• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

Tube phase splitter

Jump to Latest
#61
Magnavox AMP-142 schematic 2 (1).png


If the quality isn't good enough I can upload a better copy when I get home.
 
#62
That looks like it's a modified Magnavox 142 model amplifier circuit. That design was probably from a 1950s console like this one:
https://reverb.com/item/78923160-magnavox-142-bc-made-in-the-usa-in-1955

Is this one of those Chinese kit amps?
  • The grid leak resistors on the 6V6GT tubes are 220k ohms, with a 100k pot to the negative grid bias supply.
  • The 12AX7 is run in a current-starved condition, which was common back in the mid-1950s.
  • The 'bottom' 12AX7 in the LTP is running with only 560uA (0.56mA) plate current.
  • The 'top' 12AX7 in the LTP is running with 750uA plate current.
  • There are two capacitors connected between the 6V6 grids, presumably for stability. One is shown as being anywhere from 5pF to 20pF. The other is 5pF.
  • With only 1.3mA drawn from both 12AX7 triodes, changing that 50k resistor to 22k is going to increase the B+ by only about 30V, max.
  • The schematic shows the grid bias on the 12AX7s is at -0.53V or thereabouts. Some 12AX7s might draw grid current in that state, although it's probably OK like that. If you find that one 12AX7 sounds bad, try another. The one you're using might be drawing grid current.
Why are you looking to modify this circuit?
Is it not working correctly?
Are you unhappy with its sound?
 
Last edited:
#63
The issue was pointed out to me as being the 6V6 tubes could go unstable with 220k grid resistors due to the maximum listed grid resistor being 100k for fixes bias operation.

To use 100k grid resistors requires the phase splitter to be redesigned.

Plus as I feed more signal to the amp as it gets closer to clipping the outputs from the phase inverter start to be unequal which I know isn't right and needs to be fixed.

The amp I got as a chassis with just the tube sockets and the power transformer.
 
#65
  1. The maximum grid leak for one 6V6 with fixed bias is 100k ohms.
  2. The 220k ohm grid leak resistors in this circuit are connected to two 6V6 grids, That's like two 440k ohm resistors in parallel, or you could say that's like a 440k ohm grid leak on each 6V6. That 220k value was probably chosen to allow the starved-current 12AX7 LTP to adequately drive the 6V6 grids.
  3. If you replace those grid leak resistors with 100k, that's like a 200k ohm grid leak resistor on each 6V6.
  4. Back in the day, circuit designers used to abuse the heck out of 6V6 and EL84 limiting values. The classic Dynaco ST35 is a great example. It ran its EL84 output tubes with more than 380V on the plates and screens (using an ultralinear OPT with primary screen taps). Also look at Fender guitar amps. They ran push-pull 6V6s with 400V on the plates, into a 5k primary OPT.

Plus as I feed more signal to the amp as it gets closer to clipping the outputs from the phase inverter start to be unequal
Click to expand...
I'm sure that's the 6V6s going into class AB which upsets the precariously balanced 12AX7 LTP as the loads on its plates become unequal (when each 6V6 'side' goes briefly into cutoff).

A compromise (including solid state) would be using a mosfet for split load PI, DC-couple to the *X7.
Click to expand...
+1 this. It's simple and keeps the same number of preamp tubes (one!)
Click to expand...
I'm going to second jcalvarez and Merlinb's comments above. You can cure all these problems by installing a MOSFET split load phase inverter with each half of the 12AX7 as the input voltage amplifier. The MOSFET is tiny. You don't need to punch another hole for it. You might need to drill a hole to mount a little perfboard for mounting parts.

However, there is a problem. A 12AX7 will need at least 120V at its plate, which is almost 1/2 the supply voltage (275V). In order to DC couple the 12AX7 plate to the MOSFET split load inverter's gate, the 12AX7's plate voltage will need to be about 1/3 of the total supply voltage, or in this case 275/3 = 91.7V. Also, the 12AX7 is likely to run at about 800uA plate current, which looks to me like it will run into trouble driving the gate capacitance of the MOSFET (even in the split load inverter configuration).

The cure would be to either...
1) AC couple the 12AX7 plate to a fixed bias MOSFET split load PI, or...
2) Keep the DC coupling but use a triode that gives you decent gain with low plate voltage, such as a 6DJ8 or 6922.

For option 2 above, here's what I had in mind:

1745774652677.png


6922 plate voltage is close to 1/3 the B+.
The MOSFET is dissipating 460mW, so if it's in a TO-220 package it won't need to have a heatsink. A clip-on heatsink would be a nice precaution, though.
The MOSFET doesn't need to be an IRF710. But they're still available and they're cheap (as low as 22 cents each). ZVN0545A would be a bit better, because of its lower Crss.

OK, pick it apart you guys. Does this look like it would be recommendable?
Do you see problems with it?
Anybody have a better idea?
--
 
Last edited:
#67
For now I'll switch to 100k 6V6 grid resistors and give the split load phase inverter a try to see how that works.

One issue I had with the long tailed pair modified for the lower grid resistors is that the output at 10kHz was either like it should be and the output at 20kHz was a little higher or the output at 20kHz was right and the output dipped slightly at 10kHz, however I could see no reason for that.
 
#68
Is this amp you have a mono amp? If so, then yes, you could use a 6DJ8 as the voltage amp-split load inverter.
It will have only about 25X gain, but that might be enough.
That circuit might look like this:

1745775884364.png


You could also use a 12AT7, or a 5965, especially if you can AC couple the first stage to the phase inverter.

If you try to make a 12AX7 split load inverter, you will absolutely need to AC couple it. There's no way that's going to work DC coupled with only a 275V B+ supply.
 
#73
So the circuit then becomes this. Will the screen grid connected as is be ok or do I need a resistor?


Phase splitter.png



I don't have 47k resistors of the wattage likely needed, however I do have four 100k 1/2 watt resistors which I can use to make two 50k 1 watt resistors. Should be just about right for 100k grid resistors.

Pretty sure I have some 6U8 and maybe some 6GH8.

With the outputs staying in better balance I wonder if I'll gain any wattage at maximum power?

Also should I keep the compensation caps where they are or move them to the feedback resistor? Also can I keep the feedback connected to the 8 ohm tap or would the 16 ohm tap be better or would it be better on the 4 ohm tap since I primarily drive a 4 ohm speaker?
 
#76
Tube Radio said:
So I don't have a 620 ohm resistor, but I do have a 1k and 1.5k resistor.... I do have a 47 ohm resistor though.
Click to expand...

Keep in mind that those values are for the Dynaco OPT and the power it has and that the 47R is used in consideration of the 1K FB resistor, so your amp's FB circuit could be adjusted for any amount of FB you want to apply. That 47R resistor attenuates the FB considerably because there is a ton of it there having just a 1K FB resistor from the OPT. There is even some added from the screen/UL tap circuit.
 
  • Like
Reactions: Tube Radio
#78
Got it wired up and I forgot that there is now two phase inversions between the input and top 6V6 pair so I was greeted with an oscillation. I swapped the output transformer primary leads and that took care of the oscillation.

If the gain is set low enough the amp wants to oscillate very slightly at maybe under 10Hz. How do I fix it?
 
#79
Also the outputs of the phase inverter are not equal and the imbalance increases as I increase the input signal amplitude with the signal on the triode plate being higher than the cathode.

What needs to change to fix that?

Turns out I goofed.

The filter cap for the input and phase splitter stages had become disconnected and I didn't see it. Reconnected it and all is well.
 
Last edited:
You must log in or register to reply here.
Top Bottom