The elusive DC coupled cathodyne perfection?

[jimboyogi]

Hi all,

Please be patient with me!

I've done a lot of reading, both here, in Merlin B's pre-amp book, and all over the internet.

I'm trying to set-up a DC coupled cathodyne in a sort of hot rodded 5E3 amp, using a post PI MV (dual gang), and using the gain stage/cathodyne as the pre-amp distortion generating stage.

Because I am trying to get a lot of gain on the gain stage, I have been using a 12AX7 as both gain stage and cathodyne.

First i tried an AC coupled version. Set up the gain stage roughly centre biased, and also the cathodyne roughly centre biased using cathode bias. For the cathodyne loads I tried both 56K Ra and Rk, and 100K Ra and Rk. I tried both 470K and 1M input grid stoppers on the cathodyne. I used a trimpot for cathodyne Rb, and played with varying values in each scenario.

The AC coupled sounded great when run clean, really bright, clean and responsive. When pushed, it moved to overdrive fairly suddenly, and when pushed further it gave an unpleasant "buzz" that sat atop the note then decayed erratically, reminding me of the sound of badly biased germanium fuzzes! This "buzz" occured especially on lower strings/lower frequencies.
To reduce/remove the "buzz", i tried varying the bias of both the gain stage and the cathodyne. I tried both a 470K and a 1M cathodyne grid stopper. I tried extra bass attenuation before this stage. I have a 150K grid stopper on the gain stage. None of these things made much difference, I couldn't remove the "buzz".

So next I tried a DC coupled version. I used a 12AX7 for both stages, and a 470K cathodyne grid stopper. I tried 100K, 220K and 330K for the gain stage Ra. I tried both 56K and 100K for cathodyne load Ra and Rk. I altered the bias on the gain stage with a trimpot. And I measured voltages with my DVM, and scoped output waveforms.
The best I have achieved so far is the following -

Gain stage - 330K Ra, 1K Rk, HT = 297V, A = 78V, K = 0.65V, Voltage gain = 75

Cathodyne - 100K Ra and Rk, HT = 297V, A = 192V, K = 86V. Max output pre-clip - Anode output = 50Vpp, Cathode output = 78Vpp

My concern is that using this set-up, the clean signal sounds a bit 'muffled', not as bright or clean as using AC coupling. The overdrive is amazing though! But the ideal that i'm seeking is great cleans AND great overdrive.

I scoped the input signal to amp versus the output, with the MV all of the way up. The wavefroms tracked very well until compression began at nearing full output, so I am assuming that the output stage was causing this compression. So by the scope it looks very linear at lower signal levels. But it doesn't sound it!



So to the theory of the DC coupled cathodyne. I understand that using a 12AX7, or other high ra valve, will lead to quiescent grid current flowing from the gain stage anode to the cathodyne grid. This will cause assymetric compression, and also assymetry between the cathodyne outputs.

So am I correct in thinking that to get the "cleanest" DC coupled performance, this quiescent grid current wants to be minimised?

And what are the best methods to minimize it? Is it to lower the gain stage anode voltage down to, or even lower than the cathodyne cathode voltage?

In my current best set-up, I am using a 330K Ra on the gain stage, but even using such a large Ra, I still found that I had to bias the stage very hot to obtain the best output symmetry via a scope. So it is running only 0.65V Vk.
I am wondering if this very hot bias is in part to blame for the 'muffled' clean tone I am hearing. Or is it due to the high Ra of 330K?

If i increase gain stage Ra to say 390K, or even 470K, will I then be able to increase Vk and cool the bias a bit, and will this give me better cleans?


I greatly appreciate any help here!

[Austin]

-.65 volts is almost in grid current for a 12ax7, it begins at lower voltage than zero so is probably loading your guitar pickup. A cathodyne phase inverter has no gain so a 12ax7 will not help get more. Have you thought about trying a 12dw7? It is half a 12ax7 you could use for the gain and half a 12au7 that can give more current that you could use for the phase inverter. The master volume is going to load the phase inverter and as you know a cathodyne phase inverter should not be loaded much at all really because it will lose it's perfect balance. Can you post a schematic? I use a free program called "TinyCad" that is free and easy to use. Are you talking about the one at the bottom of the page here? http://www.freewebs.com/valvewizard/cathodyne.html

[jimboyogi]

Hi Austin,

Yes the schematic is exactly as for Merlin's version on the page, except for the changes in component values as in my post. As it's turned out, even the HT is almost the same as Merlin's example.

I have thought about a 12DW7. The reason that I'm avoiding it ATM is that I'm using the gain stage for lots of gain, so I'm keen to use a valve that has the lowest noise. I'm not sure how the 12AX7 side of a 12DW7 compares noise wise to the fairly highly developed and scrutinised 12AX7?

I've got a 12DW7 (JJ ECC832) here to try if it's well recommended though

[jimboyogi]

I just remembered another reason I was reluctant to use a 12DW7. Using the 12AU7 half as the cathodyne, my sums show that the cathodyne cathode will sit at a lower voltage than that of a 12AX7, given reasonable loads (say 100K Rk for the 12AX7, 47K for the 12AU7).
If my theory is correct, and please feel free to correct me if it isn't, then I would need to get the anode voltage of the preceding gain stage down even further than with the 12AX7 set-up.

If I used a 12DW7 I would probably have to level shift between the stages, which means more components.

[jazbo8]

Isn't it the way it suppose to work? Since you removed the coupling cap, the high pass filter is gone, so resulting in more lows with the DC coupled cathodyne, which could be perceived as muffled. Does it make sense?

Jaz

[jimboyogi]

OK, I think that I have answered my own question

I tried a 390K Ra on the 12AX7 gain stage, and adjusted Rk looking for the best symmetry of rounding/clipping on the scope. I ended up with -

Gain stage - 390K Ra, 1.5K Rk, HT = 295V, A = 82V, K = 0.86V, Voltage gain = 76 (@1KHz)

Cathodyne - 100K Ra and Rk, HT = 295V, A = 184V, K = 93V. Max output pre-clip - Anode output = 55Vpp, Cathode output = 80Vpp

So slightly more output, and output symmetry from the cathodyne.

On the gain stage, by using a higher value for Ra, I have been able to increase Vk/Rk while still biased for max headroom. The result both on the scope and by my ears is that the compression has reduced, and the sound is clearer and brighter.

I could be imagining things, and I could be less happy when I try it again tomorrow (happens all the time). But I think that this change of values on the gain stage has significantly improved the clean sound of this amp. The overdrive seems maybe slightly less compelling, and possibly a bit more 'gritty' than the previous, but the extra versatility gained outweighs this easily.

[Steve Conner]

I think the hot bias could be to blame, it could cause a lot of low-order distortion resulting in a weird-sounding clean tone.

Cathodyne PIs also suffer from what I have heard called "nipple distortion". (sounds a little sadistic :O ) When the power tube driven by the cathode output is pushed into grid current, it increases the gain of the anode output. A nipple-like peak appears on the waveform. This could be the source of your nasty buzz in the AC coupled circuit.

To bias a cathodyne PI for maximum unclipped output, you want the cathode at something like 1/4 of the HT voltage and the anode at 3/4.

[jimboyogi]

It does make very good sense Jaz.

From what I understand, with a cathode biased cathodyne, the input impedance is extremely high due to bootstrapping. I was using a 470K grid leak here, which gave me something approaching 10M input impedance! The result is that the input coupling cap only needs to be a very low value to pass down to very low frequencies. I was using a 10n cap here, which should have passed all frequencies down to sub-audible.

Not to say that there isn't some audible effect from sub-audible frequencies though.

[jimboyogi]

I think the hot bias could be to blame, it could cause a lot of low-order distortion resulting in a weird-sounding clean tone.

Cathodyne PIs also suffer from what I have heard called "nipple distortion". (sounds a little sadistic :O ) When the power tube driven by the cathode output is pushed into grid current, it increases the gain of the anode output. A nipple-like peak appears on the waveform. This could be the source of your nasty buzz in the AC coupled circuit.

To bias a cathodyne PI for maximum unclipped output, you want the cathode at something like 1/4 of the HT voltage and the anode at 3/4.

Thanks Steve.

Below I have written up my latest impressions. It seems that raising the gain stage bias voltage from 0.65V to 0.86V has made a big difference to the clean tone.
For tonight i'm really happy with this set-up!

[Wombaticus]

Thanks Steve.

Below I have written up my latest impressions. It seems that raising the gain stage bias voltage from 0.65V to 0.86V has made a big difference to the clean tone.
For tonight i'm really happy with this set-up!

Aside from a humongous grid stopper on the cathodyne itself, the other secret to good cathodyne tone is using much larger than normal grid stops on the power tubes. I would regard 22k as a minimum starting place.

Incidentally, in some recent experiments with the AC coupled cathodyne, I found it turned out to be remarkably insensitive to variations in bias varying from centre to quite hot. If this turns out to also be the case for the DC coupled circuit, this may possibly ease some of the level matching issues.

[Austin]

Have you ever tried battery bias with the battery to actualy elevate the cathode? I think it works pretty well, it ends up charging your battery with some of the juice flowing through the tube but with a 12ax7 isn't very much. You could use a 1.5v cell in the first half and a 9 volt on the second half if you used a 12dw7. Just a thought.

[jimboyogi]

Aside from a humongous grid stopper on the cathodyne itself, the other secret to good cathodyne tone is using much larger than normal grid stops on the power tubes. I would regard 22k as a minimum starting place.

Thanks Wombaticus. Currently I've got 47K grid stoppers on the output tubes, and the overdriven tone of the output stage is great, very smooth. Using EH 7591's.