LMR23625: LMR23625CFDDA fails to start up with a capacitive load

I don't support LMR23625 directly but I think you are probably entering hiccup mode current limit. The required current at start up just to charge the output capacitors is i = Cout * Vout / tss = (2200uF + 2*47uF) * 7 V / .002 sec = 8.03 A which is well above the current limit specification for LMR23625. Can you check to see that is what is happening? Moving to the WSON package would solve this if current limit is in fact the issue.

Hi,

Thanks for the quick reply.
How would I quickly check if that is what is happening?

Thanks,

Graham

Look for waveforms similar to figures 30 and 31 in the datasheet. It helps if you can monitor inductor current using a wire loop in series with the inductor and a clamp on current probe. The output voltage won't be flat zero volts like a short circuit, but should start to ramp up during the time the part is switching then fall back down during the hiccup time.

Thanks, I'll take a look.
The output isn't 0V, it's around 3V, but I haven't looked at it on a scope yet.
I'm looking in the datasheet, and which value would allow the WSON package to work do you think?

Thanks,

Graham

I do have a high power 5V supply also on the design (5V 5A), I wondered about diode 'OR'ing' the two supplies together so the 5V supply can help the 7V supply initially?
Bad idea?

Graham

Any more ideas?

Thanks,

Graham

Hi,

Does anyone have any more help or feedback on this?

Are there any tricks I can use to make it work?

Does it require a different part?

Thanks,

G

WSON package has a higher current limit and longer SS time.  You should have one third of the start up current.  I would probably want to check the corner cases to ensure it would work in all conditions.  Otherwise you will need to look at something with adjustable slow start time.

Hi John,

Thanks for the reply.

Can you please elaborate further on the one third of the startup current comment please?

Could you suggest a good part with slow startup? One that was at least a similar footprint would be nice.

Thanks,

Graham

If you look at my original response, you can see:

The required current at start up just to charge the output capacitors is i = Cout * Vout / tss = (2200uF + 2*47uF) * 7 V / .002 sec = 8.03 A

If you replace the 0.002 sec SS time with 0.006 sec, the current will reduce by one third 2.68 A.

I do not directly support that particular product line. None of the parts I support directly will have a high enough input voltage range for your application. At least at nominal conditions, the WSON package variant seems to have plenty of margin at 4.0 A min current limit. For an alternate, I would suggest using Webench design tool. You can use the advanced filters to search only for parts with SS function.

Great, thanks for the reply.
I've looked and I can't find the WSON package available to purchase anywhere :-(

Regards,

Graham

Looks like that might be in pre-production. Let me see if I can find you a marketing contact.

Thank you John. We're in the UK.

Please can you share more details on your application and potential volume. You can contact me directly at j(dash)tucker(at)ti(dot)com.

LMR23625 WSON will release around March/April18.

If you need higher current, you can look at LM73605/6 but this is more expensive than LMR23625 WSON and not even remotely in the same package.