need a smallish tube for high plate voltages, fixed bias

[imaradiostar]

Wow, it's been years since I've been on Ampage but I've got the itch to build again! Lucky for me I have a stash of tons of used transformers and stuff from things I've taken apart.

Some time ago I took apart an old PA or hi-fi amp that had (I think) a pair of 6973's and a six volt tube rectifier. I'm pretty sure it was cathode biased but I don't recall. This little transformer set is pretty nice but has really high plate voltages- higher than I'm comfortable using with an EL84 or 6V6. It comes out over 800vct if I recall correctly. It's easily over 500 volts dc once filtered and loaded a little bit. Not surprisingly one of the output tubes was shot when I got the amp.

I don't think I wanna use a standard 6L6 or el34 because I'm afraid I'm going to want to upgrade to larger transformers to see how the amp could really sound. I have a pair of output transformers and I can't remember which was from the original amp. Either way one is in the 6k:8 ohm range (from a Bogen PA?) and the other (depending on what you use the taps for) is either 7k:8 or 3.5k:8. Both are sized appropriately for a 30ish watt am. Their measurements are nearly identical to Fender Vibroverb size outputs so I figure they should be pretty safe as long as I'm not overdoing it on the output section distortion. I do want it to distort in a beautiful way though since I have a THD hotplate and I tend to use it!

So...given the above parts, what would you build? I was thinking about finding some 6973's and building a sort of a Trainwreck style circuit.

Do el34's work well at higher voltages using a higher plate impedance? Would this just be a way to cook the output tf?

Oh yeah, thought I should mention. I don't want to use a zener to kill plate voltage. To me it kinda defeats the purpose of this exercise. I'd like to be able to make it switchable (as I have in the past) between tube and SS rectifiers so it'll need to take some serious plate voltage. I usually just use a DPDT and switch between the to rects with one side and output bias settings with the other.

Wanted to add- I just looked at a kt66 data sheet- cathode biased with an ez81 rect to lower plate voltage would do about 30 watts so there's a possibility.

thanks, looking forward to some cool ideas.

jamie

[loudthud]

7868's or 7591's would be good but NOS are hard to come by ($$$) and current production 7591's don't get good reviews. EL34's would work well if you can get the screen voltage down to 350V or so to throttle them back. Maybe just using higher than normal screen resistors would do the job.

[Chuck H]

I found a few. But all NOS. Some rare, others I found were affordable...For now.

6fh6 6dq6 6dq6b 6gw6 6fw5 6gc6 6bg6ga

All use octal sockets and 6.3v heaters. All can take the voltage but have much lower G2 voltage and moderate G2 wattage (= big screen resistors). All dissapate 20watts or less. Some use a cap plate supply.

Good luck

Chuck

[Steve Conner]

How about 6146s? They have a cap plate connection and can take plenty of abuse, but their screen voltage rating is low. I've also seen a cap plate variant of the EL84, the EL81, designed as a sweep tube for small TVs:

http://www.r-type.org/exhib/aad0121.htm

You probably won't find any tube that can take the high screen voltage you want, short of a KT88. You might well need some kind of dropper for the screen supply, maybe a couple of those gas regulator tubes.

The other problem is that small tubes running off a high plate voltage end up biased cold to stop them melting. This can make the amp sound a little thin and scratchy at lower volumes.

Having said that, I once worked on a Sansui Q-55 that ran two EL34s off a 800V CT transformer with only a 1K shared screen resistor. It idled at about 550V on the plates and 530 on the screens :eek: It worked, but I was converting it to a guitar amp, and wasn't confident that it would hold up under heavy overdrive. I lowered the screen voltage to 360V with a Rube Goldberg arrangement of MOSFETs, and still got 50W out.

[imaradiostar]

Thanks for all the replies. I've done some looking around on this forum and the internet and I'm leaning toward the new EH 7591. From the looks of it they're fairly durable and with careful design and biasing they should be able to take the voltage. I may use the old transformer trick to lower the plate voltage- small transformer out of phase with the mains trans to lower the voltage by x amount. I'm also toying with using a choke input filter to lower B+ a bit. If some of the stuff I've read is true they should hold up just fine as long as I don't bias them too hot. I figure it's worth a shot and I'm really curious to see how they sound.

I think I'm going to end up using these parts to build an amp for my little brother. I don't feel very good about him having the plate lead at the top of the tube. I want any generic amp tech to be able to open this amp up and not have trouble in case it needs to be worked on (I live in Tennessee and he lives in Pennsylvania). I figure sticking with standard new production tubes might make this easier.

He's not far enough along to know or care what output distortion sounds like and I won't mind so much if he blows it up- it'll just be an excuse to build something better. I think I'm going to build him a plexi or a blackface fender circuit with more marshallish tone stack values...and maybe add reverb with a few mosfets to save PT current. I might even cheat and use a plexi circuit with just one channel and a mosfet cathode follower. Less inputs, less knobs, tough to make it sound like complete crap, and loud enough that he shouldn't need to turn it up enough to hurt things. Of course if he does and it blows up it might be a bit unsafe. I must think about this more.

Would killing screen voltage with an EL34 or 6L6 make for a more reliable lower wattage amp? Will it be more likely to hold up even when abused?

jamie

[Satamax]

Well, may be you could try some wafer base? They take extreme beating.

http://search-desc.ebay.com/6P3S-E_W0QQcatrefZC5QQfbdZ1QQfclZ3QQfposZ05220QQfromZR6QQfrpp Z50QQfsooZ1QQfsopZ1QQfssZ0QQftrtZ1QQftrvZ1QQftsZ2QQga102 44Z10425QQnojsprZyQQpfidZ0QQsaaffZafdefaultQQsacatZQ2d1Q QsacqyopZgeQQsacurZ0QQsadisZ200QQsalicZQ2d15QQsaslopZ1QQ sasltZ2QQsatitleZ6P3SQ2dEQQsofocusZbs


Or since your trany is center taped, you could do something like this

http://music-electronics-forum.com/attachment.php?attachmentid=1581&stc=1&d=1200505535

With the two bridges seried for high voltage on the plate and low on the screen as showed. Or paralleled for low voltage.

[Wakculloch]

The old M.O. valve co had an application report for KT77's, that if I remember rightly used a high voltage power transformer. Havenae got a copy of it but the web is a big place.
Loved the double bridge thingmy. Never thought of that one.

[Steve Conner]

I've never thought of it either, maybe because it doesn't work. (I'd have done it to my Q-55 if it did.)

For a start, a 400-0-400 transformer gives about 500-600VDC with the full-wave circuit, so it would try and produce 1.2kV with a bridge, with the tap at 600V.

Next, you only need one bridge rectifier to get that V/2 tap. This was covered a while ago, I forget in what thread exactly. Some Sound City amp does it, although it doesn't actually use the centre tap voltage for anything, IIRC.

Finally, if you paralleled the two bridge rectifiers in Max's circuit like he suggested, something awful would happen. Either half the diodes would be redundant and you'd still get 1200V out, or the transformer would be shorted and blow up.

[imaradiostar]

Lots of good ideas here and things to think about while I work on dinner. Please correct me if I'm wrong on any of my thoughts. I really like the smaller Russian tube- the 6p3s that's similar to a 6V6. It could probably take the voltage in stride.

I can put a bunch of voltage on the plates of any of a number of tubes. I'm going to have to cut back on screen voltage to prevent the tube from self destructing. The power and output transformers aren't huge so I don't want to abuse them by drawing too much current. If I have a highish output transformer impedance it'll keep transformer saturation to a minimum.

Does this leave me with a lifeless underbiased amp? Perhaps it would just be a gutless amp with too little bass?

How would a choke input power section sound on a guitar amp? This could get me into a much safer voltage range for normal tubes.

jamie

[Chuck H]

Look into the 6fw5

It can take 770 Vp (but only 220 G2), no plate cap, put out 18 watts max, they're cheap and you should be able to put at least some current through them at 525 Vp.

Not sure about the best way to derive a 1/2 voltage screen supply. I'd probably try to split the rail and use a pi filter with a big honkin resistor in it for the screens and run the preamp from the plate supply.

Chuck

[Satamax]

I've never thought of it either, maybe because it doesn't work. (I'd have done it to my Q-55 if it did.)

For a start, a 400-0-400 transformer gives about 500-600VDC with the full-wave circuit, so it would try and produce 1.2kV with a bridge, with the tap at 600V.

Next, you only need one bridge rectifier to get that V/2 tap. This was covered a while ago, I forget in what thread exactly. Some Sound City amp does it, although it doesn't actually use the centre tap voltage for anything, IIRC.

Finally, if you paralleled the two bridge rectifiers in Max's circuit like he suggested, something awful would happen. Either half the diodes would be redundant and you'd still get 1200V out, or the transformer would be shorted and blow up.

Hi everybody.

Steve, for the sound city, i's the 200 series, either MK3 or 4 use the dual winding, dual bridge recto. http://www.soundcitysite.com/sc_webpages/sc_200_plus.gif

Other than that, i haven't tried what's on my drawing, but i realy don't see how it can't work. I'll have to try.

And for paralleling the outputs of the bridges, Not possible? Or the diodes are redundant? How. You know i'm a dumbass. :D

[Steve Conner]

Well, look at the Sound City schematic you posted. It has two completely separate secondaries, one feeding each bridge. They probably did it that way because they wanted a lower voltage tap that wasn't V/2, or because they had a bunch of low voltage diodes in stock that they needed rid of.

With the secondaries joined at a centre tap, it doesn't work. Draw the circuit out and see for yourself.

The only thing that does work with a centre tapped winding is a single bridge rectifier, giving a tap at V/2: the same circuit as provides the split rails in a transistor stereo amp, but with the ground reference moved to the negative rail.

[Wakculloch]

Today found the application report and the secondaries are only 640-0-640(still unfeasably high), dual U19 rectifier, swinging choke, giving a HT of 550V, and an output of 60W.
800-0-800 may be a problem!

[imaradiostar]

I've tried the two bridges/one tapped transformer thing before- it's a good way to make smoke if I recall correctly. I was thinking a simple fix would be to use a different transformer for the screen and preamp supply. I have a bunch of smaller ones from an old hardware store and various ones from little tube tape decks I took apart years ago. They should be able to provide plenty of juice for the screens and leave the bigger TF to do the dirty work.

I was looking through my box of used tubes and found some interesting stuff. I actually have a pair of 7591a tubes that look to be in really good shape but we know that doesn't necessarily mean anything. Once again, probably purchased from an old hardware store. I also have some JAN 6v6's (grey bottle), some 7189's, two 7868's from a Bogen, some 6973's, an assortment of newish Sovtek 6l6's/5881s, some random EL34's and some Fender 6v6's from a Silverface Deluxe that I modded to use Tung Sol 5881's. (side note- I really miss that amp! Made a finger jointed pine cabinet and used 2 10" speakers and it sounded really good) I think I'll make up a rig and start testing and see if any of the above really strike a chord with me.

I had another thought too. I have quite a few 6SN7 A and B tubes, the ones rated for relatively high plate voltage. They're only rated for 7.5 watts plate dissipation for both units but I have a hunch that they'd put out quite a bit more than that before destroying themselves. The absolute max values for TV usage are pretty severe. Maybe I could use three triodes per leg of a push-pull output? If nothing else it might make a neat hi-fi amp for home or shop usage. I have no problem with them blowing up- it'd just be a neat experiment. I was thinking I could cathode bias them AB1 around 6-10ma idle current per triode. If I end up with B+ voltages around 500 volts this should just keep me inside the absolute max current and wattage values even though the plate voltage will be about 50 volts above spec. I could always add a little resistance in the power supply to drop the B+ just a little bit or use a tube rectifier. There should be enough current available to make it stereo!

I spent some time with Duncan's power supply simulator and measuring the TF. It's not quite as bad as I though. It's about 810 vct with no load and the DCR is around 210 ohms measured across the outside connections. I got under 500 volts DC using solid state rectifiers and drawing 100ma. That's really pretty usable- makes me think an EL34 may be the way to go.

If I were to use a readily available 150ma Hammond as a swinging choke I'd end up with really usable voltages for a class A el84 or 6v6 amp. The Hammond is only rated for 400 volts though- should I be concerned about this?

jamie

[Chuck H]

I said:

Look into the 6fw5

It can take 770 Vp (but only 220 G2), no plate cap, put out 18 watts max, they're cheap and you should be able to put at least some current through them at 525 Vp.

Not sure about the best way to derive a 1/2 voltage screen supply. I'd probably try to split the rail and use a pi filter with a big honkin resistor in it for the screens and run the preamp from the plate supply.

________________________________________________________ _____________________________________________

Seriously, Look into these tubes. They can be purchaced on Ebay for about $10 to $20 a pair. There doesn't seem to be shortage of them either.

With an 18 or 20k 25W resistor in a pi filter you would get about 230 volts for the screens. That should be fine with 550 Vp. Just locate the resistor away from the filters so it can dissapate heat safely.

IIRC the 7591 has a max VP of 525 and screens are 350. Thats good but they are expensive.

"makes me think an EL34 may be the way to go"

Didn't you mention size limitations to the OT. Unless you bias absurdly cool (Which defeats the purpose of EL34s IMHO) or, again, tank the screens, you'll stress your OT.

Chuck

[Bruce / Mission Amps]

I said:

Look into the 6fw5

It can take 770 Vp (but only 220 G2), no plate cap, put out 18 watts max, they're cheap and you should be able to put at least some current through them at 525 Vp.

Not sure about the best way to derive a 1/2 voltage screen supply. I'd probably try to split the rail and use a pi filter with a big honkin resistor in it for the screens and run the preamp from the plate supply.

________________________________________________________ _____________________________________________

Seriously, Look into these tubes. They can be purchaced on Ebay for about $10 to $20 a pair. There doesn't seem to be shortage of them either.

With an 18 or 20k 25W resistor in a pi filter you would get about 230 volts for the screens. That should be fine with 550 Vp. Just locate the resistor away from the filters so it can dissapate heat safely.

IIRC the 7591 has a max VP of 525 and screens are 350. Thats good but they are expensive.

"makes me think an EL34 may be the way to go"
...

Chuck
Right.
Any decent made EL34 can easily handle 550vdc B+ and lower screen voltage with ease when idling at around 15-17 watts each.

Here's something you can try though to get lower screen voltages and high plate voltages.
Run a separate pair of 1N4007 diodes off the PT secondary windings into a simple full wave rectifier and feed the DC from the cathodes into a 100ma to 150ma, 5Hy to 10Hy choke, followed by a healthy sized, 47uF@450v E-cap.
This will be your dedicated screen supply... which will draw current from the PT under load but the DC voltage derived will not exceed about 90% of the actual Hi-Vac when run into the choke input filter and it's filter capacitor.
In your case this will be around 360vdc.

Now I would not even bother with the SS vs tube rectifier switching but, to add to your questions and desires....
in parallel with that new screen node circuit run your Hi-V secondary leads over to a good DPDT switch, wired so that the center lugs can steer the H-Vac to either a rectifier tube or another set of SS 1N4007 diodes in FW with a capacitor input filter for higher B+ and the rest of the B+ rail, ... which of course will be the high current B+ rail for the power tube's plates and the rest of the preamp section, if desired to take it from there.
(You could take the B+ for the preamp stages from the new screen node if you wanted as the preamp will not draw that much current either and that might be better since you will already have a lower B+ node to work with.)

With the B+ lead into a capacitor input filter, the hollow state rectifier (5AR4 probably) will develop a little lower B+ then the SS rectifiers.
Depending on the duty cycle and current rating of your PT, both will be closer to around 125% to 140% of the actual Hi-vac secondary delivered to the rectifiers, or somewhere around 510vdc to 560vdc, as a guess if it is a good one.
Your DPDT will allow you to switch between the two voltages but with this configuration the screen voltage and preamp stages will remain close to the same.
Lower B+ with the same screen voltage and bias voltage will result in a colder running power tube.
Not sure how greatly the power tubes plate's B+ differences will effect the tone but if you allowed the tubes to idle aropund 18-20 watts with the higher B+ voltage settings, then the lower voltage B+ might be closer to a cooler 15 watts... I'd just experiment with that and keep the power tubes idling at under 20 watts with the HI B+ setting.

[Steve Conner]

Bruce, so you're suggesting to use two full-wave rectifiers at once? One with a capacitor input filter for the plate supply, and another with a choke input for the screen supply? That's a clever idea :D

[SlidePicker]

If you have room and don't mind an anode cap, look at 807s. (And Russian 807s are cheap now.)

[Bruce / Mission Amps]

Bruce, so you're suggesting to use two full-wave rectifiers at once? One with a capacitor input filter for the plate supply, and another with a choke input for the screen supply? That's a clever idea :DYes that is what I thought work for this application... and a builder could use the "screen node" choke input filter side to feed B+ to the preamp node with ease as it already has a lower voltage to start from.

[imaradiostar]

thanks for all of the advice. I'll have to actually start building. All of this talk is awesome but until I get out the chassis punch and heat up the soldering iron it's wasted keystrokes!

I really like your idea Bruce- I'll have to try it.

jamie

[imaradiostar]

Dunno if anyone cares but I built something with an ef86 preamp, 12ax7 LTPI and a pair of 7591 outputs. I found out after building that one of my two 7591's is bad so I've ordered a pair of the new EH7591's. I built a temporary rig with some EL34's to load the PT up and see what kind of voltage and current it produces- I'm actually pretty surprised how much current it'll put out for the size of the PT. B+ will end up around 450 volts with a normal fixed bias arrangement and no zeners.

While I'm waiting for the new output tubes to show up I'm going to rebuild the amp to use back biasing as described on Randall Aiken's web page. I think it's a neat way to kill a little voltage and have an "adjustable cathode bias" without messing with large cathode resistors. This should match up nicely with my 6k6 output transformer- it's sized appropriately to make about 35 watts so it will probably end up being a perfect match. If I decide I don't like the sag of this arrangement it'll be easy to drop a zener in place of the back bias resistor.

Not to be ignored- I think I am going to get some 6fw5's for experimentation ! Though I don't see class AB1 data on them it looks like they'd be great for this application.

Thanks for all the suggestions!

jamie