Hi Vinh,
Please let me know if you can offer any further advice.
For my next step in the investigation, I have ordered an EVM3650-QW-00A. I’ll test the EVM under the same conditions and see if I can replicate the behavior. If I can’t, I’ll remove one of the MPM3650 devices from my board and fit it to the EVM to see if the behavior follows the device. This should let me know if there is an issue with my design, or if the component has sustained damage somehow.
Hi Vinh,
Just another update…
I have replicated the fault using the MPM3650 EVM, to eliminate design factors. The only modification to the EVM were the necessary changes to R3 and R5 in order to set the output voltage to 8.4V.
The issue occurs between VIN= 11.87 --> 12.30V on the EVM, with a ~5ohm load.
1 Like
Hello danltec,
I apologize for not replying. I was not getting an updates for this post. I will need a couple days to get back to you regarding this issue.
Thank you,
Vinh Tran
Field Applications Engineer
1 Like
Hello danltec,
Is it possible for you to use an electronic load to do your test? It looks to me the device is in hiccup mode. This occurs during OCP or SCP. Your load current is too low to be OCP. However, you are using 5ohms to create 1.68A. I wonder if the 5ohms is some how triggering SCP. I don’t know why only at a certain voltage though. The hiccupping theory makes sense to me because when you removed Css the period got shorter. There is a tss between each hiccup.
Can you measure the the voltage at the pins of the device to check if Vin is as said on the on the power supply. Maybe the voltage drop on the cables
Can you measure the fsw of the device under working conditions and the bad condition?
Thank you,
Vinh Tran
Field Applications Engineer
Thanks Vinh, I’ll try to test with an electronic load and report back in the next couple of days.
I agree it appears to be hiccupping. The device “knows” it’s shutting off the output, as the PG line drops low as the output shuts off.
I’ll also try alternative input power sources, although I don’t believe there is an issue there.
I have previously observed the input power… in the oscilloscope image above the Yellow/Ch1 trace is the input (granted, it’s difficult to see in the small image). It’s holding steady at a little over 12V.
This device doesn’t have a pin on which the fsw appears, as far as I’m aware. Is there another way to check this?
Hi Vinh,
I repeated the test with a different power supply (@12V), different cables, and an electronic load.
Oscilloscope:
Yellow/Ch1: Input Voltage
Blue/Ch2: Output Voltage
Up to 1.5A, the output was stable:
At 1.6A, up to about 2A, the output was hiccupping:
Note that I had to run the electronic load in “Constant Resistance” mode for this particular load point, as setting it to constant current @1.6A caused the MPM3650 to switch off its output permanently, and thus I couldn’t capture the behavior in an image easily.
In all other cases, the load was set to constant current mode.
Hello danltec,
It is very strange to me that it shuts off only at ~1.6-2.0A, but not at the higher currents… I will need to check with the PL. This will take a couple of days.
Thank you,
Vinh Tran
Field Applications Engineer
2 Likes
Go get a resistor, they don’t have fancy control loops.
Waiting for answer
resistor not working too
Hello all,
An update to this question. I will be getting an EVM to test myself. If possible could you provide SW node captures?
Can you try with R3=20k and R4=1.54k?
Thank you,
Vinh Tran
Field Applications Engineer
1 Like
Thank you Vinh.
I can’t try SW node captures as this device has an integrated inductor, SW node is not available externally.
I’ll aim to try the altered R3 and R4 values in the coming week. Please keep me posted on your progress too.
Hello danltec,
There are SW node pins you can measure from.
Thank you,
Vinh Tran
Hello danltec,
Just an update on this. I have been able to recreate your issue on my end. Still trying to figure out what is causing this issue.
Thank you,
Vinh Tran
1 Like
I have the same problem.
With 12V input and 5V output, the shutoff is about 14mS with about 3A load.
Hello All,
The issue is related to a DCM and CCM transition issue. This issue will be fixed in the next revision.
It is recommended to switch over the MPM3650C if your output voltage is below 3.3V. The MPM3650C has FCCM.
If work is needed at no load conditions the MPM3632S or the MPM3683-7 are recommended.
If you need high efficiency at light load conditions it is recommended to use the MPM3630 or the MPM3683-7.
If your operation stays in CCM or DCM, then it is fine to keep using this device.
Thank you,
Vinh Tran
Field Applications Engineer
2 Likes
Thank you Vinh, great to hear you’ve identified the source of the issue.
May I ask when the next MPM3650 revision is planned?
I intend to switch to the MPM3630 in our next PCB revision, as I have the EVM for it, and it appears to be working very well in our application. However, it would be great to know when a revised MPM3650 would be available, as it would allow us to use our current PCB’s. It’s a great device too in terms of high current output in a small package, and ease of use, so it’d be great to have the option to use it with variable loads in the future.
Hello danltec,
Our next revision for this device will be around 2-3 months. If you would like more business/technical support please reach out to MPS NOW Remote Support.
Thank you,
Vinh Tran
Field Applications Engineer
1 Like