Experiments with power regeneration

[analog_sa]

I have been trying to build a nice power supply for my Garrard 301 for a while. It's had its ups and downs but overall the sound improvement compared to being powered from the wall is very worthwhile. Initially the sine wave was generated by a Wien bridge, but more recently i started using an 8515 Atmel which uses a look-up table to generate a sine by 100 points. The convenience of the Atmel, being of course easy changing of frequency and the option to configure a second port for 90deg phase difference, thus making it compatible with synchronous motors as well.
The generated sine is amplified by a single 3886 which in turn feeds a step-up transformer 1:12 (220v, 50Hz here). The Garrard seems to run happily at around 180v. I could make life a bit easier by setting the Garrard to 110v operation and thus reducing the step-up requirements but that will make it incompatible with my other turntables.
Today i decided to see how good is the regenerated power in objective terms and had an unpleasant shock. At 195v at output the third harmonic reached 1.8% (compared to 2.9% directly from the wall) and the 5th looked almost equal. The output of the 3886 at this stage has only 0.05% third harmonic so most of the distortion seems to be generated in the transformer core. The measurements were taken with the transformer unloaded since it seems the distortion is not influenced by the loading so much.
The transformer i use is a simple 60VA step down connected in reverse.
Now, i hate to admit, but my transformer related knowledge leaves a lot to be desired Could it be that i'm exceeding the maximum flux density for the core? I briefly tried a 300VA transformer and got even worse distortion. Can't make any sense out of this.
Any suggestions sincerely appreciated.

peter

[analog_sa]

Most of the above turned out to be bull. In my setup i had a puny step-down transformer feeding the soundcard. Of course it turned out to be the main contributor of third harmonic. For those interested here are measurements taken without the step-down (just using an attenuator)

182v - 0.13% second harmonic at o/p unloaded
177v - 0.38% second harmonic at o/p loaded with garrard 401 (12W) the o/p drops by 5v when loaded

the distortion at o/p is predominantly second harmonic and almost exclusively generated by the step-up transformer, but still quite acceptable. mains power otoh shows a max on fifth harmonic at 2.4% (it's 22.30h here so mains is realtively clean)

It seems like the power regeneration works reasonably well with a step-up transformer, but for turntable supply it's probably a better solution to use bridged STK modules and drive the motor directly. Output impedance will also be lower. Obviously this will only work if the motor is set to 110v.
How difficult would be to build a discrete high voltage power amp which can generate 200v rms @ 30W? And one which presumably does not disintegrate at the first short in the output.

peter

[P.Lacombe]

As you have seen, you cannot obtain low distortion with a mains transformer, even very big, because this devices are not designed for low distortion, they are constructed in such way that flux density in the core is high.

To obtain lower flux density, and therefore lower distortion (specially third, fifth harmonic) you have to use such transformer at lower voltage.

By instance, a specified 24 v/220 v transformer can be used to step-up from 12 v to 110 v, but at the expense of half usable power.

Low second harmonic distortion can be attained by carefully adjusting dc balance in the primary winding (=offset of the amplifier).

Regards, P.Lacombe.

[analog_sa]

DC offset is about 3mV. Would the avove mean that low step-up ratio is to be prefered? Then obtaining a higher voltage out of the amp, say by bridging, will be a sensible route.

peter