So I scored some free tube stuff...

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

Curtis

Well-known member
Joined
Aug 24, 2006
Messages
305
Location
Australia
...And I'm thinking it'd be fun to convert it into something guitar/bass amp-like.

The beast in question is this:

http://i153.photobucket.com/albums/s222/ac1176/Valvethingy1.jpg

http://i153.photobucket.com/albums/s222/ac1176/ValveThingy2.jpg

http://i153.photobucket.com/albums/s222/ac1176/Valvethingy3.jpg

It was a power amp used to test turbine governors at power stations, but now it's been saved from certain death during one of our cleanouts at work. When I first saw the size of those transformers I was thinking perhaps Ampeg SVT clone, but I think the output transformer is probably only rated for 100W or so. This thing is fecking heavy!

Without pulling it apart, I've established the following:

* 4x 6J5 triodes in the preamp, 2x 807's in the power amp

* Output transformer is unbranded, but judging by the appearance and from things I've seen previously, I'd say it's probably an old Australian-made A&R tranny. No markings on it except for "OPM21A" and "12-70". Haven't measured the ratios yet. From the labelling on the side the output terminals are 4 equal-ratio windings capable of delivering 50V @ 0.4A each. Windings can be series or parallel-connected.

* Power tranny (the massive black box) also has no brandname, badge has fallen off. Looks epoxy-encased or something - haven't seen that before. Only labelling is "PT6501". Secondaries are 500-0-500, 250-0-250, 0-10, and 3x 6.3V @ 1.5A, plus electrostatic screen connection.

* Measured the HT voltage and got 650VDC.

So, any suggestions as to what to turn it into? I'm probably not interested in keeping the 807's and I'd much rather use something a little more easy-to-get (hint - I've got loads of 7581A's that I could use :wink: )
 
> 4x 6J5 triodes in the preamp

What is the 5th small tube?

> I'm probably not interested in keeping the 807's and I'd much rather use something a little more easy-to-get

The 807 is an EXCELLENT audio tube, quite E-Z to get. It and the 6L6 are near kin. The 807 went for high plate voltage and low radio losses; 6L6 went for cheap audio. Later 6L6GC (or 7581) allows high plate dissipation and screen voltage; 807 is not drop-in for late 6L6GC amplifiers.

I strongly suspect it IS an "audio amplifier" already. Like my Ampeg servo motor amps. However the Ampegs did not like to go above 2KHz (they pulled like oxen to below 20Hz).

http://www.mif.pg.gda.pl/homepages/frank/sheets/084/8/807.pdf

Page 2 and 3 show class AB1 audio operation. Pair of tubes with 650V plate and 300V screen is near 60 Watts. No SVT.

I think the plate-cap on 807 is cool. And while a well optimized 6550 amp can do 100 Watts, the ~~60 Watts possible with 807 is just as loud for most practical purposes.

> Measured the HT voltage and got 650VDC

Under load?

6L6GC is rated 450V max; 7581 500V. While we know many 6L6GC serve OK at 500V and above, when you are looking at over 600V you want 6550/KT88. (EL34 and 7027 also have high ratings, but IMHO it is silly to run them that hard in audio applications.) (And 8417, but leave that to collectors.)

Your 4*50V,0.4A seems to imply 80 Watts. It may not be a sine rating.

There is no combination of windings to put you at a common loudspeaker impedance. All in parallel does give 32 ohms which could be 4*8R or 2*16R (or a wall of 4R in series).

Replacement OT is an option, but any likely 2-bottle power at 650V implies a 10K winding, and there are few of these above the 20W level. You could however, for guitar, use a 50W 5K or 6K6 OT, mis-label the outputs to give 10K-13K effective load.

You could run the AB2 conditions on page 4; however then you must add SVT-type grid drivers, which is too much work for a 2-bottle amplifier. (Makes more sense on the SVT's 6 big bottles.)

While it is a great base and cage, I think it will be a poor touring amp. That power supply is clearly over-over-specified, the way we used to do it when expediency and utter reliability was more important than cost or weight. You are testing a $250,000 generator, you don't strain to decide between a $100 or a $200 power tranny, you get the big one just to be sure. Especially in a shop where 6,000 pound generators and 600 pound rotors are moved.... you have a dolly for a 60 pound amp, and it never leaves the house.
 
What is the 5th small tube?


Sorry, missed that one. 6J7 pentode.


> Measured the HT voltage and got 650VDC

Under load?

That was with the standby switch in the "Run" position, all four outputs parallelled up, a 25W 4.7R resistor hanging off the output , and no input signal applied. I have yet to try running some signal through it and hearing what it does.


Your 4*50V,0.4A seems to imply 80 Watts. It may not be a sine rating.

I'd like to think the rating is for clean sinewave drive. Considering the amp was intended to reproduce high-ish power 50Hz AC for testing purposes it'd make sense.

The amp definitely looks DIY'ed, or at least heavily modded Popping the top reveals a lot of tagboard-mounted components of varying vintage/manufacture/size, even an extra 110VDC power supply on a piece of breadboard with it's own transformer. Considering the period of when this thing would have been in service by my employer it's quite possible it was put together by some in-house testing engineers. Sadly those days are long gone.

I might try and trace out the guts and get some kind of schematic up....
 
OK, I'll accept 80 Watts of "sine-like" wave. The 807 datasheets are conservative; testing dynamo toys a 10% THD would not be distressing. And there could be a grid transformer inside, but 6J5 is an unlikely driver (think 6F6).

> possible it was put together by some in-house testing engineers.

Good bet. But...

The filter caps are circa 1972 +/-7 years. However after like 1975, they probably wudda bought a Crown DC-300 and a 32V:line power transformer, rather than build with tubes. Some of the connectors and the knob could be 1960s-1970s. But everything else could be 1950s. And very conservative even for the late 1950s; octal went to mini very fast around 1950.

I'm wondering.... could this be a personal High-Fidelity amplifier, DIY-ed from spare parts in the days of Mono, put aside for Stereo, then dragged out of the home closet to work and re-capped with shop funds when a need arose for a utility AC supply?

Still the four 50V/0.4A= 125 ohm windings does not seem to suit Hi-Fi, but the OT was surely bought the same time/place as the big choke(?).

Now wait. Where is the rectifier? When those filter caps were on the market, green candy silicon rectifiers were king; but any older has to be 5U4, but no place to stick it. Unless it sat where the "big choke" is, and this iron and the OT are modifications for utility amp duty.

IAC: four 6J5 is two 6SN7/12AU7 which is a Williamson, and a low-NFB Willie has a lot of gain. One 6J7 (EXcellent tube) would be all the preamp you need for instrument pickups. If that is the case, your main obstacle is finding a 125/4= 32 ohm speaker array you like.
 
Had a chance to trace the circuit out. I've got a scratchy hand-drawn schematic, but I'll redraw it before I post it.

Input signal comes in straight to the input "volume" control (50K, 10-turn Bourns precision pot), then straight into the grid of the 6J7 pentode, 1M grid resistor. 6J7 is wired up in triode mode, 100K plate resistor.

Next is a pair of 6J5 triodes set up as a phase splitter, pretty similar setup to what Radiotron Designers Handbook 4 calls the "See-saw phase inverter" (page 526).

Output of each 6J5 feeds another pair of 6J5's as cathode followers, then feeding the grids of the 807's. I guess that indicates class AB2 drive, yes? Negative bias voltage is -110VDC via a 1K2 5W resistor to the grids of the 807's.

Negative feedback is taken from one of the four output windings of the OPT (47K and 0.47u in series) back to the cathode of the phase splitter.

Fed a 10V 1KHz sinewave into the primary of the transformer (plate-to-plate), measured 0.83V on one of the output windings (12:1 ratio). A bit rusty on my transformer theory but I think that's an impedance ratio of 144:1, which would make it OK as an 2K3 to 4 ohm transformer if I run all four outputs in parallel??? (4x4x144)


I'm wondering.... could this be a personal High-Fidelity amplifier, DIY-ed from spare parts in the days of Mono, put aside for Stereo, then dragged out of the home closet to work and re-capped with shop funds when a need arose for a utility AC supply?


It does look very "bitty". Some of it looks quite well put together - neatly loomed wiring, well-laid out tag boards - while other parts look very slap-dash. The bias supply for example definitely looks like an afterthought - a separate circa-30VA transformer feeding a bridge rectifier and cap sitting on a piece of perf board. It could very well be someone's self-made amp that's been adapted to the application.


Now wait. Where is the rectifier? When those filter caps were on the market, green candy silicon rectifiers were king; but any older has to be 5U4, but no place to stick it. Unless it sat where the "big choke" is, and this iron and the OT are modifications for utility amp duty
.


Power supply rectifiers are "SD410" diodes. Black cylindrical package, pretty similar to the usual 1N400x, but maybe 2x as long. Can't find any data on those. The choke looks like it's always been there. The base isn't covering up any holes that would indicate a previous valve rectifier. In general the chassis looks like it's been drilled and punched for all the components that are currently in it, nothing looks missing or "hidden".

Schematic to follow...
 
Schematic for those who are interested:

http://i153.photobucket.com/albums/s222/ac1176/807amp.jpg

The output transformer appears to have ultra-linear taps, but they're not used in this design.

Voltages measured are with the standby switch on, a 4.7ohm resistor hanging off the outputs, and all four outputs paralleled up.
 
> Schematic

Beautiful rendering.

> a pair of 6J5 triodes set up as a phase splitter

Yes, though it may be missing a 1Meg from the lower grid to ground?

The 0.27u+470K gives 1.25Hz bass cut. The two R-Cs at the 6J7 give 32Hz and 16Hz. (The 32 is mighty close to 50Hz power freq., yet must be deliberate.) The 0.47u+47K in the NFB loop suggest a bass rise at 7Hz. I have the feeling the output stage has been aimed at deep-deep bass and low phase shift, but the input stage has been designed as a filter. Maybe the output began as hi-hi-fi, but in test duty it was convenient to filter-off signal generator subsonic bobble.

NFB: looks like for 50V output we need 45V pk 30V rms at 807 grid, 6J6 gain is near 15 so 2V at the input to the power stage. Overall gain near 50V/2V= 25. NFB loop forces gain to 11. There is "some NFB" at full load (very like many old Fenders). The main point of the NFB may be to limit gain when load falls off: at 10X nominal load a pentode's gain may be 5 times higher, or 125. The NFB still limits this to 11. So the output sags only a dB or two from no-load to full-load.

> AB2 drive, yes?

Yes indeed. Looks like 6J5 abuse, but not to excess. With 300V across them, 6J5 can pass well over 40mA. The book suggested AB2 conditions suggest about +13V peak 807 G1 voltage. 807 G1 curves for Vg1=+13V show Ig1=20mA at Vp=100V, so it can do it. (As Vp falls below 100V at high current, Ig1 goes through the roof.)

For low distortion, driver impedance should be "less than" 807 grid impedance. If 807 G1 pulls 20mA at +13V it is roughly 650 ohms. Also infinite on the negative swing, 1K at +5V, 500 at +30V. We want driver impedance less than 650 ohms so the gain does not go away on positive peaks. At 250V 9mA the 6J5 Gm is 2,600uMho or 385 ohms, which is "less" than 650. We don't want a zero impedance driver; a little impedance cancels the rise of 807 Gm at high current.

It's a fine plan. Very near the manual's 600V 80W suggestion.

> Negative bias voltage is -110VDC via a 1K2 5W resistor to the grids of the 807's.

No. The -110V, 1K2, 22K, 22K, Gnd gives -100V and -50V taps. 6J5 grids are biased to -50V. Their cathodes will want to sit 10V or 15V higher, near -38V. When you allow for the change of 807 G2 voltage from 300V suggested to 350V measured, this is spot-on. The -110V is pull-down, to make the 6J5 cathodes drop the 807 grids to cut-off.

In guitar/bass stage work, it is liable to distort like transistors. Tube stage amps use grid current against R-C coupling to de-bias in overdrive; this one won't. True, the 300W Ampeg and Fender tube amps also work like this one, and people use them.

> 2K3 to 4 ohm transformer if I run all four outputs in parallel???

There's a dropped bead somewhere....

> (4x4x144)

Ah. No. Running the windings in parallel gives you the same voltage, the same voltage ratio, the same current ratio, the same impedance ratio, as one winding. They split the windings to give 100V and 200V choices. Using one windings "works" but resistance losses might be 8%. Using all the windings (or putting the same pounds of copper in one winding) would give resistance loss more like 2%.

Or work from what we know. The (no-load) B+ is 650V, the 807 can pull-down near 100V. Peak plate swing is 550V. That is 390V rms one plate, 780V plate-to-plate. The rated output is 50V. The guesstimated voltage ratio is 780:50 or 15.5.

You measured 12:1. Working backward from the 50V rating, this is 600V rms plate-to-plate, 300V rms one plate, 424V peak one plate. Seems a bit small for a 650V supply, but there is easily 60V drop in the B+, 100V drop in the 807, and if we assume their wall-voltage may have been 10% below yours it is spot-on. (I know it is power-station gear, but maybe like shoemaker's children, power facilities don't get the best wiring.) Oh, and a dozen V lost in the 807's plate and cathode resistors.

807 data for high voltage and big power suggests 4600 to 7300 ohms plate to plate. Say 6K. 6000/144 is 42 ohms.

How to get that to 4/8/16 ohms?

I still think it would rock on 32 ohm loudspeaker load. The 42/32 mismatch is small, and speaker impedance runs high.

If your land uses 100V for large loudspeaker systems, you can find a 100W 100V:4/8/16 transformer, and use it on the 100V series/parallel connection. But that adds cost, weight, and losses.

We could say that 42/8= 5, so 5 pair of 807 would come out 8 ohms. (And 300 Watts!) On speech/music signals, this OT will handle it. But that PT, beefy as it is, won't support ten hard-working 807s. And the 5:1 mismatch from OT design load says losses 5 times higher, maybe over 10% ("just" 270 watts actual output).

A 48VCT 40VA winding could be used as an autotransformer to give 10.5 ohm nominal load. Suitable for 8 or 16.

Beautiful rendering.
 
Thanks for all that info, PRR. Definitely straightened a few things out for me. :thumb:

I've since found out a bit more info on that OPT. Turns out it wasn't made by A&R, it was made by Ferguson (another sunken Australian company).

This scan from an old catalog has a bit more info:

http://www.ozvalveamps.elands.com/generic/fergcat4.jpg

OPM21A is the one in question. They reckon it's good for 100W.
 
Aren't an 807 and a 6L6 essentially the same tube? I think, in fact, if you check the 807 datasheet you will find that the 6L6 actually has a somewhat higher rating, slightly higher max. plate at 450V and somewhat higher plate dissipation at 30W. Neither tube holds a candle to a 6550.
 
> OPM21A is the one in question. They reckon it's good for 100W.

In Fender-bass use. In guitar, considerable more, though you get into annoyingly high voltage.

But this confirms my guesses: "50V" minimum tap voltage is 32 to 42 ohms. (4500 is a bit low loading for 807 at 600Vp 300Vg2; book suggests 6900 for these voltages.)

In speaker use, "mis-match" is not critical, but 8 ohms is a LONG way from 31 ohms. There are 32R speakers, you can array 8 or 16 ohm coils to any impedance, but it is a lot more awkward than a "4/8/16" winding.

> Aren't an 807 and a 6L6 essentially the same tube?

Yes and no. Original 6L6 is for audio, 807 is for RF. In audio we like lower impedances, which means lower voltage, and leakage is not an issue. At RF we can tune-up to high impedance, favoring high voltage for efficiency, which forces the Plate connection out of the leaky (especially at RF) base to the top of the glass. There are several other changes related to 6L6 being less-professional than 807.

Original metal 6L6 circa 1938 was rated 21W to 19W, 400V to 360V plate (ratings dropped after introduction). It was re-issued in glass at the same ratings. The 360V limit has a lot to do with the base insulation and long-term breakdown.

807 has always been CCS/ICAS 25W/30W, 600/750Vp, 300Vg2. Stage amp duty would be between CCS and ICAS duty.

6L6 will handle MUCH more for short periods. But when abuse became common, a market arose for a better 6L6. Tung-Sol introduced the 5881, drop-in for 6L6, rated just a bit higher than 6L6, but lasted a lot longer in 6L6 sockets. There are many other fancy 6L6 models.

But "today" we don't use 6L6, we use 6L6GC. This is a TV tube, derived from 6L6 but using newer techniques and materials, and then re-packaged as a 6L6 replacement. 6L6GC is rated 30W, 450V. The "450V" is marketing; the identical guts are rated 500V as a TV tube, but they wanted audio boys to buy the 6CA7 7027 6550 bottles at higher prices if they were using higher voltage.

(My observation is that 6L6m is rock-solid at 19W, and 6550 is solid at 40W, but a 6L6GC worked -at- 30W will often soften and die. Fortunately p-p amps rarely at at max dissipation.)

> Neither tube holds a candle to a 6550.

807 worked at 600Vp 300Vg2 AB2 comes close to 6550 worked at 600Vp 300Vg2 AB1. Yeah, because of the low G2 limit and smaller cathode you have to beat the 807 G1 to get the power through. The KT88/6550 series is another spillover from the TV tube racket, not 1938 technology.

I think it would be bad taste, but you could re-bottle with four 6550 and pull 160 Watts into 16 ohms. Not that 160W is louder than 80W-100W, but the 16 ohms is a more likely speaker impedance. This increases strain (both B+ and heater) on the power transformer but it looks capable.
 
Back in the day in the USA, an amp for this kind of service probably would have been filled by one of the large Altecs running 8005 outputs, or a MacIntosh MI-200. Stomberg-Carlson also made a series of big amps used for shaker tables, etc. Several other large amps were made by Western Electric and other companies for drive-in theater use and provided well over a thousand watts when the common mono hi-fi amp avaraged around 12 to 25 watts. Most of those big amps didn't really have that good of specs, although the Mac MI-200 did and was often used to drive cutter heads for disk recording and where high power PA was needed. Also used as variable speed motor drive amps.
 
Back in the day in the USA, an amp for this kind of service probably would have been filled by one of the large Altecs running 8005 outputs, or a MacIntosh MI-200. Stomberg-Carlson also made a series of big amps used for shaker tables, etc. Several other large amps were made by Western Electric and other companies for drive-in theater use and provided well over a thousand watts when the common mono hi-fi amp avaraged around 12 to 25 watts. Most of those big amps didn't really have that good of specs, although the Mac MI-200 did and was often used to drive cutter heads for disk recording and where high power PA was needed. Also used as variable speed motor drive amps.
The SVT amps originally came is two flavors; one with the 6146 and later with the 6550. The 6146 was a transmitting tube with a top cap, but about the same bottle size as the 6550. A large part of the SVT sound is the preamp, and if you can duplicate the "tone" of that section, almost any power amp takes on the Ampeg sound.
 
[quote author="Jim Zuehsow"]A large part of the SVT sound is the preamp, and if you can duplicate the "tone" of that section, almost any power amp takes on the Ampeg sound.[/quote]
Hmm, both interesting & puzzling to hear that as well.

My own rig is a Ampeg SVT-IIp preamp with two SWR 4*10" cabs each driven by a 400W solid-state amp.

Comparing that with an SVT setup (IIRC the rackmount SVT-II and an Ampeg 8*10" cab) sounded quite different imho.... while not necessarily much louder, the all-Ampeg setup had more balls, pushed more and was less controllable (dare I say relatively speaking a one trick pony ?)... all at the same time.

Should have swapped cabs & relistened though, that might have been a certain influence as well.

Bye,

Peter
 
[quote author="clintrubber"]My own rig is a Ampeg SVT-IIp preamp with two SWR 4*10" cabs each driven by a 400W solid-state amp.

Comparing that with an SVT setup (IIRC the rackmount SVT-II and an Ampeg 8*10" cab) sounded quite different imho.... while not necessarily much louder, the all-Ampeg setup had more balls, pushed more and was less controllable (dare I say relatively speaking a one trick pony ?)... all at the same time.

Should have swapped cabs & relistened though, that might have been a certain influence as well.

Bye,

Peter[/quote]
I have an original SVT along with the 8X10 cabinet. Don't use it much anymore and it sits in a corner of the studio. When I do gig, I carry around a SVT III head and a Eden 4X10 cabinet. Much easier on this old back of mine. Anyway, I have taken the preamp of both the SVT and SVT III and tun them into a Crown K2 amp, and they sounded identical. Part of the SVT sound is the cabinet which has the speaker pairs chambered from each other. That gives it lots of mid-bass projection, but not the really ballsy sound needed for a 5 string. Plus, the SVT 8X10 isn't ported, and that makes a LOT of difference in the low end, and possibly causes that "one trick pony" sound you mention. An SVT head on a really great sounding cabinet is a joy to play and a broad palette of tone is available. A friend of mine recently put new speakers in his Ampeg 6X10 cabinet and went with a high excursion speaker. Frankly it blows the 8X10 cabinet away on the bottom end. I think he cut out the inner baffles in the process and stuck a pair of ports in the back. It rocks!
 
> The 6146 was ... ... about the same bottle size as the 6550.

6146 is 25 Watt, 1.56" diameter, 3.125" off the deck to the cap, maybe 2.75" to the top of the glass.

6550 is 1.7"-2" diameter and 4"-4.2" deck to top of bottle. The A-mod carried a 42W max Pd, but older and current ones say 35W.

First SVT used 6146, and worked; common opinion is that the revision with 6550 is more reliable. That the 6146 has huge peak voltage and current, but the Pd is just enough on nominal load, leaves little reserve for misloading.

> I have an original SVT along with the 8X10 cabinet.

Original SVT used two 8-10" cabs to get enough power handling. 1980 was the single 8-10" cab rated 350W bass.

> SVT-IIp preamp

Anything "II" is an imitation of the real SVT. Anything made after the 1970s, when Ampeg crumpled toward bankruptcy, is dubious. Any very old amp is dubious. Smoke a lot of hash, time-port yourself back to 1972, and play a Real SVT.

mansawad.gif
 
I see this old amp and imagine it plugged into
4 ported cabs with 18 inch speakers and my 6 string bass.. YIKES!

Four 807's and bass yes to that.

I had an original SVT and it cranked, the new ones are not the same.

I gigged a few times with 2 complete SVT originals, 4 cabs.
vs 2 double Marshal Majors, and a HUGE Hammond B3 rig,
The drummer is lucky to be alive from those gigs.
He broke dozens of sticks, and nearly coughed up a lung.

Never want to go back to those days....
 
Back
Top