LM5008: Touchy buck converter tests

Hi Yohann,

I can answer the third part. It seems like the wire parasitics are causing oscillations at the input. You can add some extra electrolytic capacitors at input to fix this.

As for how exactly we test our devices, that information is confidential. I am not sure there would be any equipment that can test all the parameters automatically without having specialized boards and test programs. Which parameters specifically do you want to test?

Regards,

Niranjan

Hi Niranjan, nice to have reply so quickly, thank you.

There is no parameter in particular.
I 'm trying to study the best way to test this kind of component the most efficiently.

The possibility of finding an ATE for these components seems indeed unlikely.
Given that it is likely in a few months that I will be testing several buck converter references, I do not want to make a mistake in testing methods :)

Having a look at LM5008 datasheet, suppose that I want to test:

vcc regulator output + in operating and suhtdown current + curent limit threshold + ton + FB reference threshold.
If i have well understood the datasheet:

At first, I can use a digital test equipment (such as ATE Teradyne J750) to test all these parameters without using external circuitry on SW output (coil, capacitors and resistors).
(I will polarize FB in nominal with the ATE to simulate a good behaviour of regulation)

I plan to use a board with socket to test several component. If all this tests are well controled, i finally plan to add external circuitry with coil, resistors and capacitors, to verify it works fine with a vout output.

I like the idea of ​​gradually going into understanding the component

Is it a realistic approach according to you?perhaps you will tell me a socket will compromise all the tests?

Hi Yohann,

I am not very familiar with ATE. But I think it would be easier (and cheaper) to test the above listed parameters in closed loop itself using scopes, DMMs etc.. You should also be able to automate that to some extent using software like LabVIEW.

I would recommend getting the EVMs of the devices you want to test, from their respective product pages. That way you do not have to pick out the appropriate external components each time.

Regards,

Niranjan

Really a pleasure having your returns Niranjan,

In summary:

1) no ATE for buck converter exists
2) I shouldn't try to test the converter in open loop (even if I simulate feedback by forcing a voltage level).
3) I need to get EVMs, use DMMs, and an oscilloscope.

Since I have to test batches of 100 parts, do you think it's possible to route a board like the EVM with
 a test socket instead of the soldered component? Or will using a socket necessarily disrupt
the stability criteria for regulation?

And do you think it's possible that a buck converter could be destroyed if its regulation isn't controlled and starts oscillating?

Hi Yohann,

1) That is not what I meant. You might be able to setup an ATE with specialized boards (you might need different boards for different devices), but I think it would be easier and cheaper to test on EVM boards using DMMs, scopes etc.. Sorry if I wasn't clear before. I do not have expertise regarding ATE so I can't provide anything more specific on that.

2) You can force a voltage level at the feedback, provided it is within the absolute maximum ratings specified in the datasheet. You can remove the top feedback resistor from the EVM to make it open loop.

Yes, you can route a board with test socket in it. It shouldn't affect the stability too much.

Device damage ideally happens only if any of the absolute maximum ratings are violated. If very high swings happen at VIN due to some long traces or large ringing happens at switch node due to poor board parasitics etc device might get damaged. To avoid these I would suggest adding an electrolytic capacitor at the input and following the layout guidelines given in the datasheet. Other oscillations such as those resulting from loop instability shouldn't damage the device usually, just bring it out of regulation.

Regards,

Niranjan