gain differences in driver/PI stages for cathodyne, paraphrase, long tail pair

[tubeswell]

Hi there

I have all the bits a need to start building a 5e3, or a 5f11 or a 5e9A or a 5G9.

They all have varying driver and PI stages, and basically I would appreciate it if somebody could venture an opinion about the differences in tonal characteristics.

Starting with the 5E3, after the two parallel voltage amp stages, you have a driver stage then a cathodyne PI stage before the signal hits the output stage. My understanding is that you've got to have about 25V on the grid of the output tubes for them to work efficiently, and my understanding is that this is what is delivered by the 5E3 driver and inverter combination.

However I would like the advantages/sound of a 5E3, but with the benefit of a tremolo effect.

The 5F11 is basically a 5E3 circuit with the same driver and cathodyne PI stages, but with one stage of the first dual tridoe tube acting as the voltage amplifier stage, and one as the vibrato oscillator, with the plate in parallel to the voltage amp plate stage, and where the output feeds in to the plate supply of the PI.

The 5E9A keeps the two parallel voltage amplifer stages that are in the 5E3, and has a paraphrase PI, utilising two triodes, which doubles as a driver (since it doesn't have a separate driver stage). The oscillation is done using two triodes, where the cathode of the input is tied to the dual cathode of the PI, and the output goes to the screens of the output tubes.

The 5G9 is similar in several respects to the 5E9A in keeping the two parallel voltage amp stages that are present in the 5E3. The 5G9 doesn't have a separate driver stage, but the PI is a long-tailed pair, which (if I understand correctly) doubles as a driver and PI. Presumably this circuit still delivers the same sort of voltage to the output tbe grids as the 5E3?

The 5G9 oscillator circuit feeds off the front of the PI and the output goes to the grid of the output tubes via a fixed bias circuit. Bruce (from Mission Amps) has testfied to the supreme qualities of this circuit, and I am pretty sure this is what I want to build. (I have a PT with a 50V bias winding) But I was wondering if there was much difference in the 5E3 tone, and that of the 5G9 when the tremolo was disengaged? Is there perhaps some difference in tonal quality from having three sets of stages in sequence (before the outut stage) in the 5E3 circuit, as opposed to doing away with the separate driver stage in the 5G9? Is there some difference in overall resistance and capacitance for instance?

If the answer to that question is "yes there is alot of difference", then is is practical to do a 5E3 driver and separate cathodyne inverter, but with the 5G9 tremolo circuit run into a fixed bias output stage, or would that be an inferior circuit to the stock 5G9 with the long-tailed pair (and no separate driver)?

So many questions.

Cheers

[MWJB]

In a lot of cases you're not reallly comparing like with like, both 5G9 5E11 & 5F11 are fixed bias as opposed to cathode bias for 5E3 & 5E9A.

5E11 & 5F11 run their volume pots as typical voltage dividers, as opposed to the 5E3style.

I would stop and have a think about what is important to you, also as the vibrato effect is generated by different means & circuits, it would be advantageous to compare and determine which has the character that you prefer? The method of executing the vibrato will greatly affect how it sounds. If you think the 5G9 is the one for you, then go ahead.

On the other hand, if you just want to build a 5E3 inspired amp with vibrato, but aren't concerned about whether fixed/cathode bias, maybe try transplanting the first stage of a 5E3 on to a Tolex Princeton trem/PI/output, (6G2 style, but with 5E3 style volume...oops, one more option to confuse things further) tweak power supply dropping resistors & NFB to suit?

[Bruce / Mission Amps]

Very simple and works.
Build it like a one channel 5E9-A but use an N channel power Mosfet in place of the missing, DC coupled triode ... that is the triode with the depth pot connected to it's cathode. You connect the source of the power Mosfet to the parallel 100K resistor and 250K depth pot instead.
I think the total +71K source resistance will limit the current to just a +few ma and you'll have to adjust your B+ rail dropping resistors between filer caps to accommodate the extra current drain.
Use a 100 ohm resistor between the gate lead and plate of the other triode.

I've been using IRF830 series Mosfets for stuff like this. I found a couple hundred IRF830s for something like 30 or 40 cents each in a salvage electronics place.
I'm sure some of the solid state guys could come up with something better but the DC coupled TO-220 power Mosfet works really easily.
I think any 500v N channel IRF*** series would work too.
I've actually messed with these things and they're quite fun, and since cheap... easy on the pocket book if you blow one...which I have not really done much of at all.
My Aurora amps use one to drive the Reverb output transformer.

In this application, all you are doing is bringing the vibrato circuit's oscillating DC (which is really a very slow LFO and the DC appears as AC, with respect to the driver triode's cthode) through a 100K resistance (including the parallel depth pot) to the cathode biasing resistor of the 12AX7 driver triode... in front of the cathodyne driver.
The pulsing AC will dramatically change the bias, hot and cold, of the driver triode and clamp the tube down and release... which is what makes the vibrato effect.
It can be a little thumpy if the depth is set too high though... that means a little tweaking of the depth circuit so it doesn't bias the triode into cutoff and thump on-off-on-off-on-off-on ... etc.

Apply the normal power Mosfet protection tricks to keep it working right and check RG Keens website for Mosfet Follies to learn more.
That's where I got the ideas to do stuff like this with power mosfets.

[tubeswell]

Thanks MWJB and Bruce

Before I completely abandon the rush of blood to the head idea of combining different stages and PIs, and go with a stock 5G9 (or before I get into looking at using MOSFETs) I wanted to clarify what I was thinking about in terms of hybridising the 5E3 and the 5G9, and get your feedback (if that's okay)

I have attached a conceptual sketch of a schematic in this regard (I haven't filled in most of the numbers and some of my voltages are wild guesses).

Can I run the Tremolo circuit with the depth control going to the 6V6GT grid bias like that with a cathodyne PI? (come to think of it - does the cathodyne PI have enough balance to ensure that the fixed bias provides any benefit? if that makes sense?)

FWIW I have 3 pre-amp tubes to fool around with and a 5U4G rectumfrier.

Or would I be better to have a single voltage amplifer stage and driver stage (like a Princeton 6G2), but keep the long tail pair PI and the Trem circuit from the 5G9 (or is a separate driver with a long-tail pair PI an anachronism/completely unnecessary exercise in futility)?

[tubeswell]

sorry the last attachment went awry.

here it is again

[tubeswell]

Sorry if those attachments give you a crick in the neck - (don't know why they load up sideways).

Anyway - I am in unknown territory here. What are the fatal flaws with this concept?

[jeefberky]

Ha... this drawing is exactly where ended up with my logic. It seems the extra snarl of the 5E3 is locked in that extra gain stage/ PI combination, but will it be delivered as sweetly through the fixed bias 5G9 power section?

For yet another variation, check out the schematic for the Ampeg Jet (J-12B) at schematic heaven. I have the J-12D, and with the negative feedback removed, it sounds pretty darn good.

Jeff

[tubeswell]

My advice is to build it stock - that's what I did. The 5G9 is meant to have a tweed front end feeding a LTP with a bias vary trem for 6V6s. Ideal combination.

However it you want to get fancier, you could try this idea for the front end (that I came up with yesterday) - interchangeable parallel/cascaded stages for V1. You lose two input jacks, but use one of the holes for a DPDT switch. (You could even use the spare hole for a switchable ck for V1). It won't be quite the same as a 5G9 tho.

[Wombaticus]

Sorry for rezzing this old thread.

I thought I'd just point out though, that except for cost, there's not too much reason to use the IRF820/830/840 series anymore. I'm now using the things like the STF3NK80Z STF3NK80Z STMicroelectronics MOSFETs which has the advantage over the IRF series of having a fully insulated package (so you can mount directly to the chassis), and also has built in gate-source and source-drain protection diodes, which reduces your component count. Mind, it'll run you a whopping $1.27 instead of thirty or forty cents, but hey, who's counting. Not saying the IRFs won't get the job done, but these are just a more elegant solution.

I was reading this thread though because I'm fixing to do another 5E3-ish hybrid build with conversion to fixed bias and a power tube bias wiggling tremolo using a LND150. To me it seems criminal to waste a valve on a low frequency oscillator.... Won't be a deluxe of course, but should be good in its own right.