Dear MPS,
I have done all computation and study MPS COT documents.
After multiple trial and test.
There is a puzzling behavior.
When MPQ4470 or MP4470 is cold aka 25 degree C.
The output voltage will be higher and 0.25V above resulted in 12.5V.
The output is targeting 12V and input range is 28 to 36V.
Once the board is heat up by the regulator the output is drop back to 12.25V
Please help on such puzzling behavior.
Bests.
Brian
Is this gradual effect? I.e a little bit warmer a little be closer to regulation or is it some kind of threshold like in the cold stable at 12.5 and then past a critical temperature suddenly drops to 12.25? Is there perhaps condensation in the cold on the board providing leakge paths around the voltage setting resistors?
I am very puzzle and to be honest.
The COT based converter is the most tedious converter chip compared to other MPS series.
MPQ and MP4470 is having all kinds of unexpected behavior.
Q:Is this gradual effect?
A: No it is a sudden effect.
Q:I.e a little bit warmer a little be closer to regulation or is it some kind of threshold like in the cold stable at 12.5 and then past a critical temperature suddenly drops to 12.25?
A: Unable to define the answer easily. The behavior is that when room temperature the converter run at a regulated voltage and when load is applied before powering the converter will result in a higher voltage at start and drop to a point after heats up.
The temperature is causing such effect this is quite sure.
Bests,
Oh that is confusing, so there is also some load effect? If it is cold and you start into load the voltage is wrong, then as the circuit warms up it pops into correct operation. If it is cold and you start into no load does it come up at the correct voltage? If it is warm it operates correctly all the time?
Q: Do the output voltage @ no load have fluctuation under cold start and warm operation?
A: No, when no load either cold start and warm start have no different as the regulator is undergoing a PSM.
Bests,
Brian
Hi Brian,
I have to agree that this part is more on the tedious side because of the ramp compensation needed… COT converters need to get inductor phase information somehow - newer COT converters have some internal ramp compensation so the need for an external ramp compensation is no longer needed.
That said - the phenonium you are encountering is not control architecture dependent. To set the output voltage, there is an internal reference voltage that is used as a compering point. This internal reference voltage varies by temperature, so then the output varies by temperature.
Also, take a look at how the FB resistors change R value with temp - that also affects Vout. And make sure to measure Vout right at the output caps to get the real output value the part is sending out.
Hope this helps,
Regards,
Jonathan
What you are trying to suggest is that the band gap voltage reference also deviate across different temperature.
But what you case is proposing is that the temperature across 25 to 60 degree will result in an sudden jump of load and line regulation.
I have never see other MPS buck regulator doing that MP(Q)4470 is the first.
Bests,
Brian
Could be something weird like some internal timing shifts with temperature. I bet they have bandgaps down. Is this on a board of your design or do you have a demo board? No offense, but we will assume the demo board is as good as it can be.
I had printed 3 EVB on this chip and tried over 10 chip and two supplier.
I don’t think people will spend time on remarking this MPS chip.
1st the chip is flip chip packaging with very uncommon packing.
2nd regulation on 5V is acceptable. I am reusing the chip on 5V to 12V but different resistor and ramp design.
Let me include the 5V line-load error data here:
This configuration is very stable and had aged test a lot.
So it is acceptable because the PSM and CCM crossing is what COT MPS document also proposing.
I can only suspect the converter have trouble on over 5V.
Bests,
Brian
BTW the situation is about 12V not 5V or 3V3 or below. I had tested other low voltage as well. Suboptimal ramp design will also stable enough to regulate within 2% or the proposed COT design document.
So If you have 12V MP(Q)4470 EVM or EVB that can be purchased share please.
Otherwise I cannot see EVM or EVB could be the solution as inductor and all resistor must remove and replace accordingly.
So this is same as printing out custom PCB like using JLC.
Bests,
Brian
