Tube Rectifier Capacitance?

[kipman725]

I'm looking to replace the silicon rectifier in the following project:
http://www.headwize.com/projects/sho...ammer1_prj.htm
with a tube rectifier using the 5U4G full wave rectifier as 4 never used specimens have come into my possession. Looking at the data sheets apart from the increased voltage drop compared to silicon diodes they also appear to have a maximum "input" capacitance and input choke. I'm not quite sure what this is refering to, is it refering to the output of the rectifier? or to the input which is strange as I wouldn't normally put large value capacitors across an AC supply. Can I just increase the transformer voltage and directly substitute the rectifier?

[Ty_Bower]

First, go here and download PSU Designer II:
http://www.duncanamps.com/psud2/index.html

The data sheets are referring to the capacitor that is immediately downstream of the rectifier tube, i.e., the input to the rest of the power supply circuit. Most vacuum tube rectifiers won't tolerate much more than 40 uF to 50 uF. They will arc out during startup as the cap tries to quickly charge, ruining the tube.

[kipman725]

thanks for clearing that up; so with the linked HT supply I should be able to just substitute my valve rectifier in as the capacitors are initially charged through a 680ohm resistor. I am slightly confused how 230VAC(RMS) becomes 300VDC as even using silicon rectifiers it should be at maximum approximately 276VDC. However the rest of the design looks sane so I will just use a higher voltage transformer.

[leadbelly]

Originally posted by kipman725
thanks for clearing that up; so with the linked HT supply I should be able to just substitute my valve rectifier in as the capacitors are initially charged through a 680ohm resistor.

Not sure that I understand your post. If you mean you can do a straight substitution, then no. As Ty pointed out, there is no way you can put 900 uF capacitance on the 5U4G, even with the 680 ohm resistor, 40 to 50 uF is the max.

I'm not sure where you got that 276 VDC from. With silicon rectifiers, the maximum is more like 230 x 1.414 - 1.4 = 324VDC. Of course with all the extras in the schematic that is not the real world voltage of the supply.

[kipman725]

I just educated myself a bit and now I understand the issues involved. Feel like a bit of a noob now. I have always wondered why I always got high voltages than I expected out of silicon rectifiers in my power supply designs, no one had ever explained that I got the peak voltage and not the RMS. I also thought the tubes had a bridge rectifier on them not a full wave. I also now get the flash over problems etc. If I use a large enough value resistor so that the charge current is less than the maximum current of the tube at all times can I charge as large capacitor bank as I want? (I would disconnect the amplifier from the supply at this point). I ask because I guess with the additional sag of a tube rectified supply I will be at approximately 290V (although I will have to check this before hooking up the amp with a dummy load) which is an acceptable plate voltage and so would like to use a capacitance multiplier. Or if that's possible I could just go with a large capacitor bank (I could get 3200uF for $60).

Also I have some questions about the transformers, if I build a full wave rectified tube supply does the heater transformer have to be rated at greater than the current required for the heater as one of the dc power rails is connected to it? (as per diagram here http://en.wikipedia.org/wiki/Rectifier). Also If I use two 5U4G in a bridge rectifier then is the output current through each tube halved and therefore the ratings doubled?

I am having trouble finding schematics of tube rectified supplies it would be useful to see some if anyone knows where to look.

[analog_sa]

Originally posted by kipman725
If I use a large enough value resistor so that the charge current is less than the maximum current of the tube at all times can I charge as large capacitor bank as I want?

Yes, you can. I once used a 3300uF in a tube rectified PS. An easier approach is to use a choke instead of resistor or even a choke input filter as it will limit max current. In any case PSUDII will provide the fastest answers.

[kipman725]

yes I am starting to understand this more I have found this schematic here: http://www.emissionlabs.com/datasheets/EML5U4G.htm
which uses a PI filter that looks quite elegant and would avoid having to use the timer circuit I am having to use the PSU software in wine (Linux user) but it appears to work fine and be a neat bit of kit. The given schematic has too large filter capacitors and too low transformer winding resistance for the 5U4G?