Aleph P without Mosfets ?

Hi folks,

meditating about a mic-preamp circuit ba Phil Allison, published by Rod Elliott http://www.sound.au.com/project66.htm , I had the idea that one could take the first part of this preamp to form a single ended balanced in/out line gain stage according the way Aleph P or BalZen is suggesting.
The use of differential compound pairs of conventional transistors is at least an interesting thought.
I know, besides the approach of Master Pass, the single ended differential pair is done by Borbely with JFETS for example, but has anybody experiences with "normal" transistors?

Klaus
 
Pass uses JFETs himself for his later X-series amps + pre-amps. Check out some of the service manuals for preamps -- can't remember which ones.

The problem with JFETs is power dissipation. You are thus limited to a certain voltage and current range. Put complementaries butt to butt with perhaps a resistor between them.

The disadvantage of "normal" transistors (bipolar junction types) is the base current required which usually means you need to find some interesting way to set up base ("gate") bias manually. Since tubes, JFETs and MOSFETs don't carry steady state gate (or grid) current to speak of, they can normally be self-biased.

Petter
 
Petter,

when you look at the circuit mentioned above, the first part - the differential input with compound pairs - is actually an amp itself and is not included in the NFB path together with the following Opamp. So - assuming the complete circuit works - the first part should also work alone, probably with a relativly high output impedance. (And if it works with microphones it should work with other input sources too).
Finding the right output potentiometer impedance could be a problem.
BTW, I am not biased against mosfets in any way, but it was just an idea...

Klaus
 
Klaus,

Have you considered that this might make a great front end for a phono pre-amp. Since a phono cartridge is inherently balanced anyway and you could insert the eq. between the diff. amp and the op-amp.

Grey,
Have any comments? You are probably busy working on your tube amplifier's negative supply.

Jam
 
Jam,
Funny you should mention balanced phono, since that's what I'm building...
Actually, the idea of a single stage line is pretty seductive. Once I get the phono circuit laid to rest, and nail a few other coffins shut, I'm planning on doing a single stage balanced tube line preamp. (Probably 6SN7s, but leaving the door open to 6922/6DJ8s.)
*WHAT?* Grey, are you crazy? High output impedance, blah, blah, blah...
Well, yeah, actually I am crazy. Or else I wouldn't have so many people telling me so.
The card I've got up my sleeve (you knew there was a card up my sleeve, didn't you?) is that I'm going to combine the line stage with the crossover, which will save about $20 trillion in interconnects (quad-amp systems get unwieldy quickly), not to mention keep the back of my system from being any more of a rat's nest than it already is. (I'll have a full range bypass out, in addition, just in case I want one down the road.)
Anyway, the basic plan is that it will be a hybrid, thusly:
Selector->Vol->Line (prob. 6SN7s--about 15-20 dB gain)->2N5457 (JFET as follower/buffer)->Sallen-Key filters->MPSA18 (bipolar for low output impedence [as follower])
balanced all the way through.
So, as you can see, I'll end up with low Zout after all. Yes, I could use tubes as cathode followers and accomplish the same result. Eventually I will get around to trying it as an alternative just to see if it sounds better.
Back to the original question Klaus raised: Actually, I think MPSA18's might be a good candidate for a single stage balanced preamp (MPSA42/92s if you want high rails), but I tend to side with Petter, overall. Still, it shouldn't be too difficult to rig a voltage divider to bias the transistors, or...argh! don't get me thinking like that...I've got a phono stage to build!
Anyway, the idea has merit, as MPSA18/42/92s can be made to sound pretty decent. Give it a whack and let us know how it sounds.

Grey
 
Grey,

What you suggest will probably sound great but in keeping with a tube differential, why not a White cathode follower?
BAT uses multilple tubes in parellel to get a low output impedence in their tube pre-amps. They believe that followers hurt the sound. The problem is that when you are diving a low impedence the gain changes (as the cathode resistors are in parallel with the load). This could be a paticular problem with amplifiers that have balanced inputs with different impedences for the + and - legs(for example the Pass Aleph amplifiers).
I myself like Fet's and Mosfets,but what ever you choose the follower could make or break the circuit, I have found the follower is more critical to the sound of the circuit than most people suspect.

I could tell you about ways to improve followers using current sources and cascodes but that would probably make your hair curl......

Jam

PS. Is your phono stage anything like Allen Wright's RTP?

[Edited by jam on 05-30-2001 at 09:29 PM]
 
Jam,
The CJ ART line stage is also paralleled (10x6922 sections per channel, but single-ended, I believe). I had considered doing something similar, but in a balanced configuration...until I took a look at what it would require in a power supply. (I still may yet do it after I get a few other things settled.) Then I realized that there wasn't any need to worry about low Zout from the line stage since I was going to be working into a known load and could do it without interconnects and all the attendant L&C that they entail.
As far as followers and criticisms thereof. They've got their tradoffs, but there've been some damned fine sounding pieces of equipment just chock full of followers, whether cathode, emitter, or source. (As I've said before, I like solid state & tube, each for what it does well. In a pinch, I somewhat prefer tubes, particularly up front where I don't have to worry about power tubes going flooey.) The reason I'm using followers is high Zin, low Zout, which is ideal for filter math.
Current sources & cascodes with followers work well for some applications, but I don't really need them, and would prefer to keep it simple. I've already got a version 4.0 or so of the crossover running (been working at it off and on since last November or December). It's breadboarded and ugly as the girl in my sixth grade math class, but it sounds simply lovely. I've got one or two things I want to do to fine tune it, but my system already stomps anything I've heard in a month of Sundays.
Can it be improved? You betcha! The two main weaknesses at the moment are the phono stage--which is only 40dB gain--not optimal for a moving coil. I'm working on that. And lower bass power and control (15-30 Hz). I've been working on that sporadically for a while, but it's a Herculean task, as I've got to build 12 amplifiers, one for each woofer, and outfit each one with a feedback loop off the driver cone. I've got several ideas I want to pursue there, but the one I like best is optical. It'll be a bit before I get that done. For the time being the drivers are simply running in standard Thiele-Small cabinetry, good down to about 25Hz or so. It'll have to do for the time being.
I'm not familiar with the RTP. Got a website I can look at?
My phono stage:
6922 differential->6922 differential->passive RIAA->6922 differential
Total gain is on the order of 60 dB, plus or minus a few--I don't remember exactly; my notes are at home. Note no followers at the end, as I'll be running straight into the switch & a known load, etc. The way it's set up I have the option of using a 12AX7 up front to increase the gain, but don't really anticipate needing to do so. (Lyra Helicon cartridge--decent output.)

Grey
 
Grey,
This should get you where you need to be

http://www.vacuumstate.com/images/RTP5_rev_1.GIF

It sounds like you have a plan. My main concern would be noise.
The circuit that you describe looks very much like the Bat phono pre-amp.They include the option of inserting a transformer in circuit to allow for more gain in the more expensive model.
I myself might consider using FET'S in the input stage to avoid noise problems, but as you say your cartridge has probaably a high enough output to avoid this.

Jam
 
Grey,
On the subject of paralleling that many tubes, this can cause the sound to grey out (no pun intended). This could be caused by different characteristics of each tube. What if one tube drifts or one section of a tube drifts or fails and what if...........
I have played with the CJ and it can be a pain in the butt keeping it sounding consistant let alone one channel sounding like the other. No thanks, I have better things to do.

Jam
 
Jam wrote:

"Have you considered that this might make a great front end
for a phono pre-amp. Since a phono cartridge is inherently
balanced anyway and you could insert the eq. between the
diff. amp and the op-amp."

This is similar to what Ben Duncan did in his ADEQ phono
stage, albeit using the SSM2016 chip, which is basically a
pair of op-amps with a differential pair topology at its
input. Duncan points out, as you do, that cartridges are
particularly suited to a balanced input, although I've seen
very few preamps which exploit this. Using a discrete
differential pair at the input would definitely help to
keep the noise even lower, particularly for low-impedance
cartridges; the 2016 actually has a good noise figure,
despite being an integrated circuit.

In the ADEQ the 75us filter is put between the 2016 and the
first op-amp, while the latter looks after the rest of the
EQ. I'm sure Grey would be horrified at the number of
active devices in this circuit, but it does work very well
all the same.

I've just rebuilt my ADEQ with a balanced input, but since
the arm lead is still unbalanced I get a little hum pickup.
Rebuilding the arm lead is yet another of the projects on my
list of jobs to do.

Alex
 
Alex,

Could you e-mail me a copy of Ben Duncans article on the ADEQ or is there a web site that I could go to.

I plan to use cascoded J-fet (I can vision Grey turning in his grave) differential with switchable gain (for moving coils). This would be followed by a passive eq. stage and then by a second differential stage which would be buffered with followers (Grey is moving again). There will be no loop feedback anywhere in the circuit, and the circuit would be totally solid state.
This would be similar in concept to Allen Wright's RTP phono section (link above). Another thought would be to replace the second differential with a variation of Nelson Pass's X circuit.
But all this is just an idea I don't know when I will have the time to persue this,as I am working on an amplifier now.

Jam
 
Jam,
Looked at the phono stage on the link you provided. Once I saw it, I remembered looking at it at some point in the past. Looks fairly reasonable. My only reservation about using a cascode as a first stage is that cascodes have comparatively high output impedence. Another point is that he's using his RIAA across the whole differential; you can't back up and use half the circuit easily. More on this in a second.
Alex,
Yeah, I had this blinding revelation about cartridges being ideal as balanced sources about two or three years ago. The differential nature of the circuit would cancel noise, distortion, etc., so why wasn't everybody doing it? Seemed to me to be such an obvious thing. I asked and asked and asked. Heard of a few circuits out there, but not many. Finally one guy ventured that some cartridges have the negative lead tied to the body of the cartridge as a grounding/shielding scheme, and that you could run into trouble *if* you owned such a cartridge. I said *harrumph* and started planning a balanced phono stage.
(N.B.: Nelson (okay, actually Wayne), who does *everything* balanced, did a single-ended first stage in the Aleph Ono [and presumably the Xono, as it supposedly is very similar], then created a balanced output from that...)
Now, back to what I was saying to Jam about symmetrical RIAA. If you have such a cartridge on hand (or a friend brings one over) it would be nice to run the phono signal through *half* the balanced stage (thus allowing the cartridge to ground in the manner it expects), which would have to have its own, independent RIAA eq. So my circuit uses two identical RIAAs, one for each half. The only thing to keep in mind is that the 47k input has to be adjusted for balanced or single-ended operation, and you'll need some switching or jumpers to reconfigure +- to +gnd. I intend to hardwire the thing as balanced, leaving the unbalanced option as a desperation option requiring a soldering iron.
Okay, onwards: Alex, I considered, and can still arrange, a JFET first stage for noise purposes, as those little SK critters are awfully quiet. But I don't have any on hand, and wanted to try all-tube first, anyway. (Who's got the cheapest prices on those JFETs these days? I need to get some.)
Jam again,
Turning in my grave? YIKES! I hope I didn't die and not realize it...I've got *way* too much to do.
I'm in agreement on the no FB aspect. My phono has no FB as it stands now. I could easily add local, I suppose, but for the time being I'm going open loop.

Grey
 
Grey wrote:

"Yeah, I had this blinding revelation about cartridges
being ideal as balanced sources about two or three years
ago. The differential nature of the circuit would cancel
noise, distortion, etc., so why wasn't everybody doing it?
Seemed to me to be such an obvious thing. I asked and asked
and asked. Heard of a few circuits out there, but not many.
Finally one guy ventured that some cartridges have the
negative lead tied to the body of the cartridge as a
grounding/shielding scheme, and that you could run into
trouble *if* you owned such a cartridge. I said *harrumph*
and started planning a balanced phono stage."

This is why I like the logical grounding scheme SME (among
others) use in their tonearms. Their scheme is to have a
separate wire connected to the tonearm, which you can (or
maybe not!) screw into the earthing post on your preamp.
If there is a conductive path from armbase to headshell,
the cartridge body is then grounded. Mind you, if the
cartridge body itself is internally connected to the signal
return, you could have a nice hum loop. Rega, by contrast,
ground the signal return on one of the channels of their
tonearms - one of the reasons for the popularity of
aftermarket rewiring of the RB250 and RB300.

I guess the overriding reason for not doing these things
properly in the first place is cost.

Anyway, Ben Duncan claims that when a cartridge is connected
in balanced mode the hum disappears in a very dramatic
fashion.

Alex
 
Alex,

I take it that you have a Rega tonearm. I would highly recomend that you re-wire it as a soon as possible. I have heard the differences on a friend's tonearm and was skeptical of the magnitude of the differences till I heard it myself. It takes the Rega into the same league as the mighty SME V (no kidding!). In my book this makes the Rega the best bargain in audio. It is a pity the manufacturer of the Rega does not see it this way.
Ben Duncan is absolutely right, once you hear the lack of noise or hum you can't go back. The differences can be positively eerie in a high resolution system.

By the way, most cartridges that have their body tied to one of the signal leads usually can be fixed for balanced operation. Look fof a metal tab between the one of the cartridge pins and the body on the back of the cartridge, all you have to do is cut this tab and begin to enjoy the music.

Jam



[Edited by jam on 05-31-2001 at 03:21 PM]
 
Grey,

The beauty of Allen Wright's approach is that you can go into the circuit single-ended or balanced and come out single-ended or balanced and more importantly you have only one set of eq. components, which means the outputs will be totally complementary to each other and have a high Common Mode Rejection Ratio. (CMRR)
In your circuit (if I can visualise it correctly), you have to match two sets of eq. components to some great tolerence or run the risk non-linear addition and subtraction of the circuit halves and possibly end up with a non-flat RIAA equalized response.

I believe a good option in my idea is the use of an X-amp in the second stage of phono amplifier after the eq. I have to go on record and say that the X-amplifier, in my opinion, is the single most important advancement in analog amplifier technology in the past two decades. Zen-Master Pass Rules!

All this brings up an interesting question about your x-over in your quad amped. system. Are you converting the signal to single-ended before the x-oxer and then back to balanced after the x-over? If not I can see a whole host of problems (like the RIAA stage) and would like to know your solutions.

Jam

P.S. All this from a guy that don't even own a turntable.
 
Jam,
Let's see if I can keep all this straight in my head long enough to get it written down.
As for the CJ ART and other paralleled tube circuits...it's an ideal situation for a current source. As the tube(s) age, they are held to the same current, regardless. I have no idea whether CJ uses resistors or current sources in the ART. Sonically, I would vote for a resistor (somehow, you knew that...right?), but this might be one of those situations that would justify a current source.
Mis-match in my RIAA: Keep in mind that differentials tend to balance out (cancel) unequal signals, so in fact having two RIAA sections allows me to average errors, hence greater accuracy. Add to that the fact that I'm using as tight a tolerance parts as I can get my little sweaty paws on, and it should work out pretty well. (Just tossing parts in without trimming or matching capacitors got me +-.2dB accuracy to the RIAA curve, which ain't so bad. With a little fiddle factor, I can probably nudge that even closer.)
Currently my entire system is running single-ended, so the crossover is single-ended. As I go through this convulsion of revamping things, I'm making it all capable of balanced input, but not yet actually running balanced. Since I mostly listen to phono, there's not much point in balanced operation until I get the phono stage ready. Otherwise, I have to create a balanced signal out of thin air. Okay, I'm already doing so in the first stage of my tube amp, but it's semi-ready to go balanced.

Grey
 
Grey,

I believe the CJ uses resistors and the BAT is a current sourced differential. The idea of using multiple devices is to lower the output impedence and hence improve drive capability. Don't get me wrong the CJ sounds rich and lush but in my opinion is slightly fuzzy and lacks the focus of a circuit that uses one or two tubes. The problem is that when one tube starts to go bad the other tubes mask it to some degree a question a CJ representive doged when I posed it to him. The problem is keeping good tubes in pre-amp at all times, I certainly don't want to be testing tubes every time I want to listen to my system, I bet you that there are people out there using pre-amps that are not performing to their optimum.

The BAT is a far superior design and is designed to drive impedences as low as 10k which is great in most
applications,but let's take the case of an an amplifier that has unequal impedences on it's balanced input, as low as 1k on one leg and 25k on the other. This is why I like followers as long as they are done right, the exception being the case,as you mentioned, controlling the impedence of the load. I like designs that have some degree of immunity to the load presented to them (makes comparisons easier).

I am impressed with the accuracy of your RIAA curve, I was not trying to be critical but play the devil's advocate, that being said I usually prefer the simpler more direct approach (something following Nelson's designs has taught me).

Jam
 
Jam,
I haven't heard the ART, nor have I heard any BAT stuff, so I don't have an opinion on either one. (No excuse on the BAT, since there's a dealer here in town. I'm just trying to get so many things done that I haven't taken time.)
We'll see what happens with line stages and crossovers and such soon. Right now, I've got to get the phono up and running, and that in turn is waiting on me finishing the new cabinetry for mid/tweeters.
And as if that's not enough, I'm starting to think that I might want to use the Aleph 1.2 project as the excuse to begin experimentation on melding the X front onto the Aleph rear. Somebody better save me from myself...

Grey
 
Grey,

Now you are talking! I hope the Aleph X (or is it the X Aleph) is now on the drawing board. Please keep us informed on your progress.
I had a thought about a line stage using a 6SN7 in a current sourced differential configuration (I suspect this is one of your tubes of choice), followers optional. Every one else seems to be using 6DJ8's, but that was some time ago.
Another tempting project is Allen Wrights zero feedback amplifier.

http://www.vacuumstate.com/images/PP-1C_a.gif

Both projects are not practical for me due to nature of my system. You on the other hand.......

Jam