SG3525 as peak current controller

[jegandren]

hi all,

Is it possible to implement a peak current mode control (PCMC) using a SG3525 IC. As far as im concern SG is just a voltage mode control IC . please advise...

[guitar_joe]

As far as i have read because if its inherent design, peak current mode control is not possible with the sg3525.

[Tahmid]

Hi jegandren,
You cannot directly do current-mode control with SG3525 as it is a voltage mode controller. But current controlling is possible with a current sense device and a comparator which will shut down the SG3525 in over current condition.
Thanks.

[zilog]

I have implemented (working) current mode feedback using this circuit: (please disregard the nastiness of the paper) http://wintermute.csbnet.se/~zilog/schemat.jpg

[ChocoHolic]

I like the 'nastiness' of that paper.
It shows, that you had a good time working on this

Yes, I agree. It makes sense to feedback the current signal to the error amp, because your current sensing does not feedback DC information. Designs that want realize a cycle by cycle pulse cut via the shut down input, must contain the DC information of the transformer currents. So it would not work with current transformer sensing (from my theoretical pov..., not tried..)
Hm, regarding your general feedback structure:
Is there a reason for your quite special way of superimposing the current feedback signal to the reference signal for the error amp? My straight forward thinking would lead to a set up, where I would sum both signals at the error amp....

[zilog]

Originally posted by ChocoHolic
Yes, I agree. It makes sense to feedback the current signal to the error amp, because your current sensing does not feedback DC information. Designs that want realize a cycle by cycle pulse cut via the shut down input, must contain the DC information of the transformer currents. So it would not work with current transformer sensing (from my theoretical pov..., not tried..)
Hm, regarding your general feedback structure:
Is there a reason for your quite special way of superimposing the current feedback signal to the reference signal for the error amp? My straight forward thinking would lead to a set up, where I would sum both signals at the error amp....

The error amplifier arrangement is in my mind a product of the average current mode control scheme, where you apply lowpass filtering to (setpoint voltage - measured current) and decide on this value how much pulse-width you like.

Regarding the DC-component of the input current, I am quite convinced that you will retain this since either current transformer has time to reset to zero while the other conducts.

[Eva]

With some external help, SG3525A can do both average or peak current control, or both (average with cycle by cycle peak limiting). Average control requires an external op-amp (or two because the one included is not quite good). Peak limiting requires a comparator with its output tied to the shutdown input. I used both methods in my 15V 125A power supply.

[ChocoHolic]

...I see..
Your current sensing with two transformers has a fundamentally different transfer function vs a single transformer in the center tap. Good point

[ChocoHolic]

Originally posted by Eva
.....Peak limiting requires a comparator with its output tied to the shutdown input. I used both methods in my 15V 125A power supply.

That's exactly what I mean.
It should be possible to use the shut down pin of the SG3525 directly for cycle by cycle peak current control. The shut down does directly cut the pulse, that's nice.
But it also does cause a slow discharge of the soft start cap, which might cause issues if you run this control loop acting every cycle. Or not? I was considering an external current source (or in simplest way just a resistor to Vref) at the soft start pin, by this it should be possible to override the soft start discharge function and use the shut down input like a regular cycle by cycle control input.

[jegandren]

Originally posted by Eva
...Peak limiting requires a comparator with its output tied to the shutdown input. I used both methods in my 15V 125A power supply.

Dear Eva,

Can you post any schematics on how it is done..

Many thanks.