High Current Power Supply

[mr_ghz]

Hi all

I have to design a Power-Supply which can deliver a current of 150A for a time of 150us into a laserdiode (voltage =2V). This pulse is given every 1000us.

The design of the Supply itself is not the problem (multi-phase PWM design), but I have to store the energy of 1 pulse (150A @ 150us). If I use a capacitor of about 1500uF, I can work with a inputvoltage of about 12V.

My problem is the large ripple-current in the capacitor. Discharging with 150A (150us) and charging with about 25A (850us). So my questions:

- Which type of capacitor is able to handle so large ripple-currents (I found multi-anode tantalum capacitors, but they only allow <5amps)?
- Is it allowed to parallel about 10 tantalum-capacitors (or other caps)?
- Is the ripple-current limit of a capacitor a thermal limit? (so can I increase ripple-current if I allow a 'cooling' time after a couple of pulses)
- Has anyone experience with such stresses on caps (burning tantal-capacitors look nice.... ;-)

Thanks

 

[IanP]

One option are Tantalum Hybride capacitors .. here you will find more details:
**broken link removed**
Also,you may take a look at extramaly-low ESR electrolytic caps ..
In all cases connecting caps in parallel is not a bad alternative ..

Regards,
IanP

 

[mr_ghz]

Hello IanP

OK, this was also my approach to connect several caps in parallel. I have calculated that I use a ripple current of about 40-50A. So I have to parallel at least 10 capacitors. This is a problem of space.

In my opinion the max. allowed ripple-current is given because of the maximal temperature of the capacitor. Is this correct? When I allow the capacitor to cool down from times to time (approx. 5sec using, 100sec cooling) can I run with higher ripple currents? (Lifetime of the capacitor is also not a big problem, 1000h will be enough).

regards

 

[VVV]

You are right about the capacitor ripple current rating, it is simply based on dissipation (Irms^2*ESR) and can size, for a given temperature rise (usually 10°C or 20°C). But the rating applies to caps at maximum rated temperature, say 85 or 105°C. If your ambient is lower than that, you can increase the ripple current.

As for the life of the cap, the number of hours given in the datasheet (1000~7000) this refers to the cap again at the rated temperature. But each 10°C difference results in doubling the life of the cap, so life should not be a problem. But if you want, you can "cook" the caps.
In other words, if you have to, you can increase the ripple until the cap reaches 85°C at your maximum ambient and you will still get 1000 h of operation (crude approximation)
This approach would allow you to use fewer capacitors, in order to save space.

One other thing you should be careful about is the high charge/ discharge current, not just the ripple. So you may want to consider caps that are specifically designed for high charge/ discharge currents. Typical caps like these are those used in flashes, where the high discharge currents occur. Series QS caps from Nichicon, for example, are adequate. The trouble with these is that they are necessarily high voltage caps and so they come in low capacitances and large cans.
Try to see if other manufacturers rate their caps for high charge/discharge currents, as this seems to me an important point. That is because the internal contacts may not take the high inrush current, even though they can easily handle the RMS. I would pay special attention to this.

For more info, see this article, too.
www.kemet.com/kemet/web/homepage/kfbk3.nsf/vaFeedbackFAQ/D68F1B97FC4FE31385256F72006669BD/$file/Technical%20Update%20-%20Ripple%20Current%20Capabilities.pdf