Hi,
I'm putting together a HV power supply and I'm wondering if any of you have any opinions on which diode I should go for. I was planning to us IRF, since Farnell distribute a lot of their range.
I've run the filter elements of the supply through Duncan's PSUD2 and it's predicting a maximum peak of -591V on the rectifier without any soft turn on - which I'll almost certainly be fitting.
The mean is ~300 and the RMS about 350. Peak current is 6 amps. Again, that's without any soft starting, so it's not particularly relevant. The actual pulses are under 1A.
I started out looking at the Hexfreds since they come in nice and easy size packets, they're minimum voltage rating is fine for this and they can easily handle the continuous current.
But I was thinking about Shottky's as well. But I'd have to use 4 per leg of the rectifier, so 16 in total. I'm not so bothered about the cost of doing that, but it's a lot of extra components to solder and place.
I was also curious about how many of you have tried using snubbers on your rectifiers, what form of capacitor you used, whether you heard a difference and whether or not you thought it was worth it.
If I go the Schottky way, I'll also have to add bleeder resistors to balance the legs out. It's a lot of extra work when compared with the Hexfreds, about 48 components in total, where as I'd only need three with none snubberised D2PAK packages on the board.
I realise I could use a centre tapped transformer for the Schottkys, I'm just making a point, that it's still a lot of parts compared to the Hexfreds.
If it's a perhaps result, I'll just make two boards, one of each and give them both a try. But I wanted to collect some opinions from the extensive past experience available here first!
What do you guys think of the two forms?
I'm putting together a HV power supply and I'm wondering if any of you have any opinions on which diode I should go for. I was planning to us IRF, since Farnell distribute a lot of their range.
I've run the filter elements of the supply through Duncan's PSUD2 and it's predicting a maximum peak of -591V on the rectifier without any soft turn on - which I'll almost certainly be fitting.
The mean is ~300 and the RMS about 350. Peak current is 6 amps. Again, that's without any soft starting, so it's not particularly relevant. The actual pulses are under 1A.
I started out looking at the Hexfreds since they come in nice and easy size packets, they're minimum voltage rating is fine for this and they can easily handle the continuous current.
But I was thinking about Shottky's as well. But I'd have to use 4 per leg of the rectifier, so 16 in total. I'm not so bothered about the cost of doing that, but it's a lot of extra components to solder and place.
I was also curious about how many of you have tried using snubbers on your rectifiers, what form of capacitor you used, whether you heard a difference and whether or not you thought it was worth it.
If I go the Schottky way, I'll also have to add bleeder resistors to balance the legs out. It's a lot of extra work when compared with the Hexfreds, about 48 components in total, where as I'd only need three with none snubberised D2PAK packages on the board.
I realise I could use a centre tapped transformer for the Schottkys, I'm just making a point, that it's still a lot of parts compared to the Hexfreds.
If it's a perhaps result, I'll just make two boards, one of each and give them both a try. But I wanted to collect some opinions from the extensive past experience available here first!
What do you guys think of the two forms?
Don't even think about using Shottkies in series! Have a look at the reverse leakage currents in them. They are much higher than in conventional P-N junction diodes, and vary a lot with temperature. You would need quite low values for the balancing resistors with all the problems that would cause.
Use fast, soft-recovery diodes like the Hexfreds, or the excellent ones that ST Micro make.
BTW, there are some high-voltage Silicon-Carbide Shottky diodes appearing, but thet are frighteningly expensive at the moment.
Use fast, soft-recovery diodes like the Hexfreds, or the excellent ones that ST Micro make.
BTW, there are some high-voltage Silicon-Carbide Shottky diodes appearing, but thet are frighteningly expensive at the moment.
The two main advantages of Schottky diodes as I see it are low forward voltage drop and no reverse recovery.
The first point is unimportant for a high voltage supply, so that leaves just the second advantage. Without the reverse recovery of pn junction diodes you will have no need for snubbers, but that I think is more than offset by the increased number of diodes you would need, plus (presumably) their higher cost.
For applications where a suitably rated Schottky is available I wouldn't hesitate to recommend them, but otherwise not.
The first point is unimportant for a high voltage supply, so that leaves just the second advantage. Without the reverse recovery of pn junction diodes you will have no need for snubbers, but that I think is more than offset by the increased number of diodes you would need, plus (presumably) their higher cost.
For applications where a suitably rated Schottky is available I wouldn't hesitate to recommend them, but otherwise not.
Thanks for the replies so far, interesting thoughts.
If anyone else has an opinion on the diodes that'd be excellent.
I wish they hadn't started with the fast recovery, ultra fast, hyper fast, ultra hyper faster, faster ultra hyper, super duper hyper ultra fast faster fastest diode path.
Let alone the soft, super soft, super quick, super soft quick recovery forms.
I can sense years of problems coming about from those names. They describe something fast, but it always confuses me for at least a few minutes as to which is faster, hyper, softer or more ultra than the other. I always end up with a datasheet open for each reading the product description and recommended applications.
Are Stealth diodes Fairchild's version of the Hexfred? I saw Vishay advertising Super Rectifiers as well.
If anyone else has an opinion on the diodes that'd be excellent.
I wish they hadn't started with the fast recovery, ultra fast, hyper fast, ultra hyper faster, faster ultra hyper, super duper hyper ultra fast faster fastest diode path.
Let alone the soft, super soft, super quick, super soft quick recovery forms.
I can sense years of problems coming about from those names. They describe something fast, but it always confuses me for at least a few minutes as to which is faster, hyper, softer or more ultra than the other. I always end up with a datasheet open for each reading the product description and recommended applications.
Are Stealth diodes Fairchild's version of the Hexfred? I saw Vishay advertising Super Rectifiers as well.
as Planet10 said, at these low frequencies fast isn't the important thing, but a soft-recovery is. (Although soft-recovery diodes usually are fast as well).
With soft recovery, the small amount of reverse current flow you get when the diode is reverse biased, reduces to zero slowly, and doesn't suddenly stop conducting. If the reverse conduction suddenly 'snaps off' the di/dt is very fast and it can cause EMC problems.
With soft recovery, the small amount of reverse current flow you get when the diode is reverse biased, reduces to zero slowly, and doesn't suddenly stop conducting. If the reverse conduction suddenly 'snaps off' the di/dt is very fast and it can cause EMC problems.
Would you say that 'normal' slow 50Hz rectifier diodes such as those found in the big 35A blocks would have soft recovery? There never seems to be anything about this in the datasheets. Would I be better using a soft or Schottky diode in cases where I have a massive cap bank hence very small conduction angle?
Also, if the conduction angle is very small, does the reverse recovery time become more relevant? If the conduction angle is very small the diode is only on for a short time, which could be akin to operating at much higher frequency. Should we use faster diodes then?
Also, if the conduction angle is very small, does the reverse recovery time become more relevant? If the conduction angle is very small the diode is only on for a short time, which could be akin to operating at much higher frequency. Should we use faster diodes then?
It's all to do with how fast the voltage on the 'input' side to the diode drops when the diode gets reverse biased. on 50 or 60Hz full-wave rectified mains, the rate-of-change of the voltage is obviously defined by the fact that the signal is a FW rectified sine wave.
The real problem is with continuous-conduction boost power-factor-correction stages for SMPSUs like I have to design (for my paid work) at times.
Standard 50Hz bridge rectifiers diodes will not have controlled recovery characteristics. You can slug the reverse recovery by putting capacitors across each diode. This increases the reverse conduction time, but slows down the recovery time, so it reduces the unwanted RF noise caused by the fast di/dt turn-off.
The real problem is with continuous-conduction boost power-factor-correction stages for SMPSUs like I have to design (for my paid work) at times.
Standard 50Hz bridge rectifiers diodes will not have controlled recovery characteristics. You can slug the reverse recovery by putting capacitors across each diode. This increases the reverse conduction time, but slows down the recovery time, so it reduces the unwanted RF noise caused by the fast di/dt turn-off.
richie00boy said:
I'd also like to ask that question.
I'm wondering if it's fussing for the need of fussing. Theoretically, it makes sense, and you could probably measure some change on a scope. But if you were to click the snubbers in and out of place, would you actually hear a worth while difference with modern diodes like Hexfreds?
Would ceramics be suitable for snubbers?
If you are using diodes designed for SMPSU high-frequency use (like the HexFREDs) then you are unlikely to gain any benefit from the capacitors.
The only time I've had a serious HF noise problem with a transformer/bridge rectifier on normal 50Hz mains was using a bridge made of 1N4004 diodes in a low-current 18V supply.
The only time I've had a serious HF noise problem with a transformer/bridge rectifier on normal 50Hz mains was using a bridge made of 1N4004 diodes in a low-current 18V supply.
Sorry for kickin' this old thread, but rather interesting facts where discussed here.
What i have done a while ago is swapping the standard two (for dual mono) 35A diode bridges to a diy diode bridge with four HFA08TB60 hexfred diodes mounted isolated on an aluminium L profile. The new bridges are for the current section of a diy hybrid amplifier with E88CC srpp voltage stage and a old pair of hitachi K135/J50 powerfets to the speaker.
The sound compared to the standard diode bridge was more "backward', and maybe some more detailed in treble. But the bass is worsened imo, less control, more boomy, it sounds like a higher Q factor speaker, and i only swapped the diodes!
Tried a snubber on it also, but removed it, didn't help much
Anyone the same sort of sound experiences?
What i have done a while ago is swapping the standard two (for dual mono) 35A diode bridges to a diy diode bridge with four HFA08TB60 hexfred diodes mounted isolated on an aluminium L profile. The new bridges are for the current section of a diy hybrid amplifier with E88CC srpp voltage stage and a old pair of hitachi K135/J50 powerfets to the speaker.
The sound compared to the standard diode bridge was more "backward', and maybe some more detailed in treble. But the bass is worsened imo, less control, more boomy, it sounds like a higher Q factor speaker, and i only swapped the diodes!
Tried a snubber on it also, but removed it, didn't help much
Anyone the same sort of sound experiences?
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.