[SOLVED] Transformer saturation & ACM control

This has nothing to do with LEDS

Many power converters are liable to saturate the transformer under unbalanced conditions, I am looking at a full bridge with a bridge rectifier output so there are plenty of oppertunities! The standard method to reduce the problem is to use peak current control. In my case I wish to use average current mode, a typical LPF plot for a 25Khz converter is shown in the attachment (this is currently (pun) running a half bridge). What I am having trouble getting my head around is given the LPF somewhat inhibits cycle by cycle current limit will ACM still prevent flux-walking and potential saturation ?

BTW the target is 3Kw through an ETD54 at 50Khz

PS I dont mean to be asking a yes/no question, rather has anybody working knowlage in this area they care to share ?

If the LPF cutoff frequency is lower than the reciprocal of a half period, I would say that ACM would not reliably prevent assymetrical core saturation.


Think of one situation. Let's say, as an example, that when you power-off the converter, you leave core saturated in the B- direction.
If the next time you power up the converter the bridge starts in a way to impress additional B- flux, the core will saturate.

Yes unfortunatly you confirm my feelings, the reason I am using ACM is to prevent instability due to a downstream LCL filter.
Have to get the thinking cap on.....how about capacitors in series with the primary ?

Actually TBH I think I had better get the simulator out again and take a closer look though this is a very very very slow process (no access to supercomputer)!

I'm not sure I understand the whole picture but generally the modulator of a full bridge should be inherently balanced. Residual DC offset voltage in the switching waveform will be minimal. Average current mode control typically has plenty of bandwidth (>1/10th Fsw I expect) and should have no problem dealing with that residual DC offset during operation. Startup may be a little unique but still I expect average current mode control to get control in some small number of cycles.

I agree asdf44 but of course in the real world imbalances creep in, I guess it's a question of if those imbalances are suficient to eat up the reserve flux density. ACM has low bandwidth relative to Fs so I agree with schmitt it is unlikely to help. Instead I have spent some time trying to model my transformer primary & core as a saturateable reactor in LTspice, I think I have suceeded and by deliberatly inducing imbalance (by differential on times) can indeed induce saturation. I have compared the situation with & without a primary series capacitance and with defenitly reduces the severity of saturation as shown in red on the picture below, green is DC coupled.
View attachment saturatable ind_trafo1.pdf

Seems unlikely you will get 3kW continuous through an ETD54 at 50kHz - what is your calc Bpk under these conditions? Usual way to stop flux walking is a suitably sized DC blocking cap in series with pri ...

At 3 kW, LLC or at least series LC resonant converter would be preferred for lower switching loss and EMI. In this topology, the series capacitor will already provide DC balancing.

Re: Transformer saturation & ACM control

Easy peasy said:

Seems unlikely you will get 3kW continuous through an ETD54 at 50kHz - what is your calc Bpk under these conditions? Usual way to stop flux walking is a suitably sized DC blocking cap in series with pri ...

Click to expand...

According to ExcellentIT7100 0.144T @ 2882W. And yes I wondered if I could do without the caps moving from a half to a full bridge but I think the answer is clearly not.

- - - Updated - - -

FvM said:

At 3 kW, LLC or at least series LC resonant converter would be preferred for lower switching loss and EMI. In this topology, the series capacitor will already provide DC balancing.

Click to expand...

Ahh yes whilst I agree I have some peculiar reason for using hard switching, this converter is modulated and has to work dynamically over a huge power ratio, I am simply not sure a resonant job could maintain resonance and linearity. Good and tight layout minimizes EMI in my experience.

And thank you everybody for your comments you have been very helpful indeed