Tube OTL and Zotl amps

[thor.zmt]

Thor: No PWM, agreed. But still switching.
Me: It doesn't matter

Switching is the issue, sorry old man. Not PWM. PWM or not, it's still switching.

This question referred to a tube stage of any kind connected to the input of a class D amp, not the Zotl.
Yes. It has nothing to do with the Class D amp which will work with or without it.

Ok, let me find a different illustration of a CMOTL as opposed to ZOTL.



This is from John Broski es Blog.

My designs use the same principle, but are different.

All amplifiers do. The example has nothing to do with tube.

It will not work without the tube and the current modulation in the Tube is directly impressed across the load and the current from the tube is multiplied by the Amplifier.

Seen by either load or tube as "black box" my circuit (CDCMOTL) and ZOTL show no observable difference safe my circuit operating at ~1.5MHz and using a type of filter that suppresses the carrier by around 110dB, so my circuit shows no switching residue observable on an oscilloscope as it's residue is in the 10's of microvolt.

Inside the black box they are very different of course. And my circuit, unlike ZOTL acted not only as current multipler at a voltage stepdown, but operates as power multiplier and as said, has a magnite or more lower carrier breakthrough.

My first design using this principle was a 300B SE amplifier with output transformer and a 120W / 8 Ohm "boosted" output from dual LM3886 as bridged impedance multiplier.

The amp could be switched over to "direct out" from the output transformer (with lower gain and output). Many experienced listeners using very high end speakers in their own system were unable to tell which position the switch was in, until they turned up the system past the 300B SE amplifiers limits.

In a friend's system it was up against a Kondo Baransu, Kevin and Ongaku (he collects Kondo Amp's among other things).

It acquired itself very well. And when we switched from Altec VOT's to ATC 3-way 15" monitors in "boosted" configuration it was the undisputed winner.

Then again, WTFDIK?

Thor

 

[Lampie519]

Isn't it sad that with all this expensive stuff he still can't hear the difference?

https://www.tubecad.com/2009/09/blog0171.htm

 

[thor.zmt]

Isn't it sad that with all this expensive stuff he still can't hear the difference?

https://www.tubecad.com/2009/09/blog0171.htm

That presumes that there IS an audible difference.

My point in making this design (before Mr. Broskie) was to provide precise scaling of the Tubes current and to reflect the speaker impedance back into the Anode Circuit, so the transfer curve of the "boosted" amplifier is the same as that of "direct" amplifier, except shifted 10dB on the power axis.

Meanwhile ZOTL provides all the disadvantages of Class D added to a tube amplifier, without offering any benefits actual benefit. Note I do not count removing the classic output transformer as benefit, except in terms of cost.

And yes, I auditioned a ZOTL Amp in my home system when I was still reviewing gear (pre ~ 2005 BC) and declined to formally review it, as I do not enjoy writing lukewarm reviews. It was a lot worse than a pair of Croft (Small UK maker) true OTL Monoblocks I borrowed from a friend and even a fairly basic Croft Hybrid (Basically a British take on Moscode) did a better job in my system (15" Tannoy Reds - 15Ohm/97dB/2.83V - in corner horn enclosures with super tweeter).

So I simply returned it to the UK distributor noting I did not feel I could write a review that would equal to the revolutionary design.

I think Harvey Rosenberg liked the 300B SE ZOTL well enough, maybe I'd have liked that Amp better, as at the time my own amplifiers were "all silver" (including pinstripe permalloy silver wound output transformers) 300B SE Monoblocks.

Well, ancient history.

Thor

 

[thor.zmt]

Anyway, it is clear you do not welcome dissenting opinions in your thread (which may end up being your loss).

So sitting at an open air venue under palms and a tropical sky listening to a live band that is rather decent and drinking "Captain & Coke" with my 23 Year old hot girlfriend (for real) according to Captain Jack, (See below) I really have better things to do than arguing my points. I'm hardly a Masterdebater.



Thor

 

[Lampie519]

not only the Zotl technology determines the quality of the amps but also the circuit itself like in any amp

From my first entry in this thread:

True, just omitting a transformer does not make a better amplifier. BTW the Croft amps i have heard were very lame and therfor lacked dynamics (also they missed the bootstrap as Futterman did and as explained in his patent).

Zotl has nothing to do whatsoever with Class D, but that is something that is hardly understood so you can't be blamed for not understanding.

Best regards,
Frank

 

[thor.zmt]

True, just omitting a transformer does not make a better amplifier. BTW the Croft amps i have heard were very lame and therfor lacked dynamics (also they missed the bootstrap as Futterman did and as explained in his patent).

I agree, I'm not a great fan of Croft's Amps.

Zotl has nothing to do whatsoever with Class D, but that is something that is hardly understood so you can't be blamed for not understanding.

Ok, again...

ZOTL -> Switching Step-up Converter with variable load varies output current derived from the low voltage power rails

CDCMOTL -> Switching Step-Down Converter with variable load varies output current derived from the low voltage power rails

The fact that we call one arrangement "Class D Amplifier" and the other "ZOTL" (which technically should be described as "Switching with RF Output transformer" or SWOTL) is semantics.

Both rely on switching converters to impress a facsimile of the current through a tube onto the speaker load driven by a low voltage switching system.

For the rest, as with all things audio I find the end result depends more on how it is done and less on what.

I am 100% certain both ZOTL & CDCMOTL can be made to have excellent performance if sufficient effort is expended.

Seeing that tubes are rapidly disappearing, I am by far more interested in how to get excellent sound quality from completely digital systems.

So again... Have a good Friday nite. Do go down the only pub in town with a decent band and raise pint.

Thor

 

[Lampie519]

For the rest, as with all things audio I find the end result depends more on how it is done and less on what.

I am 100% certain both ZOTL & CDCMOTL can be made to have excellent performance if sufficient effort is expended.

I totally agree !

Have a great good Friday !
Frank

 

[thor.zmt]

I totally agree !

Have a great good Friday !
Frank

Thanks, time to go home.

FWIW, try using the digital audio clock to run your switching synchronised. A 4X base wordclock should do it, use a crystal oscillator for analog in only (LP).

Second use 4 quadrant Quadrature to get in effect 4 times the switching frequency.

Then really fettle your drive circuit for either the sharpest edges possible for your waveform (kinda easy, but hell on passing EMC) or design it for symmetrical controlled slew rate on falling and rising edge (easy to pass EMC but losses are up and the design becomes significantly non trivial and complex).

Using 4 phase drive is easy, reduces burden on semiconductors and limits switching speeds.

Mind you, I'd probably try a ~1.5MHz CDCMOTL as well. These too can be made multiphase. I am working on something that is 768kHz X 4 phases on 56V rails.

Thor

 

[Mr. Moscode]

Switching is the issue, sorry old man. Not PWM. PWM or not, it's still switching.



Ok, let me find a different illustration of a CMOTL as opposed to ZOTL.

View attachment 107572

This is from John Broski es Blog.

My designs use the same principle, but are different.



It will not work without the tube and the current modulation in the Tube is directly impressed across the load and the current from the tube is multiplied by the Amplifier.

Seen by either load or tube as "black box" my circuit (CDCMOTL) and ZOTL show no observable difference safe my circuit operating at ~1.5MHz and using a type of filter that suppresses the carrier by around 110dB, so my circuit shows no switching residue observable on an oscilloscope as it's residue is in the 10's of microvolt.

Inside the black box they are very different of course. And my circuit, unlike ZOTL acted not only as current multipler at a voltage stepdown, but operates as power multiplier and as said, has a magnite or more lower carrier breakthrough.

My first design using this principle was a 300B SE amplifier with output transformer and a 120W / 8 Ohm "boosted" output from dual LM3886 as bridged impedance multiplier.

The amp could be switched over to "direct out" from the output transformer (with lower gain and output). Many experienced listeners using very high end speakers in their own system were unable to tell which position the switch was in, until they turned up the system past the 300B SE amplifiers limits.

In a friend's system it was up against a Kondo Baransu, Kevin and Ongaku (he collects Kondo Amp's among other things).

It acquired itself very well. And when we switched from Altec VOT's to ATC 3-way 15" monitors in "boosted" configuration it was the undisputed winner.

Then again, WTFDIK?

Thor

This is entirely different from what I was talking about, which was Zotl vs a generic Class D switcher. What you have here is something I've never seen or heard of before. We had our wires crossed. I'll spend some time looking at it. But I have absolutely no idea of what it sounds like.

 

[Mr. Moscode]

I envy your location for sure, Thor. Good for you.

You said" Meanwhile ZOTL provides all the disadvantages of Class D added to a tube amplifier, without offering any benefits actual benefit. Note I do not count removing the classic output transformer as benefit, except in terms of cost."

The benefits claimed are in the patent. The Zotls run Class B, G1 is tied to the cathode (in pentodes) w/ screen drive so tube life is improved. It weighs a lot less. Output stage goes to DC, phase shift is entirely out of band. Efficiency is better so . . .

Your opinion is yours and Frank's is his. Frank likes Zotls, you don't. You and Frank have way more experience with Zotls as I have none. And you seem to have differing tastes and opinions about them. Fair enough. The comparisons aren't controlled A-B in identical settings so there is no way to make absolute judgement about them. I was interested to know what DB had done.

Berning seems to have done well licensing his design and making some $75,000 amps with it. Some were actually sold! The reviews are reviews and are always in a conflict of interest as mags need controversial reviews to keep their advertising and readership. The reviewer I read said it's the best, except for deep bass. That's 20+ years ago. But reviews are generally like dating. Sometimes you get what you're looking for on the first date, other times she needs to warm up before you get the sparkling high end, lush midrange and a tight bottom . . . .you dig?

Thor, it appears you are against switchers in audio in any form. I don't disagree. I do use one in my string bass amp, a 400 watt ICE power along with a car subwoofer in a Polytone 12" cabinet from the 70s. Light weight and I can drive a room with it. But I've been off the audio scene since 2009 at the Rocky Mountain Horror, oops Audio Show.

I do have 40 years of experience with Futterman OTLs and my circuit improvements and experiments including a pair of 24 tube behemoths I built in the 90s but didn't fully finish. It would have needed 2 240 volt 30 amp outlets. I got it to work at 800 W/PC with our lights dimming in my small NYC shop.

I like the efficiency of the Zotl concept with energy rates sky high. Frank, can I borrow a Zotl? Maybe we could trade for something. I have lots of Moscodes.

 

[Mr. Moscode]

Anyway, it is clear you do not welcome dissenting opinions in your thread (which may end up being your loss).

So sitting at an open air venue under palms and a tropical sky listening to a live band that is rather decent and drinking "Captain & Coke" with my 23 Year old hot girlfriend (for real) according to Captain Jack, (See below) I really have better things to do than arguing my points. I'm hardly a Masterdebater.

View attachment 107586

Thor

I'm cool with your point of view. It shouldn't be personal. Send photo of the girl.

 

[thor.zmt]

I envy your location for sure, Thor. Good for you.

Thank you. It is.

You said" Meanwhile ZOTL provides all the disadvantages of Class D added to a tube amplifier, without offering any benefits actual benefit. Note I do not count removing the classic output transformer as benefit, except in terms of cost."

The benefits claimed are in the patent. The Zotls run Class B, G1 is tied to the cathode (in pentodes) w/ screen drive so tube life is improved. It weighs a lot less. Output stage goes to DC, phase shift is entirely out of band. Efficiency is better so . . .

Well, claims in patents are always interesting. I am more interested in subjective sound quality which is commonly not claimed.

Your opinion is yours and Frank's is his. Frank likes Zotls, you don't.

I feel that ZOTL is a solution looking for a problem. We know hoe make rather good output transformers and nowadays have core materials that are on other levels.

IF Tube Amp's still were a significant market (which they are not) we would see machine wound output transformers with vertical and horizontal sectioning of at 4 X 7 sections with up to 16 X 7 sections possible. They would use metal-glass cores with a significant nickel content and very precise air-gaps, even for PP.

Even just applying 2 X 7 sectioning and modern thin steel laminations in a EL84 PP Transformer can give <10Hz - 80kHz in a full amplifier with modest NFB, low phase shift and low distortion at low levels.

So from where I stand, ZOTL doesn't really improve materially enough on standard tech. It is relatively complex and leaves us to deal with all the challenges and issues of switching systems, without, in my books, material improvements over a well designed modern technology output transformer.

Switching frequency is low. It can be made higher with a fundamental redesign. We have GAN-FET's where we could easily boost the switching frequency tenfold and use a 4-quadrant quadrature and get effective 6MHz switching. But that also means PCB Design is on another level.

And it still will not address my core criticism of a solution looking for a problem.

So for example, I would not spend my time on improving ZOTL (even if it would allow to re-patent much of the expired Berning Patent) simply because it is a tiny market.

I'd rather hybridise a Futterman like this:

Use just one pair 12B4 or even 6N6/6N30 or perhaps EL84 analogues. MCU controlled biasing and monitoring for reliability.

Add P-Channel Mosfets (say cascoded and paralleled ECW20P20) in a Sziklai circuit. Now we have a Amplifier that is a "Boosted" OTL with 160V rail and ability to output ~ 100V RMS or > 1kW into 8 Ohm.

I'm sure we can sell some $75,000 monoblocks like that....

Berning seems to have done well licensing his design and making some $75,000 amps with it. Some were actually sold!

Good on him.

Thor, it appears you are against switchers in audio in any form. I don't disagree.

My view is more nuanced. I work a lot with switchers. I feel that switchers of any kind bring many challenges.

If I make a consumer product to be sold in 1,000's, I can see good reason to design a switching Amp with tube frontend and ~ 1.5MHz switching output, switching powersupply that can adjust voltage so when not playing loud we get a higher modulation index for better resolution and clock lock everything to the Audio clock etc. and other interesting "high tech" solutions.

Like for this product:



Would I do it for what is "ultimate performance" glorified one-off custom designs?

Well, I know MY choice but I do not want to convince others to follow my choices. It would be boring if everyone made the same.

I like the efficiency of the Zotl concept with energy rates sky high.

Then CDCMOTL is even better for you.

Thor

 

[thor.zmt]

Send photo of the girl.

Look me up on facebook. You can find her via my profile.

Thor

 

[Lampie519]

There are always people for whom the best is just not good enough ( i am one of those). So i hope we can at least agree to disagree in this matter. No hart feelings.
It is just a hobby and i hope we can still learn from each other, i did.

Best wishes,
Frank

 

[Mr. Moscode]

Thank you. Now let's sing Cumbaya.

There a lot of ways to audio nirvana. We can also fall in love with a design sometimes to our detriment. I haven't been to a hifi show since 2009. Jazz bass is my focus now at 71. I do repairs/upgrades out of loyalty to my customers and it forces me think which is good for keeping Al Zhiemer away, as is jazz performance.
The difference is jazz requires in the moment problem solving results, audio doesn't, I can take my time.

What's CDCMOTL? You have a working schematic?

Hi End Audio is a niche market, tiny by nature and for the customers it's a hobby, buying and selling just to hear something new.

I once read an article about D amps about 20 years ago that claims that the then current switching speed, around 200K? could not reproduce beyond 12-14 bit resolution. I haven't done the math but bit stream audio runs at 2.1mbit/sec so I would assume that is the minimum speed to reach CD resolution. I could be wrong. Any thoughts?

I have retired my soundbar on my TV in favor of my Luxman receiver playing through Rogers LS3/5 speakers. No comparison, the top end is much better and it's amazing that those speakers can get down almost as low as my sub woofer. At least I can hear speech much better.

 

[thor.zmt]

What's CDCMOTL? You have a working schematic?

Class D Current Multiplier OTL. Yes, I do.

The principle is simple enough.

Tube drives speaker from the anode through a low(ish) value resistor.

The resistor is bootstrapped by a Class D Amplifier operating at ~ 1.5MHz, thus making the resistor appear to have a much higher value.

The fairly detailed principle schematic was in earlier on the thread.

Hi End Audio is a niche market, tiny by nature and for the customers it's a hobby, buying and selling just to hear something new.

Indeed.

I once read an article about D amps about 20 years ago that claims that the then current switching speed, around 200K? could not reproduce beyond 12-14 bit resolution. I haven't done the math

A class D amplifier is in effect a delta sigma ADC followed by synchronous delta sigma DAC. Commonly the modulator is first or second order.

Let's note that Phillips DSD used a 7th order modulator at 2.822MHz to get 120dB dynamic range and ~ 96dB noisefree bandwidth of 20kHz.

So yes, there are challenges in such systems.

On the other hand, I am working with medium power fully digital Class D Amplifiers for upper end consumer grade products.

These have a fair bit of useful DSP on board, for limiters to prevent clipping, limiter timing under software control. DSP is good enough for crossovers and limited EQ.

They can operate filter less and have post filter feedback (if an output filter is used). They only run at 500kHz but have software tweakable settings such edge rate, that, together with an active system can help optimised one channel for HF drivers, another for MF and a third for woofers.

These are combined with a powersupply that can be adjusted across 12dB (e.g. 5V - 20V) so if the volume is low so is the PSU voltage and the amplifier runs at a much higher modulation index giving more usable resolution.

Implementing such a system carefully (including good drivers) and with taste can give a consumer product with borderline high end sound quality. Excellent for "secondary" or indeed retirement systems.

Thor

 

[Mr. Moscode]

Nice idea. Can it bypass the dsp? Not a fan of digital processing, simply because of the increased latency and the same reason you don't like switchers.

 

[Lampie519]

buying and selling just to hear something new.

Funny i never had this "issue". the moment i heard electrostatic loudspeakers i knew this is the way to go (for me that is). Then i needed something to drive those (direct drive naturally and with tubes no silicon in the signal path). Now the next step is how to solve the bass ( almost wall to wall speakers?) that was solved as well. Now the premaps and then the most difficult for me: The source (CD in my case as i don't like record players as they are quirky) So i needed to "re-invent" D/A converters ( the way there were intended in 1979). Dual mono and "just"16 bit (as this is enough anyway for almost 100db dynamic range). Now i am were i want to be and still trying to get it better, but every step now is much more difficult and takes more effort.

So any idea is welcome to reach the goal of "beeing there" ... May it be OTL, Zotl or whatever name you want to give it ...

Best wishes,
Frank

 

[Mr. Moscode]

Frank,
I haven't seen Beveridge direct drive e-stat amps since the 80s. Roger Modjeski's design, I think. Are they hooked to the Quad panels directly?

What's the rest of the gear in the photos?

Thor,
re: pwr amps on a chip. I have some studio monitors I inherited that uses them. They use too much fwd gain thus they need a lot of nfb for them to work. I didn't like their sound. It's the top 2 octaves of detail that is screwed up. That's where the Futterman's shine.

You should know the Futterman's use 60 db or so of NFB, 90 db open loo gain but the input tube runs current starved so the top octaves run at much lower gain and thus NFB. I think that's the key to F-OTL magic. I've had Counterpoint SA4s in for repair and while they are very clean sounding they don't have the magic.

Which brings me to this. I hung a moscode output stage off a F-OTL just to hear it. you would think it would be the dream hybrid but it wasn't. Every thing counts, caps, resistors, wire, circuit board layout and plating, regulation, tubes, gain, bandwidth, NFB, jacks. Everything does something. You have to get the right brands of ingredients.

The trick is to make it be seductive so you don't want to turn it off.

Here's another schem of what you are talking about.



It's simple enough and it has to do what the tube tells it to, to an extent. But it does it with a high NFB chip amp + an input xfmer neither of which I'm a fan of. But I haven't heard it so I can't judge. I think you used the term "near hi end performance". For me, it's not worth the time to build it unless it gets the "magic" top octaves. Maybe a discrete amp with lower fwd loop gain would sound better in this scheme. My Luxman receiver has a simple lo nfb design in it and it has a very agreeable top end.

For a computer speaker chips are OK. I use a 2 chip amp mini system for my amateur mixing setup. It's a low power draw so I don't feel like I'm wasting $ on the power company.

Have you compared these hybrids to the chip amp alone? Any improvement? I know we tend to stand on theory when we analyze a ckt by eye but how does it sound? That's the only relevant factor. I've had some clever ideas that just didn't pan out. Of course if you're looking to find a market for these, that's another story.

We have to stay objective but customers let us know what we've got.

 

[Lampie519]

haven't seen Beveridge direct drive e-stat amps since the 80s. Roger Modjeski's design, I think. Are they hooked to the Quad panels directly?

The Beveridge amps are indeed direct coupled to the ESL’s and the other speakers are Acoustat Monitor 4’s. Also these have direct drive high voltage OTL’s designed by me about 20 years ago replacing the original ones as these were hybrids, an absolute nogo at the time for me.

Best regards,
Frank