PSE to SE - Pros and Cons

[soulminer]

What effect would pulling a valve from a PSE design have on the output transformer/circuit/sound ?? I ask, as I have a 300B PSE, and have only ever run it with both valves in. If I removed one of the 300Bs and ran it as a SE, surely this would be detrimental, as the output transformer is specced, wound and gapped to work at its optimum with the both valves fitted (ie double power), as is the power supply ?? Would it damage anything else in the circuit to remove one of the valves to try this, and has anyone actually run a 300B PSE this way ?? I know a few have run the 2A3 PSE like this. Any detrimental effects from doing so ?? If not, do you pull the valve with the humbucker, the one without, or doesn't it matter ?? I have never tried it, but wondered what it would sound like, but obvioulsy don't want to wreck my amp, or overstress the remaining valve. If it is simple and harmless and produces good results, then why aren't all SE amps built as PSE, and it is up to the owner how many valves they use, depending on music or mood ?? I suspect that if the output transformer is designed for PSE operation, then running it in SE is not a good idea (and if it does sound better, then surely there is something wrong with the design of the transformer??), although it may give the mains transformer an easier time of things. I await your educated responses to reduce my ignorance on such matters. Cheers.

[Ianm2]

AS a very general rule of thumb,

I always refer to a 300b as I am so familiar with them.

The impedance that the valve wants to see, typically is anything from around say 2 k ohms to 6 k ohms.

The lower the impedance, the more power the valve can develop.

The higher, the less.

But, the higher ,usually the less distortion.

So frequently you find 2k5, gives around 9 watts but around 6% 2nd harmonic distortion.

If you used a 5 k transformer, you may get 6 watts at perhaps 3-4% 2nd harm.

So an optimum is perhaps 3k5.

Now the effect of running pse, you would half your transformer, so if you used 3k for one 300b, you would be using around 1k5 for a pse.

normally, pse is expected to double the output power, my simulations on tube cad, really only reveal a handful more watts, which is interesting, tubecad is very accurate.
so you may only get 12 watts out instead of the expected 18 or so

Now if you remove one of your 300b's, the valve will see 1k5 instead of the optimal 3k5, so you may get a tad extra power, but distortion will now rocket up.

The extra power you would get paralleled all comes from extra current, not volts if I am correct, and I have a bizarre theory valves sound better because of volts, not current, which transistor amplifiers are more current.

So pull one valve, and you get less current, ergo sounds better, a more contentious opinion would be we get lots of nice sounding distortion.

Or perhaps we don't have a clue....

It makes life easier on the power supply as the big output stage that demands most of the current is only taking half what its specced for, assuming its adequatly designed.

It also makes the previous stage slightly happier, as a paralled output is like 2 capacitors, and needs more electrons(current) to charge(fill up the metal area) on the plates. This is what ease of drive means.

interstingly, output transformers have an operating point, and it should lower distortion in the output transformer at high powers, when it begins distorting due to saturation by current, but not sure how it will affect the air gap

The interaction of the valve with the transformer and speakers is quite fascinating, not appreciated and not understood, and I don't know the half of it.

There are other things like output impedance, damping factor, primary inductance, which iirc, will be less for a pse transformer, so normally bass won't be optimal which may be good, as large signals, ie bass cause distortion.

Its REALLY complex.

As an extra, bear in mind even if the power doubles to 18 watts instead of 9, this will only add 3 decibels to your listening level,

so if you have 92 db speakers, you will get out of a 9 watter

92db= 1 watt
95db=2 watt
98 db=4 watt
101 db=8 watt
104 db=16 watt

about 101 dbs max with the single valve, and 104 dbs with the 2 valves.

18 watts however looks great, 9 looks pretty poor, so perhaps that's why we see a few commercial pse amps, as power sells, and there really aren't many domestically acceptable sensitive, above say 95 db speakers.

[soulminer]

Cheers Ian. Quite enlightening. In theory then, pulling one of the ouput valves will have no damaging effects on the amp itself, but will increase the distortion. I may have a fiddle later on, just to satisfy my curiosity, and see what it sounds like. Thanks.

[Paul Barker]

filament voltage on remaining 300b will lift as will HT, would need to adjust if making change permanent.

[soulminer]

I realise the HT would probably go up, but why would the filament voltage go up Paul, as the ouput valves (in WAD 300B PSE anyway) are supplied by seperate filament windings, rather than being parallel supplied from one winding ?? Surely, the 5v will stay the same regradless of having one or two valves fitted, if using seperate windings ?? How much do you think the HT would go up by ?? A damaging amount ?? Could this be calculated before I do anything ??

[NickG]

The load on the transformer will decrease, partly because of HT and partly because of the missing fillament. The decreased load will mean higher secondary voltages, how high depends on the regulation of the transformer.

I doubt it will hurt to have a go. Try it, if you like the sound, then measure the voltages, and add dropper resistors as required to get the voltages back as normal. Of course before you do that you won't know if any change in sound was due to the increased voltages

I have only heard it done with a 6em7 amp, but in that case there was a improvement to my ears. There was a long a contentious thread about this on the old BB, and it was suggested there that whilst tubecad may be of use, reality will always do what it wants despite any attempt to model it.

Ian, have you any idea why two valves in parallel would fail to give increased power (ie about twice), I am not sure if I can think of a reason assuming all things being equal. What sort of models does tubecad use for its valves, are they anything like the common spice models?

So frequently you find 2k5, gives around 9 watts but around 6% 2nd harmonic distortion.

If you used a 5 k transformer, you may get 6 watts at perhaps 3-4% 2nd harm.

I think it should be pointed out that those distortions figures are at the stated power, what distortion figure does 2k5 give at 6w?

There is a good artical about the effect of load lines and the calculation of distortion from them by Steve Bench.

http://members.aol.com/sbench102/po-dis.html

In fact every thing by Steve Bench is worth reading.

Does the tubecad model use a resistive or reactive load for its calculations?