LM27342 Circuit Help

[samuelkadolph]

I've been working on a circuit to convert a supply voltage of 19V to 12V and 5V to power a hard drive using the TI LM27342 chip. But I've had a few issues.

The first issue was not getting the correct voltage but I got some help and added some resistors to add a minimum load (7.3.3 in datahseet). Now the chip on the 12V would only last seconds before dying and heating up enough to melt the solder underneath. The 5V section lasted a lot longer and I was able to power a 2.5" HDD. It worked for over an hour before I power cycled and then the chip died.

Am I missing something obvious? This chip seems really delicate. I didn't notice section 10.1.4 for PCB layout recommendations until after I needed to get help, seems unlikely it would cause my circuit to fry the chip though. And I know my inductor is a different value than TI's Webench Power Designer's reference schematic, that's because at some point they changed it and went from 3 input caps to 2 and reduced the inductor value.

• LM27342 Datasheet
• Ref Schematic
• Schematic
• Gerbers
• Board

[ledtester]

I don't have any advice for troubleshooting your circuit but you can now get a TI eval board for only $6.28 from Newark (down from $53.90):

https://www.newark.com/texas-instruments/lm27342myeval/dc-dc-buck-frequency-sync-evaluation/dp/04R9153

Might help you figure stuff out.

The eval kit's datasheet:

https://www.ti.com/lit/ug/snva381b/snva381b.pdf

[uer166]

At 2MHz switching frequency and 10ns risetime you need really good layout with low inductance switching loop. Based on the pic it looks like your loops are extremely large/long, with thin traces. 1 inch of 50mil microstrip is ~20nH, which at 2A/10ns results in 4V overshoot (surprisingly low?). In any case that would put it right at the abs. max rating.

How much current does the HDD draw from the converter? Does it fail at no load? Does it fail only with HDD attached? Did you attempt to measure/scope the switching node?

In any case such a fast converter needs a much, much better layout to function.

[samuelkadolph]

Thanks uer166, that does seem reasonable. Do you think I should make the FB loop might tigher or look for a lower frequency chip? I just went with one of the suggested designs from TI's website.

One of my full sized drive draws 1.7A@12V during spin up then between 0.1-07A after that. And a max of 0.7A@5V. The chip wouldn't work at all without a load and didn't produce a voltage on the SW pins.

I've been playing with TI's Power Designer again and this new chip LMR51420 came up. Seems like a better idea but harder to get right now. I've also been looking for other switching chips and stock is zero for most.