Inverted OTL with 6SN7/6AS7

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

#1 Inverted OTL with 6SN7/6AS7

Post by Wolfgang »

When I saw the schematics of the Inv OTL my first thought was that building this amp could be a way out of something which had become more and more part of my daily listening experience. This feeling of having reached a dead end is no doubt very subtle because I am quite satisfied with the sound of my system but it’s gnawing at me.

Over the years I have built all kind of OTLs (SEPP, PP, SE) and they all sound very good. My last project ,the 6C33 OTL, actually sounds extremely good but still everything sounds pretty similar in the end.

The Pinnacle (improved version) for example is the OTL with the lowest distortion, incredibly clear, great dynamics, with huge sound stage but still I start missing something after a while of listening with it. They all stop at some point crossing over to the reality– some sooner, some a little later – they cannot deliver this typical studio recording experience, the listening experience sitting right in front of the musicians. It’s probably even more the live atmosphere which I am missing than something that could be expressed in terms of SQ.

I also liked the idea that I could try to use a very good low voltage regulated psu for the output tubes in the final version of the Inv OTL . I remembered in this context how much the Borbely psu improved the SQ of my R-2R DAC. Most likely a discrete low voltage psu woudn’t have the same positive effect controlling only the idle current but I wanted to find out.

Many weeks later and far from being done figuring out all the details and necessary additions for feeling really comfortable with this OTL I am sitting in front of a hardly finished amp and listen in complete disbelief of what I am hearing. It’s only one channel which drives two speakers but the realism is absolutely captivating.

I had to move everything into two old amp housings because my simple breadboard construction with which I had started appeared too shaky and dangerous for 700V rail to rail.
amp one channel.jpg
Inv OTL psu.jpg
This amp burns some serious power for the output tube heaters and the psu for the output tubes and needs some substantial transformers. Right now I am looking at hefty 240W per channel.

Compared to building a “normal” OTL the calculation process was somehow harder and the construction with two voltage rails for the huge voltage swing of 260Vp-p together with the “upside down” output stage and a low voltage psu forced me to reconsider my usual ways of approaching an amp project. I definitely found what I was looking for: many challenges and new ways of seeing things and a whole new listening experience.

And here is my version of the amp with 6SN7/6AS7 tubes:
schematics p.pdf
(158.92 KiB) Downloaded 381 times
Cressy Snr
Amstrad Tower of Power
Posts: 10552
Joined: Wed May 30, 2007 12:25 am
Location: South Yorks.

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

Post by Cressy Snr »

Looks like a great project there Wolfgang.
I can’t pretend to really understand how it works, but it’s nice to see the OTL boundaries being pushed out there.
Sgt. Baker started talkin’ with a Bullhorn in his hand.
User avatar
IslandPink
Amstrad Tower of Power
Posts: 10041
Joined: Tue May 29, 2007 7:01 pm
Location: Denbigh, N.Wales

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

Post by IslandPink »

Well done indeed Wolfgang - quick work - and very interesting to hear your enthusiasm for the circuit.
What interested me was the ability to get a fairly low output impedance without using feedback.
Steve, it's a bit mad, but the explanation is all in Steve Bench's page here ( first paragraphs ) -
http://diyaudioprojects.com/mirror/memb ... erted.html
The key thing being that if you use the grid like an 'anode' for output, although it can't deliver much current, it has a very low output resistance. Once you parallel up about 6 to 10 tube sections, it starts to have proper practical application.
Perhaps Wolfgang could just update his circuit to show how the voltages are applied to the output stage ?
"Once you find out ... the Circumstances ; then you can go out"
Wolfgang
Old Hand
Posts: 229
Joined: Mon May 07, 2018 3:08 pm

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

Post by Wolfgang »

The pdf of the schematics couldn't be opened in my last post.

Don't know what went wrong but here it is:
schematics.jpg
What interested me was the ability to get a fairly low output impedance without using feedback.
This is indeed the most interesting aspect of using an amp like this. With six 6AS7 I get Zout= 9,4R. With around 12 tubes one could drive 8ohm speakers. The effect on the sound with no output cap and zero NFB is incredible. It opens a new window into OTL sound. There must be musicians/hobby musicians in this forum and maybe some have made studio recordings and know how high speed analog open reel tape recordings sound. It reminds me very much of this sound plus even more additional sparkle which I can describe in another post and when I have 2 channels ready for testing it in my main system.

In order that this amp can be used in a practical way it needs some additional circuits. S. Bench must have been an incredible guy. Made me giggle several times when I figured out some of the practical aspects and how he had presented it on his project page not mentioning a single word about it.
I have all the additional circuits, just need to build them. What it needs is the following:
1. A relay that disconnects the output
2. A sequenced start up/shut down procedure that controls the relay
3. A DC protection for speakers
4. Ideally an auto DC offset circuit

The reason for this is that the output tubes need about 10s before they conduct fully and during this time the 40V of the output psu are connected to the output, slowly ramping down as the tubes heat up. At switch off we see the same thing just with a little lower voltage but still harmfull in the long run for any speaker.
The DC protection is just an extra speaker insurance in case of a tube failure as in such a case not only the 40V of the output psu but also the -236V of the anode bias could be connected to the output. In the case of PP OTLs we have the fuses for this reason but I don't think we would find one that would protect efficiently for 40V and 236V in the necessary ms time frame in order to be effective.
I want to use an active circuit for this ( not one of those crowbar circuits) that doesn't influence at all the amp SQ or functionality.

The DC offset can be set to almost 0V! Better and more stable than with an OTL PP amp. But it drifts slowly when the amp heats up and gets finally really hot. I don't feel ok with about 200mV DC offset so I want something that automatically adjusts it. But interestingly enough the DC offset has absolutely no effect on the sound as long as it stays below 600mV and it always does. With the Pinnacles I can hear the influence already at 100-200mV .
User avatar
Ray P
No idea why I do this anymore
Posts: 6294
Joined: Thu Nov 22, 2007 5:18 pm
Location: Somerset

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

Post by Ray P »

A really interesting project Wolfgang, I will be following with interest.

Do I unerstand correctly that you're using regulated supplies?
Sorry, I couldn't resist!
Wolfgang
Old Hand
Posts: 229
Joined: Mon May 07, 2018 3:08 pm

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

Post by Wolfgang »

Do I unerstand correctly that you're using regulated supplies?
That's right, Ray. For the driver stage I use 2 Maida regulators in order to get the +/-350V. They are perfect for this job. Only down side is that it needs 2 transformers or one transformer with two separate sec. windings in order to make it work.
For the output stage I will start with the Salas regulators and find out how or if they improve the simple CRC filter/grid choke and if they can be used somehow to stabilize the DC offset better (have no clear idea yet ).
Noise floor is very low even with the CRC/grid choke ( around -85dB). And I could stabilize with the pot the DC offset around 10-60mV which is already very good.
In the schematics it still shows the choke in the filter (CLC) with the resistors as grid load but I have changed this to CRC/grid choke because it sounds better. I will soon post the corrected version.

I think this is the first tube amp where I would only use regulated psus wherever I can because I want to reduce all sources for noise and IMD to the absolute minimum because there is no NFB. It's "music in the raw" so to say.
User avatar
Ray P
No idea why I do this anymore
Posts: 6294
Joined: Thu Nov 22, 2007 5:18 pm
Location: Somerset

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

Post by Ray P »

Thanks Wolfgang. I started to have a proper read of the Stece Bench article last evening and the concept is very interesting - I don't pretend to unerstand the intricacies of the electronics but the no feeback/output cap looks like an attractive result. A set of output tubes starts to look expensive though but I see there are some Russian equivalents to the 6AS7 that might be an option. Regarding managing the DC offset, I don't know if this might be useful?

https://www.diyaudio.com/forums/swap-me ... o-pcb.html

Also, and more peripheral but for interest, I've ordered some PCBs for some active rectifiers from a DIY Audio group buy;

https://www.diyaudio.com/forums/group-b ... ifier.html
Sorry, I couldn't resist!
User avatar
Nick
Site Admin
Posts: 15707
Joined: Sun May 06, 2007 10:20 am
Location: West Yorkshire

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

Post by Nick »

I would guess a DC servo with protection diodes at the front to cope with higher than rail voltages, and then use the output of that to drive a LDR to replace the 250k resistor in your circuit may work.
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

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

Post by Wolfgang »

A set of output tubes starts to look expensive though but I see there are some Russian equivalents to the 6AS7 that might be an option. Regarding managing the DC offset, I don't know if this might be useful?
I ordered 12 of these, $10,72 per tube. They seem to be fine tubes.

https://www.parts-express.com/6as7ga-6a ... e--072-612

One of the problems is the high voltage. I looks like the board recommended in diyaudio needs+/-12V . I don't think it would work here.
I would guess a DC servo with protection diodes at the front to cope with higher than rail voltages, and then use the output of that to drive a LDR to replace the 250k resistor in your circuit may work.
Very interesting, Nick, to combine these two things. Could you sketch something or give a little more hints or links to existing circuits? I was thinking of an LDR but couldn't figure out how exactly I could integrate it in this position. I was thinking either the diff. amp or the LDR. I also tried already a diff. amp with PNP transitors which works until I connect it to the amp circuit.It sometimes works and keeps the DC offset below 10mV, most of the times it runs completely off and freezes at around +8V. My guess is that when the basis goes positive it cannot recover it's regulation process. But that's the problem: when the amp is turned on the output is always on positive voltages until the regulation process could kick in but for neg. bias output for the anode I need a PNP in my understanding.
DC offset.png
User avatar
Nick
Site Admin
Posts: 15707
Joined: Sun May 06, 2007 10:20 am
Location: West Yorkshire

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

Post by Nick »

I was thinking of a op amp low pass filter like the one in the schematic in the diyaudio link. Fed from the amp output via a resistor and back to back zeners to protect the op amp from higher than expected voltages on the output of the amp. Then use the output of the filter to drive the LED in a LDR coupler like a nsl32, and the resistive output of that to replace the 250k pot in the bias circuit. May need a bit of tweeking to get it to work and may need a additional inverter stage to get the sense of the feedback right.
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

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

Post by Wolfgang »

Thanks, Nick. I will try this thing with the nsl32. Should work and would solve the only remaining little problem with this amp.
User avatar
Paul Barker
Social Sevices have been notified
Posts: 8867
Joined: Mon May 21, 2007 9:42 pm

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

Post by Paul Barker »

I might build a raw one without the stuff you’re planning wish 6as7’s that I have no’s.... well enough for mono. I once collected enough 5687’s but foolishly sold them. Have a few loose ones, not enough.
"Two things are infinite, the universe and human stupidity, and I am not yet completely sure about the universe." – Albert Einstein
User avatar
Ray P
No idea why I do this anymore
Posts: 6294
Joined: Thu Nov 22, 2007 5:18 pm
Location: Somerset

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

Post by Ray P »

Wolfgang wrote: Wed Jun 26, 2019 2:59 pm looks like the board recommended in diyaudio needs+/-12V . I don't think it would work here.
I was meaning would it be useful as a starting point rather than as the end solution but Nick has picked up on this now so hopefully it'll help you address the DC issue.
Sorry, I couldn't resist!
Wolfgang
Old Hand
Posts: 229
Joined: Mon May 07, 2018 3:08 pm

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

Post by Wolfgang »

I might build a raw one without the stuff you’re planning
That's all you need in this case:
Inverted OTL 6SN7_6AS7 w manual DC offset p.jpg
The chokes are Hammond, 159V.
Wolfgang
Old Hand
Posts: 229
Joined: Mon May 07, 2018 3:08 pm

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

Post by Wolfgang »

I was meaning would it be useful as a starting point rather than as the end solution but Nick has picked up on this now so hopefully it'll help you address the DC issue.
It helped me to start thinking out of the box because I was too focused on this one thing I had in mind.
Post Reply