Tube/MOSFET hybrid results/comments/suggestions?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Zombie hybrid ?

Nelson Pass said:
The moon is full, and an ancient thread rises from the tomb.... :vampire:

Nelson, me thinks the thought of not dispelling the mystique around hybrids has haunted you all these years. Maybe that's why you've been drawn back here to bring the light of creativity to the dark art of hybrids? :hot:

So, what might the zombie hybrid look like? Would it be born single-ended or of bipolar parents?

Inquiring minds need to know!


Cheers,
 
Nelson Pass said:
100 dB would be radical conservative (Iverson)

50 dB would be reactionary conservative (D'Agostino)

20 dB would be center of the road moderate (Rowland)

10 dB would be moderate liberal (Curl)

0 dB is foaming-at-the-mouth, (rrrrrrrr!! Mad Dog!! rrrrrr!!)

:) :)


Over the last five to ten years I appear to have gotten more liberal; going from moderate to Mad Dog. Given that the GR-25 has at last (arguably) achieved 0dB, I suppose I'd better make an appointment for my rabies shots.
At least I now have an answer as to why I feel like biting people all the time...

Grey
 
The one and only
Joined 2001
Paid Member
Over the years everybody and his brother has fooled around with
hybrids of tube front end and transistor output stage, and they
still are.

I much prefer to play where the sandbox is not so crowded.

However, there are plenty of people here who have an interest
in such things.

:cool:
 
I love tubes. The only reason I abandoned tubes was pricing and availability problems. So I came over to solid state...only to find pricing and availability problems. Oh, well.
The idea of combining solid state and tubes is seductive. My problem with hybrids is that instead of combining the best of both worlds, they frequently tend to combine the worst. I have some ideas I want to try "one of these days" but my enthusiasm level is low.
Incidentally, the circuit linked to in the first post is, shall we say, incompletely thought out.

Grey
 
I´ve lsitened to quite a few hybrids over the years, now seeing that Grey posted while i was looking for my glasses to see the keyboard, i´ll just agree, the different tecniques flaws seem amplified, not the virtues. In many cases it looks like the hybrid way is choosen to get rid of the output transformer while a really good opt is one of the reason tubes can sound so good.

Grey!, You don´t need the vaccination, both Brora and Port Ellen works that way as long as used an a regular basis.
 
I feel that the problem is two-fold:
1) Many hybrid circuits are put together by people who are reasonably competent in either solid state or tubes, but not as well versed in the other field. The voltage gain/swing possible with tubes looks enticing to those whose background is solid state. The current delivery and low Zout possible with solid state strike similar envy into the hearts of tube folks, which leads us to...
2) It takes more current (and/or a lower impedance) to drive solid state than most people realize, regardless of which camp they come from. If the normal order of things led to solid state front ends coupled with tube outputs, we'd see a different set of problems, but for better or worse it's tube fronts and solid state outputs. If you take a 12AX7 and put, say, a 100k load resistor on it (a completely natural thing to want to do), you're waaaaay off the reservation compared to a solid state device pushing a 1k load. Tubes can take that 100k Zout in stride, in part because an average tube exhibits low capacitance at the grid (and that capacitance is linear and of very high quality compared to solid state--dielectrics don't get much better than pure vacuum, whereas petitions begging for silicon-based capacitors never seem to garner many signatures).
I'm not willing to go so far as to say that it can't be done, but I can't think of a single hybrid component that would honestly qualify as a classic. Read that line again if you have to. I didn't say that no one has done hybrids. I said that none qualify as classic. There's always some poster who thinks that because Conrad Johnson, Counterpoint, or whoever made the model X, then it automatically warrants being called classic. Not so. Note that Counterpoint is out of business (in no small part due to reliability problems) and CJ's Evolution line has long since sailed over the horizon. And with nary a tear shed by knowledgeable listeners.
It's a shame that the two don't fit together better. It would be nice to have the strengths of both and none of the weaknesses, but it's not as easy as it looks.

Grey
 
What about sand front end and vacuum tube output? There a couple of phono preamp designs that use cascoded JFET front end driving tubes. My "Shrine" amp also uses a cascoded JFET front end - it's probably on the "liberal" end of the scale in terms of loop feedback .
 
Not easy to clone

It strikes me that a good tube preamp into an F4 is kinda the ultimate hybrid in two boxes already.

But for a real hybrid that sounds stupendous (and costs the green to go with it), check out Frank Bloehbaum's Thorens TEM 3200. Fully DC coupled tube gain stage into circlotron Mosfet output stage (one giant transistor per leg only) biased from a single point.

Sonically, the tubes dominate but this thing has the control, bass, speed and linearity top to bottom tubes don't have. And it's 200 watts so it'll drive a lot of things...:hot:
 
The one and only
Joined 2001
Paid Member
GRollins said:
Note that Counterpoint is out of business (in no small part due to reliability problems) and CJ's Evolution line has long since sailed over the horizon. And with nary a tear shed by knowledgeable listeners.
It's a shame that the two don't fit together better. It would be nice to have the strengths of both and none of the weaknesses, but it's not as easy as it looks.

Apparently not. I am struck by the observation that those doing
hybrids are "jacks of two trades, master of neither", although I
don't know that this is actually true.

Didn't Harvey Rosenberg do the Moscode, and wasn't that a
tube front end with Mosfet followers, and if so, might it not also
be considered a classic?

I would like to think that Harvey accomplished such a thing, but
I've never heard it.
 
Re: Not easy to clone

Srajan Ebaen said:
It strikes me that a good tube preamp into an F4 is kinda the ultimate hybrid in two boxes already.

But for a real hybrid that sounds stupendous (and costs the green to go with it), check out Frank Bloehbaum's Thorens TEM 3200. Fully DC coupled tube gain stage into circlotron Mosfet output stage (one giant transistor per leg only) biased from a single point.

Sonically, the tubes dominate but this thing has the control, bass, speed and linearity top to bottom tubes don't have. And it's 200 watts so it'll drive a lot of things...:hot:


My definition of classic seems to be different from others'. If you look around this site you'll find a number of posts that seem to indicate that many believe mere production qualifies something as a classic. I confess that my definition of classic includes the passage of time--often so much of it that the item in question is discontinued. Another aspect--and I'm not hide-bound on this, but it helps--is that the item holds its value or even increases. The value thing gets murky when you consider a company like Audio Research, McIntosh, or Mark Levinson where the brand itself becomes a value booster. I owned a number of solid state McIntosh pieces back during the '70s and can attest that they are not sonically worth the prices they get on ebay today. That doesn't mean that they're not nice to look at, but that's not the same thing as good sound.
(I wish I had enough discretionary income that I could afford to buy another 2205 amp just to look at--it need not play music. In fact, I'd prefer that it didn't.
Given that I include time as a criterion, the Thorens TEM 3200 doesn't qualify (to me) as a classic--yet. Whether it is or is not may become obvious in another five or ten years.
Your point regarding tube preamps driving solid state amps is a good one; it's something I struggle with, myself. I own a Conrad Johnson Premier Three preamp (somewhat modified--don't tell Bill and Lew) and use it to drive both tube and solid state amplifiers.
So what's the difference between a tube preamp driving solid state and a hybrid? I'm reduced to hand-waving, so bear with me. If I ever follow through with my hybrid ideas, then perhaps I'll be able to prove or disprove my point(s).
Obviously tubes are quite capable of amplification all along the audio chain. Ditto for solid state. Tube preamps can work very well with solid state amps, so let's draw a line in the sand and say that all is well at least up to that point. But how much farther can we take tubes?
Let's talk about something simple, like world peace, okay?
No?
Well, all right, if you insist...
Again, I have no hard facts to back this up, but my feeling is that the problems arise when tubes are asked to begin the transition from voltage amplification (which they obviously do very well) to current amplification (where they are, to put it charitably, challenged). So where does this transition I speak of actually occur? Arguably in the first stage of an amplifier. Before that, you're speaking in terms of milliamps of current, and I mean single digit numbers, often fractions of a milliamp. But a normal power amp is mainly a current amplifier, by which I mean:
Assume 1Vrms to full power
Assume a 47k input impedance
This works out to .02mArms to drive this hypothetical amplifier to full power. However, you're looking at (and here I'll use the GR-25 simply because the numbers are fresh in my mind) 1.77Arms at the output...and that's for a mere 25W into 8 Ohms. That's nearly 90,000 times gain for the current. Ouch!
So where does this current gain actually come to pass? That's a harder question to answer because there are so many designs and some of them demonstrably do nothing in the first stage; the first stage is unity gain or something close to it. In a case like that, the current demands are trivial and you might get away with substituting tubes...except for the fact that the true function of a front end like that is level-shifting and that is not a function tubes perform comfortably.
In the GR-25, I go from a 100k input impedance (single-ended--even less current demand than the hypothetical 47k above) to 5.5mA in one fell swoop. Bang. Fait accompli. This is within the capability of a tube like the 6922, but the problem that crops up is that there are only "N" tubes--no "P" tubes are available. Whereas in the front end of the GR-25, I'm not just running one active device, I'm running four, so depending on how you choose to look at the circuit, I'm actually getting something like 22mA from a .01mA input.
Yes, I'm drastically oversimplifying. For instance I left out the fact that I'm using a differential, and we could use the same circuit with tubes, but then we'd run smack into the lack of a "P" tube again. Just as a thought experiment, let's assume that we were to use two differentials comprised of 6922s, cap-coupled to make up for the lack of P-tubes. How much farther down the road can we go?
The next stage in the GR-25 is a folded cascode running at 100mA.
100mA?
Yikes!
The tube amps I designed used 6SN7s for drivers. A quick glance in my RCA Receiving Tube Manual tells me that the 6SN7 is only good for 20mA per unit (it's a twin triode), so we'd need 2 1/2 6SN7s running in parallel to deliver the same current that I'm getting from a single MOSFET. But remember that the 6SN7s are running flat-out, damn the torpedoes, full speed ahead. The MOSFETs I'm using (Toshiba 2SK2013/2SJ313) are loafing along at 1/10th of their rated current.
I think we've just hit the limit.
Again, I'm glossing over things. The first one that will crop up is that I'm a fiend for drive current. According to the math, I don't need to be running 5.5mA and 100mA at these points in the circuit. Only...dammit, the thing sounds better this way. For those who want specs, I'll note that the distortion dropped a full tenth of a point as I ramped up the driver current from ca. 50mA to 100mA.
So where does this leave us? Honestly, I'm not sure I have a good answer for you. I like tubes. I worked with tubes for years. But short of using a 6550 for a driver (plate current= 175mA, design center), we run into trouble either in the first stage or shortly thereafter.
Is this an answer everyone can agree on? Absolutely not. In fact, I'm in a minority regarding my design philosophy and I've heard some very nice sounding equipment that didn't meet my objectives. As I said above, take this with a grain of salt. I'm not trying to say that a decent hybrid can't be done, only that the ones I've heard left me wanting more.

Nelson Pass said:


Apparently not. I am struck by the observation that those doing
hybrids are "jacks of two trades, master of neither", although I
don't know that this is actually true.

Didn't Harvey Rosenberg do the Moscode, and wasn't that a
tube front end with Mosfet followers, and if so, might it not also
be considered a classic?

I would like to think that Harvey accomplished such a thing, but
I've never heard it.


For the sake of argument, let's stipulate that Bill Conrad & Lew Johnson and William Zane Johnson know a thing or two about tubes. Having heard quite a few pieces by each company (Conrad Johnson and Audio Research, respectively), I'm willing to stand behind that statement. Both companies have produced solid state gear, and at least CJ has attempted hybrids (Has ARC? I can't think of an ARC hybrid at the moment). In neither case have the circuits achieved "classic" (see my definition above) status. Respectable performance and a few good reviews, then...nothing. Zip. They dropped out of sight. If it weren't for the name on the front of the gear, no one would give them a second glance.
I can't think of a company known for solid state that's gone in for tubes or hybrids. Maybe I'll be able to come up with one later.
Over and over again, the consensus--over time--seems to be, "Well, yeah, they're okaaaaay," with that long, descending, drawn-out vowel that indicates that the speaker is damning with faint praise.
I have not heard the Moscode, so as per my long-standing policy all I can say is that I have no opinion. Let's say the Moscode is the 'exception that proves the rule' (a dumb expression if ever there was one). On the one hand, all it takes is one white raven to prove that such a thing can happen. On the other hand...why so few examples of successful hybrids? Is it really that hard to do?
Does anyone have access to a schematic of the Moscode? Now that Nelson has brought it up, it would be interesting to see how Harvey solved the problem.

Grey
 
wrenchone said:
What about sand front end and vacuum tube output? There a couple of phono preamp designs that use cascoded JFET front end driving tubes. My "Shrine" amp also uses a cascoded JFET front end - it's probably on the "liberal" end of the scale in terms of loop feedback .


tubesguy said:
There's also the Zenkido, referenced in Audio Express last year - I'm especially interested in the output configuration in that amp, which uses an Aikido for the front end. - Pat


tms0425 said:
Frank Van Alstine is another that comes to mind with the Fet Valve Ultra amplifiers. Been doing it quite awhile with success and staying power. Tom


Again...my definition of classic clearly differs. "Instant classic" is obviously an oxymoron. Mere production doesn't impress me in the least. Van Alstine has been around a long time, but I'm more of a mind to note that he hasn't taken the audio world by storm--at least with his electronics, though his cartridges have a loyal following...but that's not what we're talking about. If there are hordes of people clamoring for his gear, they're certainly not making enough noise to be heard above the usual background chatter. Does he make enough to live off of? Presumably so. Has he bought out a defunct grocery store to use as a manufacturing facility? Haven't heard about it. If that's happened, I'd be interested to learn of it.
I'm not aware of any reviewers or manufacturers using older Van Alstine stuff as reference/demonstration equipment. On the one occasion I heard a Van Alstine amp, my impression fell into the "it's okaaay..." category. It certainly didn't make me want to run right out and buy one, even used.
It's not impossible to conceive of a DIY "classic" but you run into severe problems. The basic schematic may remain the same from one person's effort to the next, but differences in parts selection, PCB layout, chassis, routing of the wiring harness, etc. render any attempt to label the end products "the same as" problematical. You might point at something like Dynaco or early Hafler and try to make the case that they were DIY to some extent, but then we're faced with the fact that the kits were either tube or solid state, not hybrid.

Grey
 
Re: Re: Not easy to clone

GRollins said:

But a normal power amp is mainly a current amplifier, by which I mean:
Assume 1Vrms to full power
Assume a 47k input impedance
This works out to .02mArms to drive this hypothetical amplifier to full power. However, you're looking at (and here I'll use the GR-25 simply because the numbers are fresh in my mind) 1.77Arms at the output...and that's for a mere 25W into 8 Ohms. That's nearly 90,000 times gain for the current. Ouch!
So where does this current gain actually come to pass? That's a harder question to answer <snip>

Again, I'm glossing over things. The first one that will crop up is that I'm a fiend for drive current. According to the math, I don't need to be running 5.5mA and 100mA at these points in the circuit. Only...dammit, the thing sounds better this way. For those who want specs, I'll note that the distortion dropped a full tenth of a point as I ramped up the driver current from ca. 50mA to 100mA.<snip>

Grey

Here in Colorado, we know that one can scale the 12,000 ft mountain passes in a VW bug with 4 adults aboard. Unlike the bumblebee, the math does work. BUT, the experience is entirely different in a GTO with a mega-CID power supply under the hood.

Which is to say that it seems that rate-of-change capabilities are more significant then any steady-state parameter. dI/dt rocks my world... yeah, we be slew'n !

Just my $.02

Cheers,
 
Official Court Jester
Joined 2003
Paid Member
:rofl:

painting splitted hair tonight , grey ?

:clown:



ok - just write that missing sentence - "you can't think in same boundaries when you are mixing two sides of coin" .......

look at coin from his edge ; sorta schizo , but certainly workable ;

and - classic ?
from engineering point - one person remembering - that's enough ;

from sales point - ......... who cares - yesterday's story ....


btw - current mirrors can do wonders in toob - sand mixes


Gizmo is certainly laughing as crazy now ......... :clown:
 
And I for one can't follow Grey's argumentation about current gain... what has "input" current through a 100% arbitrarily(!) chosen resistor to do with the current levels inside an amp? If you're lucky to have a preamp that can deliver without problems into say, 100 Ohms this doesn't change things inside the amplifier a single bit, especially not with a FET or tube input.

As for the "classics" thing, I'd say there is more than the sheer fact that a design has "survived", which is based on many more things, most of them well outside actual circuit performance. Take the mentioned Thorens TEM 3200 (which I think is a very good amp, if not a masterpiece), it will most probably not survive and most probably not for any technical or sounding reasons...

- Klaus
 
lgo51,
Slew rate may be part of the problem, but that leaves begging the question of why all-tube amps sound as good as they do. A partial answer might be that tubes generally have very low capacitance and very high input impedance compared to solid state, hence they require less current to drive. Fair enough...but when you test the slew rate of tube amps you often find atrocious numbers...and yet they still sound good. Why? How? It's a mystery.
Zen Mod,
I'd be glad to reply, but I'm completely unable to make sense of your post. It seems caught mid-way between a drug trip and stream of consciousness. I'll grant that I don't always grok the posts made by members of the Off-Topic Gang, but I'm not even clear if you're making a statement or asking a question. If you'll rephrase, perhaps we can try again.
KSTR,
Congratulations! You completely, utterly, absolutely, positively, unquestionably missed my point.
Of course a good tube preamp has no problem driving a reasonable input impedance on an amp. That was never in question. My point was, and remains, that there seems to be a line of demarcation before which tubes successfully mate with solid state, but after which things don't work as gracefully. The question is: Where is that line? How far down the amplification chain can you go before you need to commit to solid state in a hybrid system? The amplifier's input stage? The driver/VAS stage? The output stage? If you carry all the way through with tubes it works wonders...but then it's not a hybrid anymore. Likewise, it's obvious that you can use solid state all the way through...but that's not a hybrid, either.
Since a tube preamp can work equally well with both tube and solid state amps, it's clear that tubes can make it at least that far down the signal path and still hand off the signal gracefully to solid state. This is essentially a generalized restatement of Srajan Ebaen's observation about a tube preamp into an F4 being the ultimate; he and I are in complete agreement at least that far. But how much farther can tubes be counted on to mate well with tubes?
Ay, there's the rub.
Srajan Ebaen feels that the Thorens amp (a hybrid) is a splendid example of the genre. I haven't heard it, so I can't comment directly, but my past experience with hybrids doesn't lead me to be particularly hopeful.
I have an operating hypothesis that the problem is related to impedance and current and I traced that idea step by step through the stages of an amp in an attempt to guess where the last reasonable place to use tubes might be. If my line of reasoning is correct, then it seems that the dividing line for conventional tube and solid state circuits is about where you would expect it to be, i.e. between the output stage of a tube preamp (usually, but not always, a cathode follower), and the first stage of a typical solid state amplifier, where the gain first begins to slant more towards current gain (90,000 x) as opposed to voltage gain (assuming an average 26dB gain amp, this works out to 20 x). And, sure enough, that's pretty much where things stand in the marketplace. Over the years things have converged on that cusp as being the right place to make the transition between tubes and solid state.
So much for conventional preamps and amps. Is there a topology that would allow tubes to penetrate deeper into the amplification chain before handing off to solid state? Maybe. Maybe not. If my hypothesis holds water, we'll need more current. You could use a power tube (e.g. 6L6, 6550, etc.) and push the dickens out of it current-wise and see what happens. Not once, ever, have I seen a hybrid attempt serious current drive at the transition between tubes and solid state. I have several things I would like to try along those lines if I ever get the time. The problem being that I generate ideas faster than I can work them and I've got the GR series and a discrete phono stage that I want to work on right now.
Did that make things any clearer?

Grey
 
How can it be?

GRollins said:
lgo51,
Slew rate may be part of the problem, but that leaves begging the question of why all-tube amps sound as good as they do. A partial answer might be that tubes generally have very low capacitance and very high input impedance compared to solid state, hence they require less current to drive. Fair enough...but when you test the slew rate of tube amps you often find atrocious numbers...and yet they still sound good. Why? How? It's a mystery.
Grey

As long as we're wading in 60's jargon; could it be simply that electrons dig streaking in a vacuum over a crystal maze? As for measured SR, absolute comparisons aside, does one observe the same decade drop in distortion with a doubling of output current? If the trend holds, then there's a clear message for me. I'm not sure that VT and SS measurements should live on the same scales, just similar domains.

Personally, I'm glad for the mystery. If all was known, then there'd be much less musing / experimenting / promoting / reviewing / re-creating -- and likely even less listening. Vive les différences! :joker:

Cheers,
 
them Thorems

I know how to write but not how to decipher a circuit schematic. That said, I had the Thorens engineer talk me thru his circuit in a way even my thick skull could follow. As it turns out, there is no voltage gain whatever in his transistor output stage. Just like the F4, it handles current and impedance conversion exclusively which, presumably (?), is one reason why the valve contributions seem to dominate the final sound.

The same approach seems at work in the Tenor monos, however their output stage runs massively paralleled transistor banks unlike the Thorens' super-beefy "singles"...

As to the classic thing, Klaus makes a valid point I think that some deserving products go the way of the dodo not for reasons of performance, reliability or circuit ingenuity. Cite market realities, mismanagement, an owner's unwillingness to lose more money on audio (ha, there's a concept)...

At the same time, classic does conjure up something tested by time and proven valid and kicking still. On that front, the Thorens certainly doesn't make it. But seeing the current economy, the fact that Thorens isn't known for electronics and may not have the distribution necessary to sell expensive amps and preamps... it's possible those amps won't remain available long enough to make classic status.

Does that mean they didn't deserve it?

Here's another one to throw into the mix: Mike Elliott's Aria Audio hybrids, a revisit, one assumes, of his earlier work at Counterpoint...
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.