Why is the MP2225 output unstable?

[daisizhou]

Hallo everyone

I used the MP2225 to create a 15V to 12V DC-DC power conversion circuit.

I found that using the MP2225 with a 12V input and an output below 5V works normally, and the performance is very good.
However, when I select 15V input and 12V output, the voltage conversion seems to malfunction; the output voltage can no longer remain stable at 12V and instead changes with the input voltage.

I don't know why this is.

Can someone tell me why? Thanks.

https://www.alldatasheetcn.com/datasheet-pdf/view/1034237/MPS/MP2225GJ.html

[AnalogTodd]

The part isn't meant for output voltages above 5V, if you look at the ABS MAX rating on the BST pin it is VSW + 5.3V. When you try to go to 12V output you will get that trying to bump to 12V.

[tooki]

The part isn't meant for output voltages above 5V, if you look at the ABS MAX rating on the BST pin it is VSW + 5.3V. When you try to go to 12V output you will get that trying to bump to 12V.

I don’t think that’s correct?

The datasheet (current one from MPS’s website, rather than an annoying third party website) says it’s Vsw+5.5V. Vsw’s max is VIN+0.3V. So BST’s max voltage is VIN+0.3V+5.5V => 15V+0.3V+5.5V = 20.8V.

So where’s the issue?

Note that the MPS website has a simulator, and it doesn’t raise any errors when configuring 12V output.

[daisizhou]

The part isn't meant for output voltages above 5V, if you look at the ABS MAX rating on the BST pin it is VSW + 5.3V. When you try to go to 12V output you will get that trying to bump to 12V.

I don’t think that’s correct?

The datasheet (current one from MPS’s website, rather than an annoying third party website) says it’s Vsw+5.5V. Vsw’s max is VIN+0.3V. So BST’s max voltage is VIN+0.3V+5.5V => 15V+0.3V+5.5V = 20.8V.

So where’s the issue?

Note that the MPS website has a simulator, and it doesn’t raise any errors when configuring 12V output.

I actually built a 12V output circuit and found that when I set the output to 12V, the output voltage became related to the input voltage.

For example, when the input is 18V, the output is 12V; when the input is 15V, the output becomes 9V.

In other words, when outputting 5V, the entire circuit is completely different.

[tooki]

I suggest you play around with the simulator on the MPS website and see what it does. A circuit designed for 5V uses different component values from one for 12V — not just the feedback resistors. What might work for both voltages is to design it for a voltage in between the two limits (for example, 9V).

[daisizhou]

Thanks.

It seems I used too small an output inductor. I'll try it again.

[mtwieg]

The part isn't meant for output voltages above 5V, if you look at the ABS MAX rating on the BST pin it is VSW + 5.3V. When you try to go to 12V output you will get that trying to bump to 12V.

No, the peak voltage on the BST pin is completely independent of Vout.
The datasheet doesn't mention any upper limit on Vout, except the maximum duty cycle Dmax of 90%. Getting 12V from a 15V source shouldn't be an issue.

I actually built a 12V output circuit and found that when I set the output to 12V, the output voltage became related to the input voltage.

For example, when the input is 18V, the output is 12V; when the input is 15V, the output becomes 9V.

In other words, when outputting 5V, the entire circuit is completely different.

I'm wondering if the circuit is suffering from some instability. Can show some operating waveforms with a scope?

[daisizhou]

Thanks

I found that when adjusting the configuration resistors R4 and R5, which are 12V input and 5V output,Regardless of whether the input is 12V, 15V, or 10V, the output always remains at 5.0V.

However, when I adjusted R4 and R5 to set them to 15V input and 12V output.The output voltage of 12V is correlated with the input voltage of 15V; it does not always remain at 12.0V.

I don't know why.

[AnalogTodd]

I had missed seeing in the data sheet that the BST has a 5V regulation point on it, my bad.

Reviewing the original schematic, R4 and R5 form the resistor divider to set the output voltage. An important thing to look at is also the capacitor C9 that goes across R4. That capacitor acts as an AC low impedance at frequency to change loop gain (it acts as a zero in the system). If you haven't adjusted that capacitor value while adjusting the resistor divider, you may be getting your phase crossing zero while there is still gain in the loop and the circuit oscillates. If the loop is oscillating, all bets are off when it comes to the regulation point.

As was pointed out, there are a number of different component values in the circuit you showed from the MPS web simulation compared to what your original circuit had. Post some oscilloscope shots of different circuit nodes and people can likely help you better.

[daisizhou]

OK.

I tested the waveform of each pin

[daisizhou]

.......

[mtwieg]

None of those waveforms make any sense at all. Are those waveforms taken with Vout set to 5V or 12V?
Vcc should be a steady 5V.
EN/SYNC should be a steady logic high.
FB should be a steady 0.6V.
AGND and GND should be zero.
Are you sure you're connecting the scope properly?

[daisizhou]

The oscilloscope waveform was measured with a 15V input, 12V output, and no load. The oscilloscope's GND and the negative terminal of the 15V circuit are connected together.

[mtwieg]

Ok but something must be wrong, either with the scope, the probe, or the connections.

Starting with the Vcc pin, that should have a clean 5V on it. But your scope shows Vrms=1.159V, with 0.913Vpp of ripple/noise. If you measure Vcc with a DMM, what does the DMM say?

[AnalogTodd]

I immediately see some things on the board that concern me. First is the fact that you have a ceramic capacitor on the input. At 5V, that cap is going to be a much higher value than at 15V, you need to check what you have on there. Add that to the inductance of the wires and you have a high-Q circuit that will want to ring and oscillate.

First suggestion whenever you have a circuit and the input voltage is unstable like that is to add more capacitance (that you KNOW will be the value you think it is) until it is much more stable. Your output loop may be good for phase margin, but if there is that tank circuit on the input that has a resonant frequency close to that of your loop, the entire system will oscillate. I've seen it with linear regulators before.

[daisizhou]

I retested and replaced the capacitor

[AnalogTodd]

Clean waveforms, operating in discontinuous mode, that looks a lot better than what you posted before. Where does the output voltage now sit?

[daisizhou]

It still uses a 15V input and a 12V output, but the output voltage varies with the input voltage and is not a fixed 12V.

[AnalogTodd]

Well, you're in DCM and no load. This is supposed to be a 500kHz switcher and you're getting pulses at 20kHz, so it's entered its PFM area for light load. It's quite possible there isn't enough load on the output to overcome the 800mA switch current that runs in that mode--your resistor divider is only puling 60uA if you have the 10k resistor on the bottom where the design example that @Tooki showed uses a 200ohm resistor for the bottom resistor of the divider, giving a 3mA load. Try adding a load to the circuit and see if the results are the same.

If you're still getting the output following the input during CCM when you're not pulse skipping, then something else is happening.

[daisizhou]

Thanks.

I connected a 150-ohm resistor, and the output remained stable at 12.27 volts.

Thank you once again