Just pre-ordered my copy today.
Thanks to Bob for working on a very exciting publication.
Thanks to OP for bringing this to our attention.
Thanks to members for the good discussions in this thread.
Thanks to Bob for working on a very exciting publication.
Thanks to OP for bringing this to our attention.
Thanks to members for the good discussions in this thread.
As far as i understand the AD797 the cap just bypasses the ouput buffer at higher frequencies. There is another distortion reduction mechanism inside that comprises a bootstrapped current mirror that is a bit less easy to comprehend.
Hi Charles,
I agree, the Locanthi T circuit is hard to beat. Apparently some have had some trouble stabilizing it against HF parastic oscillations, but that goes with the territory of using 3 EF's, especially when the transistors are fast. Also, sometimes thermal stability can be a little more tricky because fully six junctions are in series to make up the bias spread. Of course, the use of ThermalTraks helps here, but the pre-driver and driver TC's must be considered.
Double output stages, like the Darlington and standard CFP, just don't have enough current gain, and they are very vulnerable to beta droop, especially when driving low-Z loads. It is remarkable to note that Doug still uses the CFP as an example in his latest article in Linear Audio on TMC compensation. I also have very strong reservations about using CFP output stages in general, which I explain in my book.
BTW, I had the pleasure of knowing Bart Locanthi - he was a brilliant character.
Cheers,
Bob
Sorry, not a Transformer Volume Control. It's a fully discrete circuit with variable gain. The gain setting resistors are on a rotary switch driven by a stepper motor.
Not sure how you do your math. For a given transconductance, a lateral part will have about 1/10 Cgd as a vertical part. Cgd is what matters for a follower output stage.
Thank you. You've managed to create a very worthwhile thread -- just by announcing your book!
But you have to remember one thing -- VERTICALS SUCK.
The comparable P-channel lateral part is only 40 pF at 10 volts. Furthermore the rise at low values of Vds is 5x to 10x lower than verticals. The only reason to use verticals is because the designer is too cheap to spend a few extra dollars on additional lateral devices if they want higher transconductance and/or SOA.
What you say is true about vertical parts. But there is really very little reason to use a vertical part.
You have to remember 30+ years ago when power MOSFETs became available. One of the attractions was that the input impedance was so high that the amplifier topology could be radically simplified. But this was only true for lateral parts.
When the vertical parts came along, the silly audio designers thought that they could be blindly substituted for the lateral parts and got poor results with questionable sound quality. But the truth is that vertical MOSFETs have basically no advantage over bipolars.
-Positive tempco that requires compensation circuitry.
- High interelectrode capacitances that require complex drive circuitry
The only advantage is lack of secondary breakdown, which is easily avoided if the designer is aware of it. But there are other disadvantages for vertical MOSFETs:
- High turn-on voltage.
- Extremely poor consistency from part-to-part.
I say if you are looking at MOSFETs, pass the vertical parts by completely. Either use lateral parts (which can be driven directly by the Vas stage), or else use bipolars, which require complex output stages but at least everybody already knows that.
I'm working on it....we'll see!
All true. But that's what separates the men from the boys.
I can't believe anybody would use a double output stage. As you point out, the simply don't have enough current gain. In my tests with a zero-feedback circuit, going from a triple to a double increased the overall distortion of the circuit tenfold. For what? To save a dollar or two in parts?
I'm looking forward to reading your reasons. I don't like CFPs because I don't like the way feedback sounds. But that probably isn't your reason!
I envy you that. He was a true renaissance man. He developed mechanical analogies for loudspeakers, designed some of JBL's best transducers, switched fields and developed the JBL amplifier with the "T-circuit" output stage (in 1965!), and then when the bean counters who bought JBL in the late '60s or early '70s fired him for coming at noon without wearing a tie, he joined forces with Pioneer to create a full line of pro drivers for TAD that are still considered some of the finest ever developed. And I'm sure that there is much more that I don't know about... RIP, Bart.
T for Transconductance in TVC, i haven't gotten used to the variable gain term VGT yet.
Oh, sorry!
The folks that make tapped autotransformers for volume controls call those TVC's.
We had to come up with something for a name. I don't know that "Variable Gain Transconductance" volume control was the best possible one, but it has the advantage of sounding reminiscent of a high-performance automobile -- "GT" for "Gran Turismo" or "Grand Touring". Ferrari made this very popular in the early '70s.
Say it a few times -- it starts to roll off your tongue!
Hi Pete,
Yes, that was quite awhile back, but I seem to recall it may have been a DC convergence problem when I used the part in a VAS and pre-driver. I think I solved it at the time by connecting a 10 meg resistor from the VAS output node to ground. I don't remember the details, but I recall tweaking the model and simplifying it some - might have even had some help from Andy.
Anyway, I never did figure out exactly what was wrong with the original model. I've dug out the tweaked one I believe I have been using and posted it below. It is the 2SC3503.
Cheers,
Bob
Mr Hansen,
the all-aluminum body 350-GTV of the early '60s tickles the fantasy more, still (Gran Turismo Veloce).
Try to say Ferruccio, rolls-off much easier.
As you've returned again to be bugged, any news on the x-r p ?
(rp-x, rx-p, pr-x, xr-p, px-r)
Attachments
Last edited:
Hi Charles,
I guess I have to disagree with you about vertical MOSFETs. They don't suck. They have their advantages and disadvantages compared to both BJTs and laterals. With any of these technologies, it often has more to do with how the technology is applied. Some guys will be able to make a better amplifier with BJTs, while others can do better with MOSFETs.
You are correct that many naive designers made bad amps with MOSFETs - both laterals and verticals. It is also true that a less experienced designer can get into more trouble with vertical MOSFETs than with BJTs or laterals. There are many good reasons to use verticals, but they are certainly not without their own shortcomings. It really amounts to what dragons you want to slay.
As you know, I really like the ThermalTrak BJTs - I think that they have made BJTs competitive with MOSFETs for achievable sound quality by greatly reducing the thermal stability problems with BJTs. MOSFETs are about 10 times faster than even the ThermalTraks, thanks to no minority carriers. But yes, they do need more care and feeding, and often carry with them some added costs for the same power. For example, they may idle a bit hotter than a corresponding BJT design, and they usually need a bit more power supply headroom as a result of the higher turn-on voltage that you mentioned.
It is true that laterals tend to have less gate-drain capacitance, but they are pretty pathetic in the transconductance department. They also require a very high turn-on voltage to achieve high current. MOSFETs have no beta droop or ft droop, and they will put out tens of amps from a single device without breaking a sweat.
I'll be the first to agree that there are a lot of bad MOSFET amplifiers out there, but that is not the fault of the MOSFETs. The same thing can be said for negative feedback 🙂.
Cheers,
Bob
Vertical MOSFETs will put out tens of amps without breaking a sweat. Laterals have such high Rds(ON) (which increases with temperature) that their saturation voltage can be as high as 12V for a mere 7A drain current. Poor transconductance of laterals is only part of the story - unless they're paralleled (and laterals are more expensive than verticals) power supply utilization is going to be fairly poor even with boosted rails for the driver stages.
I agree with you Bob - the matter is one of an individual designer's preference and their own personal armory of tricks and techniques for taming the devices and extracting the best from them.
Hi Charles,
Actually Bart went on to play a big role in the early days of digital audio. He was certainly a very multi-talented guy.
Cheers,
Bob
Another one pre-ordered from amazon.co.uk but, unfortunately, not delivered before December. I'm eager to receive it.
Thanks for your interest, massimo. The book actually rolled off the presses last Thursday, so it is pretty much on schedule at this end. I'd be surprized if there was actually a 2-month delay before availability in the UK. Maybe Amazon is just being conservative, or got wrong or conservative information from McGraw-Hill. BTW, I have not even seen the physical copy yet; maybe mine will come in the mail today. I can't wait; I'm like a kid on Christmas Eve.
Cheers,
Bob
Hi Abraxalito,
Another concern that I have about lateral MOSFETs is their gate resistance, which can be on the order of 50 ohms or more. I believe that this may form a distributed R-C network with the gate capacitance as a result of the physical origin of the gate resistance.
This can result in a poor transconductance frequency response. Take a look at the Renesas 2SK1056 datasheet and you will see a plot of the frequency response of transconductance. It is down 3 dB at only 3 MHz. This is not a fast transistor.
Cheers,
Bob
It's here!
Not to tease you guys too much, but when I arrived home last night there it was in the mail - my first physical copy of the book! My wife Angela was duly impressed, even though she'd be the first to admit she doesn't understand a word of it.
We celebrated with a bottle of champaigne. Do I sound excited? You bet!
I hope it will make its way to Amazon and B&N soon.
I'll look forward to getting feedback from you guys, both good and bad and errors. MGH gave me "tear sheets" to allow me to bring minor errors to their attention to fix in the second printing (I don't know when that will be, and I don't know how big the first printing was).
Cheers,
Bob
Not to tease you guys too much, but when I arrived home last night there it was in the mail - my first physical copy of the book! My wife Angela was duly impressed, even though she'd be the first to admit she doesn't understand a word of it.
We celebrated with a bottle of champaigne. Do I sound excited? You bet!
I hope it will make its way to Amazon and B&N soon.
I'll look forward to getting feedback from you guys, both good and bad and errors. MGH gave me "tear sheets" to allow me to bring minor errors to their attention to fix in the second printing (I don't know when that will be, and I don't know how big the first printing was).
Cheers,
Bob
Congrats Bob! I know from experience how this must feel!
Looking forward to actually hold your Magnus Opus!
jan didden