Electronics > Beginners
MC34063 high voltage dc-dc boost converter
David Hess:
--- Quote from: dazz on January 03, 2020, 09:02:38 pm ---OK, so you mean only the Ton time matters here, because that's when the spikes happen, right?
--- End quote ---
Exactly, and even slow diodes can turn on very quickly. Some diodes do have problems turning on quickly under some conditions but I think this is a manufacturing problem. It happens even with fast recovery diodes.
MagicSmoker:
--- Quote from: dazz on January 03, 2020, 07:53:46 pm ---
--- Quote from: MagicSmoker on January 03, 2020, 04:52:27 pm ---R2 in your circuit is a bit of kludge in that it provides some damping whenever the Zener/TVS conducts, but it has to be a realistic value given the expected peak current because the voltage drop across it adds to what the switch must withstand (keep in mind the leakage inductance acts like a current source).
--- End quote ---
I've tried anything from 200 up to 10k for R2 and a bunch of different zeners from 15V to 100V with no apparent luck, not sure if I'm doing something wrong, but no combination reduced the power dissipation in the mosfet significantly. :-//
--- End quote ---
Yep, you're doing something wrong... namely, throwing random values at the schematic to see if anything works. Sometimes you can learn a lot that way, but most of the time you just end up jogging in place (ie - going nowhere). >:D
In this case, first look at the peak primary current then pick a resistor value that will result in around 10-20V of drop at that current. For example, if the peak current is 2A then a 5-10R resistor would be more appropriate, not 100R or higher! Keep in mind the Zener/TVS (D5 in the above) diode needs to be able to handle this peak current, too, and it absolutely must have a breakdown voltage significantly higher than the supply voltage (more specifically, it needs to be higher than the reflected voltage from the secondary, as I already explained).
dazz:
--- Quote from: David Hess on January 03, 2020, 10:31:30 pm ---
--- Quote from: dazz on January 03, 2020, 09:02:38 pm ---OK, so you mean only the Ton time matters here, because that's when the spikes happen, right?
--- End quote ---
Exactly, and even slow diodes can turn on very quickly. Some diodes do have problems turning on quickly under some conditions but I think this is a manufacturing problem. It happens even with fast recovery diodes.
--- End quote ---
Cool, thanks David.
BTW, I wasn't getting very good results in the sim with any combination of components in the snubber network, so I went back as I usually do, to re-read the thread. Tim said a few pages back that my tranny inductors were far too large. Trying much lower values now to reduce the leakage inductance, and efficiency/power loss in the mosfet are already much improved. When I find the optimal values of inductance for the transformer I'll go back to testing snubbers in the sim. I know this trial and error strategy is probably not the best way to go about understanding how thing actually work, especially since these are no real world tests, but I hope this will give me a general idea of what works and what doesn't, and then I can better understand why once I google the concepts. I dunno, we'll see.
dazz:
--- Quote from: MagicSmoker on January 03, 2020, 10:49:36 pm ---
--- Quote from: dazz on January 03, 2020, 07:53:46 pm ---
--- Quote from: MagicSmoker on January 03, 2020, 04:52:27 pm ---R2 in your circuit is a bit of kludge in that it provides some damping whenever the Zener/TVS conducts, but it has to be a realistic value given the expected peak current because the voltage drop across it adds to what the switch must withstand (keep in mind the leakage inductance acts like a current source).
--- End quote ---
I've tried anything from 200 up to 10k for R2 and a bunch of different zeners from 15V to 100V with no apparent luck, not sure if I'm doing something wrong, but no combination reduced the power dissipation in the mosfet significantly. :-//
--- End quote ---
Yep, you're doing something wrong... namely, throwing random values at the schematic to see if anything works. Sometimes you can learn a lot that way, but most of the time you just end up jogging in place (ie - going nowhere). >:D
In this case, first look at the peak primary current then pick a resistor value that will result in around 10-20V of drop at that current. For example, if the peak current is 2A then a 5-10R resistor would be more appropriate, not 100R or higher! Keep in mind the Zener/TVS (D5 in the above) diode needs to be able to handle this peak current, too, and it absolutely must have a breakdown voltage significantly higher than the supply voltage (more specifically, it needs to be higher than the reflected voltage from the secondary, as I already explained).
--- End quote ---
Yeah, you're right, I need to stop trying to brute force things by mindlessly trying values when I don't know what they'll do. And thanks for those pointers. May I ask why the 10-20V drop in the resistor? I would have thought that it would need to sink the same voltage as the zener clamp, or about 2x the reflected voltage in the primary, in my case 2x(300V/12.5)=2x24V=48V
MagicSmoker:
--- Quote from: dazz on January 03, 2020, 11:21:10 pm ---Yeah, you're right, I need to stop trying to brute force things by mindlessly trying values when I don't know what they'll do. And thanks for those pointers. May I ask why the 10-20V drop in the resistor? I would have thought that it would need to sink the same voltage as the zener clamp, or about 2x the reflected voltage in the primary, in my case 2x(300V/12.5)=2x24V=48V
--- End quote ---
There's no hard-and-fast rule for sizing the resistor - in fact, it can be eliminated if an RC damper is also present - so don't get too hung up on the voltage drop; I just ballparked something that looked like it would be about right based on experience.
Navigation
[0] Message Index
[#] Next page
[*] Previous page
Go to full version