hello all 
i'm making a new smps to run from a car battery - for amlifier
the pwm chip is uc2825 .
i've tested the transformer today and i get big spikes
i'm using snubbers but its not helping to supress 'em
its a push pull circuit , i'm testing only with one part of the traffo , maybe all will be ok when i will connect the second primery ?
or maybe my core is bad?
i'm making a new smps to run from a car battery - for amlifier
the pwm chip is uc2825 .
i've tested the transformer today and i get big spikes
i'm using snubbers but its not helping to supress 'em
its a push pull circuit , i'm testing only with one part of the traffo , maybe all will be ok when i will connect the second primery ?
or maybe my core is bad?
partially fixed
i used too low gate resistors (20ohm), so when the mosfets were turning off there were a big spikes .increasing the value of the gate resistor solved the problem ,partially though .
now i got very small spikes when no load ,but a 10V-15V spike at the primary when load is connected , no spike at the secondary though.
my main problem now is 20-30V spikes doring turn on .i'm using mtp50n06 mosfets , so the max voltage is 60V and i get very close to that
also i didnt test with full load yet (gotta buy more bulbs ), but the spikes increase with the load , again , no spikes at the secondary
any help would be apprisiated
i used too low gate resistors (20ohm), so when the mosfets were turning off there were a big spikes .increasing the value of the gate resistor solved the problem ,partially though .
now i got very small spikes when no load ,but a 10V-15V spike at the primary when load is connected , no spike at the secondary though.
my main problem now is 20-30V spikes doring turn on .i'm using mtp50n06 mosfets , so the max voltage is 60V and i get very close to that
also i didnt test with full load yet (gotta buy more bulbs
), but the spikes increase with the load , again , no spikes at the secondary any help would be apprisiated
Yikes! Spikes!
Hi, the spikes getting worse as the load is increased is caused by the coupled inductance kicking back harder. One way to lessen the problem is to add more turns to your transformer--both primary and secondary. I think the problem is minimized when the primary inductance is just high enough to get all the power you need and no more.
I am working on a quasi-ZVS push-pull idea which uses capacitor connected from the drain-source of the MOSFETs. A (Schottky?) diode reverse biases the capacitor when the MOSFETs turn on. Another pair of windings in phase and having an equal number of turns as the primaries pulls the capacitors back low a little after the primary is pulled down.
Hi, the spikes getting worse as the load is increased is caused by the coupled inductance kicking back harder. One way to lessen the problem is to add more turns to your transformer--both primary and secondary. I think the problem is minimized when the primary inductance is just high enough to get all the power you need and no more.
I am working on a quasi-ZVS push-pull idea which uses capacitor connected from the drain-source of the MOSFETs. A (Schottky?) diode reverse biases the capacitor when the MOSFETs turn on. Another pair of windings in phase and having an equal number of turns as the primaries pulls the capacitors back low a little after the primary is pulled down.
so the spikes are when the mosfet is going to off mode ...
i was thinking the next circuit will help
i'll make R1 "small" so the turn on time will be fast as normal and R2 "big" so the turn off will be slower .... hopefully this will supress the spikes , whatcha think?
i was thinking the next circuit will help
An externally hosted image should be here but it was not working when we last tested it.
i'll make R1 "small" so the turn on time will be fast as normal and R2 "big" so the turn off will be slower .... hopefully this will supress the spikes , whatcha think?
Here is the basic concept. It may be interesting to play around with, anyway.
thanks ImageShack
thanks ImageShack
An externally hosted image should be here but it was not working when we last tested it.
I assume that your transformer winding is "standard" S-P-P-S, ie. split secondary and center-tapped primary between them?
Edit: And what kind of like snubbers you use? if they dont help there might be room for improvement in component values, layout or topology.
Picture of your project would tell a lot if you have done some "silly things"
Edit: And what kind of like snubbers you use? if they dont help there might be room for improvement in component values, layout or topology.
Picture of your project would tell a lot if you have done some "silly things"
mzzj said:
i'm using 68ohm +100nf snubbers
tomorrow i'll buy lower value resistors , like 56 or 47ohm...
the transformer is 2 primeries with CT and 2 secondaries withCT
also i'll post pics tomorrow if my brother will bring me back my camera
btw ,the circuit is on a prototype board , no pcb yet ...
i know , this can cause problems but i dont think thats the case here.
later...
here are my toroids
i've made the tests with the big one , then i made the second one . the small one is from wilco -itl100 toroid ....
i made only one secondary with only one wire on the small toroid just for testing ....
and both generate spikes , the smaller toroid actually making bigger spikes
snubbers - 49.9ohm resistor with 100nf caps
MOSFETs - MTP50N06
my "solution" of putting a bigger value gate resistors make the transistors hot ,i dont know what else might be the problem
everybody is using the same core from wilco , and they dont get spikes
i've made the tests with the big one , then i made the second one . the small one is from wilco -itl100 toroid ....
i made only one secondary with only one wire on the small toroid just for testing ....
and both generate spikes , the smaller toroid actually making bigger spikes
An externally hosted image should be here but it was not working when we last tested it.
snubbers - 49.9ohm resistor with 100nf caps
MOSFETs - MTP50N06
my "solution" of putting a bigger value gate resistors make the transistors hot
,i dont know what else might be the problem everybody is using the same core from wilco , and they dont get spikes
Hi sss
mzzj had some good thoughts about how if the transformer is wound with much care, the leakage inductance can be lowered enough that the kickback is so rapid that the turn-off rise time and the snubber are able to absorb it. Technically, his suggestion is the accepted way.
I have noticed that your bigger toroid has more turns. The result may be that the coupling coefficient is higher, thus lowering the leakage inductance.
I, however, have decided to try to find an alternative method where leakage inductance is the builder's friend. I have been working on various ZVS circuits, but have not found a functional way to make push-pull work that way.
We have discussed half-bridge for 12v on this forum, and have mentioned that it requires a rather large coupling capacitor or a high switching frequency. I have not really tried taking ZVS to its maximum operational frequency, but I think it would work.
Well, now for some other ideas to try on your circuit. I think it should be possible to really sock your circuit with snubbing, if you would like. Maybe a 10w 2ohm resistor paired with a 1uF capacitor. Basically you may like to try to lower your resistance a lot while increasing the capacitance. Just trying different values to see the result is a suggestion.
Another possibility is to try to speed up the turn-off by using an emitter follower at the outputs of the PWM chip. The best way is to use two complementary transistors. Place turn-on slowing resistor between the two emitters and connect the junction between the PNP transistor's emitter and that resistor to the MOSFET gates. But, the transformer and circuit layout are more likely to help, though, I think.
Honestly. I too am mostly just a tinkerer, so I can sympathize with your frustration. Try to relax, and just keep at it. I think you will get it solved.
Can you see any big differences between their power supplies and yours to account for this difference?
mzzj had some good thoughts about how if the transformer is wound with much care, the leakage inductance can be lowered enough that the kickback is so rapid that the turn-off rise time and the snubber are able to absorb it. Technically, his suggestion is the accepted way.
I have noticed that your bigger toroid has more turns. The result may be that the coupling coefficient is higher, thus lowering the leakage inductance.
I, however, have decided to try to find an alternative method where leakage inductance is the builder's friend. I have been working on various ZVS circuits, but have not found a functional way to make push-pull work that way.
We have discussed half-bridge for 12v on this forum, and have mentioned that it requires a rather large coupling capacitor or a high switching frequency. I have not really tried taking ZVS to its maximum operational frequency, but I think it would work.
Well, now for some other ideas to try on your circuit. I think it should be possible to really sock your circuit with snubbing, if you would like. Maybe a 10w 2ohm resistor paired with a 1uF capacitor. Basically you may like to try to lower your resistance a lot while increasing the capacitance. Just trying different values to see the result is a suggestion.
Another possibility is to try to speed up the turn-off by using an emitter follower at the outputs of the PWM chip. The best way is to use two complementary transistors. Place turn-on slowing resistor between the two emitters and connect the junction between the PNP transistor's emitter and that resistor to the MOSFET gates. But, the transformer and circuit layout are more likely to help, though, I think.
Honestly. I too am mostly just a tinkerer, so I can sympathize with your frustration. Try to relax, and just keep at it. I think you will get it solved.
Can you see any big differences between their power supplies and yours to account for this difference?
i'm using uc2825 , all the others use the sg3525subwo1 said:
i'm using the uc2825 because higher current capability , i dont think there is a big difference between the 2 chips , anyway , tomorrow i'll try the sg3525 , who knows , maybe it'll work better
, i think i tryed all the other things ....i'll show the waveform tomorrow also ...
Spikes during turn-off in push-pull circuits are quite usual, and RC snubbers with low value resistors may be used to damp them a bit (try 2 to 10 ohms and a suitable capacitor). With high loads these spikes may reach the Vds rating of the switches thus entering avalanche mode, but this is not destructive at all as long as the avalanche lasts only for some dozens of nanoseconds.
Eva said:
thanks for your reply Eva , u help me allot as always
i'll try lower resistors and play with the snubbers / freq
but those low value resistors will dissipate allot of power , arent they?
i saw many smps designs and only i got those spikes , or maybe only me paying attention and all the others just leave it that way?
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.