Actually to be accurate I was being both sarcastic and ironic. Having to explain that takes the humor out
Single-Ended Triode (SET) Amplifier Revolution of the 1990s. Did it really happen? Was it Successful?
- Thread starter caesar
- Start date
Lol We’re talking music reproduction. You’re deflecting…. to carpenters. HahahahahaMeasurements are used by people whos hearing is not good
Same as a carpenter who constructs a house without a measure , same as a mechanic who has to align a coupling without a dial indicator or laser.
All the people who use tools are basically incompetent , thats what you re saying .
Luckily in the high end problems can be solved with cables , lol
Just about the only industry in the world , where you can get away with that
Never said you didn’t...but to rely only on measurements is not right either...They emerge every once in a while again because they are true , you cant keep them hidden
I designed speakers in the beginning by ear as well .
But to design a accurate speaker (or horn ) you need measurement equipment
No, go back and read what you wrote. It was flat statement...Actually to be accurate I was being both sarcastic and ironic. Having to explain that takes the humor out
The irony is that part where while I said that in jest, I really can show measurements to back up my claim. So it looks like a flat statement, but if you've been watching how I post, I don't say such a thing so blatently- and I back it up my comments with technical descriptions that are easily supported since its my assumption that people are reading along. If I had to guess, by the way he responded, @Cableman got it
Ralph… In the interest of transparency and of course validate your comments could you please publish these intelligences for the benefit of interested members ? Myself among them .
Cableman was making fun of youThe irony is that part where while I said that in jest, I really can show measurements to back up my claim. So it looks like a flat statement, but if you've been watching how I post, I don't say such a thing so blatently- and I back it up my comments with technical descriptions that are easily supported since its my assumption that people are reading along. If I had to guess, by the way he responded, @Cableman got it
In part yes , however i took your comment “I really can show measurements to back up my claim” to mean that you also have empirical evidence to support your assertion that your topology measures in a superior way to SET topology and I seem to read that you are also inferring transformer output topology in entirety ?
Last edited:
OK. Which amp? In general the larger amps tend to have lower distortion. But they all have similar bandwidth; full power to 2Hz on the low end, past 100KHz on the top. Our early Mk1 MA-1 amp, which was one of our first, measured at 3% THD at full power (independently measured). That was about 1988 or so. We've done a lot to reduce distortion since then and they are typically about 0.5% to 1% depending on the actual tubes used. This is an amp employing no feedback. Distortion decreases fairly linearly as the output power is decreased to zero.
The noise floor is very usable with horns- I have horns in my own system at home. Gain varies with the model, the lowest power amps have less gain- about 23 dB and the bigger ones are more like 26dB.
Common Mode Rejection tends to be in the high 90s if the tubes are set up properly. We try to get that number up as high as possible since it reflects on distortion, bandwidth and noise.
The distortion signature is primarily the 3rd, often followed in amplitude by the 2nd unless the matching in the amp is very good. Both harmonics are present in enough quantity to mask higher orders. The harmonics fall off as order is increased a bit faster than seen in SETs owing to the simple fact that the amps are fully differential, so distortion cancellation occurs at every point throughout the circuit. So distortion is compounded less from stage to stage throughout the amp. Ideally distortion falls off along an exponential decay based on a cubic nonlinearity.
There is only one stage of gain in any of the amps. The bandwidth derives from this gain stage; the output section has bandwidth from DC (being direct-coupled) to well past 50MHz (we've not measured it past that frequency). Hence the output section rise time is somewhere in the neighborhood of 600V/msec.
The amplifier measurements must be conducted with instruments that are un-grounded. This has been a problem in three cases in the past, one resulting in the review (Glass Audio) being moved to a different reviewer. The other two were published, one in a French magazine (S-30) and the other being the MA-1 measured by Bascom King about 22 years ago. There is a second Glass Audio review of the M-60 as assembled from a kit which we used to offer. If either speaker terminal is grounded, the drive to the output section will be unbalanced, resulting in high distortion and low power prior to clipping. It can be quite tricky to un-ground the measurement instruments; an input transformer is one means but introduces distortion of its own which has to be nulled. Alternatively a differential probe can be used.
Regarding output transformers I see them as a bit of a mixed bag. The smaller you make an OTL the less practical it becomes- more and more power is dissipated in the output section as opposed the load and distortion goes higher. So at some point a transformer fixes that but introduces tradeoffs of its own. There is an outboard autoformer device called the Zero that loads the amplifier at a higher impedance and provides taps for 4, 3 and 2 Ohm loads. It works nicely with our stuff but since we don't use feedback, the amps tend to act as power sources, much as SETs or any tube amp without feedback. The Zero is unusual as it has bandwidth wider than our amps and any other, being good up to about 1 MHz if my information is correct. Usually the OPT is the defining factor in amplifier bandwidth; with the Zeros this isn't the case.
I've built smaller amps employing output transformers; I started with a pair of SET 45 based amps I got from a customer. They were very nice but limited to low volumes owing to about 0.75 of a watt output. But they sounded better than the 2A3 amp I had at the time. I became curious how they would sound if PP but using the same parts. The resulting type 45 amp made about 6 watts, had more bandwidth and was more transparent (it was easier to make out vocals for example). I didn't find any downside. It occurred to me about that point that usually when PP amps are compared to SETs its done almost entirely apples to oranges; the PP amp having vastly more power and using entirely different kinds of parts (including tubes).
So that first PP amp was an attempt to limited some of those variables. My second attempt was to design a small PP amp that only made 5 watts, using pentodes rather than DHTs and see how that did. It employed a UL connection in the output stage; if the taps are set right you should get linearity similar to that seen in a good power triode. I chose the EL95 for the job as its a small pentode and it was easy to find an OPT product that was already designed for that load with UL taps. Since that tube is meant to be easy to drive I designed a 12AT7 based differential amplifier to be the voltage amp and driver circuit. I used a 2 stage CCS in its cathode circuit. like the prior type 45 amps, it too was biased class A.
So this is an amp with only 2 stages of gain- a simpler signal path than most SETs.
It wound up having more gain than needed. Since I drove it single-ended by driving only one grid and grounding the other, the other grid was thus available for feedback, eliminating the possibility of intermodulations at the feedback node. The amp really didn't seem to need it so the feedback was only used to reduce gain (which IMO is the only way it should be applied unless you are really able to support a lot of feedback at all frequencies and such isn't possible with tubes owing to Gain Bandwidth Product issues). This little amp sounds smoother than the PP type 45 amps and has full power from 5Hz to 100KHz. The problem you usually run into with smaller tube amps is if its lower power, its likely SET. There really aren't any truely low power PP amps that are also built to a high quality build standard although there are plenty of SETs that are. Once you level that playing field then you can make more valid comparisons.
I think one issue that you run into over and over again is people make the mistake of combining PP and single-ended in PP amps. This results in a more prominent 5th harmonic and was discussed by Norman Crowhurst a good 60 years ago so should be no surprise. Why people keep doing it is beyond me; IME/IMO you either do single-ended from input to output or you do fully balanced (and probably differential) but you don't mix the two. You need that exponential decay of harmonics if the amp is going to sound musical.
Well said sir. I don’t for a moment believe that Amir’s measurements tell us all we need to know about whether an amp or a DAC will sound “good” to most people (a lot more research into this area is required and that includes investigating what “good“ means in this context) but anyone who thinks that measurements are irrelevant to hi fi clearly doesn’t understand the importance of science and the scientific method and as such their opinion is entirely without value. And that’s me being extremely polite . . .Measurements are used by people whos hearing is not good
Same as a carpenter who constructs a house without a measure , same as a mechanic who has to align a coupling without a dial indicator or laser.
All the people who use tools are basically incompetent , thats what you re saying .
Luckily in the high end problems can be solved with cables , lol
Just about the only industry in the world , where you can get away with that
Unfortunately in such technical questions we need to see the impedance and phase curves. Guesses are worthless - as Ralph explained it all depends on how the crossover is implemented.
Unfortunately when I listened to them I was not carrying my measuring system!
Hilarious. The Audiopax SETs matched beautifully with AVs. That’s why Jim imported both.
Oh and the Audiopax’s also work beautiful with NON HORNS which makes GTs comments about speakers borderline nonsense IMHO
Last edited:
IME Trios needed more juice than the 15-20 watt SETs available to me then. Never talked to Jim about the combination did you?
david
Similar threads
