Just for the history of it, I might comment on what we found over the decades in solid state power amplifier design.
Back in the early 1960's, the solid state 'revolution' began. This freed up some tube power amps for the 'poor student market', like me at the time, as there are always a group of people who want to try the latest technology. I got my first DYNA MK3 power amp for $40, back in 1963. Don't underestimate the cost, that might have been a month's rent share for an apartment.
These original amps were most certainly made with germanium transistors. Interesting devices, but very difficult to build with and no real complements to use in push pull. They usually used a 'totem pole' output stage that required a driver transformer. I thought it was pointless to just remove the output transformer and have to put in a driver transformer at the time. I stayed away from germanium transistors for audio, because of this and other limitations.
Silicon came in about 1964-5, and even DYNA, while making a significantly popular line of tube kits, decided it was time to go solid state. This is when they introduced the ST-120, the PAT4, and who knows what else? Even the Dyna engineers told me to stick with tubes at the time. However, it was now practical to do direct coupled output stages with quasi-complementary output stages, and this was a real improvement in design elegance.
By the end of 1966, Motorola offered the first POWERFUL TO-3 based complementary output transistors, the 2N3716-2N3792, 10A, 80V, 150W devices. I decided this was the opportunity to make a power amp to drive my K-horn, if I could make something better than the Dyna MK3, what was pretty good, already, only being surpassed by the Marantz tube power amps, and triode amps of the long past.
Now, I was not the only one to attempt to make good power amps, once these devices were introduced. JBL, Marantz, Crown, and several other companies, started to design and develop solid state power amps.
My personal design did NOT have to have the need for really high power, 15W into 8 ohm's was enough. This made my design easier, because a commercial power amp would have to have 40W or more, in order to drive direct radiator loudspeakers. This caused serious design challenges for these commercial companies, because protection circuitry could be just as difficult to implement, and potentially destroy the sound quality of the power amp as any other factor. (more later)
Back in the early 1960's, the solid state 'revolution' began. This freed up some tube power amps for the 'poor student market', like me at the time, as there are always a group of people who want to try the latest technology. I got my first DYNA MK3 power amp for $40, back in 1963. Don't underestimate the cost, that might have been a month's rent share for an apartment.
These original amps were most certainly made with germanium transistors. Interesting devices, but very difficult to build with and no real complements to use in push pull. They usually used a 'totem pole' output stage that required a driver transformer. I thought it was pointless to just remove the output transformer and have to put in a driver transformer at the time. I stayed away from germanium transistors for audio, because of this and other limitations.
Silicon came in about 1964-5, and even DYNA, while making a significantly popular line of tube kits, decided it was time to go solid state. This is when they introduced the ST-120, the PAT4, and who knows what else? Even the Dyna engineers told me to stick with tubes at the time. However, it was now practical to do direct coupled output stages with quasi-complementary output stages, and this was a real improvement in design elegance.
By the end of 1966, Motorola offered the first POWERFUL TO-3 based complementary output transistors, the 2N3716-2N3792, 10A, 80V, 150W devices. I decided this was the opportunity to make a power amp to drive my K-horn, if I could make something better than the Dyna MK3, what was pretty good, already, only being surpassed by the Marantz tube power amps, and triode amps of the long past.
Now, I was not the only one to attempt to make good power amps, once these devices were introduced. JBL, Marantz, Crown, and several other companies, started to design and develop solid state power amps.
My personal design did NOT have to have the need for really high power, 15W into 8 ohm's was enough. This made my design easier, because a commercial power amp would have to have 40W or more, in order to drive direct radiator loudspeakers. This caused serious design challenges for these commercial companies, because protection circuitry could be just as difficult to implement, and potentially destroy the sound quality of the power amp as any other factor. (more later)
Scott, what happened to the your previous avatar, the cute black "demon" ?
The new avatar is cute too, but the previous one was also a bit weird, which in today's culture means cool (as far as I can tell).
Batz maru was so au currant. Shirley is in sync with my time.
Power amp design continued:
By 1968, Ampex had a Crown DC-300 power amp with some JBL monitors for our subjective listening. It seemed OK at the time, but later I found out more about the DC-300's and their limitations. In 1968, as well, Marantz came out with their first solid state power amp, the Marantz 14. I bought one, immediately, first for Ampex Research, for a convention that they needed a power amp for, and later, one for myself. This is where I truly learned about feedback stability in power amps, because the Ampex unit would oscillate with certain cap loads, but my personal unit would not. Still, I learned how to measure for it with a square wave generator, a scope, and a multiple cap value switchbox, at Ampex Reseach, by engineers older and wiser than me.
I measured the Marantz 14, in every way that I knew what to do, at the time, including IM at virtually every level to max output. Looked good to me. So I took the Marantz 14 home to replace my own, homemade unit, and of course, the Dyna Mk3.
What I found when I did this was that the Marantz 14 did NOT sound as good, over time, as the Dyna Mk3 or my power amp. It actually gave me listener fatigue. This surprised me, and it was the first time that measurements did NOT tell me what the amp sounded like. What was going on? (more later)
By 1968, Ampex had a Crown DC-300 power amp with some JBL monitors for our subjective listening. It seemed OK at the time, but later I found out more about the DC-300's and their limitations. In 1968, as well, Marantz came out with their first solid state power amp, the Marantz 14. I bought one, immediately, first for Ampex Research, for a convention that they needed a power amp for, and later, one for myself. This is where I truly learned about feedback stability in power amps, because the Ampex unit would oscillate with certain cap loads, but my personal unit would not. Still, I learned how to measure for it with a square wave generator, a scope, and a multiple cap value switchbox, at Ampex Reseach, by engineers older and wiser than me.
I measured the Marantz 14, in every way that I knew what to do, at the time, including IM at virtually every level to max output. Looked good to me. So I took the Marantz 14 home to replace my own, homemade unit, and of course, the Dyna Mk3.
What I found when I did this was that the Marantz 14 did NOT sound as good, over time, as the Dyna Mk3 or my power amp. It actually gave me listener fatigue. This surprised me, and it was the first time that measurements did NOT tell me what the amp sounded like. What was going on? (more later)
In those days I wonder if *anybody* was thinking in terms of monotonicity or of inherent linearity. Lots of folks to this day discount the importance of either. They could be right, but haven't convinced me yet.
Thanks,
Chris
Well, what did I find about the Marantz 14 that really made a difference? Not much, but I DID find that they were using a high feedback op amp type design, similar to what many use today, some power diodes in series with the output devices, and a class B output stage. I, then, at the time, used 1/2A, they might have used 50ma or less. The amp always ran fairly cool. Could it have been HIGHER ORDER HARMONICS, generated at relatively low listening levels, that made the difference? That is my opinion. Yet, the test equipment of the day could not measure it easily. Too much residual noise and distortion to see any detail at 1W or less, where the K-horn operated. (more later)
Right, I still don't quite understand exactly what Dick did, no comment in any case.
At times I tire of this anachronism. This dabbling becomes tedious and boring.
Last edited:
Scott, this is the situation. Dick Sequerra made a phono preamp that is competitive with my designs, with your AD797, several years ago. He is still happy with the design. He told me that he ran it 'open loop' because he does not like the sonic compromises of global negative feedback. As a competitive designer, I have never personally asked him exactly how he did it, and then I came to a tentative conclusion that your 'extra' pin must be, as well, an open loop frequency comp connection. Then, all that would be needed is to use the pin to create an accurate RIAA contour, and I left it at that. However, later looking into that 'pin' it is NOT located properly to do that. So what did Dick Sequerra do? I thought that you might be interested enough to ask him, as you used to be good friends, wine buddies, and all. If I were you, I would ask him. Perhaps, if you won't, I will, sometime. I am curious about it, but I thought it was your place to ask first, perhaps.
One approach is to deliberately put a DC offset of about 80mV across the input pins. That will reduce the open-loop gain to about the right amount for an MC front end, and make even-order distortion strongly dominant.
DC offset can be fixed by a servo acting on pins 1 and 5 via high impedance. I presume those pins are connected to the emitters of Q3 and Q4 or something equivalent in the real circuit.
The response will roll off 6dB/Oct across the entire audio band so RIAA Eq can be implemented with a simple RC network after the AD797. Pin 8 could be used in the normal way to reduce output stage distortion, if desired.
If the DC offset at the input is made variable, that could be used as a gain control to cater for different sensitivities of cartridge.
DC offset can be fixed by a servo acting on pins 1 and 5 via high impedance. I presume those pins are connected to the emitters of Q3 and Q4 or something equivalent in the real circuit.
The response will roll off 6dB/Oct across the entire audio band so RIAA Eq can be implemented with a simple RC network after the AD797. Pin 8 could be used in the normal way to reduce output stage distortion, if desired.
If the DC offset at the input is made variable, that could be used as a gain control to cater for different sensitivities of cartridge.
Attachments
Before I lose track, I might go on about solid state amplifier problems.
By now, complementary output stages were more or less, standard. Then, Motorola made some devices that had a higher voltage breakdown. Crown stayed with quasi-complementary output, as did Phase Linear so that they could just just NPN power transistors.
The REAL problem with these newer designs was that their protection circuitry was continuous and self recovering, so that the sound quality would significantly suffer with REAL loudspeaker loads, but would not come on with rated RESISTIVE loads, so that a fair amount of power could be advertised, yet it could not really put it out with a real load.
The reason for this was that the SAFE AREA of the power transistors was continually being exceeded with reactive loads. So while increasing the power supply voltage made good specs, as far as maximum power output was concerned, the necessary protection circuits reduced the possible peak current that reactive loads required. This made a 150W into 8 ohms solid state amplifier sound less loud, before clipping than a 75W tube amp. Quite a surprise at the time. Later, in the 1970's we tried to extend the working range of solid state amps with more multiple output devices, for example, and more sophisticated V-I limiting circuits. (more later)
By now, complementary output stages were more or less, standard. Then, Motorola made some devices that had a higher voltage breakdown. Crown stayed with quasi-complementary output, as did Phase Linear so that they could just just NPN power transistors.
The REAL problem with these newer designs was that their protection circuitry was continuous and self recovering, so that the sound quality would significantly suffer with REAL loudspeaker loads, but would not come on with rated RESISTIVE loads, so that a fair amount of power could be advertised, yet it could not really put it out with a real load.
The reason for this was that the SAFE AREA of the power transistors was continually being exceeded with reactive loads. So while increasing the power supply voltage made good specs, as far as maximum power output was concerned, the necessary protection circuits reduced the possible peak current that reactive loads required. This made a 150W into 8 ohms solid state amplifier sound less loud, before clipping than a 75W tube amp. Quite a surprise at the time. Later, in the 1970's we tried to extend the working range of solid state amps with more multiple output devices, for example, and more sophisticated V-I limiting circuits. (more later)
I also made the AD745 null network in a very eccentric way so that the feedback to the null pins trick that Walt was very fond of works optimally. It makes a nice open-loop amplifier or RIAA.
Scott,
Yes, it does.
Now, how about you come clean and show to the community all the different neat ways within which we can abuse al sorts of Op-Amp's you either designed or know very well.
For example, I am completely peeved that Burr Brown did not bring out the two transconductance nodes in the OPA1632. Maybe we should start a petition to get them added?
Ciao T
Yes, it does.
Now, how about you come clean and show to the community all the different neat ways within which we can abuse al sorts of Op-Amp's you either designed or know very well.
For example, I am completely peeved that Burr Brown did not bring out the two transconductance nodes in the OPA1632. Maybe we should start a petition to get them added?
Ciao T
(Allow me to misinterpret what you meant.)
Most of the older US post offices already have over built lobbies! Marble everywhere and very high ceilings often arched. As they are closing many in the near future you might look at buying one of your own.
Any resemblance to some posters here who miss the intent of the issues is entirely satirical.
The AD744 works well with an external PNP connected to output, comp, Vee (e,b,c). This was before complimentary processes, we actually made this up for a customer in Germany, Dr. Hell. RIP
http://www.hell.de/topics/etopics/archive/en/hell_etopics_en_2002_6_28.pdf
Last edited:
- Status
- Not open for further replies.