We had one of those SAE power amps in our Swiss lab in 1974, we used it for the frequency range of 400-4KHz in our 3 way horn loaded speaker design. OK for this application, but full range, the Electrocompaniet Lohstroh-Otala prototype sounded better with Stax SR-3 Headphones. Just my opinion, of course. I bought the Electrocompaniet and rejected the SAE, also available. It just goes to show that a fancy input stage is not EVERYTHING in a successful amp design.
VERY INTERESTING, Wrinkle. I was only a boy of 12 years when this patent was applied for, playing with relays found in a junk yard. It just goes to show that the PREPARED MIND is able to reach ahead of the rest of society. I am sure that some experimental transistors were available even in the 1950's that could duplicate these circuits, but they were VERY expensive, and outside the reach of regular people. We just had 12ax7's, etc to play music with.
Last edited:
Hi John,
The reviewer in question did this to a 300 wpc power amp with 125 V rails at full power. There most probably were no speakers or load connected at the time. I was tasked with undoing the damage. I think the resistor power rating was marginal as long as music was being reproduced. As you know, a clipped sine wave has an equivalent DC heating power that is above the normal RMS computation. So while the resistor was under-rated for the power it could see in use, the damage was caused by an abnormal operating condition. I had to replace the PCB (I still have the damaged one). I look at it now and again to remind myself how much energy some components may have to withstand.
The trend to lower impedance feedback networks may visit this unfortunate event on any unsuspecting designer. Even with lower output power devices, such as a preamplifier maybe.
As for that early patent, I'm not surprised. Human beings are far too curious to have not tried that arrangement, as you can see. I don't know that many of us would be well motivated to search for prior art for "our baby". An attempt at a patent may have brought the prior art to the surface.
Still, if you hadn't run into it before, it is still an idea that you had and made work. There's high value in that. From my viewpoint, your circuit makes perfect sense if you're trying to reduce or eliminate input bias currents or increase the standing current in order to increase the bandwidth of the circuit. Input bias current reduction would have been my first guess.
You may find it extremely rewarding to finish those projects yourself John. If not, Burning Amp would be a great venue to get them into interested and capable hands. I have to get back down there one year soon!
-Chris
I wouldn't be so harsh. The designer was either Bob Carver (but he left before the amp was completed) or maybe a guy I talked to a lot there. If anything, it was an oversight I think.
The reviewer in question did this to a 300 wpc power amp with 125 V rails at full power. There most probably were no speakers or load connected at the time. I was tasked with undoing the damage. I think the resistor power rating was marginal as long as music was being reproduced. As you know, a clipped sine wave has an equivalent DC heating power that is above the normal RMS computation. So while the resistor was under-rated for the power it could see in use, the damage was caused by an abnormal operating condition. I had to replace the PCB (I still have the damaged one). I look at it now and again to remind myself how much energy some components may have to withstand.
The trend to lower impedance feedback networks may visit this unfortunate event on any unsuspecting designer. Even with lower output power devices, such as a preamplifier maybe.
As for that early patent, I'm not surprised. Human beings are far too curious to have not tried that arrangement, as you can see. I don't know that many of us would be well motivated to search for prior art for "our baby". An attempt at a patent may have brought the prior art to the surface.
Still, if you hadn't run into it before, it is still an idea that you had and made work. There's high value in that. From my viewpoint, your circuit makes perfect sense if you're trying to reduce or eliminate input bias currents or increase the standing current in order to increase the bandwidth of the circuit. Input bias current reduction would have been my first guess.
As a young lad, all electronic parts were expensive. Everything we had came from old TVs. Capacitors were highly prized and relays (good ones) were trophies.
$6 to $10 for a single signal transistor in the 60s and early 70s. Every discarded electronic item was the equivalent to a treasure. That's probably why so many of us have trouble throwing away any part that might be good (as opposed to might be defective - so throw it out!). I'm a bit surprised that you allowed that P&G volume control to collect dust in the closet John! The top part at the time, and there it sits.
You may find it extremely rewarding to finish those projects yourself John. If not, Burning Amp would be a great venue to get them into interested and capable hands. I have to get back down there one year soon!
-Chris
Whose 'baby' Anatech? This is no documentented evidence that the COMPLEMENTARY DIFFERENTIAL input stage was designed prior to 1968, as of yet. The ONLY supporting evidence to the contrary is the patent shown to me by Dimitri about some Bell Labs patent, but that was not a design for normal amps, either.
Please look closer to the schematics. The 1952 patent is for the 'balanced bridge' complementary output stage. The 1975 patent is a little late, don't you think?
Please look closer to the schematics. The 1952 patent is for the 'balanced bridge' complementary output stage. The 1975 patent is a little late, don't you think?
For the record, everyone. When complementary transistors were contemplated back in the early 1950's, circuit theorists 'went to town' anticipating future applications of USEFUL complementary devices. They apparently attempted to PATENT every application that they could think of, as well. Have you ever seen the patent of the DARLINGTON transistor configuration? Look it up, yes even an obvious connection like that, at least to us, today, was patentable. It was at this time that complemetary output stages, complementary Darlingon pairs and complementary feedback pairs of output transistors were all invented. However, they did not become practical until the mid 1960's, at least 10 years later. This was due to the COST and the AVAILABILITY of complementary transistors. The military may have used similar circuits in the '50's for special applications, but they have not seen the light of day, as of yet, in any publication that I have ever seen. Anyone out there with some really old actual complementary circuits?
Hi John,
Pretty innovative stuff, but a totally different market aim from what you do. I'm afraid that this amp wasn't cheap, nor shoddy in design. The units made as Carver went out of business did get shoddy though, I had to look after those failures as well.
I haven't gone on a search to discredit your claims in any way. I just recognize that if I had come up with a cool idea, I would pretty much expect the very bright lads at Bell Labs to have already developed any worthwhile concept I could think of. Is this so difficult to imagine for you as it applies to what you've thought of?
The IP rights clause is pretty common for US companies to include in employment contracts. I had a few myself. In response, I decided that if I came up with a good idea, unrelated directly to my work, that I'd quit right away. If it was related to my work, well then they are paying me to do exactly that. Extra compensation would be nice (but unexpected).
As for the commercial usefulness of the complementary differential input stage that is so popular, it's not for mass production unless these can be produced in integrated format to ensure temperature tracking and close matching parameters. You are well aware of this as your own products use hand matched parts. I've made a fair amount of money over the years by going over mass produced products using this design and doing what they should have done. Matching those parts by hand. I'm actually very tired of matching all four transistors, but the design pretty much demands it. From my point of view, and many out there might agree with me on this, this is called being both cheap and shoddy. Wouldn't you agree with this? It's as if they all completely missed the point and basic operation of how this circuit is supposed to operate.
-Chris
Ouch! What do you have against Carver?
Pretty innovative stuff, but a totally different market aim from what you do. I'm afraid that this amp wasn't cheap, nor shoddy in design. The units made as Carver went out of business did get shoddy though, I had to look after those failures as well.
It doesn't matter where the circuit was applied in the patent world, and that will not take away from the fact that other people came up with similar ideas either.
I haven't gone on a search to discredit your claims in any way. I just recognize that if I had come up with a cool idea, I would pretty much expect the very bright lads at Bell Labs to have already developed any worthwhile concept I could think of. Is this so difficult to imagine for you as it applies to what you've thought of?
Haven't looked as yet. I may. Bell Labs is the first place I would tend to look though.
Absolutely! What is interesting about this granted patent is that is was, granted. Since you hadn't taken the steps to patent your work, even after leaving Ampex, the question forms. Why not? After leaving Ampex, they had no IP rights for anything you could develop after you left. (you hid your work, you naughty boy!)
The IP rights clause is pretty common for US companies to include in employment contracts. I had a few myself. In response, I decided that if I came up with a good idea, unrelated directly to my work, that I'd quit right away. If it was related to my work, well then they are paying me to do exactly that. Extra compensation would be nice (but unexpected).
As for the commercial usefulness of the complementary differential input stage that is so popular, it's not for mass production unless these can be produced in integrated format to ensure temperature tracking and close matching parameters. You are well aware of this as your own products use hand matched parts. I've made a fair amount of money over the years by going over mass produced products using this design and doing what they should have done. Matching those parts by hand. I'm actually very tired of matching all four transistors, but the design pretty much demands it. From my point of view, and many out there might agree with me on this, this is called being both cheap and shoddy. Wouldn't you agree with this? It's as if they all completely missed the point and basic operation of how this circuit is supposed to operate.
-Chris
Figure 1 and figure 2 of the 1962 patent do not preclude being used at the input, only figure 3 is definitely an output stage.
not sure how to attach a png of it though?
Wrinkle
Anatech, I designed and used the complementary differential input stage at Ampex, both in 1968 for a prospective portable monitor amp, and in 1969 at Ampex Research for a 2000 W, complementary differential input balanced bridge power, current out motor drive amplifier. IF I put it in my patent notebook, then it would be Ampex's property, not mine or anyone elses. I didn't do it on purpose. However, that did not mean that I could talk about it to just anyone. However, at Alembic, in 1970, we made a number of power amps and line drivers using the complementary differential input stage. If you want to follow up, just call Alembic Inc in the Santa Rosa area in CA as ask for Ron Wickersham, my boss at the time. He is still alive and available to tell you what you need to know. His number is: 707-523-2611. Please, either follow up or hold your peace.
Hi John,
Or cascode, or current mirrors ....
I'd like to see the Bell work actually. Can you post it John? Dimitri?
Thankfully, these days a patent isn't supposed to be granted for any circuit change that is obvious to "practitioners of the art". Or anything that quite naturally follows an existing patent. Thank goodness for some common sense in the patent office!
-Chris
Or cascode, or current mirrors ....
I'd like to see the Bell work actually. Can you post it John? Dimitri?
Early ones? Of course not! They were not economically feasible in the consumer entertainment marketplace. Quasi-complimentary output circuits are still in use even though complimentary devices are no longer hideously priced. This is only to lower the cost of production, except in cases where a designer uses it to be different, or to create their "corporate sound".
Thankfully, these days a patent isn't supposed to be granted for any circuit change that is obvious to "practitioners of the art". Or anything that quite naturally follows an existing patent. Thank goodness for some common sense in the patent office!
-Chris
Hi John,
I'm not trying to discount anything you've done. I'm only being a realist about this. No attack in your direction.
If I wished to pick at you, I could simply point out that you actively withheld a useful invention that had application for your employer, contrary to the legal obligation that existed due to the employment contract that you signed in good faith. They could have sued your butt off for that. Not really the most honest thing you could have done back then.
Would you have had the patent (with Ampex to back you up) and Ampex as the Assignee? This is how most of the professional engineers that work for semiconductor companies dealt with this issue. I'm sure application engineers could really rack up a portfolio of patents simply by doing their job.
Anyway, no one is out to get you. You're seeing threats in the bushes that don't exist.
-Chris
I could if I were a moron with something to prove. I also wouldn't dream of interrupting anyone for such a petty reason, even though I currently have a plan that allows me to call anywhere in North America for no extra charge. This is also true for the countless times you have invited people to call others that they don't know. That's just plain rude for one!
I'm not trying to discount anything you've done. I'm only being a realist about this. No attack in your direction.
If I wished to pick at you, I could simply point out that you actively withheld a useful invention that had application for your employer, contrary to the legal obligation that existed due to the employment contract that you signed in good faith. They could have sued your butt off for that. Not really the most honest thing you could have done back then.
Would you have had the patent (with Ampex to back you up) and Ampex as the Assignee? This is how most of the professional engineers that work for semiconductor companies dealt with this issue. I'm sure application engineers could really rack up a portfolio of patents simply by doing their job.
Anyway, no one is out to get you. You're seeing threats in the bushes that don't exist.
-Chris
For the record, I was working at the Ampex Audio Division, designing a phase locked servo capstan motor drive, and doing advanced electronic research on the next generation tape recorder.
I made this complementary differential power amp, that many on this website saw at Burning Amp to drive my single K-horn that I had at the time.
I tried to interest the Ampex audio department to use it, but it was just too radical in 1968. Then, in early 1969, my boss, (and friend) at the audio division was pushed aside and some new guy came in. Most of us resigned, but I was able to move to the Research Division as I had worked with people there who helped me with the servo design, and we got along. The amp that I officially designed at Ampex, was to deliver 50V-50A current, instantly reversing, and current drive, for a special project.
I used a total of 4 2N5684 and 4 2N5686 300W,50A each devices as a complementary bridge output with complementary differential input, fully 4 quadrant symmetrical, with both positive and negative feedback to generate voltage in, current out.
To understand how to do this, you might look at 'Current Driving Loudspeakers' by Esa Merilainen, who is a member here, to get the general idea.
I DIDN'T put the circuit in my patent notebook. Later, when I had returned to audio design, elsewhere, I was told that they dropped the amp, because they couldn't make sense of it, and it was too complicated anyway. It probably was, for the time, and the application, but I was looking to the future, for audio design. Apparently I was on track.
I made this complementary differential power amp, that many on this website saw at Burning Amp to drive my single K-horn that I had at the time.
I tried to interest the Ampex audio department to use it, but it was just too radical in 1968. Then, in early 1969, my boss, (and friend) at the audio division was pushed aside and some new guy came in. Most of us resigned, but I was able to move to the Research Division as I had worked with people there who helped me with the servo design, and we got along. The amp that I officially designed at Ampex, was to deliver 50V-50A current, instantly reversing, and current drive, for a special project.
I used a total of 4 2N5684 and 4 2N5686 300W,50A each devices as a complementary bridge output with complementary differential input, fully 4 quadrant symmetrical, with both positive and negative feedback to generate voltage in, current out.
To understand how to do this, you might look at 'Current Driving Loudspeakers' by Esa Merilainen, who is a member here, to get the general idea.
I DIDN'T put the circuit in my patent notebook. Later, when I had returned to audio design, elsewhere, I was told that they dropped the amp, because they couldn't make sense of it, and it was too complicated anyway. It probably was, for the time, and the application, but I was looking to the future, for audio design. Apparently I was on track.
Well I tried that and for some reason I do not get that button.
Transistor signal amplifier circuit
That should take you straight to the pdf, you will see a transformer coupled input stage, a capacatively coupled input stage and then an output stage, all on the first page. I think the attorney who helped the writer was quite smart, a lot of care was taken not to limit the scope.
Wrinkle
PS Merril, John B. seems to have some disparate patents:
Transistor gain control - Patent 2941153 General Dynamic corp...
Transistor temperature regulator - Patent 2932714
Zinc silicate phosphor - Patent 2554999 a phosphor patent for sylvania...
Last edited:
Scott,
I suspect a not very good feedback resistor would be noticeable and you don't. PM me a shipping address and I will send you a mono amplifier, left over from my recent AX article, with a make before break selector switch selecting from closely matched but different types of feedback resistors.
I suggest doing the test with friends and beer. You could do a controlled test, but I doubt it is worth the bother.
ES
For everyone, wondering what the 'fuss' is all about and what we are parsing about:
The complementary differential input stage first appeared, at least in audio designs, in the 1960's. The only people that I know to have invented it, was John Iverson and myself, both completely independently. With use, and exposure to others, it becomes 'obvious' like so many circuit topologies, but this was not so, back in 1968.
First, what is the reason that the complementary differential input stage has any added usefulness at all? In my case, it was the EASE in biasing the input stage and REMOVING problematic elements. You see, I started by making an ALL COMPLEMENTARY power amp, from input to output. That was my first design, a year before, using diode biasing of a complementary single pair consisting of an NPN and a PNP transistor, much like many video amps, today, except they replaced the diodes with complementary transistors. (I wish that I would have thought of that then). It was the diode biasing that was the real headache, and the relatively lousy thermal stability due to the fact that the diodes were not thermally connected to the devices they were 'tracking'.
So what I wanted to do was to increase the effective input impedance of the power amp, get it more thermally stable, and make the circuitry as linear as possible.
Of course, the first thought is a differential pair of devices. This gives EASY input biasing, a relatively high input impedance, and thermal stability BECAUSE the 'current source' sometimes just a large value resistor, forces the idle current to remain about the same, even over temperature. Therefore, no input bias diodes and their feed resistors, that potentially distort and lower the input resistance.
However, this does NOT give me everything that I needed. I also needed a complementary drive to the VAS second stage transistors, so that the drive would be symmetrical. This is where the second differential pair came in and it would use a PNP pair of transistors, complementary to the NPN pair that would typically be used in a single differential pair. Putting these two complementary pairs together, biased merely by identical high value resistors to the opposite supplies to each other, and tying the bases together, gave me a SIMPLE input loading scheme of one resistor that could be much higher in value than typical, IF I balanced the betas of the input pairs so they would cancel out each others bias current. This was great, because my tube preamp was straining with the original power amp's input loading.
Still, in 1968 terms, I suspect that only Barrie Gilbert would not scoff at the 'overkill' aspect of the complementary differential input stage.
This was new to me, to anything that I could find in the literature, and useful, if not rather exotic. That is why I hold it dear.
The complementary differential input stage first appeared, at least in audio designs, in the 1960's. The only people that I know to have invented it, was John Iverson and myself, both completely independently. With use, and exposure to others, it becomes 'obvious' like so many circuit topologies, but this was not so, back in 1968.
First, what is the reason that the complementary differential input stage has any added usefulness at all? In my case, it was the EASE in biasing the input stage and REMOVING problematic elements. You see, I started by making an ALL COMPLEMENTARY power amp, from input to output. That was my first design, a year before, using diode biasing of a complementary single pair consisting of an NPN and a PNP transistor, much like many video amps, today, except they replaced the diodes with complementary transistors. (I wish that I would have thought of that then). It was the diode biasing that was the real headache, and the relatively lousy thermal stability due to the fact that the diodes were not thermally connected to the devices they were 'tracking'.
So what I wanted to do was to increase the effective input impedance of the power amp, get it more thermally stable, and make the circuitry as linear as possible.
Of course, the first thought is a differential pair of devices. This gives EASY input biasing, a relatively high input impedance, and thermal stability BECAUSE the 'current source' sometimes just a large value resistor, forces the idle current to remain about the same, even over temperature. Therefore, no input bias diodes and their feed resistors, that potentially distort and lower the input resistance.
However, this does NOT give me everything that I needed. I also needed a complementary drive to the VAS second stage transistors, so that the drive would be symmetrical. This is where the second differential pair came in and it would use a PNP pair of transistors, complementary to the NPN pair that would typically be used in a single differential pair. Putting these two complementary pairs together, biased merely by identical high value resistors to the opposite supplies to each other, and tying the bases together, gave me a SIMPLE input loading scheme of one resistor that could be much higher in value than typical, IF I balanced the betas of the input pairs so they would cancel out each others bias current. This was great, because my tube preamp was straining with the original power amp's input loading.
Still, in 1968 terms, I suspect that only Barrie Gilbert would not scoff at the 'overkill' aspect of the complementary differential input stage.
This was new to me, to anything that I could find in the literature, and useful, if not rather exotic. That is why I hold it dear.
- Status
- Not open for further replies.