Hi, I have problem with MPM3650 shutdown.
When it powers up, there’s no problem.
But about half hour or more time elapses, it suddenly powers down and about 15ms later it powers up again.
I attached waveform.
Setup conditions are as below.
Regards
cmoh
I would bet output overload, how do you know it is 0.7A or worse a commanded act.
I measured current with external power supply and multimeter after removing mpm3650. Multimeter showed 0.7A value.
So if the load suddenly increased to 2A for 10 msec would your meter catch that? What is the nature of your load? I would try your circuit on a resistor alone and see what happens.
I used Fluke 87 V multimeter and I will try to find out whether it can catch up or not.
I’m not sure about meaning of nature of load, but if I understood right, load doesn’t change radically.
I would be trying to eliminate the possibility of load changes. It makes no sense to me, that the chip would malfunction after 1/2 hour for 5 msec. That speaks of something else is happening.
‘I would be trying to eliminate the possibility of load changes.’
Does it mean to test by connecting constant current load or resistor?
constant something power whatever load you have separately and load the MP to 0.7A and wait.
Alternatively you could power the application with a bench power supply with a fast current limit at 0.75A or something and see if that crashes after 1/2 hour
When I use a bench power supply, application has no problem.
was the bench supply current limit set to 0.7A?
A bench power supply isn’t good enough… different supply = different results. Put a series resistor in the DC feed and measure the volt drop with the scope, trigger off the 3.3V or PG and see what you capture. I agree, the issue seems to be current, you may even want to measure the temperature of the chip during this process.
You may also want to consider a Joulescope JS220. It measures current up to 3A over long duration’s and stores to a PC (designed for battery operated equipment).
Note I am not affiliated with joulescope, and have a NRF Power Profiler instead (way cheaper but only 1A max)
Simon
Hi, jshannon.
Sorry for late reply. I tested with bench power supply current limit set to 1A about 8 hours, but there was no problem.
Hi, speacock.
I am curious about the reason you think problem is related to output current. MPM3650 can supply current up to 5A and as I mentioned above, system’s nominal consumption current is less than 1A. I don’t understand it’s related to overcurrent.
And also I considered about temperature problem, so I used cooler but the result doesn’t change.
Are you darn sure this isn’t a commanded shutdown? This thing can be controlled via I2C could the software be telling it to do something stupid? This is perplexing
Generally SMPSU either starts or not. There can be issues with load capacitance which reduces the output voltage to (generally) the reference voltage for the error amp. But running for 30 minutes… that points to a load or temperature issue. Even if its capable of delivering 3A, the inductors should be chosen for a lower output current or it will run inefficiently (which can’t be done in an integrated device. It is probably running DCM mode and a current spike is causing the glitch. OR the 30 minutes is how long it takes to turn on at very low current (guess here)
Its quick and easy to use an integrated BUCK, but there are a lot of 5 pin, 1A buck converters which can be tailored for your load. Most can use a 3mm or 4mm inductor and 2 x 10uF caps. I have built a lot of devices using 1A BUCKs and apart from the error amp issue, none have failed (and they run continuously for up to 10 years)
Simon
Hi jshannon.
I’m sorry but have you reviewed datasheet of MPM3650? There are no I2C pins.
Sorry, busy toasting Joe Biden. Is your 5V source solid?
Hi, speacock.
I know that in case of 0.7A load, there is possibility that MPM3650 runs DCM mode and as you mentioned above, I want to know MPM3650 DCM mode characteristics.
And no offense, but speacock and jshannon, both of you are MPS engineers?
Yes, other ICs using 5V don’t have problem and if you see waveform I attached above, 5V waveform doesn’t go down.
Have you tried extra cooling, like blow a fan on it and see it that helps, desperation move I don’t think it will.