diyAudio

diyAudio (http://www.diyaudio.com/forums/)
-   Tubelab (http://www.diyaudio.com/forums/tubelab/)
-   -   6AV5GAs Re-Visited -OR- Pretty Darn Good SE Amp (http://www.diyaudio.com/forums/tubelab/204661-6av5gas-re-visited-pretty-darn-good-se-amp.html)

hareynolds 16th January 2012 11:15 PM

6AV5GAs Re-Visited -OR- Pretty Darn Good SE Amp
 
3 Attachment(s)
After a long interlude, I decided to clean the wood/metal/uncontrolled coulombs shop and see what's here.

I resurrected my "pretty" Simple SE with the blown diodes, plus the prototype Simple SE (currently set-up for 6v6s or 6BQ6s, the one with the plate cap), and then dusted-off the point-to-point-wired Simple SE Clone on which I had been experimenting with the 6AV5GAs (George's famous screen-tube SE discovery).

On the first go-round, the 6AV5 hadn't wowed me like it had George, so I intended to get to the bottom of that issue (at least within the limits of the old EL84 P/S that's on there, good for about 380V B+)

First photo is the top of the amp, made from scrap plywood and an old Porter Cable tool box. You can see it's been thru a lot, what with the extra ("vent") holes. Driver tube is a 6sl7 just cause I have some nice ones.
Second photo is the guts. Note the wire nuts and big loopy wiring solder-tacked in place. Notice the little TubeLab-style constant board in the upper right; forget what the resistor values are, guess I should look. Note also the AA NiMH battery bias for the 6SL7 with "helper" resistors to prop the voltage up a little.

When the smoke settled, I had ~365 V plate and ~57V at the cathode with those massive 50W 750 ohm resistors for a cathode current of ~76mA and plate dissipation of about 23 W. Remember, these tubes are "rated" for 11W.

This is slightly LOWER than Geo. recommends as a "sweet spot" with the cheapie Edcor OPTs (~315V at 80mA, or 25W), but I started adjustments at about 17W, so I'm emotionally exhausted, and resting here for a few days.

MAN, does this amp sound good all of a sudden. Somewhere around 20W dissipation, the 6AV5-Edcor combo started to come alive.

I've built an uncomfortably large number of SE amps, including EL34s, EL84s, 6L6S and 6V6s and their cousins, a 6AS7 thing, a 6BM8 thing or two, 6AQ5s, and about a dozen 1626 "Darling" amps of all stripes, plus the usual DHT suspects (45,46,too many 2A3s, 71As, and working on my first 300B).
Except for the shop, all the speakers I use are >95dB/W.

On the junk-yard shop speakers (about 90dB Radio shack 8" clams plus some surplus Foster horns in open-back "semi-baffle" cabs), the 6AV5s biased-up this way sounds like a 45 with an SRPP driver and James OPTs, except the bass is borderline fantastic. WAY better than any 2A3 that I've built. Certainly more detailed than any other pentode/strapped/ultra linear combo I've tried.

$2 Power tubes (I have an old stash), <$20 Edcors, $1 Russian PIO coupling caps (stash) and Scrap Iron P/S. Coincidentally, the bias resistors are real close to what you'd use on a 2A3 (which is why I had them). Just sayin'.

Ella Fitzgerald's right in the room. You can hear Keith Richard's fret-work clearly on the el-cheapo Rat-Shacks. Even the wife, who claims not to hear anything, said "ooo, that DOES sound good." And the BASS. Did I mention the BASS??

Other than the relatively high voltages, THIS is the SE amp that everybody should build first.

Now, of course, the trouble starts; home-brew success is nearly as costly as failure, I've noticed. Time for tinkering!

(1) Anybody else cobbled together a 6AV5 SE? What plate/bias/dissipation are you using? Any downside (e.g. tube life) to 25-26W dissipation that George recommends? I should mention that my 6AV5s are all the nice Sylvanias, with NO hints of visible incandescence at 23 W (after midnite with ALL the lights out).

(2) Anybody tried 6AV5s with "better" OPTs?? The angel on my right shoulder says to leave well enough alone; there's MAGIC in them cheap Edcors. The devil on the other shoulder says "imagine what she'll sound like with GOOD IRON!!!"

(3) Hybrid Rectifier
The last picture shows a couple of diodes slapped on the rectifer terminals, which I THINK are leading to the choke. Gotta shut-off the amp long enough to double-checek that. I think that's called a "Hybrid" rectifier set-up, but for the life of me I can't remember what the heck I was thinking. What did I do that for, anybody have any idea? What happens if I by-pass the diodes? Can I get the voltages up (and then tune with the first cap)?
I have a mental picture of evaporating some P/S caps if I touch them diodes.

(4) I wonder what this puppy would sound like with a 417A front end, let alone the Tubelab SE constant current grid thingy. No, no, bad devil.

tubelab.com 16th January 2012 11:34 PM

Quote:

The angel on my right shoulder says to leave well enough alone; there's MAGIC in them cheap Edcors. The devil on the other shoulder says "imagine what she'll sound like with GOOD IRON!!!"
I haven't had much time to do any tube tinkering lately, but I have one observation from my past. If you found "magic" in this amp don't screw with it. Dig deeper into your stash and find, buy, or steal the parts to make another one. Then experiment with the second one. You will always have the magic amp for a direct "better or worse" comparison and you don't have to worry about messing up success.

I have a 300B push pull amp that I built about 8 years ago from total junk quality parts except for the tubes and even they were 4 for $100 used Sovteks. I have been experimenting off and on for years trying to duplicate that magic and I haven't found it yet. The original amp is now dead, but I refuse to scrap it or even steal parts from it......I just got to figure out where the sound comes from.

hareynolds 17th January 2012 03:37 AM

1 Attachment(s)
Good idea, George. There's definitely some voodoo in this one. Gotta leave it alone until I can figure out what it is.

For example, I went back to look at the "hybrid" rectifier; turns out I had tacked a couple of high amp diodes to the tube socket, and run a jumper over to the first cap; the wire from the 5V fils on the tube rectifier was CUT. No wonder switching tubes didn't change the Plate voltage *snort*.
I guess I was looking for higher B+ voltage, but generally I NEVER use S/S rectifiers if I can help it. Cue Twilight Zone theme.

I SUSPECT the following voodoo is in play:
(1) "synergy" between over-loaded 6AV5 and the little Edcors. Wish I knew what that could be; suspect it might be "psycho-acoustic", e.g. a pleasant distribution of distortions. Don't have an FFT set-up yet, or I would have taken a look-see. [does anybody have one of those cheap new 'Scopes with the built-in FFT? That would be nice.]

(2) a FAST, simple p/s; between the diodes and a large, low DCR choke (about twice as heavy as the recommended Simple SE choke, but lower DCR) seems to work great. Very open sound, snappy transients. Again, I normally filter the B+ within an inch of it's life, and consequently regularly TOAST PSTs. This one is running pretty cool, even drawing >150mA.
Haven't looked at the ripple, but I suspect it isn't tiny,

(3) A Totally Constant Current 6SL7 (with CCS plate and Battery Bias).
IIRC it's a std set-up and the settings are conventional, but there might be something special there; if there is, it's quite accidental.
BTW I'm using NiMH AA cells, NOT the lithium cells in the photo; the LI cells buzzed and crackled too much.

Now all I need is a regulated filament supply (I have a couple of Pete Millett's boards) and change the battery to a Zener, and the 6SL7 would be completely surrounded by SAND.

[I've been wondering: what's on your Test Music List, George? If it isn't too esoteric, it might be good to be playing selections from the same list.]

I've misplaced my spare SE boards so I'll have to order more (plus a TubeLab SE while I'm at it), but in the meantime I can convert the 6V6 / 6BQ6 LabRat into a 6L6 / 6AV5 amp with a PST & OPT swap. The 6AV5 sockets (diiferent pin-out) will have to go outboard (paralleled or with an octal plug). Plus I'm guessing I'll want the bias resistors OFF the board due to the heat.

BTW here's the 6V6 / 6BQ6 LabRat with the 6BQ6's in it:
IIRC the OPTs are 8K Edcors. It's OK with 6V6s, but with the 6BQ6s, meh. Definitely nothing special, except that the tubes are cheap.
I like the 1626 ("Darling") amps better than either the 6V6 or 6BQ6s, regardless of which version of Darling (DC, cap-coupled, single or paralleled).

wrenchone 17th January 2012 03:48 AM

Interesting thread. I just grabbed a bunch of clean used GE 6AV5s not too long ago. I was thinking of using them triode strapped.

alexg 17th January 2012 04:09 AM

hareynolds,

What is the primary impedance of the Opt you used on the 6AV5?

Just like you, I did some work on the 6AV5 and I was not very impressed, buy unlike you, I did not experiment.

This thread will inspire me to go back to the 6av5!

hareynolds 17th January 2012 06:04 PM

Quote:

Originally Posted by alexg (Post 2865589)
hareynolds,

What is the primary impedance of the Opt you used on the 6AV5?
Just like you, I did some work on the 6AV5 and I was not very impressed, buy unlike you, I did not experiment.
This thread will inspire me to go back to the 6av5!

5K. they're the cheap open-frame Edcors. I bought a bunch, mostly 5K but also 8K (for 6V6 etc) and 3.5K (don't know why; EL34? 2A3? DC Darlings? Dunno). Very highly recommended, ESPECIALLY for advanced tinkering.
The bass is typically a little weak, but still excellent for its mass (at least, that is, until I heard the over-powered 6AV5s, whose bass sound like a Dynaco ST-35 PP EL84, no joke) but the clarity above 100-200hz is always astounding for the price.

Before you do anything with 6AV5s, be sure to read George's experiments:
6AV5GA testing
All 6AV5s are NOT created equal, although MOST seem like they will survive 23-25W.

My prototype is Triode strapped with 1500uF cathode feedback; don't yet have switches on these two features, but this combo seems to sound best with the junkyard Rat Shacks, at least so far.

I'm thinking that I'll scrounge-up a PST and do the socket and bias component mods to the LabRat Simple SE board this weekend. I'll revert when I've got something. My principal interst is seeing what difference the Octal 6SL7 makes (and the battery bias, for that matter). I'll start with the 12AT7, but perhaps ought to try say a 5751 as a quasi 6SL7 in a noval tube. Any thoughts out there?

PS I have a couple of Heathkit power supplies, too weak (100mA) to power two 6AV5s, but with good bias supplies. Fixed bias would allow me to adjust the bias continuously, without stopping to tack-solder bias resistors all the time. HOWEVER, I don't know how to "convert" a cathode bias to a fixed bias. If George could offer a 10-line tutorial, I'll be off and running; otherwise, I'll have to dive into Morgan Jones' Valve Amps again.

hareynolds 18th January 2012 05:21 AM

Fixed Bias for 6AV5
 
I'm trying to figure-out how to apply fixed bias to the 6AV5s so that I can vary the biasing continuously (as George did in his tests)

I'm guessing it isn't as simple as it looks, since:

(1) I've got >300V across the tube, and, according to George, some 6AV5s start to runaway at 295V, others are OK up to about 350.
and
(2) the 6AV5GA datasheet has a little note at the bottom: 3. In stages operating with grid leak [fixed]bias, an adequate cathode bias resistor or other suitable means [I assume this means a FUSE] is required to protect the tube in the absence of excitation.

YIKES.
George recommends a 100-150mA fuse in the cathode leg to prevent the OPTs from turning into little Friolators. Elsewhere I've seen it suggested to use a low value (say 10 ohm) resistors suggested as quasi-fuses. I suppose that a 10 ohm cathode resistor would also be a swell place to measure the plate current.

Further, runaway can apparently be INCITED by a Too-High Grid Resistor value (makes sense; it starts to look like an oscillator at some point, excited by the grid voltage).

The 6AV5 data sheet is no help, but there are several rules-of-thumb and design considerations suggested here: http://www.diyaudio.com/forums/tubes...-question.html

I'll revert when I've digested most of these, but in the meantime, for what it's worth, the 2A3 spec sheet lists the MAX grid resistor for Fixed Bias as 0.05 Megohm = 50K ohms, which looks like a reasonable starting point.

How about:
(a) 47K grid resistor, 10 ohm cathode resistor
(b) make sure the bias supply is cap-coupled to ground ust before the grid resistor
(c) start with bias voltage of about -65-70 Volts, then slowly RAISE it (make it less negative)
(d) measure plate current across the 10 ohm cathode resistor

Assuming 150 mA plate current "protection", the voltage drop across a 10 ohm resistor would be 0.150 x 10 = 1.5 Volts. 1.5V x 0.150A = 0.225W
Use a 1/4W cathode resistor? Or is that too much, and I should use 1/8 W or use a fuse?

"Runaway Tubes". Fun Stuff.

wrenchone 18th January 2012 05:27 AM

Well it means that you'd better not take away the negative bias while you have full plate voltage, that's all. Given the current capability of the average sweep tube and the fact that these are depletion mode devices after all, it's sound elementary advice. Either you put the bias in series with the cathode, or you supply it at the grid - either way, you need it.

Fuling 18th January 2012 11:16 PM

6AV5 + Edcor... been there, done that, will probably be there again soon.

Years ago I made and amp with 6AV5GA output tubes, a single 6SN7 (shared between channels) input tube (CCS loaded) and Edcor 5k/8R 15W XSE OPTs. I *think* the output stage was UL connected with cathode feedback but it might have been triode wired, I donīt know. Being a little paranoid I ran the tubes quite cold, maybe 17W or so on the plates. The PSU was mosfet regulated and based on a power transformer from a butchered Philips EL6400 public adress amp.
The whole thing sounded good enough but in the end I didnīt use it much so it was traded away for some stuff that I thought I needed for another project... Stupid thing to do, in retrospect.

Just a few months ago I bought a stash of Edcor OPTs for a multi-channel amp project related to my 4 way horn system. Four 15W 4,5k/8 for the mids and highs plus two 25W 3,5k/8R for the midbass horns (still cheap open frame stuff, not the slightly less cheap ones with end bells). After ordering the transformers I just happened to make a good deal on a pair of midbass drivers that proved to be much better than the ones I had, only problem is that the new ones are 16 ohms and thus my 8R OPTs are no longer suitable for the job.

Well, having a pair of 3,5k 25W OPTīs and 3-4 pairs of NOS Sylvania 6AV5GA keeps me up at night (which is a good thing as I work the graveyard shift...) I have this wild idea about building an semi-permanent experimental platform employing all the parts needed to listen to practically any form of SE tube amp connection possible. Triode, UL, pentode, zero NFB, adjustable NFB, cathode feedback, E-linear feedback, Schade feedback, triode driver, pentode driver, resistor vs CCS loaded driver, anything.

hareynolds 19th January 2012 03:38 AM

Quote:

Originally Posted by Fuling (Post 2868459)
6AV5 + Edcor... been there, done that, will probably be there again soon.

Years ago I made and amp with 6AV5GA output tubes, a single 6SN7 (shared between channels) input tube (CCS loaded) and Edcor 5k/8R 15W XSE OPTs. I *think* the output stage was UL connected with cathode feedback but it might have been triode wired, I donīt know. Being a little paranoid I ran the tubes quite cold, maybe 17W or so on the plates. The PSU was mosfet regulated and based on a power transformer from a butchered Philips EL6400 public address amp.
<snip>

I did almost exactly the same thing, except for the selling part; fortunately, mine wasn't sales-worthy, so I parked it on a shelf in the shop. Mine was VERY similar, including the 17W dissipation, except I used a CCS-loaded 6SL7 (for the higher stage gain) rather than the 6SN7.

When I looked at it again last week, I finally decided to see what it sounded like at "proper" (that is to say CRAZY) loadings.

If and when you (re)build a 6AV5 SE, strongly suggest aiming for an operating point pretty close to what George recommends as the "sweet spot": ~315Vp and 80 mA (~25W dissipation). What I found (your results may vary) is that things start getting good when you approach about 20W dissipation.

That being said, the transconductance of individual 6AV5s appears to vary pretty widely at these high power levels (I guess all screens are not created equal), so I suggest (a) setting a lower operating point (say 17W, like you had before) (b) finding a pair of 6AV5s that draw roughly the same current, (c) THEN start dropping the cathode resistor values.

Curiously (considering that the 6AV5 was rebadged as a 6B4GA), the cathode resistors end-up VERY similar to those used on a 2A3 SE amp (I started with ~950 ohms and ended-up with ~750 ohms). Admittedly, the 6AV5 B+ is a lot higher than the typical 2A3 (woohoo!) but the sound (at least triode-strapped) is remarkably similar to a DHT (only with BASS). To my ears, it sound about half-way between a 45 and a 2A3; sort of a 1-1/2 plate DHT (*snort*, LOL)

Try it, you'll like it.


All times are GMT. The time now is 04:15 AM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright Đ1999-2014 diyAudio


Content Relevant URLs by vBSEO 3.3.2