I blew a Buck measuring the current in the AC Current mode :-(

[jwhitmore]

So I made a stupid mistake and toasted the BUCK (ADP2302ARDZ), as well as the PCB Under it. All I did was have the 12V Supply going through the DMM to measure the current but the DMM was in AC Mode.

I'm now trying to understand what exactly happened. Obviously don't do that again, but I'm surprised at the outcome of the mistake. Other things were powered off the 12V Supply rail so maybe it's to do with those, but I don't think they were drawing current at the time. I've been searching around but I can't see anything that answers my curiosity.

If anybody knows why that would happen I'd love to learn why.

[Muxr]

AC or DC mode shouldn't make a difference, you're just not going to get the readings you were after. I would trace your steps and make sure you didn't somehow create a short.

It is really handy to have a current limiting PSU when working on power circuits.. it can save you lots of components.

[Twoflower]

Can you please explain what you've done? My understanding is that you replaced the load by the DMM. If that's true you shorted the PSU. Because the amp-range of your multimeter has a very low input resistance.

[hiperco]

A meter connected in current measuring mode is essentially a wire (well, a low value resistance to be more accurate).  The meter goes in series with the load, and then the meter measures the voltage drop across the internal resistance to calculate current.  AC or DC shouldn't matter from a connectivity standpoint.  How did you have things connected?

[capt bullshot]

AC or DC current setting shouldn't matter.
What matters is the additional inductance (just the wires, the DMM shouldn't have too much inductance) that you place into the ICs power supply input by connecting the DMM.
A common thing with these small buck converter chips is that they don't have any headroom for input overvoltage - AFAIR this is a 20V abs max specified chip and so may blow up at maybe 22V.
By placing inductance into its power input, the switching causes more or less ringing, then you have your usual ceramic capacitor here which doesn't really dampen the ringing - this way you can generate enough voltage at the input to blow the chip.
This is my guess what happened, there's an app note by Linear Technology adressing this issue.
In short: If you want to measure the input current of such a converter, or have a somewhat longer input power wiring, always place an additional electrolytic capacitor in parallel to the ceramic one, this does a good job to dampen the ringing. Never insert the DMM between the capacitors and the ICs input.

[Damianos]

So I made a stupid mistake and toasted the BUCK (ADP2302ARDZ), as well as the PCB Under it. All I did was have the 12V Supply going through the DMM to measure the current but the DMM was in AC Mode.

I'm now trying to understand what exactly happened. Obviously don't do that again, but I'm surprised at the outcome of the mistake. Other things were powered off the 12V Supply rail so maybe it's to do with those, but I don't think they were drawing current at the time. I've been searching around but I can't see anything that answers my curiosity.

If anybody knows why that would happen I'd love to learn why.

AC mode is irrelevant, as already mentioned; know your instruments to understand what are you doing.
If another thing had an over-current then that would have the problem!

Because the device has over-current and short-circuit protection, a first thought would be an over-voltage. But you said that the PCB has also burned and this "needs" a high current.
The mystery can be revealed if you remember where you connected the DMM.
   
BTW: Did you verify that the device was destroyed or just does not work due to missing traces?

[jwhitmore]

The connections were simple. Pull the positive from my Power Supply and connect from the +ve of PSU to the Amp Input of the DMM and then from the Com of the DMM go to the load. So the DMM is in series with the load.

Somebody asked about current limiting and that would be a great idea but I was testing a 65W load switched by an Infineon Profet chip. My current limiting PSU limtis to 1.5 Amps so it wasn't going to give me the power I wanted. The Infineon Profet is controlled by a uC hence the 12V is supplying the Profet but also going through a Buck to give me lower DC Voltage for uC.

Somebody asked if I'd really damaged the BUCK. I'll attach a photo. I really damaged it. ;-)

[glarsson]

Tis but a scratch.

[hiperco]

Do you "hot plug" the power (through the meter) to the supply?  If so, maybe you got a spark which overvoltaged the supply.  Or maybe the inrush killed it?

[Belrmar]

Well gottas say that looks nasty. Its weird that the buck failed if the load was the same , obviously there is something wrong that we didnt´have in account.

Having it in ac mode will only tell you the ac component of the signal across the current shunt of your dmm

[capt bullshot]

I absolutely second blueskull. Your picture confirmed my guess, you've got a tiny ceramic capacitor at the input pin, and used a long wire. Read AN88: http://cds.linear.com/docs/en/application-note/an88f.pdf This is an explanation of what you've done.

[jwhitmore]

Thank you, thank you, & thank you.

AN88 is the answer to my issue. I'm not sure how to proceed given that education but at least now I have learned something. The datasheet for the ADP2303 says to use a ceramic X5R because of it's low ESR but AN88 suggests that a low ESR will increase the problem. A further problem is that I added 2 x 10uF and should probably only have had one in there. I think I added two footprints to the prototype in case I needed to change things and populated it with two 10uF.

I was testing this on a mains PSU giving me 12V and I must have hot plugged the circuit. The circuit will ultimately be going into an automotive 12V supply, so I think I'll always be "hotplugging" the device. It's not like the 12V will gradually build it's just there.

I've only read AN88 once so it's early days. I have too much capacitance there is one problem but I'll re-read this a few times and consider my options.

Thanks again for taking the time to point out that document. 

(Software is easier ;-)