Low power of 5D3

[Leevi]

I got a 5D3 wired and made some measurement and wonder the low power.


http://ampwares.com/schematics/deluxe_5d3.pdf


The amp is loud and I wonder why I get so low rates when measuring the power:


Plate voltage 326V
Cathode voltage across 250R 15.7V
Measured voltage with 1000Hz sine wave across 8 Ohm load on clipping is 7V. So the power I get is 7*7/8 = 6.1W.


/Leevi

[HotBluePlates]

Try measuring the 6V6 grid voltage when the speaker is at 7v.  If the 6V6 grid peak voltage is much less than your ~15v of bias, maybe the paraphase inverter is distorting before the output tubes.

[Leevi]

If the 6V6 grid peak voltage is much less than your ~15v of bias, maybe the paraphase inverter is distorting before the output tubes.

I noticed that I had a wrong grid leak resistor 220K instead of 270K.
After I changed it the situation improved a bit but the power measurements show still very low values like ~7W.


The grid voltages at 7V across load have the following values:


-5.8V DC and -2.2V DC (I wonder the difference)
17.8 AC and 15.8V AC


/Leevi

[PRR]

Try going to the *edge* of clipping, not heavy clipping.

The G1 voltages should be nearly zero up until clipping.

[92Volts]

Are you measuring AC voltage and checking for clipping with an oscilloscope? All but the best multimeters suck for measuring AC, and the DC readings you're getting are likely influenced by the prescence of AC (as PRR says they really should be zero)

[Leevi]

Try going to the *edge* of clipping, not heavy clipping. The G1 voltages should be nearly zero up until clipping.

Yes, that's true, the voltages are on mV level.
If I put more gain the grid voltages start to go up and also the difference between the grids is then visible.

Are you measuring AC voltage and checking for clipping with an oscilloscope? All but the best multimeters suck for measuring AC, and the DC readings you're getting are likely influenced by the prescence of AC (as PRR says they really should be zero)

I'm using oscilloscope. I also tried with two DMMs and noticed that one of those starts to show wrong values when increasing the frequency.


/Leevi

[HotBluePlates]

If your a.c. volt figures were peak voltages, then you applied all the driving signal you can for the tubes to give clean output.  That makes me think the tubes are old/weak, there is something amiss with the OT, or the 6V6's are biased wrong.

Plate voltage 326V
Cathode voltage across 250R 15.7V
Measured voltage with 1000Hz sine wave across 8 Ohm load on clipping is 7V. So the power I get is 7*7/8 = 6.1W.

326v - 15.7v = 310.3v
15.7v / 250Ω = 62.8mA
62.8mA / 2 = 31.4mA
310.3v * 31.4mA = ~9.75w

Even if I misunderstood and your tubes have 326v plate-to-cathode, we're only up to ~10.2w and that assumes that all cathode current is plate current.

Idling cool will constrict the possible power output.  This is the answer when people ask, "but why are all Class A amps idling at 100%?"

[92Volts]

Idling cool will constrict the possible power output.  This is the answer when people ask, "but why are all Class A amps idling at 100%?"

Class A amps clip the instant the tube(s) cut off. But a push-pull amp can usually operate in class AB or B.

Bias might impact power, but it shouldn't be true that  idle current directly equals the available output swing (because hitting 0ma on one tube doesn't stop the other tube from continuing to conduct more).

I guess there might be issues with entering class AB operation, typically crossover distortion. but I'd lean towards an incorrect load impedance

[Leevi]

That makes me think the tubes are old/weak, there is something amiss with the OT, or the 6V6's are biased wrong.

Tubes are new and matched but of course they can be faulty/weak.
I have measured the OT primary couple of times and it seems to be fine about 90 Ohm/side.
Primary plate-to-plate impedance is 8000 Ohms.

326v - 15.7v = 310.3v15.7v / 250Ω = 62.8mA62.8mA / 2 = 31.4mA310.3v * 31.4mA = ~9.75w

So this is plate dissipation per tube?


This is not the first time I get lower power values.
I measured recently a Vox AC50 that should have 50W output power. I got at clean signal only 30W
with full gain (clipping) more than 50W. Does it tell something about my setup?


/Leevi

[92Volts]

If those AC voltages are peak-to-peak (not RMS or peak-to-ground) that's actually not enough driving voltage.

I think imbalance between the driving voltages technically creates distortion, but it wouldn't look like standard "clipping". So if you're looking at the waveform (rather than using a distortion analyzer) this wouldn't be what you're running into.

[Leevi]

If those AC voltages are peak-to-peak (not RMS or peak-to-ground) that's actually not enough driving voltage.

AC voltages I have stated are all RMS measured with DMM


/Leevi

[tubeswell]

Maybe try a 220R cathode resistor (or even 180R)

[Leevi]

Maybe try a 220R cathode resistor (or even 180R)

I tried with 200R but the effect was minimal.


Plate dissipation is ~10W/tube. What is its relation to output power is probably a more complex question?


/Leevi

[Leevi]

If I increase the screen voltage the output power reaches the nominal 12W power level.
The resistor between B+ and screen marked in the schematic is 10K. By lowering the
resistor value close to 1K I reach 12W or a bit  more.

/Leevi 

[tubeswell]

If I increase the screen voltage the output power reaches the nominal 12W power level.
The resistor between B+ and screen marked in the schematic is 10K. By lowering the
resistor value close to 1K I reach 12W or a bit  more.

/Leevi

So have you tried fresh tubes with the 10k screen dropper?

[Leevi]

Yes, the tubes are brand new. There was no difference with another tube set I tried.
If I leave this 10K there the output power will stay at 7W instead of 12W.


/leevi

[PRR]

> The resistor between B+ and screen marked in the schematic is 10K.

OK, that makes sense. At full power, screen current rises. Large resistor causes screen voltage drop. This reduces maximum plate current and power to load.

Remember this amp was not sold on "steady" power but "pluck" power. You say "The amp is loud". I assume on guitar. Between plucks the screen voltage is fairly high. You are probably getting 12W-14W at the start of a pluck. The 10K resistor wants to sag, but the 16uFd cap holds it up for a while (0.1-0.2 seconds). The power starts to sag, but also the plucked note is decaying to lower power, so the amp sag is not important in "loud but nice" (barely clipped plucks) playing.

On steady test-tone, yes, it sags. And this may also relate to low power rating speakers of 5D3 days, which might take 15+ Watt plucks but not 10 Watts steady.

And if you wanted a 25 Watt monster, you would not be building a 5D3, one of the older and lower-power DeLuxes.

As you see, a smaller screen dropper will give more power on steady overload, such as with fuzzes and compressors. This could be a switch option. Silicon rectifiers and fixed-bias can take you to even more power. That's the way Leo was going to stay ahead of his competitors. But that was all after 5D3.

[Leevi]

As you see, a smaller screen dropper will give more power on steady overload, such as with fuzzes and compressors. This could be a switch option. Silicon rectifiers and fixed-bias can take you to even more power. That's the way Leo was going to stay ahead of his competitors. But that was all after 5D3.

The power is not the most important parameter in this kind of amp. I was just interested in that because the amp has probably been specified for 10-14W?


http://ampwares.com/amplifiers/fender-wide-panel-tweed-deluxe/


I have to listen the amp more carefully with both power options and analyze the differences.
My fist feeling is that the low power option sounds better and is still loud enough.


The power switch is a good idea.


Thanks
Leevi

[PRR]

> specified for 10-14W?

Even into the late 1960s, many guitar amplifier power numbers were just guesses.

If the voltages were stable, 10-14W is what you expect for P-P 6V6. And you got this much from many hi-fi amps. (Hi-fi amps were often *tested*, not guessed.)

The 10K resistor gets the idle voltages near the tube-book voltages, but when driven MAX the G2 voltage drops below what the tube-book wants. And nobody cared. Because it IS loud in guitar use.

[Leevi]

Even into the late 1960s, many guitar amplifier power numbers were just guesses.

That's my feeling as well since the measurements I have done show almost always lower numbers than what they should be.


/Leevi

[Leevi]

I installed the high/low power switch on the down side of the chassis.
It changes the resistor as follows:


high power:


1st resistor 10K ->1K, 2nd resistor 10K ->22K


low power:
vice verca


I must say that the result is more or less theoretical.
The power increases by 5W but the change is difficult to hear.


/Leevi