LM5123-Q1: [UPDATED] Disasters with LM5123 boost converter design

Benedek Tomka

Part Number: LM5123-Q1
Other Parts Discussed in Thread: CSD18514Q5A, CSD17573Q5B, CSD18540Q5B

Dear Colleauges,

We are between a rock and a hard place with this chip.

This is our second attempt after, in the first one we had problems with VCC regulation. We now redesigned the converter with the help of WEBENCH, manually verifying value selections. The board was also rerouted following the official layout guidelines of the datasheet.

Please look at the attached design, and WEBENCH document. Our schematic is a simultaneous design of a 19V and 28V optimized design which share a PCB.

Neither of them works.

On both designs bottom mosfets are driven to be an effective short (as observed in our input power measurement setup) after a short grace period, and they overload the input supply as burning to a total short. After the mosfet is removed, the chip boots up and provides the expected 5V clean VCC and 1V reference, and the feedback voltage on the TRK pin is totally right. If a new mosfet is placed, it burns out just like the previous.

Could anyone explain this behaviour?

3252.power-module-boost.pdf 19v.pdf 28v.pdf

All of our chips and parts came from Mouser.

over 1 year ago

0 Niklas Schwarz over 1 year ago

TI__Mastermind 20510 points

Hi Benedek,

Thanks for using the e2e forum.

First I would check if there are any issues within the design.
I would recommend to use our LM5123 quickstart calculator tool for this. The tool can be downloaded from the product page:
https://www.ti.com/product/LM5123-Q1#design-tools-simulation
Please check the calculation and see if the design can be optimized in some way.

If the failures still persist, it would be helpful to take a closer look at the schematic and parts.
For this, could you please send us the schematic with the according input/ouput parameters? (Vin range, Vout, Iout)

Thank you and best regards,
Niklas

0 Niklas Schwarz over 1 year ago in reply to Niklas Schwarz

TI__Mastermind 20510 points

Hi Benedek,

I hope customer able to improve their system.
As there were no further comments, I will close this thread for now.
If you have any additional questions, simply reply below to re-open the thread.

Otherwise, feel free to click on the resolved button.

Thank you and best regards,
Niklas

0 Benedek Tomka over 1 year ago in reply to Niklas Schwarz

Prodigy 10 points

Updated the post as the problem evolved

0 EM over 1 year ago in reply to Benedek Tomka

TI__Genius 11215 points

Hi Benedek,

Could you please point out what has been updated? Did you modify the initial thread? Has anything been modified after the suggestion of double checking the circuit with the Excel Quick Start calculator?

Kind regards,

0 Benedek Tomka over 1 year ago in reply to EM

Prodigy 10 points

The values has been crosschecked before the suggestion already.

The original thread was on the same batch of this IC bought from Mouser.com, some of them showed unexpected VCC regulation voltages and were useless.

A new layout was created by me because I've actually looked up what my colleague done on the internal layers and it was "interesting" to say the least, and we ordered a new batch of IC's from the same distributor.

Now this batch has the specified 5V and 1V reference rails, but it wants to drive the suggested mosfet to death :)

0 EM over 1 year ago in reply to Benedek Tomka

TI__Genius 11215 points

Hi Benedek,

What are the saturation and RMS ratings of the inductor? It might be that the inductor is saturating and therefore tends to behave like a short consequently damaging the FET. What is the voltage of the Miller plateau of this mosfet?

Kind regards,

0 Benedek Tomka over 1 year ago in reply to EM

Prodigy 10 points

In both circuits the RMS and saturation values were selected with careful design.

On the 28V model we have a 74435582200 WE inductor, which is rated for 15A and we are pulling 4.25 through it + 0.6A pp, and int this Ipp percentage range the saturation is somewhere between 5.5A and 18A, which seems more than enought. The fets are also from TI, specifically CSD18514Q5A which has a miller plateau voltage around 3.2V as far as I know.

The 19V model has a 74435580680 WE inductor, rated for 28.A RMS (our circuit is around 12.2A + 1.5App) and the saturation shall be more than 9A. I thinks this part is fit for the job too. We are using two differents fets as per the recommendations of the WEBENCH, the lower one is a CSD18540Q5B with a threshold of around 3V, the top fet is a CSD17573Q5B which is around 2.5V.

Sadly we dont have the equipment to measure actual values and have to go with the advertised manufacturer ratins, but with these safety margins we should be safely in range, I believe.

0 Brigitte over 1 year ago in reply to Benedek Tomka

TI__Guru 59715 points

Hello Benedek,

Did you fill in the calculator for your 19V and 28V designs as asked by Niklas? Please attached the filled in files in this thread.

The above layout is not good for a power supply converter. So it is possible that the whole measurement system of the controller does not work because of the layout.

In these 4 short articles, you can find recommendations for a good layout for a buck-boost converter. Most of these recommendations are generic and can be used for boost converters as well. Please make sure that you take special care on the power and return paths as well as the sensitive nodes. Please go through these articles:

https://e2e.ti.com/blogs_/b/powerhouse/posts/four-switch-buck-boost-layout-tip-no-1-identifying-the-critical-parts-for-layout
https://e2e.ti.com/blogs_/b/powerhouse/posts/four-switch-buck-boost-layout-tip-no-2-optimizing-hot-loops-in-the-power-stage
https://e2e.ti.com/blogs_/b/powerhouse/archive/2017/03/01/four-switch-buck-boost-layout-tip-no-3-separating-differential-sense-lines-from-power-planes
https://e2e.ti.com/blogs_/b/powerhouse/archive/2019/02/20/four-switch-buck-boost-layout-tip-no-4-routing-gate-drive-and-return-paths

My recommendation at the moment after going through the calculator for both converters, buy an EVM and replace the components there with the ones you want to use. Then you can exclude the layout from the problems.

Best regards,
Brigitte

Power management

Power management forum

LM5123-Q1: [UPDATED] Disasters with LM5123 boost converter design