| aliazhar |
I have just added snubber on my smps...(330 ohm 4w +0.1uf)...
after adding snubber, the input filter capacitors are puffing within a matter of couple of minutes.
any idea?
(i am using P200nf04 x4, 20khz, E-core - total watts 350 @ 44v. no input/output filters. single winding with bridge output. bridge is FEP16B) |
|
|
| aliazhar |
problem fixed !!!
thanks guys. |
|
|
| luka |
Hi
What was the problem? |
|
|
| aliazhar |
Luka,
to be honest with you, i have removed the snubber and everything works almost fine.
but, there is something i still dont understand....
when i connect the dc output to oscilloscope, i still see waves....its not pure dc....like a straight line....
why is that so?....
and one more thing is, whenever i use this smps, my circuit which i am running this smps for is getting more hot then compare to ac powersupply i use.
any suggestions? |
|
|
| AndrewT |
| The HF pulses from the SMPS could be upsetting the circuit and causing it to oscillate. |
|
|
| aliazhar |
| any remedy for this ? |
|
|
| luka |
Hi
Can you show it on pic?
Has nothing to do with how you drive your smps...I had most problems with ground, I should say nuteral, which we don't have, and everything runnig from same phase, no insulating transformers...
if there are spikes on pretty much flat DC line then you are ok, but if you see anything else send pic...hope your osc isn't acting as load to smps |
|
|
| aliazhar |
I have checked again....and there are no noticeable spikes coming out at the output dc....its almost very clean.
now, here is whats happening without using the snubber.
after using this smps for about 10 minutes, i see a very little puffed filter caps...
i am using 2200uf 25v x2 105-C....do you think thats enough?....or are they more?
using scope probes at the input of traffo (output of mosfet) i see lot of spikes....
my scope frequency is fluctuating from 20khz (which its suppose to be) upto 225khz...
when i used the snubber it was way better but, the caps was puffing....
i think there is something which is out of balance....either my snubber values are wrong or the something else....dont have any idea.
can you help? |
|
|
| DigitalJunkie |
| Maybe try a smaller snubber cap. 0.01uf or so? :confused: |
|
|
| aliazhar |
i did....i have tried different combination...out of which the best wave was out of 330r/0.1uf....but, using that puffs the caps...
what are the best caps values i can use for filter ?
2200uf or less ? |
|
|
| luka |
Hi
2200 is ok, you can use even more or less...shouldn't be a problem
post pic od primary side on fets |
|
|
| aliazhar |
| here is the mosfet output |
|
|
| megajocke |
At what output power is that? The duty cycle is just 5% or so... Does "puffing" mean they explode or what is it?
Are you using some kind of feedback and regulation? This won't work without an output filter - RMS currents in transformer and ripple current in caps will be very big. If you don't have an output filter you need to run the supply at maximum possible duty cycle. |
|
|
| aliazhar |
i took with a very very little load...maybe 0.5 a or less
when i was using the snubber, input filter caps exploded within 2 minutes...
after removing the snubber, they are puffed from top...
i am using a regulation using zener diode.
after a full bridge, i am using 2x 2200uf 63v.
does this info help ? |
|
|
| megajocke |
| It is strange that the snubber would affect this. But in any case - trying to have regulation without an output filter will cause lots of stress on both switches and capacitors. The transformer will get much hotter too due to increased RMS currents. |
|
|
| aliazhar |
| maybe my snubber is wrong value ? |
|
|
| luka |
Hi
Forget about filter!! If you have waveforms like that when you load your smps just few watts, there is no point in searching for right value for snubber coz there is nothing to filter out, by that I mean you should have square wave and some oscilation, that would be ok, but here you can't even tell what is main signal
Your trafo is ******* if you know what I mean...you WILL have to rewind it, coz you can't do much with this...You can't filter hundert and hundred watts of oscillation, to get main waveform out, unless you want efficiency of only few %
When you say output of fet I hope you mean Drain
I think it is time for new trafo if you are willing, you pretty much have to wind it only one general way to get best performance....what do you say?
One more thing, some me pic of you trafo and how did you wind it exactly
PS:Are those pic with or without snubbers on? if with, show pic without (Your filter should be *and depends on freq. used* about 10-33R +1n-10n, but that is what I used. But I have square waves with small oscilations(compared to yours) of around 5MHz) |
|
|
| aliazhar |
well... the traffo i am using is made by a traffo manufacturing company...they are rated 350w....primary 12+12, secondary 55v single...
winding is professionally done using machines on E-Core...almost same size as etd39...but, simpler. so, i cant say anything about these...
btw: mosfet output means Drain.
let me tell you what i experiment till now...and you will get even better picture...
i can get nice and clean wave both on the output of sg3525 and at the drain... only when i remove the bridge rectifier...
once i remove the rectifier, these output looks really neat and clean...
but, once i connect a bridge rectifier (FEP16B) they turn out to be this way...
i have no doubt on the traffo at all.
i am actually using two traffos.....with single sg3525....and there are two bridge rectifiers....the output from both is connected parallel... to get more amps.
there are 4 mosfets on each traffo....
and two bridge rectifiers.
any help? |
|
|
| luka |
Hi
Are those diodes ok, they are just 100v, so you have to have less then this on both secondarys together, if you use only single then it is
And if we assume trafos are ok... what is your input current idle? 12 turns for primary is kinda big, unregulated toroind use aout 6 only...
If you get nice waves with no rectifiers and everything that is after it, check there first if everything is connecter right, no bridges or something like that... is your secondary waveform pretty much the same when without rectifiers?
This is one strange problem and you only connect output stage.. |
|
|
| guitar_joe |
a photo of your design and a schematic might help track down the problems in this circuit
| quote: | | primary 12+12, secondary 55v single... |
do you mean 12v - 0v - 12v primary?
or... 12+12 turns on the primary? |
|
|
| luka |
Hi
Dual primary, normal push-pull |
|
|
| megajocke |
There is nothing strange about the waveform - it looks like that because you are trying to have regulation without an output filter. The duty cycle is really low because the transformer is good (low leakage inductance) - not bad.
Remove the regulation or add an appropriate output filter! Make sure the diodes can take the voltage. |
|
|
| aliazhar |
thanks for your help guys.
traffo is 12v-0v-12v....and output is 0v-55v
i am using two bridge rectifiers (one for each traffo)....diodes are fine....they do get hot when my output load is around 12amps.... thats normal....
but, under less load, they run cool....i am using heatsink for them.
at the moment, i am using 2 x 2200uf - 63v output filter...how much more i should increase it to ?
i have to use regulation...coz if i change the output volts, it will toast my circuit which i am going to run it with this smps.
the only two concerns i have is:
1. why my input filters are puffed after around 10 minutes of use
2. whenever i run my circuit using this smps, it runs hotter then compare to if i run it on AC powersupply.
i am using a nice heatsink on the mosfets....with fan....the temperature on the mosfet, after using this smps for around 10 minutes goes upto 70 C.....
is there a way to fix these problem...and have them run a bit cooler?
Thanks guyz ! |
|
|
| areza |
| seriously “a photo of your design and a schematic might help track down the problems in this circuit” |
|
|
| megajocke |
"at the moment, i am using 2 x 2200uf - 63v output filter...how much more i should increase it to ?"
You will need an inductor (coupled if you have multiple outputs) to be able to have a regulated output. |
|
|
| aliazhar |
coupled?.....means on a single toroid?.....
which one?.....or where to get one?......(how to make - any info?)
i will post the photos in a little bit.....(charging my camera) |
|
|
| aliazhar |
here is my smps....
lower line are the mosfets.....upper left line are the rectifiers and two filters....
on the upper right, thats my circuit which i am going to run on this smps....with another daughter board...
sg3525 is at the back of the board on the lower right side....
i am using a 4oz pcb board...with solder filled layer on the power lines.
|
|
|
| luka |
Hi
You don't need inductor to have regulation, it just helps a lot at low loads...and those waveform shouldn't be like that with or without regulation, low duty cycle or not...you should be able to see it...maybe not if only pulses...you could try to add coupled inductor, or if you don't know how to make them just use 2...after rectifiers and before output C
PS: you board looks cool, good job |
|
|
| aliazhar |
Thanks for your comments...
i have tried using output filters....but, results are same....no change. |
|
|
| megajocke |
| How was these output filters connected? Have you calculated the ripple current on the input? |
|
|
| luka |
Hi
I hate to say be schematic would help here a lot |
|
|
| aliazhar |
| quote: | Originally posted by megajocke
How was these output filters connected? Have you calculated the ripple current on the input? |
output filters are connected in parallel to each bridge rectifier and then joined the output + to + and - to -
i really dont know how to calculate the ripple.
| quote: | Originally posted by luka
Hi
I hate to say be schematic would help here a lot |
i am using a schematic from Risto80:
http://www.diyaudio.com/forums/show...1614&highlight=
with a little modification of dead-time and frequency.
if you remember, you helped a lot on that thread to me.... |
|
|
| mag |
Hi aliazhar
the schematics seems fine but it seems that your SMPS is running
in discontinuous mode.
Have you placed two inductances between the diode bridge and the output capacitors?
If there is no inductance the regulation becomes nearly impossible with no load.
If there is an inductance (better only one core with two windings for the two rails) which value is it? Are you sure that the inductance core is suitable for the frequency and Bmax you need?
BTW: 220kHz swithching frequency seems quite high... If you have 12+12 windings for primary you can probably decrease that frequency without saturating the transformer core. |
|
|
| areza |
| may be feedback waveing like output, have a look |
|
|
| luka |
Hi
One thing you could do is if you have supply that you can set to <10v just to start SG and would try to get full duty cycle, or even without FB...then you would see how bad oscilations are and try to fix them then...any other duty cycle should be good as well then |
|
|
|
