LM3488Q-Q1: Voltage drop under high temperature test

Hi Jass,

Thanks for reaching out. We'll get back to you with some feedback by today.

Thank you and best regards,
Bryan

Hi Jass,

Thanks for providing the details of this case.
Please allow me some additional time to review the schematic.

I will get back to you beginning of next week.
Best regards,
Niklas

Hi Jass,

Based on the waveform measurement, it is not fully clear what caused the drop of Vout initially.
If the device shut down because of overtemperature protection, Vout would drop and ramp up again after the device resets.
To confirm this, a measurement of the DR signal would show that the device stops switching.

Stability can be an issue as well. If the system is barely stable under normal conditions, worst case scenarios and small drifts due to high temperature can lead to instability, which can cause resets as well due to overcurrents or voltage peaks.
If there is margin phase margin during normal conditions, the device will still run stable under worst case scenarios.

As the behavior did not reoccur after changing the compensation, it is likely that instability was the root cause.
If the resets happens again, please let me know and we can check if there can be another root cause, like overtemperature reset.

Best regards,
Niklas

Thanks! How to calculator the phase margin of 3488, I didn't see calculator tool as LM5155 at ti.com.

The waveform didn't show oscillating, only dropping. Will compensation cause this? I think lower phase margin will cause oscillating waveform.

Hi Jass,

We have a Power Stage Designer tool that comes with a compensation calculator.
https://www.ti.com/tool/POWERSTAGE-DESIGNER

This calculator works for LM3488 as well.
If you have a Bode Analyzer available, measuring the stability on bench will give the best result, but if this hardware is not available, low stability can also be visible in load transient measurements.

Best regards,
Niklas

Hi Niklas,

My customer is using 3488 for sepic converter, how can I use this tool? Many parameters are not shown in 3488 datasheet.

Hi Jass,

The SEPIC topology is not specifically listed in the control scheme, but I would assume a CMC Flyback would be the most accurate replacement. (I will need to check this with a compensation expert)
For alternative calculation guides, we have an app note on SEPIC compensation:
https://www.ti.com/lit/an/snva405a/snva405a.pdf

Reference designs of similar devices can be found here:
https://www.ti.com/tool/LM3481SEPICEVAL
and here
https://www.ti.com/tool/LM5155EVM-SEPIC

Best regards,
Niklas

Hi Jass,

Quick follow-up on the power stage designer tool.
It is indeed possible to use the Power Stage Designer for calculating a SEPIC topology.
Either use a Flyback with 1:1 winding ration or an inverting Buck-Boost topology. Both with give an acceptable estimate.

Best regards,
Niklas

Hi Niklas,

We find that it may be the bandwidth of the system.

The load may rise suddenly in 100us. With old version of compensation parameters, the loop bandwidth is low.

I see there is a huge change of Gm in different temperature. Can you confirm that will high temeratures cause a low Gm?

Thanks!

Hi Jass,

I will check with design if there is a correlation and let you know till end of this week.

Best regards,

 Stefan

Hi Stefan,

There is another possible reason.

The load is changed from 10mA to 400mA. 10mA is a extreme light load, if LM3488 keep 200kHz switch frequency and a hundreds ns minimum pwm time, the output may not stable at 5V. During high temperature test, the minimum PWM time will be over 500ns, causing 5V output to be 5.2V,,meaning the compensation system may be saturated. If a 400mA load comes suddenly, will the compensation system response slowly to have this big voltage drop(under 4V).

In this theory, after they add a fake load(25mA), the system will have strong control ability and pass high temperature test.

Hi Jass,

in this light load the device should go into skip mode, but this should be identical to lower temperatures.

So, I am not sure why this helps or changes for high temperature.

Best regards,

 Stefan

Hi Stefan,

I didn't see skip mode in LM3488's datasheet, can you show me where it is?

Thanks!

Hi Jass,

sorry, used the wrong name - it is not skip mode but over voltage protection.

So if in one switching cycle more energy is transferred then the output does consume the output voltage will increase and hit the over voltage protection level and the controller will stop until the output voltage has reached the lower hysteresis level again.

Best regards,

 Stefan

Will 3488 have low transient response in this over voltage mode?

HI Jass,

yes, the response will be lower as the controller first needs to get back into operation and regulation.

Best regards,

 Stefan