It mostly has to do with how the transformer is wound, and comes from differences in the resistance and lekage inductance of the winding halves (somewhat less depends on interwinding capacitances). There will be very little if the secondary is bifilar-wound for a FWCT but for high voltages this may be impossible beacuse of the voltage differential between the adjecent windings - it is fairly easy to do on low voltage transformers. On EI transformers a split bobin can be used to get around that problem although it increases the stray magnetic field somewhat.
On a FWB set-up, it's always the same winding used for both halves of the waveform, so assuming the bridge diodes are all reasonably matched, there is less difference between the halves of the mains sinewave.
On a FWB set-up, it's always the same winding used for both halves of the waveform, so assuming the bridge diodes are all reasonably matched, there is less difference between the halves of the mains sinewave.
May I ask maybe a stupid question ?
When I have look into the 5u4G, I find typical max AC Vrms specs of 550V.
Well, I need 700VDC...and would like to used the 5U4g still if possible.
Now, when reading the RCA data sheet a bit more in detail, I find as well a spec. which says 1100V plate to plate voltage....
Hmmm...what does this mean if I would use a normal FW, but with two 5u4g with their Anodes in parallel ? Would this not take a lot of stress away from the individual tube and allow for a higher voltage than 550Vrms AC input ?
When I have look into the 5u4G, I find typical max AC Vrms specs of 550V.
Well, I need 700VDC...and would like to used the 5U4g still if possible.
Now, when reading the RCA data sheet a bit more in detail, I find as well a spec. which says 1100V plate to plate voltage....
Hmmm...what does this mean if I would use a normal FW, but with two 5u4g with their Anodes in parallel ? Would this not take a lot of stress away from the individual tube and allow for a higher voltage than 550Vrms AC input ?
Plate-to-plate voltage is for a CT transformer secondary, which was the only option when tube rectifier specs were written. 1,100 VCT gives you 550 input volts per plate.
For 1,000 V you should use TV damper diodes.
For 1,000 V you should use TV damper diodes.
Well, I was hoping that Vinvp would double if one dual diode is used as half-wave by paralleling the anodes...
...when looking at the Ongaku, the GZ34 delivers 960VDC, but this is in a bridge configuration, not two-phase...
...when looking at the Ongaku, the GZ34 delivers 960VDC, but this is in a bridge configuration, not two-phase...
bridge wins hands down....
since transformer utilization is higher....
i stopped using the full wave center tapped traffos....
an not going back there...
imagine a winding doing nothing at each half of the electrical cycle....
my NS audio handbook puts it a 30 percent more available power...
I opt for bridge too. But I still like the tube rectifiers, even though it's less practical.
By adding 2 carbide diodes I make it a bridge and keep that naturally delayed B+ feature.
The further tweak is to shunt the tube rectifier with 2 diodes and a 20K resistor in series, that eliminates the cap charge surge and well-serves for a standby mode. Just change that 20K to a value where the output stage is a bit beyond conducting the quiescence. Then, for the standby mode you interrupt the rectifier's filament by e.g. an optorelay.
By adding 2 carbide diodes I make it a bridge and keep that naturally delayed B+ feature.
The further tweak is to shunt the tube rectifier with 2 diodes and a 20K resistor in series, that eliminates the cap charge surge and well-serves for a standby mode. Just change that 20K to a value where the output stage is a bit beyond conducting the quiescence. Then, for the standby mode you interrupt the rectifier's filament by e.g. an optorelay.
No, because you still have to keep the anode-cathode reverse voltage low enough.Blitz said:
Putting anodes in parallel may increase the current handling, although not by much.
A hot turn-on would still typically cause cap in-rush current, and so impact the choice of valve rectifier (for its turn-on surge current rating), and of protection fusing on the power transformer primary and secondary sides.
Blitz, you can 'extend' the max secondary Vac rating of a valve such as the 5U4 by inserting at least 2 1N4007 diodes in series with each 5U4 plate. This provides more assurance that the 1550V PIV voltage rating of the valve diode can be sustained, and can make the power supply more bullet-proof in years to come when the 5U4 goes a bit gassy (and loses its PIV capability).
I wouldn't push the Vac rating too much higher than datasheet 550Vac, but 600Vac would appear reasonable, and would provide circa 700Vdc.
You can also move to the 5AR4, to get that bit higher PIV rating, and a slow start cathode - but maybe you don't have that option. Even with that option, the addition of series ss diodes to each valve plate provides more assurance that the valve will survive for longer.
Of course, the valve base and all associated wiring and insulation etc has to also be rated for such levels, as for example 250Vac cable insulation is not pedantically enough.
I wouldn't push the Vac rating too much higher than datasheet 550Vac, but 600Vac would appear reasonable, and would provide circa 700Vdc.
You can also move to the 5AR4, to get that bit higher PIV rating, and a slow start cathode - but maybe you don't have that option. Even with that option, the addition of series ss diodes to each valve plate provides more assurance that the valve will survive for longer.
Of course, the valve base and all associated wiring and insulation etc has to also be rated for such levels, as for example 250Vac cable insulation is not pedantically enough.
- Status
- Not open for further replies.