A true comparative distortion test would need to include thermal distortion, different types of harmonic distortion, TIM, IMD, crossover, and many more. It would require a well thought out and thorough effort.
-125 db s/n is about as low as possible for a power amp. I got my amp to -108db at 1w by pre-driving the volume pot to 15v p-p. 2-5 PPM distortion also appears to be a limit.
I don't use sims as this is what can happen, build it and test it.
I don't use sims as this is what can happen, build it and test it.
Wouldn't this kind of accuracy would be much more valuable in industries other than audio?
I made a simple test amp about 20 years ago. It was a somewhat typical KT88 push pull tube amp, wired in triode mode for about 30 WPC. I had built several of these using mostly surplus parts when I bought budget 200 OPT's rather cheap (still have a few).
I put a simple knob on the front in addition to the usual volume knob. It was a 4 gang pot that increased the GNFB from the OPT secondary as it was turned up, and also decreased the input attenuation to make up for the gain reduction as the GNFB was increased. The idea was to attempt to make the volume constant as the knob was turned from 0 to 10, but it wasn't perfect.
I loaned this amp out, in fact it was passed around for nearly a year before someone liked it enough to buy it from me. My only request during that time was to write down a brief description of how it was used, what kind of music, how loud, and what number you liked on the unmarked knob. In most cases I knew the people who did the testing.
The results were a typical bimodal distribution. Most people liked the amp set on 10 (maximum GNFB), they were typically younger than 30, liked loud dance or rap music, and their typical listening environment involved white wires coming from their ears (iPOD). Some of this testing involved multiple listeners at the same party, and the knob setting was by group consensus.
There was a few responders who liked a setting between the extremes, but nearly none from 4 to 8. There were about half as many responders who liked low settings than "10's", with 1 or 2 on the knob, being the most popular setting in that group. These people tended to be older than 30, but there were quite a few younger people. Many tended to already have some component audio equipment, but none owned a tube amp. Several also had experience playing a musical instrument.
Note that this testing was done among people at work in a cell phone engineering design company, or among their friends in the late 1990's. It was NOT a normal cross section of society, as I often had to point out during phone design sessions. Most of the people had a college education, usually in a technical field, mainly engineering or software development. The average age of our facility was 27 and almost half of the people were Hispanic.
Updated the first post with some more complete data after rummaging through some old notes in old PMs.
I know people aren't interested in it but I made an early partially completed PCB draft for the super triode I posted earlier for my future self if nothing else since it's the only tube thing I want to build that I haven't tried.
Before I finish it I want to know peoples thoughts on what tube I should use that has potential for the tube "Sound".
I'm planning on using the 6sn7 because it will quickly and easily begin reacting to the load and quickly allow a significant output impedance shift during current sharing adjustment since I'm using its plate as the output and it's a low current high RP tube.
Also because it's linear and shares a socket with 6as7/6080.
Any thoughts on alternatives?
Before I finish it I want to know peoples thoughts on what tube I should use that has potential for the tube "Sound".
I'm planning on using the 6sn7 because it will quickly and easily begin reacting to the load and quickly allow a significant output impedance shift during current sharing adjustment since I'm using its plate as the output and it's a low current high RP tube.
Also because it's linear and shares a socket with 6as7/6080.
Any thoughts on alternatives?
Last edited:
6CG7 / 6FQ7 because it's a complete waste of time using Octals, the early 6SN7 was upgraded during its lifetime, and it's far more difficult to get good samples at sensible prices.
12FQ7/12CG7/8FQ7
6CG7 / 6FQ7was widely used in TVs and radios and is an exact equivalent of the last 6SN7GTB but likely better quality control.
Once you use modern Novals it gives you far more options to be able to swop in less linear / different dual triodes into a test circuit.
do you have any pdf files of the schemes you posted in post #1? these tired eyes have difficulty looking....thanks
Hmmmm. Can't say I've used any Novals before although I think I have some.
Do Novals really offer the most options?
I don't but it should maximize if you click on it or right click and open image in new tab.
6CG7 / 6FQ7 because it's a complete waste of time using Octals, the early 6SN7 was upgraded during its lifetime, and it's far more difficult to get good samples at sensible prices.
12FQ7/12CG7/8FQ7
6CG7 / 6FQ7was widely used in TVs and radios and is an exact equivalent of the last 6SN7GTB but likely better quality control.
Once you use modern Novals it gives you far more options to be able to swop in less linear / different dual triodes into a test circuit.
I agree, the 6SN7 started like inside a television set as vertical osc. output tube, the 6CG7/6FQ7 is a small brother, brought about by attempts at miniturizations, and the 12au7 is still its smaller brother, so here we have big medium and small...
the 6GU7 is almost the same but has a tad lower mu..
I know people aren't interested in it but I made an early partially completed PCB draft for the super triode I posted earlier for my future self if nothing else since it's the only tube thing I want to build that I haven't tried.
Before I finish it I want to know peoples thoughts on what tube I should use that has potential for the tube "Sound".
I'm planning on using the 6sn7 because it will quickly and easily begin reacting to the load and quickly allow a significant output impedance shift during current sharing adjustment since I'm using its plate as the output and it's a low current high RP tube.
Also because it's linear and shares a socket with 6as7/6080.
Any thoughts on alternatives?
you tell us your choice and then explain why you choose a tube, and then we can offer advice...unless we understand what you are up to, we can only opine...guess, i guess....
what you think in your mind, may not always be what we take it from your written words, english is not that hard but.....
It was no "attempt" it worked!
(Even the 6V6 was turned into the 6BW6).
Another pain in the A with octals...you break off the spigot- very common and it's a real mess.
Of course if you want to do it properly....you can always go to 12pin compactron sockets which are easily available for PCB off the shelf....but there again who cares!
If I were you I would go for wire ended - forget sockets altogether - dirt cheap non microphonic USA 6111 or russian 6N16B-V being a tiny fraction of the size and just as good.
The whole family
Don't look at me, I have no clue. I'm making assumptions on what will sound better. I was hoping others with more knowledge and experience could do a deeper analysis.
At that point I might as well use a nutube or a transistor strapped into triode mode. It's defeating the purpose of bothering to do the super triode. If nothing else was a factor I'd just use the coolest looking tube.
I have no clue what the effects are of driving the load from the plate vs the cathode using X or Y tube under A or B bias.
This is something I wanted to discuss in the "do tubes actually sound like anything" thread before it got hijacked and derailed.
Surely if anyone claims that tubes have a sound there must be a way to induce it with this circuit. It is effectively a SET amp that is directly driving a speaker from the plate.
Last edited:
Everybody likes different tubes. My preference would be a graphite plate 801a at 15ma 350V. Since you have no clue, it is better to prototype with what you have already, a 6SN7, 4P1L or 6V6 is just fine. Make sure everything works as designed and try several bias points, measure and listen to get the idea. Better can be found later as an option or upgrade.
I guess I'll just use my 6sn7 then. Not that I'll be able to build it in the foreseeable future but it's good to plan ahead.
I have to admit the lack of interest, technical discussion, criticism, or even mention of the circuits I posted that this thread was supposed to be about is disappointing. I at least expected some more interest in discussing the super triode considering how much effort people go through to use tubes in their system.
How boring. I'll take my leave I think. In 4 days I'll be forced to abandon this crap for a few years anyway. If anyone with the means to measure that super circuit I discussed earlier wants to hit me up send me a PM. If not
I have to admit the lack of interest, technical discussion, criticism, or even mention of the circuits I posted that this thread was supposed to be about is disappointing. I at least expected some more interest in discussing the super triode considering how much effort people go through to use tubes in their system.
How boring. I'll take my leave I think. In 4 days I'll be forced to abandon this crap for a few years anyway. If anyone with the means to measure that super circuit I discussed earlier wants to hit me up send me a PM. If not
Last edited:
There have been a few attempts to demonstrate the supertriode topology (including from me) but there seems to be a distinct lack of interest. I have built a 50W push-pull version. I think it sounds great but I built it for my own interest and enjoyment and don't really care what others think.I thought 0.25% distortion (mostly second and third) at 36 watts with no global negative feedback was pretty good, output impedance 3 ohms.
I found it difficult to follow the circuit you posted, it was laid out in an unusual way. Can you please explain how each part works? I've posted it again here for reference.
I found it difficult to follow the circuit you posted, it was laid out in an unusual way. Can you please explain how each part works? I've posted it again here for reference.
Attachments
As explained,-
I suggest speaking to "wavebourn" on here who has frequently recommended this solution.
This gold grid device is in the 1-3USD bracket.
They are also some of the last valves ever made in high quality production 12/84-86
:-
(Typically Moscow region MELZ plant - now still alive as a factory making ultra high quality precision military optics today)
some links
1/
2/
3/
The whole family 1/ mu 20-30 2/ mu 60-90 3/ mu 20-30
This one from 2018:-
Think it sounds right up your street!
No kidding.
Working simulation? Really? More like a back of the napkin scribble only using a PC.
Similarly impressed with the pcb layout.
But it's the thought that counts 😀
There's a rumour that some of the best race car people like Chapman (Lotus) used to scribble most of their good ideas on pub beer mats.
Perhaps it's how they killed Jim Clark? 😎
However as you rightly say, actually making a prototype that doesn't crash or burn is one thing, going from prototype to production is the REALLY REALLY hard bit.
Been there, done that more times than I can count. 🙄
Perhaps it's how they killed Jim Clark? 😎
However as you rightly say, actually making a prototype that doesn't crash or burn is one thing, going from prototype to production is the REALLY REALLY hard bit.
Been there, done that more times than I can count. 🙄
The 6SN7 has a well deserved reputation for being quite linear for most of it's useful operating range. The 6080 / 6AS7 does not since it was designed as a power supply pass tube with very low gain. The 6SL7 has more gain than the 6SN7, but gives up some linearity. The 6BL7 and 6BX7 are designed for TV vertical sweep, and are essentially 6SN7's made BIGGER. The gain is about half of a 6SN7 and the linearity is a bit lower. All share the same pinout. The 6SN7 or 6FQ7 / 6CG7 is a good place to start for something like this.
The definition of REALLY HARD scales with the run rate. One stupid mistake on an automated line cranking out 10,000 cell phones per shift will make for a lot of scrap.....Make sure all of your part numbers are correct, and the parts stock is actually the right part before starting up that SMD gang placement machine!