What determines output power in a tube amp design?

[johnblue]

I have a pretty basic question that has been haunting me that I can't seem to find answers about.

Is the output power of a push pull amplifier determined/influenced/changed by the size/rating of an output transformer? If I were to use a 50w rated transformer (with the same resistance) in what is a 25w specified amplifier circuit would the actual power output increase at all or would I just get more headroom/fidelity ?

My gut tells me NO, the physical size or rating of transformer wouldn't "add output power" by itself.

In designing a custom PP amp, how much do you guys overshoot the the output transformer ratings? Let's say the data sheet for a tube lists output power a PP ab1 circuit At 25w would you feel comfortable with a 25W output transformer?

What happens if you slap a 75 w transformer into the same circuit?

Thanks in advance

 

[Selmerdave]

It is the choice of power tube and its operating conditions (power supply, output transformer impedance, screen operation, and bias) that determine the output power, assuming an adequate driver stage that drives the outputs to clipping. An OPT should not be operated past it's power rating or saturation or failure will result. It's power rating will not affect power output, just capability.

 

[kward]

how much do you guys overshoot the the output transformer ratings?

It depends on the make of the output transformer. Some are very conservatively rated, others are not. I've used Hammond output transformers quite a bit. They are accurately (not conservatively) rated. In the Hammond line, I like to overshoot the power handling capacity by 1.3x to 1.5x. If you want an amp that will do 25 to 30 Watts, get Hammond iron that can handle 40 watts.

The low end seems to suffer the most if output transformers are driven over wattage rating.

Larger transformer cores typically become harder to keep high frequency extension up there where you want it for a PP amp--unless extraordinary measures are taken in the winding technique. A good example of this is the Dynaco iron and Dynaco clone iron. The specs on the A431 for example say they will handle 60 watts down to 20 Hz. I think they will actually handle a good deal more than this. The amp I built using these transformers put out 60 watts, and low end extension was down at 16 Hz. Literature says they will do 120 watts from 30 Hz to 15 KHz!

 

[thorpej]

What determines the output power of the amplifier is Ohms law . Specifically, what is the voltage swing into a given impedance. For an example, let's consider a push-pull 6550 in AB1 ultra-linear mode at 450V idling at 75mA, with a 4.3K plate-to-plate load. The Class B portion of the load line reaches clipping at a plate voltage of 75V, representing a 375V swing for 1/2 of the output stage... which is 750V peak-to-peak when you factor in the swing of the other half, which is 530Vrms. Using Ohms law to compute the power: (530 * 530) / 4300 ~= 65W.

Now, because at the max output power only one tube of an AB1 push-pull pair is engaged at a time, computing this for the secondary is a little tricker. Assuming a 4.3K->8 transformer, the swing of just one half of the pair is 265Vrms, but it's coupled to the entire secondary winding... the equation is: SWINGsec = SWINGhalfpri * (Sqrt(Zsec / Zpri)) ... 265 * (Sqrt(8 / 1075)) ... 265 * 0.086 = 22.8Vrms, which into 8 ohms is: (22.8 * 22.8) / 8 ~= 65W

(Why 1075? Each half of a 4.3K primary OT is 2150 ohms when both tubes are conducting, and when only one tube is conducting, the conducting tube sees an impedance of half that again, or 1075 ohms.)

Over-rating the output transformer mainly gives you frequency response headroom. If, say, an OT is rated at 30Hz-20KHz @ 50W, you would expect the frequency response at 30W to be better (especially at the low end). Don't use an under-rated transformer for your expected output power -- the transformer will saturate or overheat and destroy itself.

 

[thorpej]

It depends on the make of the output transformer. Some are very conservatively rated, others are not. I've used Hammond output transformers quite a bit. They are accurately (not conservatively) rated. In the Hammond line, I like to overshoot the power handling capacity by 1.3x to 1.5x. If you want an amp that will do 25 to 30 Watts, get Hammond iron that can handle 40 watts.

Yes, I do the same with Hammond, and also Edcor. Specifically, with the smaller Edcors, I have had very good luck simply picking the "next higher" power rating and having good frequency response across the entire audio band.

 

[maxhifi]

Thing is; the output transformer is the most expensive single part of an amp, so over sizing it doesn't make a whole lot of sense for a new design. I like to design an amp to get the most out of an output transformer.

For a half assed automotive analogy, think of the output transformer as playing a similar role to a transmission.

 

[trainbuftony]

Technically the power tubes determine the maximum output power. Their rating sheet will give guidelines for how much power you can get through a specific load. You can usually get a little more. Thats not the whole story. What controls the output power in most of our old tube amps is the max current of our single rectifier. About 250 milliamps from a 5AR4. In my experience you rarely see dual triple or quad tube rectifiers on a power amp, and in order to get any real power, (above 20 watts per channel stereo) you need to have more than one, or go solid state... as far as dropping a 75 watt transformer into a 25 watt circut, you will notice more bass response but will most likely have to tweak the feedback to get it to sound best. Adding a 75 watt output transformer by itself will not get you any more real power.

 

[BinaryMike]

The Class B portion of the load line reaches clipping at a plate voltage of 75V, representing a 375V swing for 1/2 of the output stage... which is 750V peak-to-peak when you factor in the swing of the other half, which is 530Vrms. Using Ohms law to compute the power: (530 * 530) / 4300 ~= 65W.

1500Vpp is 530Vrms. There is 750Vpp at each end of the primary winding, but in opposing phase, so the total is 1500Vpp across 4300 ohms in this example.

Transformer low-frequency power handling capability is proportional to the square of frequency. For instance, a transformer rated at 100W at 80Hz could manage 25W at 40Hz with similar distortion.

 

[johnblue]

Thanks guys, I have a project that i put on the back burner because i didn't have the right output transformers, and now i have literally everything else, 2 Big power traffo's and 2 big chokes, and 10 NOS 1625 tubes and a bunch of 12sn7s and 12sl7s and all required hardware that is destined to become a pair of monoblocks.

so from here the next step is matching some output transformers to this power traffo and these tubes.

thank you =)

 

[GordonW]

It depends on the make of the output transformer. Some are very conservatively rated, others are not. I've used Hammond output transformers quite a bit. They are accurately (not conservatively) rated. In the Hammond line, I like to overshoot the power handling capacity by 1.3x to 1.5x. If you want an amp that will do 25 to 30 Watts, get Hammond iron that can handle 40 watts.

The low end seems to suffer the most if output transformers are driven over wattage rating.

Larger transformer cores typically become harder to keep high frequency extension up there where you want it for a PP amp--unless extraordinary measures are taken in the winding technique. A good example of this is the Dynaco iron and Dynaco clone iron. The specs on the A431 for example say they will handle 60 watts down to 20 Hz. I think they will actually handle a good deal more than this. The amp I built using these transformers put out 60 watts, and low end extension was down at 16 Hz. Literature says they will do 120 watts from 30 Hz to 15 KHz!

My experience with Hammond transformers almost exactly mirrors this. My most recent example was my Dyna "Mk 6L6" amps- which used a MK III chassis and driver board, with Hammond 1650P (6.6K UL, 60 watt rated), with 6L6GC output tubes, under about 450v B+, fixed bias, for a usable output of about 40 watts full-range. That worked VERY well.

FWIW: Comparing the MK III to the MK VI- I think that may be the biggest drawback of the larger Dynaco Mk VI amp (PPP 8417s, which are about like the driving force that could have been done with doubling up the KT88s in the Mk III)- the OPT seemed to NOT be as over-rated for its power (120w output) as the Mk III output transformer. And, under the loading it was in (lots of idle current, for one), it just, as a result, doesn't have the "effortless" character of the Mk III, within its rated power range. It's almost like they slightly upgraded a Mk III output transformer, and just said "good enough for government work", in essence. It works, but it just doesn't work as well as the smaller amps...

Regards,
Gordon.

 

[gadget73]

You also have to consider winding losses. Figure the transformer is somewhere in the 80-85% efficient range, so your 25 watts at the speaker terminals is closer to 30 at the plates. Makes the "over-rating" of the transformer seem a little less extreme.