Not Enough Signal in Cathodyne PI to run KT66s?

[David Root]

I'm debugging a modified 5E7 Bandmaster circuit and I have a problem, basically not getting any significant power out of it.

I'm using the Fender stock cathodyne PI with a grounded cathode stage feeding it, exactly the same as the 5E7 arrangement. The KT66s are cathode biased, not fixed, using a single 150 ohm resistor getting 25.0 VDC at the cathodes, and 83 mA cathode current each.

The following numbers were obtained with vol pot maxed, the VAC are RMS and all component values Fender stock.
With a 400Hz sinewave at 1.0 VAC into the input jack I get max. 4.4 VAC at the input to the gain stage, pin 2, and 18.5 VAC out, off pin 1. Out of the PI triode I get 8.0 VAC on pin 6 and 6.8 VAC on pin 8. I get 24 VAC and 16 VAC on the KT66 plates, very low output. B+ is 334V, Vp is 315, but I'm getting only 16VDC on the PI cathode, instead of the Fender stock 55V, so the current is low.

The prior gain stage is also 334 B+, but I get nowhere near the Fender schematic numbers. Vp is 289, cathode reads 0.62V, about 0.45 mA.

Is my PI not capable of driving the KT66s to full power? They have 345V on both the plates and screens. I ran the PI on my Tubecad software which indicate Vo max as -11 +19 V and near as I can tell the KT66s need close to 60V pk-pk to maximize Po.

If I change the 56K resistors to 82K or 100K will that help. Also why are my values so different from the Fender schematic values, and outside the +/- 20% tolerance Fender allowed?

I am also getting funny numbers on the second stage preamp triode and cathode follower. For example, I get current in the gain stage 0.71 mA based on 257 B+ and 186 Vp across 100K, but the cathode voltage is only 0.59, which across 1k5 is only 0.39 mA!

It's almost like I have a false ground that's pushing up the plate voltages and dropping the cathode voltages. Is that possible?

[David Root]

Here's what it looks like.

[Andy Le Blanc]

I found something in this data sheet..... theres a precaution about useing a
common or shared chathode resistor for bias.... and power supply regulation
hope it helps..... these do have a large anode dissapation..... might want to
go back over the power side of the project

[David Root]

Thanx, I have seen that caveat in the KT66 sheet before.

At 320V (345 plate less 25V at cathode) and 82 mA plate current, Pa at idle is 26.2W, well below 30W. How much that might increase with full power operation I don't know, I can't get any power out of it now, but up until close to full power this circuit should be in Class A or very high AB, so I wouldn't expect a huge increase in Pa.

[Andy Le Blanc]

whats the impedance of the opt primary plate to plate?......
whats your anticipation of watts measured at the secondary?......

with these and the transconductance of the tube type we should be able to
figure out what the tube needs applied on the grid..... then we should look at
the gain figures of the inverter......

[Andy Le Blanc]

so... assumptions...... a bandmaster, thats 35w rated with a 4k primary at the
appropriate load..... and a kt66 has a gm of up to 7 ma/v or 7000 micromhos
useing....

Vg = (1/gm) SQRT{Po/Rp}

Vg = (1/.007) SQRT{35/1000}

Vg = (142.86) SQRT{0.035}

Vg = 142.86 x 0.1870

Vg = 26.7

now the impeadance of the primary is going to be lower than this pushing the
needed V to drive the grids of the kt66 up a few more volts
so your inverter WILL need to put out around 60v rms to drive these to a
full output.... good so far....... what about the inverter type?

[David Root]

Thanx Andy. I am not expecting to get 35W because plate and screen are at 345V, not 400, and it is cathode biased too. Maybe 25W.
THe OT is 4k primary with the 4 ohm tap feeding three 12 ohm Webers in parallel for 4 ohms.

The inverter is a cathodyne, exactly as per the Fender schematic, with 56k on both plate and cathode. Vp=334.

When I run my observed numbers thru Tubecad, I get Vo max +11 /-19 V, which is way low. For some reason I cannot get anywhere near the Fender schematic values, which give Vo max -29 + 57. Pin 8 should be 55VDC but I get only 16, so the DC current is too low.

Max AC signal feed into the PI from the prior gain stage is 18.5 VAC, RMS, as measured with a Fluke 187. AC signal on pins 1 and 8 are 8.0 and 6.3 VAC respectively. This is with 1.0 VAC into the amp's input jack.

[Andy Le Blanc]

gain loss?........ is the inverter properly biased?......

[David Root]

Well the inverter is supposed to drop gain 2 or 3%, but not as much as I am seeing.

I think I'm going to have to check the wiring again.

[max_lwedge]

I used this DC biased Cathodyne on a SF Princeton that I added BF Deluxe iron to; I wanted to make sure there was plenty of swing on the PI, the output tube plates now have 435 volts. The supply voltage to that stage is 385v. Tube CAD says the the swing is -51/+99. It seems to hang in there just fine.

The 5E3 guys call the DC bias a "Paul C" mod, but without changing the static voltage or plate/cathode resistors. The 95v at the cathode and the 100k plate/cathode resistors are what I arrived at to suit this amps need for more clean PI swing.

[Andy Le Blanc]

bias and operation point stability?.... with an inverter?......is that a valve or a junction?.......
its neat to see just how close transistors really are.... lower the voltage and
change the device and pooof...... its a gain stage useing the hunter-goodrich
method of bias stabilization......

that looks like a good way to mod. zero bias tube stages for use at higher gain levels....
put the bias where-ever its needed........

[David Root]

Thanx Max, I have seen the Paul C. variation but this is a bit different. I have 334V B+ on this stage, which gives a lower Vo swing.

[rfgordon]

I would recommend you change the KT66 grid load resistors from 220k to 470k. Remember that in cathode bias these resistors can safely go to fairly high values.

[David Root]

Thanx rfg, I had forgotten that. Will definitely plug that in. I think I need to re-read on cathode bias!

I changed the 56k resistors in the PI to 100k, only a slight change resulted, then I replaced the EH 12AX7 with a used '62 Tungsol 12AX7. Huge difference! DC currents increased to more normal levels, plate voltages dropped 25V or so, cathode voltage on the PI almost doubled to 48V. AC signal doubled to 23-24VAC (RMS) out of each side of thePI.

I guess that says something about "UOS" vs. current production, no?

However, this got no voltage gain in the power tubes (KT66) at all.