Electronics > Projects, Designs, and Technical Stuff
LTC3895 step down converter burns off
(1/6) > >>
Srijal97:
Hello all,

We're a group of 3 undergraduates working on a 140V to 32V DC-DC step down converter using the LTC3895, with a current requirement of 6-7A nominal and up to 20A burst. Circuit simulations were done in LTPowerCAD and LTSpice and everything seemed alright.

We went on to design the PCB following the guidelines for SMPS PCBs that we found online, mainly: https://micro.rohm.com/en/techweb/knowledge/dcdc/dcdc_pwm/dcdc_pwm03/2734
The OVLO(Overvoltage Lockout) pin was used to set to 150V. EXTVCC was not used, but there is a MOSFET for NDRV. The picture of designed PCB and the Eagle files are attached below.

We got all the components(major ICs from Digikey and others from local vendors) and assembled the PCB, and upon testing with 120V DC input, we had no output on the output side. After some time the LTC randomly blew up on the pin 30(Vin) pin with the magic smoke out. The 120V input DC supply was a really old one and was acting strange(voltage not changing smoothly when turning knobs) and we thought that might be the issue. So we assembled a new PCB with all new ICs. This time for the input, we used a step up module we had lying around to get 80V DC from a 48V SMPS. Even with the 80V input, the new board instantly had the same pin on the LTC blow up as soon as the input was turned on. A picture of the blown IC is attached below.
 
At this point we think the following points might be the issue:
- Both the above tests were done without any load attached on the output and we realised afterwards that this is not to be done for certain SMPSs. Is this an issue in this case?
- The input ceramic capacitor C6 was not soldered because the ones we purchased had high ESRs of 1.45 to even 14.65 Ohms, and we thought that we'll wait till we find better capacitors.
- The board layout. Even though we tried our best to follow the guidelines, this is a first for us and we expect some issues with it.


Now, we have 2 more LTC3895 ICs left to try out and wanted to make sure that we have success with them. Any help would be greatly appreciated!

Thank you for your time!
Miyuki:
And what other capacitors? C3 and output capacitors what was used during test
Or was it powered without capacitors ? (then it must blow)

And then ground layout > Ground bounce  especially as you take it from both sides, you need meeting ground point at switching node not far as you have
Are current sense resistors really calculated right ?
Srijal97:

--- Quote from: Miyuki on January 06, 2019, 04:29:20 pm ---And what other capacitors? C3 and output capacitors what was used during test
Or was it powered without capacitors ? (then it must blow)

--- End quote ---

This was connected during the tests:
C3 - 47uF, 250V; Electrolytic
C6 - not connected
C12, C13 - 22uF, 50V; Electrolytic
C14 - 150uF, 35V; Electrolytic
C15 - not connected
C8 - 0.47uF, 250V; Ceramic
These were soldered on the bottom side of the board because we had plans for adding a heatsink later on.


--- Quote from: Miyuki on January 06, 2019, 04:29:20 pm ---And then ground layout > Ground bounce  especially as you take it from both sides, you need meeting ground point at switching node not far as you have
Are current sense resistors really calculated right ?

--- End quote ---

Do you mean that the GND trace from the IC to the GND pad is long? Can you please elaborate more on this?
The current sense resistors have been calculated to accommodate for current limiting values a bit more than 20A as far as I know, though I'll get that verified. But that shouldn't have been the problem because there was no load connected right?
Miyuki:
Not sure how this control ic react to no load, but shall be ok

You need to take ground from one point near switch as is in some schematics marked and look to some reference buck design for ground

Like this modification or even better to have ground plane you have lot of space
Miyuki:
You have problem with this loops:
Navigation
Message Index
Next page
There was an error while thanking
Thanking...

Go to full version
Powered by SMFPacks Advanced Attachments Uploader Mod