Designing and Building
Audio Amplifiers.

Ultra Linear or Pentode?

I have begun to have doubts about the results presented in this article. Now that I have built a flexible test bed for power amplifiers and purchased an assortment of output transformers I intend to do this work over again. I will leave the old work up but watch this space for an update. This message will disappear and the "last updated" date at the bottom of this page will change.

With triode connected 6L6s having been shown to produce meager amounts of power the only other choice is whether to use the ultra linear or conventional pentode connection. All of the push-pull output transformers made by Hammond, and available from Antique Electronic Supply, have primaries tapped for ultra linear connection. So lets compare the two for power and distortion.

The Ultra Linear Connection.

Schematic diagram

For a verbal description click here.

Balancing for both AC and DC is important to minimize distortion. To adjust the balance controls connect a sensitive DC voltmeter between the two cathodes, pin 8, of the 6L6GCs. Connect a load resistor to the output and turn on all power. Adjust the DC balance control for zero reading. Now apply a 1000 cycle sine wave to the input and bring up the power to approximately 20 watts (12.5 volts AC across 8 ohms). Leave the meter set to DC and adjust the AC balance control for zero. The voltmeter must have very good AC rejection to work in this application. You are adjusting for equality of average currents in the two tubes.

The plate to plate load should be 6800 ohms. The bias was adjusted for a zero signal plate current of 110 mA. Lower current produced visible crossover distortion. To cut to the chase the power and distortion figures are as follows.

WELL??? I guess ultra linear isn't that linear after all. The distortion is too high for my comfort. 4 tubes in push-pull/parallel would give 40 watts but the distortion is still too high. The conventional pentode connection provides much lower distortion and gives 10 more watts using 4 tubes.

Conventional Pentode Connection.

As I tested this circuit I kept adding refinements to it to make it easier to adjust and work better. Here is the schematic diagram.

Schematic diagram

For a verbal description click here.

Why Things Are.

The series 68 pf capacitor and 56 k ohm resistor on the plate of the first stage are to compensate the amplifier at high frequencies so it won't oscillate. The 0.47 microfarad capacitor used as coupling may seem more than a little excessive especially to those who know that the 470 k ohm resistor in the grid of the phase splitter appears about 10 times larger than it actually is. (See Phase Inverters and Phase Splitters.) If there is more than one high pass RC network in an amplifier with feedback all but one must have a very low cutoff frequency. With a 0.1 in place of the 0.47 the amp had a tendency to ring at about 2 cycles.

6L6s with fixed bias require a maximum grid resistance of 100 k ohms. That is too low for the simple phase splitter to drive. I intend to build two versions of this amplifier one using 2 6L6s per channel and the other using 4 6L6s per channel in push-pull/parallel. The second version would lower the grid resistance to 50 k ohms on each side. I decided to go with direct coupled cathode followers which seems to be a popular approach with modern day tube amplifier designers.

The two one ohm resistors in the cathodes of the 6L6s and the two 100 k ohm resistors in the cathodes of the 12AV7 are selected to be equal in value. If you want to know the cathode current of the 6L6s the one ohm resistors should be selected to be as close to one ohm as possible. It should be pointed out that the cathode current is not equal to the plate current.

The DC balance is adjusted for equal cathode currents in the 6L6s. This is not necessarily equal grid voltages. The two 100 k ohm resistors from cathode to cathode of the drivers permit the average voltage to be found in a single measurement of the voltage at TP1.

The amplifier requires about 5.2 volts to drive it to full power. If you want or need more sensitivity than that you will have to add another stage between the pot and the grid of the 12AX7.

There is enough information here so you could build this amplifier if you were so inclined. I recommend a transformer with a plate to plate impedance of 6800 ohms. The Hammond 1650G comes pretty close at 6600.

Distortion.

Without feedback connected.

With 26 dB of feedback applied.

Now, that's more like it!

The Power Supplies.

B+1 can be anything between 390 and 500 volts but the lowest distortion is obtained at 405 volts. B+2 must be regulated and adjustable from 250 to 350 volts. The negative supply of -300 volts was used because my bench power supply delivers that fixed voltage at 50 mA. When I build a permanent version I will likely use the same value regulated by two gas voltage regulator tubes (VR150) in series. If you feel the design of an adjustable voltage regulator is beyond your capability then I suggest you wait for the finished project. It will be along some time.

Did You Build It?

Let's say you did, and are firing it up for the first time. Temporarily connect a current meter capable of measuring 200 milliamps full scale in series with the red lead, primary center tap, of the output transformer. Connect an 8 ohm 25 watt resistor across the 8 ohm secondary of the output transformer. Start out with the bias set as negative as it will go and B+2 at minimum.

Connect a high resistance DC voltmeter to TP1. Turn on the power and let it warm up for about 2 minutes. Adjust the bias control for -22.5 volts at TP1. Adjust the voltage of B+2 for a plate current of 90 mA, assuming you are using two tubes not four. Connect the voltmeter between the two cathodes of the 6L6s and set it to the 200 mV DC range. Adjust the DC balance control for zero.

Temporarily disable the feedback by shorting out the 510 ohm resistor in the cathode circuit of the 12AX7. Connect an audio oscillator to the input set to 1000 cycles. Bring up the power so it is about 10 watts, that's about 9 volts AC across 8 ohms. Return the voltmeter to the connection between the two 6L6 cathodes and change it back to 200 mV DC. That's right DC. Adjust the AC balance control for zero. Remove the short from the 510 ohm resistor

If you have distortion measuring equipment set it up and adjust the AC balance control for minimum distortion with feedback enabled.

Burn in the tubes by using the amp for about 72 hours, not necessarily continuously. Then repeat the above adjustments after the amplifier has warmed up for about an hour.

Next; Designing an Amplifier with Feedback.

Previous; Experiments with a Triode Power Amplifier.

Home

Everybody's entitled to one dumb thing in life so
Throw away your transistor amplifier and get one with tubes in it.

This page last updated March 20, 2005.