It's possible but I wouldn't recommend it. When the amp is turned on after it's been off the reservoir cap (while empty) looks like a short and causes the rectifier to conduct heavily until it's full. While the current through the rectifier during this period might not exceed the max values it's better to avoid that because it shortens the lifespan of the tube. What's your goal for a bigger capacitor? To achieve more filtering without burdening your power supply you can use multiple RC etc (or LC) stages before your plate supply. Or even use a choke input supply (where the choke comes before the reservoir) to reduce that in-rush current. Other arrangements using SS transistors are possible too but get more complicated quickly.
As for the NTC thermistors, I'd recommend a value between CL-60 and 90 (it's not strict, bigger ones will just slow things down a bit more). Make sure that they can handle the current of your amp. Wire them in series with the primary of your PT.
As for the NTC thermistors, I'd recommend a value between CL-60 and 90 (it's not strict, bigger ones will just slow things down a bit more). Make sure that they can handle the current of your amp. Wire them in series with the primary of your PT.
Thanks RTF671, I have three 680µF in series that mean 230µF after a bridge of 4 GZ34 with both plates in parallel (that mean that the current is =double) I got 48 units and I want in avoid to get more capacitors for the duty. I am not valve rectifier fan, when I discover silicon diodes (long time ago) I start to jump in one leg, that's mean that never thank again in uses an historical 5Y3 or 5U4 that was the valves that I knew in those times. Now a big and wisdom friend, recommend hardly in it use with my new project (Ongaku clone), because the project come with four of these demoniac devices, but recently in a Double mono valve amp from the fifties (Bogen db130a) that's come with two 5y3 in parallel, i change for only one 5ar4, and after two-three months one of the 5AR4 blew up in the starting (arcing hard) when I inspection the schematic looking for the failure, I discover that the in cap was replaced for 100µF on step the original one. BTW now I looking for the better choice, I got the valves, but they are they occupy considerable space in relation to what a simple solid state rectifier does not. Okay I deal with the principle in respect at maximum the circuit of the venerable Kondo San, and if I put rectifiers solid state have bound to change the power supply and the peak voltages would rise carrying out work on a different points from that assumed by the original designer.
My question is: If I uses a in rush like NTC or, a resistance in series between the rectifier and the reservoir, and this resistance bypassed for a time relay yielding charge the reservoir with out stress the rectifier by a current peak? In other words, with this measure, could me violate this limitations in the capacitor at the end of the rectifier?
Considering that the device work in Class A2.
Best Regards friends.
My question is: If I uses a in rush like NTC or, a resistance in series between the rectifier and the reservoir, and this resistance bypassed for a time relay yielding charge the reservoir with out stress the rectifier by a current peak? In other words, with this measure, could me violate this limitations in the capacitor at the end of the rectifier?
Considering that the device work in Class A2.
Best Regards friends.
Hi,
Max. filter capacitance allowed per GZ34 is 60 microF. (I've seen a lot more than that being used without problems but I wouldn't risk it with new production run GZ34).
Since your circuit (Ongaku?) uses 4 units that would make for 4*60 = 240 microF. At 230microF you should be safe.
I'm no fan of running PS caps in series, even with voltage sharing R it's rarely rock stable and that translates in the performance of the amp.
I'm well aware this is sort of unavoidable with amps a la 211, 845 and consorts.
A possible workaround may be to handpick the caps so they're all of the same C value and ESR.
Cheers,
Max. filter capacitance allowed per GZ34 is 60 microF. (I've seen a lot more than that being used without problems but I wouldn't risk it with new production run GZ34).
Since your circuit (Ongaku?) uses 4 units that would make for 4*60 = 240 microF. At 230microF you should be safe.
I'm no fan of running PS caps in series, even with voltage sharing R it's rarely rock stable and that translates in the performance of the amp.
I'm well aware this is sort of unavoidable with amps a la 211, 845 and consorts.
A possible workaround may be to handpick the caps so they're all of the same C value and ESR.
Cheers,
Thanks Frank for the information. I see more clarify the problem and the solution. BTW the caps are all the same, 680x400vcc, I need to put three in series with 510k in parallel. I hope they have the same ESR. But like you ask is hard to get a cap with this voltage rate.
A king regard for you
A king regard for you
As I understand it the limitation on the capacitance is not only for switch-on. It is more to limit the charging current peaks at every conducting half-cycle with a C-R-C filter. Thus a PTC resistor will ease matters at switch-on, but still allow charge peaks while on - limited of course by its hot resistance, which is usually small.
Also of some limited consolation is the fact that a transformer has inductance - this acts in series with the switch-on high charging peak to limit it. But again this inductance is damped as the transformer works under load. The limiting capacitor value is meant to be a safe figure, so yes - some tubes will work quite satisfactory with higher capacitances, but it is not recommended to abuse a design safety factor for that purpose.
Also of some limited consolation is the fact that a transformer has inductance - this acts in series with the switch-on high charging peak to limit it. But again this inductance is damped as the transformer works under load. The limiting capacitor value is meant to be a safe figure, so yes - some tubes will work quite satisfactory with higher capacitances, but it is not recommended to abuse a design safety factor for that purpose.
Hi,
In the old days series a resistor of a few hundred Rs was added to each leg of the secondary to limit the inrush current somewhat.
PTC are much better in that respect.
Another trick, which can't be applied here as you have the sharing Rs across the caps which double as bleeders, was to NOT use a bleeder across the cap.
A bit unsafe if you weren't aware of it and had to work on the gear but it was much kinder on the rectifier at turn on as the cap was often still charged to say 75%. Even after weeks.
I know, I shouldn't be telling all this so be careful.
Cheers,
In the old days series a resistor of a few hundred Rs was added to each leg of the secondary to limit the inrush current somewhat.
PTC are much better in that respect.
Another trick, which can't be applied here as you have the sharing Rs across the caps which double as bleeders, was to NOT use a bleeder across the cap.
A bit unsafe if you weren't aware of it and had to work on the gear but it was much kinder on the rectifier at turn on as the cap was often still charged to say 75%. Even after weeks.
I know, I shouldn't be telling all this so be careful.
Cheers,
Hi,
Esteban, in your circuit the caps after the rectifiers are mounted in series with a voltage divider resistive network so each cap sees a fraction of the total B+.
At the same time, once the amp is switched off, these resistor will discharge the caps since they're reference to chassis ground.
Short of replacing the series caps for a single one of appropriate voltage rating and value, there's nothing you can do to avoid this cycling process.
I doubt such a cap even exists so in your case (well, in case of the Kondo) you just have no choice.
So, yes, ideally each of the caps should always see the same voltage as the others do but that rarely happens.
This can lead to over voltage on one of the caps which in turn can lead to failure.
One way to minimize this risk is to use caps that when combined in the series string arrangement, still have a high enough voltage rating to be on the safe side.
For example, supposing total B+ across the series of caps is 1200 VDC, you then rather use 3 x 450VDC rated caps than ones rated for 400VDC.
Either way, the voltage dividing network is mandatory.
Cheers,
Esteban, in your circuit the caps after the rectifiers are mounted in series with a voltage divider resistive network so each cap sees a fraction of the total B+.
At the same time, once the amp is switched off, these resistor will discharge the caps since they're reference to chassis ground.
Short of replacing the series caps for a single one of appropriate voltage rating and value, there's nothing you can do to avoid this cycling process.
I doubt such a cap even exists so in your case (well, in case of the Kondo) you just have no choice.
So, yes, ideally each of the caps should always see the same voltage as the others do but that rarely happens.
This can lead to over voltage on one of the caps which in turn can lead to failure.
One way to minimize this risk is to use caps that when combined in the series string arrangement, still have a high enough voltage rating to be on the safe side.
For example, supposing total B+ across the series of caps is 1200 VDC, you then rather use 3 x 450VDC rated caps than ones rated for 400VDC.
Either way, the voltage dividing network is mandatory.
Cheers,
Thanks Frank, I guess so, but I understood that you found another way. We agree in put the series equalizer voltage resistors.
The highest voltage of the transformer is 760 AC, and maximal DC voltage is 1072 but supposing that we uses silicon diodes. With the rectifier and a constant idle current of 140mA the voltage of them never rise until the 1200, Do you consider anyway in rise the capacitance voltage? The original Ongaku uses two caps of 550voltDC.
Best regards and thanks for break you head in my poor cause.
The highest voltage of the transformer is 760 AC, and maximal DC voltage is 1072 but supposing that we uses silicon diodes. With the rectifier and a constant idle current of 140mA the voltage of them never rise until the 1200, Do you consider anyway in rise the capacitance voltage? The original Ongaku uses two caps of 550voltDC.
Best regards and thanks for break you head in my poor cause.
Hi,
I Just looked at a copy of the Ongaku's diagram.
How exactly are these 2 caps wired?
If I read it correctly then there are two dual section caps of 100microF/500VDC each.
Assuming both sections are //ed then that would give a total input C of 200microF/100VDC, correct?
The plates of each 5AR4 are //ed as well. No idea how that would allow high input C though.
http://musicangel.ru/messpic/mess124pic03.gif
Cheers,
I Just looked at a copy of the Ongaku's diagram.
How exactly are these 2 caps wired?
If I read it correctly then there are two dual section caps of 100microF/500VDC each.
Assuming both sections are //ed then that would give a total input C of 200microF/100VDC, correct?
The plates of each 5AR4 are //ed as well. No idea how that would allow high input C though.
http://musicangel.ru/messpic/mess124pic03.gif
Cheers,
Hi Frank again, Thanks
Yes this is the same copy, in mine, the mid point in the secondary of high voltage is not connected to the half, In my case I rearrange another bank of caps for the +460.
The rest is the same. I will put many soft start and the rod coleman regulator (current) for the DHT valves. The essence of the circuit is a clone. For the maximum current in the rectifiers the plates are in parallel and the bridge is in series for two valves per cycle. I will use 680*400/3=223*1200. In the original circuit the rectifier see 100µf. (100+100*500/2=100*1000).
Respect to the maximum capacity supported for the rectifiers depends basically in the first charge of the in capacitor, the question is, if I put a NTC or a resistor (bypassed by a relay) with the object of avoid the initial peak could resolve the problem? The circuit work in class A, for this reason there are no peaks of currents that will discharge the caps.
Yes this is the same copy, in mine, the mid point in the secondary of high voltage is not connected to the half, In my case I rearrange another bank of caps for the +460.
The rest is the same. I will put many soft start and the rod coleman regulator (current) for the DHT valves. The essence of the circuit is a clone. For the maximum current in the rectifiers the plates are in parallel and the bridge is in series for two valves per cycle. I will use 680*400/3=223*1200. In the original circuit the rectifier see 100µf. (100+100*500/2=100*1000).
Respect to the maximum capacity supported for the rectifiers depends basically in the first charge of the in capacitor, the question is, if I put a NTC or a resistor (bypassed by a relay) with the object of avoid the initial peak could resolve the problem? The circuit work in class A, for this reason there are no peaks of currents that will discharge the caps.
Last edited:
Hi,
In theory the PTC thermistor should protect the rectifiers from seeing too high inrush current but I just feel a bit uncomfortable recommending them in someone else's amp as I don't have any real life experience with such a type of circuit as this Kondo amp.
Fact is that you're close to twice the max. allowed value which is a lot.
Also, keep in mind that if you're going to use three caps instead of two you'll need to use 2*510k= 1200K > 1200/3 = 3*400K resistors.
This is a pretty high value so the caps will retain their charge for awhile after switch off.
All in all wouldn't you be better of sticking to the original 500V caps? (JJ or F&T should carry these values)
You could still use the PTCs which are likely to prolong the rectifier's life anyhow.
Best,
In theory the PTC thermistor should protect the rectifiers from seeing too high inrush current but I just feel a bit uncomfortable recommending them in someone else's amp as I don't have any real life experience with such a type of circuit as this Kondo amp.
Fact is that you're close to twice the max. allowed value which is a lot.
Also, keep in mind that if you're going to use three caps instead of two you'll need to use 2*510k= 1200K > 1200/3 = 3*400K resistors.
This is a pretty high value so the caps will retain their charge for awhile after switch off.
All in all wouldn't you be better of sticking to the original 500V caps? (JJ or F&T should carry these values)
You could still use the PTCs which are likely to prolong the rectifier's life anyhow.
Best,
Basically people are very gullible, many tubes are way over priced and if someone wants to pay what the seller is asking then the seller is happy for sure, but there are limits of price for everything. For the price of 1900.00 you could easily re-design the power supply using solid state components c/w a time delay and soft start system for applying the HV and still have money left over.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.