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

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

Post by Wolfgang »

Last night I connected the Inv OTL to my main system. Left channel the 6C33 SE OTL, right channel the Inv OTl. Their input sensitivity is almost the same so there was no real need for adjustments. The 6C33 sounds excellent and I thought this would be a fair comparison.

Normally when I do these tests I have to listen really hard in order to hear some rather subtle differences between amps like for example with the 300B SE and the 6C33 SE OTL. Although they sound in the end very distinctively different it doesn’t stand out so obviously at first. In this case with the Inv OTL and the 6C33 SE OTL it was so obvious that it was almost annoying to listen to both amps hooked up to one system.

Everything was very different starting with the bass of the Inv OTL, which was so much better defined, more subtle, giving instruments their individual space rather than mixing them together. However, the Inv OTL didn't have the same powerful punch around 100Hz that the 6C33 OTL can deliver.

The clarity and refinement starting from the upper mids upwards was outstanding with the InvOTL. My limited listening impression from my shop with the PM6A was correct. It just got a lot better in my main system.

Dimension and shape of the virtual sound stage changed also dramatically on the right channel. The nice well defined half circle sound stage which the 6C33 OTL creates with each recording was bulging out with the Inv OTL to more complex and irregular shapes depending on the recording, changing from recording to recording, which is for me a sign for accuracy of an amp. It was kind of irritating for the brain to assemble these different ways of reproducing the same recording on the left and right channel into one coherent picture.

The dynamics, the capability to reproduce true level jumps, was at least as good as with the Pinnacles. But in comparison the Pinnacles seem to “pump”a little between very dynamic peaks followed by silent passages like a badly adjusted compressor for audio signals . That is probably one of the most impressive things the Inv OTLs can do: to deliver real dynamic jumps without any compressing or holding back. Must be the missing NFB.

Overall I got the impression as if the right channel with the Inv OTL was playing always a fraction of a second ahead of the left channel and the music was played at a little higher tempo without of course changing the pitch. It reminded me a bit of out of phase issues when one adjusts a sub to the rest of the system. Could this be the influence of the output cap vs. no cap?

Finally I wanted to find out how these listening impressions would translate into measurements and was a bit surprised at what I found.
I would have expected rather a rising level in the frequency response than a slightly falling one (1,6dB 10Hz-20kHz):
InvOTL frequ resp 10Hz_20kHz.jpg
I also would have expected a more perfect square wave shape from how it sounds:
InvOTL 1kHz square.jpg
And I somehow must have mixed up something when I wrote that the noise floor was at -85dB. It’s around -70dB but still dead silent.
InvOTL noise floor.jpg
User avatar
IslandPink
Amstrad Tower of Power
Posts: 10041
Joined: Tue May 29, 2007 7:01 pm
Location: Denbigh, N.Wales

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

Post by IslandPink »

Very interesting indeed. I can well believe the soundstage expands - have heard the reverse whenever global feedback is applied. Also the 'higher tempo' effect seems to come when bass time constants are reduced or eliminated. On the FR declining slowly - seems odd, only something like a miller capacitance effect would be present here - is there enough to account for this amount ? The output stage capacitance is very low of course due to the very low mu.
And is the rounding of the 2kHz square wave enough to be explained by the FR ?
I would love to build one of these amps ( eventually ) or at least hear one.
There is someone reading this thread who has a stash of the higher filament voltage 5687's saved just for this circuit !
"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

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

Post by Ray P »

Interesting developments Wolfgang, thanks for the progress update.
Sorry, I couldn't resist!
Wolfgang
Old Hand
Posts: 229
Joined: Mon May 07, 2018 3:08 pm

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

Post by Wolfgang »

Also the 'higher tempo' effect seems to come when bass time constants are reduced or eliminated.
I don't fully understand what you mean with "reduced bass time constants" in this context. Maybe the following makes it clearer what I mean. The bass seems first "less" powerful compared to the (very powerful) 6C33 but after some minutes it becomes obvious that it is not "less" but simply more refined and even deeper, much deeper, when the music signal goes there. It's on one side like the bass has lost some extra weight but on the other side it - consequently - becomes more defined, refined and quite a bit faster. I very much prefer this. One of the biggest problems with bass reproduction (lower mids- upper bass)is that everything which happens in the same frequency band (piano, bass, drums, tenor sax) appears as "clumped together" if the amp/speakers and room are not in very good shape acoustically speaking. As soon as the bass notes become audible as single notes and can be clearly assigned to the instruments where they are coming from it is a good system IMO. This exactly happens with the Inv OTL. Not a tiny bit, but in a big way.
On the FR declining slowly - seems odd, only something like a miller capacitance effect would be present here - is there enough to account for this amount ?
Yes, Miller effect. And with a typically connected tube I would know what to do or at least where to look. I didn't think that this could even become a problem as S.Bench writes in his Inv OTL article: "There is also essentially no Miller effect capacitance to worry about, as the voltage gain is much lower than unity." So what happens here is clearly beyond my scope.

I definitely would like to understand better what's going on but I know already for sure : This is the amp I was waiting for. 1,6dB are not audible especially when it starts mostly above 10kHz . Below 10kHz it's marginal.
Wolfgang
Old Hand
Posts: 229
Joined: Mon May 07, 2018 3:08 pm

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

Post by Wolfgang »

As I try to avoid grid stoppers - and I don't think the 2k which S.Bench uses in his original circuit would change anything because of the low value - would an anode bypass cap parallel to the 60k anode resistor of the second 6SN7 triode help to shape the high frequency response a little?
User avatar
Nick
Site Admin
Posts: 15709
Joined: Sun May 06, 2007 10:20 am
Location: West Yorkshire

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

Post by Nick »

I would check and measure what you are using to measure first. The square wave looks fine. Nice and equal +ve and -ve slew, thats always a good thing. The noise would be better with a log frequency scale and some reference signal. You need a siignal to measure the noise against. Not sure what the frequency response is showing, whats all the noise? Again a log x would help.
Whenever an honest man discovers that he's mistaken, he will either cease to be mistaken or he will cease to be honest.
User avatar
Ray P
No idea why I do this anymore
Posts: 6294
Joined: Thu Nov 22, 2007 5:18 pm
Location: Somerset

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

Post by Ray P »

I might have to have a go at one of these inverted OTLs. 5687 version looks interesting but they're getting rarer and pricier. What would the panel say about subbing Russian 6N6Ps - claimed to be equivalent based on a quick bit of research (apart from the different pin layout) and much easier to find at a good price.
Sorry, I couldn't resist!
Cressy Snr
Amstrad Tower of Power
Posts: 10552
Joined: Wed May 30, 2007 12:25 am
Location: South Yorks.

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

Post by Cressy Snr »

Ray P wrote: Fri Jun 28, 2019 12:46 pm I might have to have a go at one of these inverted OTLs. 5687 version looks interesting but they're getting rarer and pricier. What would the panel say about subbing Russian 6N6Ps - claimed to be equivalent based on a quick bit of research (apart from the different pin layout) and much easier to find at a good price.
Can’t see any problem using 6N6P. Cheaper than NOS 6AS7/6080 and far less power needed for the heaters that’s for sure. JJ ECC99 might be good too.
Sgt. Baker started talkin’ with a Bullhorn in his hand.
Wolfgang
Old Hand
Posts: 229
Joined: Mon May 07, 2018 3:08 pm

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

Post by Wolfgang »

Not sure what the frequency response is showing, whats all the noise
I used the spectrum analyzer with a sweep signal 10Hz-50kHz/10ms/10Hz intervals and flat top/peak hold.What you see on the bottom of the curve are the measurements/probes in very short time intervals and without "flat top" so it looks like noise. Every peak is a measurement at a different frequency of the sweep signal.Another way of measuring would be using the scope and checking the amplitude but that doesn't show a very defined curve even if I use the storage function. I guess you use more expensive stand alone equipment (not PC based) and this will give you of course a nice better looking curve.
Will try the log setting.
Wolfgang
Old Hand
Posts: 229
Joined: Mon May 07, 2018 3:08 pm

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

Post by Wolfgang »

What would the panel say about subbing Russian 6N6Ps - claimed to be equivalent based on a quick bit of research (apart from the different pin layout) and much easier to find at a good price.
In this case it's a little more complicated, Ray. "Equivalent" refers to regular use. You need to find tubes which will give you in the inverted connection low enough impedance, can handle high enough current through their grids. I at least would feel that it is important to plot a power dissipation line for the grids in connection to the load line if I would use different tubes. S.Bench goes into details about this topic. Also don't forget that certain tubes can "fringe", as he calls it, at both ends of the bias. But you won't find out/know unless you plot the curves as he did. Considering all this the 6AS7 seemed to me a very good starting point. And they don't show the "fringe" tendency like the 5687.
User avatar
Ray P
No idea why I do this anymore
Posts: 6294
Joined: Thu Nov 22, 2007 5:18 pm
Location: Somerset

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

Post by Ray P »

Wolfgang wrote: Fri Jun 28, 2019 3:30 pm In this case it's a little more complicated, Ray.
Thanks Wolfgang but now I'm confused, or maybe I'm just exposing my technical ignorance, but if the specifications are essentially equivalent why would they behave differently?

Also, what are fringe affects?
Sorry, I couldn't resist!
User avatar
thomas
Old Hand
Posts: 910
Joined: Sat Sep 26, 2009 1:52 am
Location: N.W.Kent

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

Post by thomas »

Ray P wrote: Fri Jun 28, 2019 12:46 pm I might have to have a go at one of these inverted OTLs. 5687 version looks interesting but they're getting rarer and pricier. What would the panel say about subbing Russian 6N6Ps - claimed to be equivalent based on a quick bit of research (apart from the different pin layout) and much easier to find at a good price.
I switched from 5687s to 6n6ps for my tower of power, as they are cheap and can take a bit more current. They were not a straight swap (aside from heaters) as they biased up slightly differently. Also tolerances varied somewhat between sections when measured on my utracer.... but they sounded very sweet.....
Wolfgang
Old Hand
Posts: 229
Joined: Mon May 07, 2018 3:08 pm

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

Post by Wolfgang »

Ray, this is what S.Bench calls "fringing". At both ends the tube looses its linear characteristics .
fringing.jpg
This shows the "plate characteristics" of the 6AS7.
no fringing.jpg
S. Bench explains it in his article. If we start using grids as anodes they have to be of a certain strength/sturdiness. Some tubes are better suited than others. If a tube is used in its normal way this is not an issue. But there are no specific parameters in the tube data that would reflect the "grid as anode conditions". That's why he plotted these curves. He calls them consequently "plate characteristics".

Look in comparison at the original characteristics (grid):
6AS7  characteristics.jpg
Everything looks quite different. Shape of the curves, voltages, current.
User avatar
Ray P
No idea why I do this anymore
Posts: 6294
Joined: Thu Nov 22, 2007 5:18 pm
Location: Somerset

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

Post by Ray P »

Thanks Wolfgang. That's helped and I can understand your steer towards the 6AS7. As, like you, I use efficient 15ohm speaker units it maybe the safer bet.

I need to have a clearout to gather some funds!
Sorry, I couldn't resist!
User avatar
Paul Barker
Social Sevices have been notified
Posts: 8874
Joined: Mon May 21, 2007 9:42 pm

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

Post by Paul Barker »

But the 6as7 inverted curves look high in 2nd harmonic yet you say how great the amp sounds. Same story for usual orientation 6as7 se. Sounds real good with all it’s 2nd!

The 5687 inverted lines look so linear you would think you accidentally mixed them up with 211.

I’d prefer to build 5687 if only I’d kept them. But I have enough nos 6as7 for mono. 6as7 looks to need higher load impedance. autoformer might improve things or bank of drivers in series.
"Two things are infinite, the universe and human stupidity, and I am not yet completely sure about the universe." – Albert Einstein
Post Reply