Mains supply results (50Hz) are included but not likely to be relevant as IIRC only early Gyrodecks had a mains supply from a 24V transformer. All other tests are at 25Hz as this was what Voyd used to halve the nominal 4 pole speed of 1500rpm to 750rpm. Plots captioned.
.
Clean 25Hz PSU with 23.5uF capacitor and reduced voltage (16Vrms for Voyds)
.
As has been mentioned on a forum it appears we did not adjust the phase capacitor correctly for early units (1985-1987). The voltage was correctly reduced for 25Hz operation but the capacitor should have been increased. It makes a small difference of ~10dB (Right or 3rd plot) but nowhere near how the split phase improves rotational speed. This behaviour correlates well with the absolute shaft speed measurements below.
.
Split Phase PSU – phase shifted by ~90 electrical degrees at the small signal level and then driven to the motor (16Vrms for Voyds, 14Vrms for Valdi’s). The smoothest operation was with the 90 deg shifted phase at approx 1V less than the main phase. This is how we shipped all units from 1988 on along with many PSU upgrades for early customers. It seems the split phase approach was the correct way forward.
.
Adjustable Phase PSU – small, 1 degree increments of phase allow a new ‘null’ to be reached at around 85 degrees. Notice how the peak at 50Hz is reduced by over 10dB and also a 25Hz reduction. Z axis – stray flux is remarkably low.
.
We now have the absolute shaft speed measurements. First is a combined plot showing how the B112 motor improves from 23.5uF phase cap to 47uF to Voyd Split Phase PSU to a further Tuned Phase approach. Also included are FFT’s of the two best approaches. The plots speak volumes and confirm the accelerometer plots but now with an absolute measurement of speed variation within one revolution…..
.
In the zoomed in plot below there is some small lower frequency variation (12.5 & 50Hz) – possibly due to small residual distortion in the sine wave or drive amplifier. The periodic nature of the Tuned Phase plot is 100Hz but at a very low level.
.
FFT’s – Amplitude scale is number of counts (50nS = 1/3200counts per encoder pulse) 1 count = 0.23rpm
.
.