The reason that the current output is dependant on the Cap value is because two of the caps are in series with the load. The DC is pushed around the cap through the diode in parallel with the cap by the next 1/2 cycle of that same phase. This circuit is generally used for high voltage and low current and at high frequency. At 60 HZ ac line and the voltage and current you are talking about the cap values required would be VERY LARGE. Remember that each diode in series (3 on the + side and 3 on the - side) with the load will have a .7 to 1 volt drop as well. Not a problem when your talking about 1000 volts or so but suddenly a big deal at 18 volts.
Have fun any way you can
BZ
Have fun any way you can
BZ
Well, it IS much easier, if it is a toroid. It's a bit tougher with an E-I core!
Anyhow, the voltage drop is maybe 3 volts yes? less for bigger diodes... but the page claims Vout = 2.8xV, which makes it less of a big deal. Without even putting the secondaries in series you get 47V. Watch out with secondaries in series!
(100V! )
anyhow, I know they would have to be some big fat caps, but 10,000uF is something some of us don't think twice about. For 10A (Nice round number) what do you need for a low-ish ripple voltage? If it's 10,000uF, that's fine. If it's 100,000uF, that may be a bit much for some people. Now I'm getting curious.
Anyhow, the voltage drop is maybe 3 volts yes? less for bigger diodes... but the page claims Vout = 2.8xV, which makes it less of a big deal. Without even putting the secondaries in series you get 47V. Watch out with secondaries in series!
(100V! )anyhow, I know they would have to be some big fat caps, but 10,000uF is something some of us don't think twice about. For 10A (Nice round number) what do you need for a low-ish ripple voltage? If it's 10,000uF, that's fine. If it's 100,000uF, that may be a bit much for some people. Now I'm getting curious.
For 10 A 100,000uf would not be out of line. Then you need to worry about the ripple current spec for the cap as well because the full 10 amp load current will plus ripple current will be passed by the cap. The diodes will need to be able to handle the current spikes, and they will be very large. The DC resistance of the transformer will be the limiting factor for the current. I would not be surprised if these spikes are more than 50 amps for a 10 amp supply in this configuration. The diodes will have a higher voltage drop because of the high current.
So, lets see here 2 18 volt AC secondaries in parallel to handle the current spikes x 2.8 = 50.4 - 6 diodes x 1volt, voltage drop at the diodes = 44.4. The other option is 2 18 VAC secondaries in series = 36 x 1.4 = 50.4 - 2 diodes x 1 volt each in series for a fullwave bridge = 48.4.
It just doesn't seem worth it to me. That circuit is for high voltage applications, the example was for a 2800 volt power supply. 1000 volt transformers are very expensive and hard to find and 2000 volt ones for the supply in the example will put you in shock (more ways than one) so this circuit makes sense for a big power supply for a high power tube RF amplifer, but not for SS audio. By the time you get the caps and diodesto make it work you could have gotten a transformer that would do the job for much less money.
So, lets see here 2 18 volt AC secondaries in parallel to handle the current spikes x 2.8 = 50.4 - 6 diodes x 1volt, voltage drop at the diodes = 44.4. The other option is 2 18 VAC secondaries in series = 36 x 1.4 = 50.4 - 2 diodes x 1 volt each in series for a fullwave bridge = 48.4.
It just doesn't seem worth it to me. That circuit is for high voltage applications, the example was for a 2800 volt power supply. 1000 volt transformers are very expensive and hard to find and 2000 volt ones for the supply in the example will put you in shock (more ways than one) so this circuit makes sense for a big power supply for a high power tube RF amplifer, but not for SS audio. By the time you get the caps and diodesto make it work you could have gotten a transformer that would do the job for much less money.
4 times the caps
I like the input filter doubler circuit, input filter has to be 1 vac peak wvdc and the output filter has to be twice. That circuit also leaves one side of the transformer grounded. It is only 30 cycle so you need 4 times the capacitance over a FWB. Assuming that 60hz is your mains.
Or just toss the doubler out and go with a PWM DC to DC converter running at 100kHz and you only need 1/800th of the cap value over a 120 cycle PS. And a big azz choke.
Here's a diagram:
I like the input filter doubler circuit, input filter has to be 1 vac peak wvdc and the output filter has to be twice. That circuit also leaves one side of the transformer grounded. It is only 30 cycle so you need 4 times the capacitance over a FWB. Assuming that 60hz is your mains.
Or just toss the doubler out and go with a PWM DC to DC converter running at 100kHz and you only need 1/800th of the cap value over a 120 cycle PS. And a big azz choke.
Here's a diagram:
Attachments
Stocker
With any voltage doubler the problem of ripple current comes up. Just any good cap is not the issue. It takes a cap designed for high ripple current. There is no two ways about this. Never under estimate the stress high ripple puts on the caps in the circuit. The big caps that you find around these days are very good and you are correct that the cost is much less than in days gone by. That does not change the point of this posting however.
RIPPLE CURRENT, RIPPLE CURRENT, RIPPLE CURRENT
Have you ever seen a 47,000uf 63VDC cap vent? I have. It was in a voltage doubler (the 2 diode type) power supply with a 25 volt transformer and about 35 VDC on each cap. Nice little class A amp with about 2.5 amps idle current per channel. The cap were rated for 5000 hours at 105 C. Temp in the case of the amp was about 45 C. After replacing the first set of caps that sprayed sticky goo all over, I measured the temp of the caps at .5 amps with a load resistor and the rise was less that 10 degrees. Then I fired up the amp again and measured the temp of the caps. After 1 hour the caps were at 115 degrees C but the case temp was still 45 C. Ripple current will get you. I found some smaller value caps with a high ripple current rating and paralleled those and now the temp rise in the caps is much better, about 15 degrees C above the case temp.
Have fun: BZ
With any voltage doubler the problem of ripple current comes up. Just any good cap is not the issue. It takes a cap designed for high ripple current. There is no two ways about this. Never under estimate the stress high ripple puts on the caps in the circuit. The big caps that you find around these days are very good and you are correct that the cost is much less than in days gone by. That does not change the point of this posting however.
RIPPLE CURRENT, RIPPLE CURRENT, RIPPLE CURRENT
Have you ever seen a 47,000uf 63VDC cap vent? I have. It was in a voltage doubler (the 2 diode type) power supply with a 25 volt transformer and about 35 VDC on each cap. Nice little class A amp with about 2.5 amps idle current per channel. The cap were rated for 5000 hours at 105 C. Temp in the case of the amp was about 45 C. After replacing the first set of caps that sprayed sticky goo all over, I measured the temp of the caps at .5 amps with a load resistor and the rise was less that 10 degrees. Then I fired up the amp again and measured the temp of the caps. After 1 hour the caps were at 115 degrees C but the case temp was still 45 C. Ripple current will get you. I found some smaller value caps with a high ripple current rating and paralleled those and now the temp rise in the caps is much better, about 15 degrees C above the case temp.
Have fun: BZ
What you are telling me is: I am on the right track for reasons I didn't know about!
My little (maybe 80-100VA ) transformer goes BAAAAAA when it's plugged in... I got the expected output voltage, but I wound the secondaries with some FAT (like 2mm dia. ) wire, and didn't secure it at all... guess what the noise is...
anyhow I didn't like the idea of using a buzzing secondary for a power-ON indicator, so...
so instead of dropping $$ on some new high ripple current caps, I'll just do the sensible thing, get some thinner wire and rewind to the appropriate voltage, obviating the need for a doubler.
Thanks, you just saved me watching my work go up in smoke!
(literally)
My little (maybe 80-100VA ) transformer goes BAAAAAA when it's plugged in... I got the expected output voltage, but I wound the secondaries with some FAT (like 2mm dia. ) wire, and didn't secure it at all... guess what the noise is...
anyhow I didn't like the idea of using a buzzing secondary for a power-ON indicator, so... so instead of dropping $$ on some new high ripple current caps, I'll just do the sensible thing, get some thinner wire and rewind to the appropriate voltage, obviating the need for a doubler.
Thanks, you just saved me watching my work go up in smoke!
(literally)
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.