DC TO DC FULL BRIDGE CONEVRTER OUTPUT SPIKES

[darla1]

conevrter

DEAR ALL

Hope everybody is fine...

i am working on full bridge dc/dc converter...when i loaded this upto 4A circuit is fine and dc/dc output getting pure dc with less riplle....

when i increase the current beyond 4 A getting spikes at output dc like shown in fig..

I have designed the transformer and inductor for 10A, using snubber also...

can anybosy suggest anything

thanks
darla

 

[TekUT]

dc waveform + spikes

Could you post/share your schematic?

Thanks

Pow

 

[darla1]

u can see the schematic is like this....but real sechenatic i can not upload...

thanks
darla

 

[TekUT]

Thanks for the schematics, well, do you've setup some simulation in order to see if the behaviour can be replicated on simulation?
From your first post, about the waveform, spike are present only when the secondary is fired or in other word when the output diode take to start in conduction, may be this ringing is due to the turn-on diode...

Bye
Pow

 

[eugen kus]

Hello darla ,
I suspect that at currents greater than 4A , the output inductor goes into continuous conduction and the reverse recovery of the output diodes will become important . Use Schottky diodes and plenty snubbing .
Cheers,]
Eugen

 

[darla1]

Hello darla ,
I suspect that at currents greater than 4A , the output inductor goes into continuous conduction and the reverse recovery of the output diodes will become important . Use Schottky diodes and plenty snubbing .
Cheers,]
Eugen

helo kus i am using schoottky diodes only and snubbing also good....

i am thinking that the problem is at output inductor......inductance is creating the problem at higher current..

darla

 

[eugen kus]

Hello darla ,
please post details of frequency and max dc output current .
Also give details of present output incuctor , i.e number of turns, core size and size of air gap .

Cheers,
Eugen

 

[FvM]

The circuit is missing the buck-converter free-wheeling diode. This must not necessary cause the spikes but is bad anyway, because all output current loads the transformer.

 

[darla1]

hello Kus

Hello darla ,
please post details of frequency and max dc output current .
Also give details of present output incuctor , i.e number of turns, core size and size of air gap .

Cheers,
Eugen

the switching frequency is 50khz and output inductor frequency is 100khz,
max dc output current is 10A
i am using toroidal micrometal core with OD 2.5 and ID 1.5 with inducatnce value 10UH.

thanks for ur reply

darla

Added after 2 minutes:

The circuit is missing the buck-converter free-wheeling diode. This must not necessary cause the spikes but is bad anyway, because all output current loads the transformer.

hello FVM

my problem is when spikes coming my tarnsformer is giving some noice so that i am not able goto high current i.e 8A.

can you suggest anything

thanks
darla

 

[fsmyth]

A bit hard to tell, since you posted no values, but
it is most probably core saturation.

 

[TekUT]

@darla1

Could you please report to me the transformer detail like magnetizing inductance, core material and so on and the output diode data in order to perform some simulation?

Also could you please post a screenshot of the waveform before the output inductor? May be the spike came from primary to secondary?

Bye
Pow

 

[darla1]

@darla1

Could you please report to me the transformer detail like magnetizing inductance, core material and so on and the output diode data in order to perform some simulation?

Also could you please post a screenshot of the waveform before the output inductor? May be the spike came from primary to secondary?

Bye
Pow

hello pow

my transformer primary inducatnce is 1.8mH and using ETD 44 N87 epcos make...

diode i am using APT30D60B ...

the waveform shape after diodes is good but after the inductor only getting the spikes..

i am attaching the waveforms please see.....

darla

 

[FvM]

The spikes at primary side are caused by a specific problem of your design, I already mentioned, see above. The spikes at the output are visible either cause the output capacitor has a high series inductance or are possibly measurement artefacts. The current waveform is basically O.K., also the primary voltage (apart from the spikes). One may ask however, if the current ripple should be designed lower.

 

[TekUT]

Yes this is exactly what I like to know, spike seems originated from the primary side of transformer, but as FvM said could you make a try by using the free-wheeling diode on the output buck section? Also the ripple might be a issue, you've sized the trasformer for 10 A and ripple is comparable, may be a ripple reduction can solve your trouble, but if you can please make a try with the diode.

Bye
Pow

 

[darla1]

The spikes at primary side are caused by a specific problem of your design, I already mentioned, see above. The spikes at the output are visible either cause the output capacitor has a high series inductance or are possibly measurement artefacts. The current waveform is basically O.K., also the primary voltage (apart from the spikes). One may ask however, if the current ripple should be designed lower.

thank you very much FVM

the input spikes at transformer primary is due to long dead time of pulses.....

out spikes is causing due to output series inductor....but my problem is with the transformer, transformer is giving sound if i go higher than 4A....

thank you very much teuker

can u suggest anything for this....

darla

 

[darla1]

heloo FVM / tuekur

i found the root cause of problem for sound in transformer...

this is caused by loop compensation.........can you suggest anything for thsis

darla

 

[rauol]

i think your DC-Dc has become unstable, increase phase margin. maintain minimum 50 phase margin.
improper layout also will be responsible for this. use single point grounding. and see that the controller is not near the ETD core and that the IC ground is returned at the output cap negetive at a single point

 

[ghydda]

If further input is appreciated I will chip in my 2 cents as well.

The transformer primary waveform spikes are completely normal as darla1 has stated. However they can be (and should be) dampened to:

A) Reduce the spike generation on the secondary side (notice the primary turn off spikes correlate with the spikes on the secondary side, which most certainly are common mode spikes).
B) Reduce the EMI emissions.

The audible noise generated by the transformer is most likely due to inadequate phase margin in the control loop compensation. This should be fairly easy to solve, of course depending on the transient performance required of the design.


Best wishes
/Ghydda

 

[darla1]

thank you Mr Ghydda

now i am able to run my circuit from 4Amps to 7Amps, but after 7Amps the output dc is generating the spikes, due to these spikes the PWM IC is not sable i.e at this condition transformer is giving sound......

here i need to do good loop compensation......can any body siggest here

darla

 

[ghydda]

I don't think the spikes are disturbing your controller.
If my assumption is correct only your oscilloscope can see these spikes.

To verify this, do the following: place both the scope probe tip and the scope probe return clip on the secondary ground. The scope should measure zero volt (obviously) but any presence of common mode noise will still show up as sharp needle spikes.

Common mode noise only disturbs surrounding equipment (a big generalization).




Regarding the loop compensation.
Have you tried a PI-regulator?
Attached is a principal schematic with a transfer function shown graphically.

C1+R1 form the purple line (upper cut off frequency).
C2+R2 form the green line (integration corner frequency).
R1+R2 form the blue line (proportional gain).

Together they form the red line.


Then it's just a matter of selecting the appropriate corner frequencies and proportional gain.

I suggest trying:
C1R1 = 1000Hz
C2R2 = 100Hz
R1R2 = 0dB

Keep lowering until loop stabilizes.
If already stable, increase gain/corner freq until unstability and then back of a bit.


This is the quick and dirty method. It is assumed that signal integration has not been compromised (poor shielding of sensitive signals and so on).



Let us know how it turns out.



Cheers,
/Ghydda