Schematic check request: Rocket-style KT88 EF86 project

[rbryerton]

OK, first off, a little background on this project, and where the idea came from. I've built several amps this past year (liverpool, rocket, 1987, 2004, 18watt w/ef86, etc) and really dig the 'wreck thing. All of these projects are either EL84 or EL34 in the power dept., and I'm looking for something a bit different.

Seeing a few projects other members here built that are based on the Rocket intrigued me (KT66 power amp section etc), so I decided that I wanted to do something similar, but with some big bottles

On doing research on KT88's, I noticed a lot of designs and data sheet info are spec'd for ultra linear operation, and seeing as I have yet to experiment with KT88's or UL, I picked up a Triode A-431-S OPT, which is a hefty 4.3k tranny with UL taps.

In order to house these beasties and provide a lot of flexibility and room to build with, I went with a 100W Plexi-style chassis. I'm planning on wiring the 4 output tube sockets, but will only use only 2 of them with the KT88's; this leaves me flexibility in the future to drop 2 or 4 EL34's in there to mess around with. I'm also going to wire the OPT so I can switch between pentode & UL operation (powered off, of course) and a switchable voltage divider to provide the KT88 screen grids with the proper lower voltage in pentode mode.

To top things off, I'm adding a parallel 1st gain stage w/an ef86: this will be wired up like a jumpered Plexi, so each paralleled 1st gain stage (12ax7 & ef86) has it's own volume control so they can be mixed in any proportion.

I have drawn up a schematic, and would really appreciate if you could take a look at it and see if you find any errors or less-than-optimal connections/values. There's 2 areas that I'm not fully comfortable with:

Voltage divider on B+2: is this divider a good way to wire things to drop the B+2 to 250-300VDC for the screen grids in pentode mode, or am I way off? (this creates a bleeder for the caps also....) Any value/sizing suggestions, or alternate ways to achieve this voltage?

Mixer resistors/caps for V0 (ef86) and V1b (12ax7): this is reminiscent of how a Plexi (1959/1987/etc) achieves this function, but any ideas here would be cool. I don't know if the mixer setup as I have it will mess with the tone too much. I'd like the mixer to be as transparent as possible.

Any other input/suggestions are welcome as well. Thanks for your time!

[img966]http://www.bryerton.com/misc/kt88rocket ... et-v01.gif[/img]

[Jana]

A few comments:

You need some equalizing resistors on the caps right after the choke.

You may need to switch/adjust the bias when you switch the various pentode/UL modes and voltage levels on the screens.

You should have a filter cap at the junction of the divider resistors for the screen supply.

Is that transformer 380-0-380? If so, you are going to be getting about 1000 volts on the plates with the FWB rectifier.

[Cliff Schecht]

He would be alright by switching to a two diode FW rectifier. He would benefit by putting multiple diodes in series because the voltage is relatively high. Also, a 1A fuse seems a bit too small.

The amount of highpass filtering and "treble bleeding" you have on the first stage is a bit ridiculous. Do you know why you are doing it or are you just doing it because you want it to be transparent? What you have will be the opposite of transparent, it's going to filter the shit out of your signal and perhaps even cause instability. If I were you, I would kill all of the excessive filtering and move the volume controls to the inputs of the first stage, not the outputs. As you have it now, even with the isolation resistors, the "mixers" are going to interact pretty badly. OTOH, the plate impedance of the first triode and pentode stay fairly constant as you change the input signal so turning down one channel won't affect the other. Just a thought and maybe something to play with.

The way you are dropping your B6 for the screen grid is pretty common but in guitar amps, the 0.1uF cap you have there is typically left off completely. Leaving it on gives you a "tighter" sound by keeping the screen supplied with a constant voltage even under large transient loads but this isn't necessarily what you want for a distorting amplifier. I'd play around with this cap in and out of the circuit and see what you like the most.

[VacuumVoodoo]

That resistive divider for B2 screen grid supply won't work.
2 KT88 screen grids draw ca 3.5-4mA current at idle. This will drop 350-400V on the 100k resistor. At full power output those grid will want to pull around 30mA and thus drop 3000V !

Ditch the resistive voltage divider altogether and get your reduced screen grid supply straight from point between the two filter caps to the left of the now not needed resistive divider.

Tube data sheets are your friends even when designing a rule breaking non-hifi guitar amp.

[rbryerton]

OK, thanks for your input. I know just enough about theory to be dangerous, as you no doubt have noticed

I made a few changes with the filtering on the input stages, and changed the voltage divider. I must admit that I don't understand the voltage division function of two stacked caps

Also, the power tranny is 190-0-190 400mA, hence the FWB. Cliff, about your comment about the 1A B+ fuse: is this because of the inrush that you think it may be too small? I put these KT88's in my 1987, biased at 70% w/~475V on the plates at idle, same choke, and 1K 5W on the screen grids and that amp uses a 500ma B+ fuse, and it held cranked through an attenuator. The screens were most likely drawing way too much current also. I only played it for about 5 minutes, but it sounded monstrous with these tubes

OK, enough of my babbling; here's an updated schematic:

[img971]http://www.bryerton.com/misc/kt88rocket ... et-v02.gif[/img]

[VacuumVoodoo]

I made a few changes with the filtering on the input stages, and changed the voltage divider. I must admit that I don't understand the voltage division function of two stacked caps

Looking some more at the schematic I see that taking low screen grid supply from the midpoint of the first filter cap stack is an even better way. This point is hardwired to 1/2 B+ through PT HV center tap. Just take a look at Fender Vibro King schematic and you'll see what I'm talking about.
http://www.fender.com/support/amp_schem ... ematic.pdf

see page 8 or 9, lower left, terminal "Y" 232V. Works very well.

[rbryerton]

Looking some more at the schematic I see that taking low screen grid supply from the midpoint of the first filter cap stack is an even better way. This point is hardwired to 1/2 B+ through PT HV center tap. Just take a look at Fender Vibro King schematic and you'll see what I'm talking about.
http://www.fender.com/support/amp_schem ... ematic.pdf

see page 8 or 9, lower left, terminal "Y" 232V. Works very well.

OK, that makes more sense to me. What is the purpose of the diode before the Y junction I wonder? Is that a half-wave rectifier?

Also, I assume the caps at both Y junctions are for smoothing ripple as they're taking the B+ from the reservoir caps after the FWB?

Maybe it would be cleaner to just make a separate FW rectifier w/2 diodes and a couple lower voltage filter caps for the low voltage? What do you think? Do the screens need to be nicely filtered, or can they tolerate moderate ripple?

edit: it would be nice if I could ditch the voltage divider all together, it would save on components and complexity. But I'm afraid I'd melt the screens on the KT88's, and they're relatively expensive...

Thanks!

[VacuumVoodoo]

OK, that makes more sense to me. What is the purpose of the diode before the Y junction I wonder? Is that a half-wave rectifier?

Also, I assume the caps at both Y junctions are for smoothing ripple as they're taking the B+ from the reservoir caps after the FWB?

"Y" terminal feeds screen grid of the 6V6GT reverb spring driver, look at what follows after "Y", it's an RC ripple filter to provide screen grid with well regulated voltage.

Do the screens need to be nicely filtered, or can they tolerate moderate ripple?

edit: it would be nice if I could ditch the voltage divider all together, it would save on components and complexity. But I'm afraid I'd melt the screens on the KT88's, and they're relatively expensive...

Thanks!

KT family of power tetrodes usually like a well regulated i.e. non sagging low ripple supply.
I would either adopt the Vibro King supply or be the "stupid engineer" and use a MOSFET based voltage regulator to set a stable ripple free screen grid supply. One from hallamplification.com could just do the job. Something like the VVR for cathode biased amplifier used only to set screen voltage, not as an output power reducer.