Stereo Power amp

[Bob-I]

A friend of mine used to use a tube power amp for his PA and asked if it was possible to build one. I have iron that I happened on cheep a few years ago, 2 Hammond 1650R UL output xformers and a Hammond 287CX, a monster 400-0-400 capable of 465ma

I'm thinking that I can use PP 6550's or KT88's, a single gain stage and a cathode follower into the LTP PI.

I have concerns of course.

1) Is the B+ going to be too much? If I use a full wave like I drew I can expect 600V. Can 6550's handle that kind of voltage? I know most are rated at about 500V and I have 4 Sovtek 6550's that I can use for this.

2) Where can I insert a gain control? After the CF?

3) I'm questioning the filtering. The largest 220uF cap I can find is 385V so 2 would be 770V.

4) I drew up 2 rails, once for each pair of 12AX7's. Will the single B+ be ok, or will there be crosstalk between the 2 OT's.

I only drew up one channel, of course I'd duplicate this.

http://www.patmedia.net/bob-ingram/SPAschem.jpg

Thoughts?

[MarkB]

Design Center max. is 600V for the plates and 450V for the screens. I don't know how current production tubes would hold up in practice.

Why the cathode follower? I don't see any benefit to it before the PI. The gain stages needed would depend on the signal that's coming into the inupt.

[Bob-I]

Design Center max. is 600V for the plates and 450V for the screens. I don't know how current production tubes would hold up in practice.

Thx, the spec sheet I read said 575 plate. I should check Sovtek's sheet.

Why the cathode follower? I don't see any benefit to it before the PI. The gain stages needed would depend on the signal that's coming into the inupt.

No real reason for the CF. I just had 1/2 12AX7 free. Maybe I'll just take it out and out a gain control between the gain stage and the PI.

The gain stage will be needed because the intended use is out of an SS mixer and the signal will be too small for direct into the PI.

[MarkB]

In TUT3, K. O'C. talks about putting 704V on the plates of 6550s for an SVT build. He also uses two different PTs for B+ - one for the plates, and a lower voltage one for everything else. To run the screens near the plate voltage, you'd need KT-88s.

From his description of what's needed to run 6550s, I think I'd stick to 4xEL34s or 6L6s for that application.

[Bob-I]

In TUT3, K. O'C. talks about putting 704V on the plates of 6550s for an SVT build. He also uses two different PTs for B+ - one for the plates, and a lower voltage one for everything else. To run the screens near the plate voltage, you'd need KT-88s.

So I should increase the size of the screen resistors to lower the voltage... cool.

From his description of what's needed to run 6550s, I think I'd stick to 4xEL34s or 6L6s for that application.

Nah, I want to run only 4 power tubes, not 8. Just too many tubes in one build. I'll adjust the voltages.

[skyboltone]

So I should increase the size of the screen resistors to lower the voltage... cool.

Errr. I dunno here. Those are for parasitic oscillation Bob. I would adjust the voltage at the PS stack for your target at 6MA (4 tubes?)or so and let 'em sag with signal supplied. When I used to fool with AM modulators we would regulate the screens to keep them from flopping around so much, but I gather here that we want them to move. (edit: Damn I should read all the posts before I jump in the discussion, I thought we were talking about guitar amps. Where's my sign?) I would regulate the screens like we did with the modulators for a PA amp. I have some high voltage transistors that were originally designed for solid state ignition systems. You can make a simple darlington with zener reference with one and model the screen supply rail with the voltage about 60 VDC high at the input to the regulator. There are tricks with Darlington transistors though. No positive voltage relative to input on the output side of the regulator for one. So no caps! Or use diode protection. This just is not a problem with a screen supply though. Alas you can't use one on the plates. There is just too much current and heat to dump. You could however buck the secondary windings with a 36V transformer on each side of the center. Too much trouble I know but possible. Those little Triads transformers are megged at 1500V minimum and some are rated at 4000V. If you use a separate filament transformer you can buck the input side of the power transformer, (much simpler) and just build your bias voltage normally or from the main windings if it's too low. Buck transformers only need to be current rated at their share of the load not the whole load. PM me if you want.

Dan

[mlp-mx6]

I thought screen resistors were primariy there to limit screen dissipation (screen current).

[skyboltone]

I thought screen resistors were primarily there to limit screen dissipation (screen current).

Hi Michael:
The 1965 RCA tube manual in part: "When power pentodes and beam power tubes are operated under conditions such that there is a large change in plate and screen-grid currents with the application of signal, the series resistor method of obtaining screen-grid voltage should not be used. A change in screen-grid current appears as a change in the voltage drop across the series resistor in the screen-grid circuit; the result is a change in power and an increase in distortion. The screen grid voltage should be obtained from a point in the plate-voltage supply filter system having the correct voltage, or from a separate source."

Then, each of the information sheets of the beam tubes under consideration note some such as the following from the KT66 tube information sheet: "For the prevention of parasitic oscillation a series resistor (values vary from 300ohms to 1K) should be connected close to the screen grid terminal of the valve socket."

Please note the discussion below:

http://www.raw-sewage.net/jtm45ri_mods.html

Dan

[skyboltone]

In TUT3, K. O'C. talks about putting 704V on the plates of 6550s for an SVT build. He also uses two different PTs for B+ - one for the plates, and a lower voltage one for everything else. To run the screens near the plate voltage, you'd need KT-88s.

From his description of what's needed to run 6550s, I think I'd stick to 4xEL34s or 6L6s for that application.

6L6s are not suitable for Ultra Linear application. EL-34's are only slightly usable. KT-66 or KT-88 would be OK.

Please read below:

http://www.webace.com.au/~electron/tubes/ul.html


Dan

[Bob-I]

I would adjust the voltage at the PS stack for your target at 6MA (4 tubes?)or so and let 'em sag with signal supplied.
Dan

Where I appreciate the time and efort you put into this reply Dan, you missed that this is UL so there's no supply for the screens.

Thx though.

[skyboltone]

I would adjust the voltage at the PS stack for your target at 6MA (4 tubes?)or so and let 'em sag with signal supplied.
Dan

Where I appreciate the time and efort you put into this reply Dan, you missed that this is UL so there's no supply for the screens.

Thx though.

Did you get your PM?

[Bob-I]

I would adjust the voltage at the PS stack for your target at 6MA (4 tubes?)or so and let 'em sag with signal supplied.
Dan

Where I appreciate the time and efort you put into this reply Dan, you missed that this is UL so there's no supply for the screens.

Thx though.

Did you get your PM?

No, I don't have any new PM's

[skyboltone]

Hi Bob:
I did initially miss the UL because: a) I failed to read all of the posts from the start before chiming in completely off topic, b) I didn't open your scheme to take a look before chiming in completely off topic, and c) I'm a dunce.

Behold the following document:

http://www.webace.com.au/~electron/tubes/ul.html

BTW, it supports your cathode follower driver stage.

I stand by my original assertion that a series screen resistor is the wrong way to adjust screen voltage on a beam power tube.

And, with a UL scheme, why do it anyway? By memory, KT66, KT88, and EL34 tubes all allow screen voltages equal to plate maximums in a UL arrangement. For a PA amp, I can't imagine a better scheme than UL. You will get power slightly less than a Pentode hookup and will cancel nearly all IM and Harmonic distortion in the process. The Amp will "Sound" louder because you can push it so much harder before clipping sets in. The sound is said to be somewhere between Triode and Pentode sound.

As far as the 400-0-400 tranny goes, it is simpler than you think to buck the primary voltage down so that the secondary is more in the ballpark you are looking for. With a 117V primary the turns ratio is 1/6.84. You need to loose 15 volts on the primary side. We can choose Hammonds 266J14 or 266J16 and wire it like an autotransformer on the input side of your PT. The technique is to series the secondary windings with the primary, in subtractive polarity to buck the income house current. The buck transformer does not need to be rated for the full load, only that portion of the load that it bucks or boosts. Unfortunately this renders the filament windings useless unless you intend to rectify, filter and regulate it back to 6.3 VDC. Easier in my mind would be to buy a separate filament transformer and tape off the original PT windings.

If there is interest in the Buck/Boost trick, I can write something up for the Wiki, but information on the subject abounds on the net.

My apologies for all the confuseion and general weirdness of my posts on this topic, (others as well?). It's a reflection of my nature I fear. A bit impulsive.

Cheers
Dan

[Bob-I]

I'm a dunce.

Me too.

http://www.webace.com.au/~electron/tubes/ul.html

BTW, it supports your cathode follower driver stage.

What am I missing? I don't see where it supports the CF, then again, see above.

I stand by my original assertion that a series screen resistor is the wrong way to adjust screen voltage on a beam power tube.

I see your point.

As far as the 400-0-400 tranny goes, it is simpler than you think to buck the primary voltage down so that the secondary is more in the ballpark you are looking for. With a 117V primary the turns ratio is 1/6.84. You need to loose 15 volts on the primary side. We can choose Hammonds 266J14 or 266J16 and wire it like an autotransformer on the input side of your PT. The technique is to series the secondary windings with the primary, in subtractive polarity to buck the income house current. The buck transformer does not need to be rated for the full load, only that portion of the load that it bucks or boosts. Unfortunately this renders the filament windings useless unless you intend to rectify, filter and regulate it back to 6.3 VDC. Easier in my mind would be to buy a separate filament transformer and tape off the original PT windings.

See above... I'm a dunce because I'm not following this at all.

My first thought is if the primary is connected to the secondary, this becomes dangerous as AC input is directly coupled to the HV.

Maybe a schem can explain this better, I'm just not following.

Thx for the reply.

[skyboltone]

Bob,
Well here goes. I may buck my input voltage down a little, I'll be running about 500 VDC on the plates of KT66s, 460 on the screens. I've got a UL tranny but because Scott recommends pretty strongly against UL for guitar amps I'll be using a regular screen supply. Using Glasman's (I think) modeling software. Somebody's anyway.

With regards the CF question, the comment occurs later in the sited document under the heading High Power, and goes as follows.

High Power Ouput:
Where more power is needed it is preferable
to use a larger tube that is known to be more
suited for ultra-linear circuitry, such as
the KT88, KT90 or 813.
Another useful option is to run multiple
pairs of tubes in parallel push-pull, such as
the arrangement used in the GEC 400W KT88
Amplifier. It goes without saying that normal
precautions against instability and parasitic
oscillations are essential in layout, lead
dress, use of grid stopper resistors wired
directly to the pins, and keeping inputs away
from outputs. The output transformer must be
of high quality with low leakage capacitance
and low leakage inductance between windings.
Tubes should be mounted close together to
minimise inter-connecting lead length. Grid
#1 circuit resistance must be held within the
manufacturer's ratings. A low-impedance
driver, such as a cathode-follower or
transformer is recommended.

Then here are the relevant schemes for buck/boost transformer hookup

http://tinypic.com/20fwz1d.jpg

http://tinypic.com/20fx1df.jpg

http://tinypic.com/20fx3tl.jpg

If these are too small to read, send me a PM with e-mail addy and I'll send them to you full size.


Cheers
Dan