- Posted by

In Reply to: ripple values in B+ supply posted by Antonio Tucci on September 2, 2006 at 14:07:19:

It depends on the power supply rejection of each stage's design. If you power the filament from AC, you can expect about 3mV peak to peak hum (1mV rms) at 120Hz on the 8 ohm tap from the heater. Assuming a 20:1 transformer (3200 ohm primary impedance) That would be 60mV peak to peak (20mV rms) at the plate. You would want to match that with the high voltage supply, at a minimum:1) 2A3 with series feed output transformer: the transformer impedance (3200 ohms) makes a voltage divider with the 2A3 plate resistance (800 ohms) so that 80% of the power supply ripple appears across the transformer. So the power supply needs less than 75mV pk-pk (25mV rms) ripple.

2) Parallel feed with the smallest practical choke, 10 henries: The choke impedance is 7500 ohms at 120Hz, and forms a voltage divider with the tube (800 ohms) in parallel with the load (3200 ohms). Maximum ripple would be 750mV pk-pk (250mV rms)

3) Parallel feed with a serious, high-inductance plate choke of 40 henries: Choke impedance - and power supply rejection - is four times higher, so you can get away with 3v pk-pk (1v rms) ripple.

A similar analysis holds for the driver stage, except the driver plate ripple is amplified by the 2A3. An amplification factor of 4 is close enough for estimating purposes.

Of course there are finer details if you want to get fancy, e.g. phasing the power supply ripple to cancel some of the filament ripple. This analysis just gives a general ballpark idea of where you want to be. Hope that helps!

This post is made possible by the generous support of people like you and our sponsors:

Follow Ups

Re: ripple values in B+ supply-Paul Joppa22:37:09 09/02/06(1)

- Many thanks -
Antonio Tucci09:51:26 09/04/06(0)