Inverted OTL with 6SN7/6AS7

What people are working on at the moment
Wolfgang
Old Hand
Posts: 229
Joined: Mon May 07, 2018 3:08 pm

#31 Re: Inverted OTL with 6SN7/6AS7

Post by Wolfgang »

My decision to use the 6AS7 was based on several things. First I wanted a cheap and easily available output tube.

Secondly, I was a little irritated why S.B. would use a tube with MU=15 as he comments at p.2 that “An upper practical limit is achieved with devices with a normal MU of about 15.” The 5687 has MU of 15. So why using a tube that would already operate at the upper limits of what’s possible? And is the “fringing” just something typical for this particular tube or a general effect that other high MU tubes would also show? We could only find out by drawing plate characteristics for each tube.
The “more even distortions” on the other hand which can be predicted by looking at the plate characteristics of the 6AS7 was something that I actually wanted to have (if it wouldn't get too much) because this is the stuff that is mainly responsible for the beautifully “blooming” sound colors of SE amps. I miss that a little with the CCS SE OTLs which are operated in the more linear part of the tube. I think that CCS is better for “reacting” to the changing speaker impedance at different frequencies but it lacks a bit the “warmth” for my feeling.

The only reasons I could find why S.B. would have chosen the 5687 for this project ( as S.B. obviously didn’t have any comparison with different tubes regarding sound quality at this point) was that with smaller sized tubes (20!) the amp housing would be considerably smaller too and he would need less power for the heaters. The biggest problem with the 6AS7 is indeed the demand for the heaters and space. For driving an 8ohm speaker this would be already quite something. But I think my solution works well (3 in series and 2x3 in parallel) and gives exactly 6,303V at 120V AC with the 36V CT /2x18V toroid. 250VA should be enough to handle 12tubes without getting too warm (so far I have only tested one section with 6 tubes which is no problem).
And last, I wanted a higher sensitivity than 1,9Vp-p. For use with a passive pre/volume control one could get this version down to 0,6Vp-p by adding a bypass cap to the first cathode resistor of the 6SN7. Maybe that’s already too much… but just saying.

How it all turned out soundwise with the 6SN7/6AS7 was maybe a lucky coincidence and the final word about sound quality etc is only possible with two channels and maybe after 1-2 weeks of listening. But what I have heard so far with one channel and different music styles (Classical music, Jazz, Electronic, solo instruments, Smooth Jazz) is very promising and solves lots of minor problems and points of criticism which I had regarding my existing system when it comes to correct reproduction of trumpets, violins, dynamic jumps, lower frequency band, separation of instruments etc.

Lowther speakers are often criticised for "shouting" or being too bright around 1-3kHz. While I partly agree regarding the PM6C and PM6A I don't agree when it comes to the DX4. But it can sound a bit too bright with certain instruments (loud violins and trumpets and soprano sax)) and it shows this tendency with all OTLs I have.For some time I tried using a dynamic EQ which helped to solve this problem but also with some disadvantages. The Inv OTL is the first amp that wouldn't need any help in this respect and sounds absolutely neutral and realistic with the Lowthers.
User avatar
Paul Barker
Social Sevices have been notified
Posts: 8879
Joined: Mon May 21, 2007 9:42 pm

#32 Re: Inverted OTL with 6SN7/6AS7

Post by Paul Barker »

Sounds like a nice recipe
"Two things are infinite, the universe and human stupidity, and I am not yet completely sure about the universe." – Albert Einstein
Wolfgang
Old Hand
Posts: 229
Joined: Mon May 07, 2018 3:08 pm

#33 Re: Inverted OTL with 6SN7/6AS7

Post by Wolfgang »

The noise would be better with a log frequency scale and some reference signal. You need a siignal to measure the noise against.
Nick,
I tried to figure out if we mean the same with “noise floor” because for me it simply means sum of all noise without any signal. Noise compared to a signal would mean in my understanding signal to noise ratio (SNR). Did you mean SNR?

I was looking for more clarification online but I could only find comments like:

“Signal to Noise Ratio (SNR) is the relative strength of the signal to the noise. It's a comparison. Noise Floor is the absolute measurement of noise.”

“The noise floor is just a measure of the noise itself; that is, you can think of it as P(fn); how much noise do you have, regardless of whatever signal there is. This is like asking the question "how loud is the crowd I'm in?", and doesn't depend on any signal at all.”

“The noise floor is the sum of signals that is always there without the desired signal. The signal to noise ratio is the ratio between the desired signal and the noise floor. Think of an analogue AM radio where you did not tune into a station: the rain-like noise you hear is the noise floor.”

So if I measure with a signal as you have recommended wouldn’t I get the SNR and not the noise floor I was interested in ?
User avatar
Ray P
No idea why I do this anymore
Posts: 6294
Joined: Thu Nov 22, 2007 5:18 pm
Location: Somerset

#34 Re: Inverted OTL with 6SN7/6AS7

Post by Ray P »

I've reread the Steve Bench article several times and I think I have a basic grasp of the concept if not all the technical details, I wouldn't have a clue how to measure an plot valve characteristics for altenatives for example so will need to go with the 6AS7 flow I think.

I've only just registered the almost throw away last senetence about EL509s - did you contemplate that route Wolfgang?
Sorry, I couldn't resist!
User avatar
Paul Barker
Social Sevices have been notified
Posts: 8879
Joined: Mon May 21, 2007 9:42 pm

#35 Re: Inverted OTL with 6SN7/6AS7

Post by Paul Barker »

Plotting characteristics is easy, measuring is hard because you have to give a short burst of power (i.e the specific volts/current at the specific grid voltage) so as not to damage the weaker component parts such as grid. Best get or borrow a curve tracer. Mine broke and I’m so far from electronics in my head, I would send it to the Supplier for repair if I could give a fig, which atm I can't. Bigger fish to fry.

But for 6as7 Stephie has given us the characteristic curves so we don’t have to trace them.
"Two things are infinite, the universe and human stupidity, and I am not yet completely sure about the universe." – Albert Einstein
User avatar
Paul Barker
Social Sevices have been notified
Posts: 8879
Joined: Mon May 21, 2007 9:42 pm

#36 Re: Inverted OTL with 6SN7/6AS7

Post by Paul Barker »

Addendum I see the up dated 400 max bolts utracer is only £283 built by Ronald. Worth every penny, but he’ll also send the upgrade to 400v for 15 💶 (they used to be 300v as is my one.)
"Two things are infinite, the universe and human stupidity, and I am not yet completely sure about the universe." – Albert Einstein
User avatar
IslandPink
Amstrad Tower of Power
Posts: 10041
Joined: Tue May 29, 2007 7:01 pm
Location: Denbigh, N.Wales

#37 Re: Inverted OTL with 6SN7/6AS7

Post by IslandPink »

Paul Barker wrote: Sun Jun 30, 2019 5:46 pm But for 6as7 Stephie has given us the characteristic curves so we don’t have to trace them.
and for 5687 -
https://jacmusic.com/techcorner/SBENCH- ... nv5687.gif
"Once you find out ... the Circumstances ; then you can go out"
User avatar
Ray P
No idea why I do this anymore
Posts: 6294
Joined: Thu Nov 22, 2007 5:18 pm
Location: Somerset

#38 Re: Inverted OTL with 6SN7/6AS7

Post by Ray P »

Yes, for the 6AS7 and 5687s we have Steve Bench's curves; I was thinking out load about measuring and plotting inverted curves for others, such as the 6N6P or EL509.
Sorry, I couldn't resist!
User avatar
Nick
Site Admin
Posts: 15711
Joined: Sun May 06, 2007 10:20 am
Location: West Yorkshire

#39 Re: Inverted OTL with 6SN7/6AS7

Post by Nick »

Nick,
I tried to figure out if we mean the same with “noise floor” because for me it simply means sum of all noise without any signal. Noise compared to a signal would mean in my understanding signal to noise ratio (SNR). Did you mean SNR?
You are measuring in dB, that is a relative measurement. Without having an absolute we have no idea what that graph means. You can have the absolute by using a display in a calibrated dBv, or you can compare to a known output voltage.

Also, the interesting stuff will happen I think between DC and 1kHz
Whenever an honest man discovers that he's mistaken, he will either cease to be mistaken or he will cease to be honest.
Andrew
Eternally single
Posts: 4206
Joined: Thu May 24, 2007 2:18 pm

#40 Re: Inverted OTL with 6SN7/6AS7

Post by Andrew »

Wolfgang, I agree with Nick, in fact we had a chat about this over the weekend at a quiet point in a show we were doing...

1) To see the interesting stuff you need a log scale on the x-axis or, if you can't do that, zoom in more or 20Hz-1kHz

2) We need a baseline to compare your relative measurements against, there's lots of ways of doing this, Nick has suggested as few, another one simple way is to present a before and after picture. What does you noise floor look like with just the measuring system (loopback more) and then with the amp in the loop? Then you/we can compare.

Great work, by the way :)
Analogue, the lost world that lies between 0 and 1.
Wolfgang
Old Hand
Posts: 229
Joined: Mon May 07, 2018 3:08 pm

#41 Re: Inverted OTL with 6SN7/6AS7

Post by Wolfgang »

I was considering a uTracer which Thomas had mentioned. But after reading more about the details and how it works I got the impression that for tracing inv. tube operation one would not only need to rewire a lot and use an additional low voltage psu but it would be impossible (for me at least) to predict the ramifications for the PC program/interface.
Plotting characteristics is easy, measuring is hard because you have to give a short burst of power (i.e the specific volts/current at the specific grid voltage) so as not to damage the weaker component parts such as grid
The power bursts at much higher voltage than for normal tube tracing operation could become a problem.
I was thinking out load about measuring and plotting inverted curves for others, such as the 6N6P or EL509.
The EL509 was next to the 6AS7 the other tube I would have been interested in. But I wouldn't know how to approach this without having plate characteristics.
2) We need a baseline to compare your relative measurements against, there's lots of ways of doing this, Nick has suggested as few, another one simple way is to present a before and after picture. What does you noise floor look like with just the measuring system (loopback more) and then with the amp in the loop? Then you/we can compare.
Ok, now I understand better what's the problem. Thanks!
User avatar
Nick
Site Admin
Posts: 15711
Joined: Sun May 06, 2007 10:20 am
Location: West Yorkshire

#42 Re: Inverted OTL with 6SN7/6AS7

Post by Nick »

What does you noise floor look like with just the measuring system (loopback more) and then with the amp in the loop? Then you/we can compare.
Well, sort of no. That will show you how much noise in dB that the amp is producing, but without knowing what the noise floor of the measuring kit is all you can say is the amp adds a extra 15dB of noise (for example). But you don't know if the measurement kit has a noise floor of (say) -80dB where 0dB is 1v RMS, or -150dB against 1v RMS. You can know that by using a 1kHz signal, and adjusting 0dB so the 1kHz is your 0dB. Once you have done that, you know (for example) that -60dB is 1mv, -80db is 100uv and so on.
Whenever an honest man discovers that he's mistaken, he will either cease to be mistaken or he will cease to be honest.
Wolfgang
Old Hand
Posts: 229
Joined: Mon May 07, 2018 3:08 pm

#43 Re: Inverted OTL with 6SN7/6AS7

Post by Wolfgang »

This whole topic seems to be a problem for many picoscope users (and in general). I could find in many different blogs questions referring to the measured noise and the absolute noise (subtracting the noise of the measuring device) over and over and no real answers.

Most important for me at this early point of testing(having just finished one channel) is a more general overview about any noise that would stand out from the rest. And, yes, it should be done with log scale.

Measuring with SNR mode with the picoscope would also show the ratio between signal and noise but the noise of the picosscope itself would still be included in this test. What Nick has proposed makes most sense under these circumstances.
User avatar
Nick
Site Admin
Posts: 15711
Joined: Sun May 06, 2007 10:20 am
Location: West Yorkshire

#44 Re: Inverted OTL with 6SN7/6AS7

Post by Nick »

Measuring with SNR mode with the picoscope would also show the ratio between signal and noise but the noise of the picosscope itself would still be included in this test. What Nick has proposed makes most sense under these circumstances.
Yes, but its not as bad as it sounds. If you measure the scope itself with a 1kHz 1v RMS signal and make that 0dB, the the noise floor itself will be seen relative to that reference. Then the amp can be tested with a matching 1kHz tone, and the difference in the noise floor will either be nothing, in which case the amp is quieter than the scope, or more, in which case you are measuring the noise of the amp.
Whenever an honest man discovers that he's mistaken, he will either cease to be mistaken or he will cease to be honest.
Andrew
Eternally single
Posts: 4206
Joined: Thu May 24, 2007 2:18 pm

#45 Re: Inverted OTL with 6SN7/6AS7

Post by Andrew »

Nick wrote: Mon Jul 01, 2019 7:50 pm
What does you noise floor look like with just the measuring system (loopback more) and then with the amp in the loop? Then you/we can compare.
Well, sort of no. That will show you how much noise in dB that the amp is producing, but without knowing what the noise floor of the measuring kit is all you can say is the amp adds a extra 15dB of noise (for example). But you don't know if the measurement kit has a noise floor of (say) -80dB where 0dB is 1v RMS, or -150dB against 1v RMS. You can know that by using a 1kHz signal, and adjusting 0dB so the 1kHz is your 0dB. Once you have done that, you know (for example) that -60dB is 1mv, -80db is 100uv and so on.
OK, so what you're suggesting is that you inject 1Khz into the DUT at x volts and adjust until 'x' until the screen says 0db at the 1KhZ spike?
Analogue, the lost world that lies between 0 and 1.
Post Reply