Hi.
In a typical LLC smps the main resonant cap is 22nF-100nF at 1000Vdc.
Thats not a common size to be found in a normal guys junk-bin 🙂
So, could I use 2 caps in series with some current-sharing resistors? Caps would surely need to be exact same types and maybe losses would be a little higher.
A typical X2-cap is 100nF-220nF@275Vac. 2 in series would total 50-100nF@550Vac (probably with a huge margin as its X2-caps).
Not saying its an ideal situation, but would it be considered safe? - the potential for damage is pretty large if the caps do not share the high voltage evenly.
Kind Regards TroelsM
In a typical LLC smps the main resonant cap is 22nF-100nF at 1000Vdc.
Thats not a common size to be found in a normal guys junk-bin 🙂
So, could I use 2 caps in series with some current-sharing resistors? Caps would surely need to be exact same types and maybe losses would be a little higher.
A typical X2-cap is 100nF-220nF@275Vac. 2 in series would total 50-100nF@550Vac (probably with a huge margin as its X2-caps).
Not saying its an ideal situation, but would it be considered safe? - the potential for damage is pretty large if the caps do not share the high voltage evenly.
Kind Regards TroelsM
Of course you can run caps in series. But 1000V ?
And current sharing resistors, in a series connection?
I havent seen so large voltages. When doing the math, typically a few hundred volts over the resonant cap.
I wuold stick to high pulse, low esr ,low esl , high ripple current variants, typically some polypropen cap like Wima MKP/FKP series. They do take a beating, repetitivly, not intermittently like an X2 cap.
Unless you are in a money pinch and high series where every cent counts 🙂
Aldo, check out the split cap topology for LLC. Se app note AND 8311 from Onsemi fex. Most semi conductor manufacturers have similar app notes.
/ regards Rickard
And current sharing resistors, in a series connection?
I havent seen so large voltages. When doing the math, typically a few hundred volts over the resonant cap.
I wuold stick to high pulse, low esr ,low esl , high ripple current variants, typically some polypropen cap like Wima MKP/FKP series. They do take a beating, repetitivly, not intermittently like an X2 cap.
Unless you are in a money pinch and high series where every cent counts 🙂
Aldo, check out the split cap topology for LLC. Se app note AND 8311 from Onsemi fex. Most semi conductor manufacturers have similar app notes.
/ regards Rickard
Oh, and BTW , I recommend for experimenting purposes not trying to integrate all inductances in the leakage inductance.
Adding a small external inductor alleviates the problem of getting the leaking x-former correct first try.
You have more controll that way.
Otherwise risk of magic smoke rising from Mosfets 🙂
There has been quite a few posts about LLC converters at rhe forum. Check those out.
Adding a small external inductor alleviates the problem of getting the leaking x-former correct first try.
You have more controll that way.
Otherwise risk of magic smoke rising from Mosfets 🙂
There has been quite a few posts about LLC converters at rhe forum. Check those out.
Hi
Thanks for the answer.
Maybe I was not clear on the resistor-part, but I would install the resistor across the caps to help them share equally.
Something like the attachment.
Most app-notes and simulations I have seen/done suggest around 1Kv peak across the cap with a mains voltage of 230Vac.
I have tested diy transformers with integrated magnetics, but it never became good enough. With the ferrite I had for testing it was close to impossible to get the right ratio of inductances and the resulting fringe-flux and leakage (I think...) resulted in rather large local heating in the windings.
Thanks for the app-note. Had not seen that one before. I have tried simulating the split-cap and diode solution before, but I found it to limit the regulation-range too much.
On the other hand it looks very simple and when limiting the voltage I can use "smaller" caps.
Kind regards TroelsM
Thanks for the answer.
Maybe I was not clear on the resistor-part, but I would install the resistor across the caps to help them share equally.
Something like the attachment.
Most app-notes and simulations I have seen/done suggest around 1Kv peak across the cap with a mains voltage of 230Vac.
I have tested diy transformers with integrated magnetics, but it never became good enough. With the ferrite I had for testing it was close to impossible to get the right ratio of inductances and the resulting fringe-flux and leakage (I think...) resulted in rather large local heating in the windings.
Thanks for the app-note. Had not seen that one before. I have tried simulating the split-cap and diode solution before, but I found it to limit the regulation-range too much.
On the other hand it looks very simple and when limiting the voltage I can use "smaller" caps.
Kind regards TroelsM
Attachments
Last edited:
Ah, ok. Well considering you have two 400VAC/630VAC caps, and a 10% , likely 5% tolerance I dont really see the need of voltage equlising resistors. They have to match the caps, so each RC is identical.
But you will complicate matters with this solution. Wouldnt the resistors have to be so small for to be effective that they will affect the LLC turning it into something like a LLC/R converter, how now such a beast might work?
Anyhow massive dissipation in resistors for sure.
But you will complicate matters with this solution. Wouldnt the resistors have to be so small for to be effective that they will affect the LLC turning it into something like a LLC/R converter, how now such a beast might work?
Anyhow massive dissipation in resistors for sure.
Hi Luka.
I dont know actually. When simulating different designs from app-notes I come close to 1000Vpp.
I think I will give it a try with the split-cap and diode-protection. That will probably give me a mush larger range of possible caps.
Kind regards TroelsM
I dont know actually. When simulating different designs from app-notes I come close to 1000Vpp.
I think I will give it a try with the split-cap and diode-protection. That will probably give me a mush larger range of possible caps.
Kind regards TroelsM
- Status
- Not open for further replies.