How to Adjust Output Voltage of a Boost Converter

[Imperf3kt]

Your setup is what I tried to achieve myself, but there were some drawbacks to the MP2636 that prevent it.

There is a pin on the IC which, when shorted to 0v will halt the boost module and put it in an ultra low power state (effectively off). Unfortunately, this pin is not accessible, so you need to disconnect the battery whenever the Pi is intended to be fully off.

The switch between the boost module and the Pi is not strictly necessary, I use it as a "master switch", you could always use the dt shutdown overlay and a push button to turn the Pi on and off - if using a switch like I have, always power the Pi off via a proper shutdown.


The MP2636 is not the best best choice for recharging while powered.
While it can do it, I found that some modules caused the Pi to reset upon insertion or removal of the charging cable that was later determined to be caused by inrush current.
You'll also need to either change the charging current by replacing an SMD resistor, or use a ~5A power supply so the Pi can still use 2.5A while the battery charges at the default 2.5A

[pixelpop]

I should have my MP2636 on Tuesday, so I'll sort things out then. But it would seem like, with the 115V PS connected and the switch closed, the Pi and the battery will receive power from the MP2636 (probably split, so I'll have to figure that out). If you disconnect the 115V PS then the Pi should get 5V from the battery. If you open the switch but leave the 115V plugged in, it should charge the battery. And if you disconnect the 115V PS and open the switch, the battery and MP2636 should be isolated.

At least that's what it looks like on paper. I'll report back in a couple of days when I get my MP2636.

[Imperf3kt]

Yes that's exactly what happens according to the datasheet of the MP2636.
The issue is that during the switching (600kHz, IIRC) there is a sudden spike of current on either the battery or the attached psu and this creates a phenomenon known as inrush current. This inrush current is not good for the Pi and usually causes it to reboot.

I gave up trying to fix that particular issue for my use case - I simply isolate the battery when charging now.

[pixelpop]

A few days ago I was battling the current issue and finally said "screw this" and bought a 10A buck converter. It should be here tomorrow. Then I'll use a 115Vac-12Vdc wall converter to feed it and then take the output from that to the MP2636. Maybe the buck converter, combined with the step-down rather than step-up, will smooth out the transients. I'll know more by mid-week.

[pixelpop]

Well, a slight setback. My MP2636 arrived today and I got exactly what I ordered...two MP2636 chips about 1/8" square. Not at all what I expected! LOL

So I went back to the site I ordered this from and took another look. Sure enough, if I scroll on down the page, there it is, the much needed DFR0446 board that incorporates the MP2636. I was soooo close.

So it will be a couple more weeks before I get the board because it's backordered until 10/11 (of course!).

[Imperf3kt]

So you got just the IC itself?

Can you point me to where you bought that? I've been interested in obtaining just the IC without the module for a while but didn't find anywhere selling it by itself.

[pixelpop]

So you got just the IC itself?

Can you point me to where you bought that? I've been interested in obtaining just the IC without the module for a while but didn't find anywhere selling it by itself.

Yeah, just the IC...or I should say two of them. If I knew how to get these to you I would. They were nice enough to refund the purchase price and told me just to toss them.

If you are in the US, this is where I got them (and the same place that has the DFR0446 on backorder):

https://www.mouser.com/Search/Refine?Keyword=mp2636

Good luck!

[pixelpop]

This looks interesting too:

https://www.monolithicpower.com/en/ev2636-r-00a.html

According to the product description:
When input power is present, the board charges a single-cell battery with a programmable charge current of up to 2.5A.

In the absence of an input source, the board switches to boost mode through the MODE pin to power the SYS pins from the battery with a current of up to 2.5A

Still have some figuring out to do, though.

[Imperf3kt]

Looks to be the same product in a different design.

Thanks for pointing it out though, it might be worth looking at as the IB pin is broken out.

[pixelpop]

Looks to be the same product in a different design.

Thanks for pointing it out though, it might be worth looking at.

It's more expensive, but at least it is in stock. Still need to figure out a third mode...power source applied but going to the Pi (battery bypass).

[Imperf3kt]

Still need to figure out a third mode...power source applied but going to the Pi (battery bypass).

The DFRobot module already incorporates this feature. Any power applied to the VIN (micro USB-B port) will bypass the battery charging circuitry and directly power the device while the battery is full.
(by the way, the actual term is passthrough)

[pixelpop]

Still need to figure out a third mode...power source applied but going to the Pi (battery bypass).

The DFRobot module already incorporates this feature. Any power applied to the VIN (micro USB-B port) will bypass the battery charging circuitry and directly power the device while the battery is full.
(by the way, the actual term is passthrough)

"while the battery is full"...does this mean that it applies full power of Vin to the device and completely bypasses the battery (no trickle charge)? Or does it charge the battery to full before transitioning to the pass-through state?

[saraparker]

I tried all these but nothing is working for me.

[pixelpop]

I tried all these but nothing is working for me.

We shall see. I should have both in the next couple of weeks. It may take some engineering (for instance, the Lipo Rider Plus has a shutdown output I/O command) but I trust creativity and perseverance to find a way.

[Imperf3kt]

"while the battery is full"...does this mean that it applies full power of Vin to the device and completely bypasses the battery (no trickle charge)? Or does it charge the battery to full before transitioning to the pass-through state?

I had to recheck the datasheet as it has been some time since I last studied it.
Going by what the datasheet says, while the device is charging, power will passthrough from VIN to VOUT and the charging current will lower from the default 2.5A if the voltage sags. I don't know how much by, just that it does.
Once the battery is fully charged (after its trickle charge stage), the full current is directed to VOUT and no charging occurs.

[pixelpop]

"while the battery is full"...does this mean that it applies full power of Vin to the device and completely bypasses the battery (no trickle charge)? Or does it charge the battery to full before transitioning to the pass-through state?

I had to recheck the datasheet as it has been some time since I last studied it.
Going by what the datasheet says, while the device is charging, power will passthrough from VIN to VOUT and the charging current will lower from the default 2.5A if the voltage sags. I don't know how much by, just that it does.
Once the battery is fully charged (after its trickle charge stage), the full current is directed to VOUT and no charging occurs.

Both of these modules (DFR0446 and Lipo Rider Plus) look promising but the proof will be in the performance. I just hope the spec'ed output amperage is true. It's going to a bit frustrating if I have to put a boost module at the output from these to get the right voltage/amperage to the device.