Push pull dc to dc converter inefficient / General Science and Electronics / Forums

Forums

4hv.org :: Forums :: General Science and Electronics

« Previous topic | Next topic »

Push pull dc to dc converter inefficient

previous 1 2 3 next

Move Thread

LAN_403

furnace

Sat Oct 05 2019, 03:04PM

Registered Member #4992 Joined: Wed May 23 2012, 03:57PM
Location:
Posts: 108

I don't know why I can't get the pictures uploaded, the files size might be to big.
I think you are right Patrick, but what measurements can I take to correct this, as I've tried to
Increase and decrease dead time also, but it just burns the mosfets when dead time is to little.

furnace

Thu Oct 17 2019, 04:51AM

Registered Member #4992 Joined: Wed May 23 2012, 03:57PM
Location:
Posts: 108

Can anyone else please try and shed some more light on this. I've also read that the push pull
Topology isn't the best way to go with higher power levels and is prone to inductance leakage with higher power levels. Is this true ? And would full bridge be a better topology to use with my low input voltage and high output of 380?

Patrick

Thu Oct 17 2019, 04:58AM

Registered Member #2431 Joined: Tue Oct 13 2009, 09:47PM
Location: Chico, CA. USA
Posts: 5639

It sounds like it is saturating. Push-Pull tends to be pretty good at 500w and less. Have you read the many pages here :

Are you sure you have the phasing right ?

furnace

Thu Oct 17 2019, 11:25AM

Registered Member #4992 Joined: Wed May 23 2012, 03:57PM
Location:
Posts: 108

I don't know if the ferrite transformer is saturated, because It's an ee65 core and I've a
Also tested the smaller ee55 core and both act the same, I just don't know what type of material it is. I can draw large current with the 65 core , but it's just that the efficiency gets less, that's why I thought of limiting the current on secondary.
I need to use over 2000 watt so I'm thinking I that I've chosen the wrong topology to start with.
The sg3525 has a fixed output at pin 11 and 14 for driving the mosfets , so I'm sure the phasing is right.

Patrick

Thu Oct 17 2019, 05:35PM

Registered Member #2431 Joined: Tue Oct 13 2009, 09:47PM
Location: Chico, CA. USA
Posts: 5639

If you reach excess heating and low power at 400 w, then limiting current wont help you reach 2000 watts. Youll need to post pics and a schematic of you current and past work, modifications etc. if you want more help. Also it sounds like your just needing a bigger ferrite core. if you have an oscilloscope post the gate drive too.

furnace

Fri Oct 18 2019, 11:18AM

Registered Member #4992 Joined: Wed May 23 2012, 03:57PM
Location:
Posts: 108

This is my circuit but only with bigger mosfets
I did contact the forum for assistance so that I can try an post pictures of my scope trace
Of the gate and drain. I also ordered new ee55 core pc40 material to eliminate any core issues, but it's still going to take a few weeks to get it.
In the meantime can anyone suggest a full bridge circuit that I can use with my driver circuit?

Sulaiman

Fri Oct 18 2019, 02:27PM

Registered Member #162 Joined: Mon Feb 13 2006, 10:25AM
Location: United Kingdom
Posts: 3140

Things that I would check;

gate drive signals, maybe at 300W the duty cycles are already approaching 50% each

drain-source ON voltage may be increasing at higher current levels

supply voltage to transformer centre-tap may be decreasing at higher power levels

to check for transformer core saturation,
if possible, increase the number of primary turns (e.g. double) and operate at lower voltage output (e.g. half) for testing purposes.
If you can get full design output current at half the output voltage then the cores are probably saturating in the present setup.

Try operation at a higher frequency, e.g. 1.5 times present operating frequency,
the core will heat up more but maybe the output will hold up to 1.5x 300W before displaying the present heating problem ?

For efficiency, the combined copper weight of the primaries should be more than the weight of copper in the secondary,
from memory, at least by a factor of sqrt(2).

furnace

Sat Oct 19 2019, 06:01AM

Registered Member #4992 Joined: Wed May 23 2012, 03:57PM
Location:
Posts: 108

Sulaiman wrote ...

Things that I would check;

gate drive signals, maybe at 300W the duty cycles are already approaching 50% each

drain-source ON voltage may be increasing at higher current levels

supply voltage to transformer centre-tap may be decreasing at higher power levels

to check for transformer core saturation,
if possible, increase the number of primary turns (e.g. double) and operate at lower voltage output (e.g. half) for testing purposes.
If you can get full design output current at half the output voltage then the cores are probably saturating in the present setup.

Try operation at a higher frequency, e.g. 1.5 times present operating frequency,
the core will heat up more but maybe the output will hold up to 1.5x 300W before displaying the present heating problem ?

For efficiency, the combined copper weight of the primaries should be more than the weight of copper in the secondary,
from memory, at least by a factor of sqrt(2).

Thanks Sulaiman, these are great tips.
I think that the duty cycle might be the problem because I disconnected feedback for testing purposes so I'm running at 50% duty cycle from very low watts already.
But how do hold 380volts from all the way down at 50 watt to 2000 watt on my core?
Would this mean that I would need to wind my secondary for a lot more higher output up to 1000volts or more and adjust resistance on feedback to keep duty cycle low up till 2000watt?

Drain- source voltage I have checked and it stays the same

Supply volts stays above 11 volts with high current draw.

I have increased frequency, than it becomes a little more efficient with lower voltage but burns the mosfets at higher power levels due to parasitics and switching losses .my core never even gets a temperature it just stays normal.

My primary definitely weighs less than my secondary so this should also be a problem which can be dealt with.

Patrick

Sat Oct 19 2019, 06:55PM

Registered Member #2431 Joined: Tue Oct 13 2009, 09:47PM
Location: Chico, CA. USA
Posts: 5639

I think wire and construction geometry may be contributing to these issues. Though all of sulaimans diagnostic tests should be completed first.

What are the diameters of the primary and secondary wires ? did you wind transformer this yourself ? Skin effect and or proximity effect could be contributing to these problems if no other solution presents itself. core area and the B-H curve become essential and non trivial in transformer design.

I still think since you point out your at 50% duty cycle that you may have shoot-through condition at the transistors. i try to ensure 3% non-conduction through the primary at max power on my designs (this 3% rule, may mean 45% max DC). Avoiding "flux walking" from unbalanced current is essential as well.

Sulaiman

Sun Oct 20 2019, 08:44AM

Registered Member #162 Joined: Mon Feb 13 2006, 10:25AM
Location: United Kingdom
Posts: 3140

Mainly as a learning exercise;
Try adding feedback to get a dc voltage less than 380V, e.g. 300V.
The duty cycle at no load should be approximately 50% x 300/380, about 40%.

As you increase the current from the regulated 300V you will see the duty cycle increase to compensate for (mainly) resistive losses.
(power supply to centre tap of primary, Rds(on), Rpri, Rsec, and the resistance of your output inductor)
_________________________________________________ _________________
When you run/ran the inverter open-loop (fixed max. duty cycle) did you notice much voltage drop at the output as you increased the current/power output ?

Lastly (for now), try connecting the SG3525 output pins (11 and 14) directly to the mosfet gates with no resistors in the gate circuitry.
This will speed up the switching on and off of the mosfets, and apply a slightly higher gate drive voltage resulting in lower Rds(on).
Check that the SG3525 does not get too hot as this mode of operation does put more stress on the IC outputs.

P.S. as you were operating at near 50% duty cycle per transistor and the core was not overheating at low output current,
(maximum integral(Vpri.dt) = volt.seconds applied to primary winding)
you can be confident that the core is not saturating,
the problem lies elsewhere.
For a similar reason, you can assume that you do not have a significant 'flux-walking' condition.

P.P.S. in your original post you said that you are using a totem pole configuration (half-bridge)
I assume that you mean that you are operating push-pull as in the article that you pointed to

?
(The internal SG3525 outputs are totem pole, but you are operating your inverter in push-pull mode)

previous 1 2 3 next

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob