of millivolts, and is of little interest in itself; it is the current passed from the input stage to the VAS that counts. This topology has many advantages, including simple compensation."
Fig 1a
So the Lin amplifier in Fig 1a has an input stage, VAS and output stage - looks like 3 stages to me. In fact, the same 3 stages as in a Blameless amplifier. Hence my confusion.
C'mon, you do got to be kidding. 1 or 2 degrees, easy in any event to thermally couple. I think this is a non issue - you just have to take care with the execution of the design - as with all else.
All in all I must say Doug's book is a good read, and the blameless is an excellent amplifier to start out with but Golden eared or not it's definately far from the end all be all. The fact of the matter is that there was very little if nothing submitted of much value in the book to cfa. Since this is a technology rapidly gaining praise in new designs it's well worth a look into, the Lin Bailey has been perfected
Colin
Colin
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.