Raspberry Pi Forums

I purchased a boost converter for my Pi (to boost a battery's 3.7V to 5.2V), but I need some clarification. The instructions say to connect the load (my Pi) to the output of the converter before connecting the input (the battery). But a review of this converter says it comes out of the factory set to 12V output. Wouldn't applying 12V to my Pi be a very bad thing? And it would be 12V until I could adjust the output down to 5.2V. So how the heck to I adjust the output to 5.2V before I fry my Pi (nice rhyme) by applying 12V?

You are right, you must not connect your Pi before the output voltage is set.

Beyond that, I can think of two reasons why the load should be connected before powering the converter.

1. The regulator needs a load to work correctly.
2. The regulator reacts badly to sudden large step changes in load.

Both are due to poor design, there may be other reasons I haven't considered.

If either of these apply, you should not run the regulator without a load. So:

1. Connect a 1K resistor as a temporary load.
2. Connect the battery and set the output voltage.
3. Disconnect the battery.
4. Connect the Pi.
5. Connect the battery, check the output voltage.

Which model Pi do you have? All except the Pi0 have some form of overvoltage protection, some more than others. In particular the Pi4 is only protected by a TVS (transient voltage suppression) diode D1 that will sacrifice itself to protect the Pi. So the Pi itself probably survives, but D1 will need replacing if you apply too much voltage.

Using a resistor as a load seems like sound logic, but do you think 1K is enough of a load to get an accurate output voltage measurement?

I'm using a 3 A+, but I would rather not test its voltage protection.

I ran one of my Pi off a PSU brick that was adjustable in steps up to 18v.

One day the inevitable happened, the PSU got connected to the Pi when it was set to 12v. And smoke came out.

The Pi survived. Protected by the protection diode that jumped off the board.

Until I did it the second time...

All in all I would advise against using such a supply.

Heater wrote: ↑

Sat Sep 28, 2019 11:23 am

I ran one of my Pi off a PSU brick that was adjustable in steps up to 18v.

One day the inevitable happened, the PSU got connected to the Pi when it was set to 12v. And smoke came out.

The Pi survived. Protected by the protection diode that jumped off the board.

Until I did it the second time...

All in all I would advise against using such a supply.

Indeed. Note that if the TVS bails out physically, the Pi is lucky if it is not toast/fried/ex/dead.

pixelpop wrote: do you think 1K is enough of a load to get an accurate output voltage measurement?

A 1K resistor should be enough load to set the voltage, unless the regulation is so poor that the supply is useless anyway. You could use a lower value resistor, but make sure it can dissipate the heat. Don't go below 600R at 12V with a 0.25W resistor, for example.

pixelpop wrote: ↑

Sat Sep 28, 2019 9:08 am

I purchased a boost converter for my Pi (to boost a battery's 3.7V to 5.2V), but I need some clarification...So how the heck to I adjust the output to 5.2V before I fry my Pi (nice rhyme) by applying 12V?

Those I had until today were all adjustable at no-load, but the 10MΩ voltmeter.

Heater wrote: ↑

Sat Sep 28, 2019 11:23 am

I ran one of my Pi off a PSU brick that was adjustable in steps up to 18v.

One day the inevitable happened, the PSU got connected to the Pi when it was set to 12v. And smoke came out.

The Pi survived. Protected by the protection diode that jumped off the board.

Until I did it the second time...

All in all I would advise against using such a supply.

Did the boost converter fail or did the adjustment screw get moved somehow?

These Pi's are so (relatively) inexpensive that I don't mind experimenting "a little." I've already smoked one 3 A+ because I was on heroin (sounds better than just plain stupid) and had my multimeter set on Ohms when I placed the leads on the 3.3V and GND pins of the GPIO. So I headed down to Micro Center and got two 3 A+'s for $20 each (figured it would be smart to get a spare).

I don't mind using a boost converter. I tried using the PowerBoost 1000 from Adafruit, but was limited to 1A output.

davidcoton wrote: ↑

Sat Sep 28, 2019 1:54 pm

A 1K resistor should be enough load to set the voltage, unless the regulation is so poor that the supply is useless anyway. You could use a lower value resistor, but make sure it can dissipate the heat. Don't go below 600R at 12V with a 0.25W resistor, for example.

Interestingly, I found a web site that does Watts/Volts/Amps/Ohms calculations (https://www.rapidtables.com/calc/electr ... lator.html) and if you plug in 2.5A and 5.2V, you get a resistance of just 2.08 ohms. Seems like a pretty light load.

pixelpop wrote: ↑

Sat Sep 28, 2019 4:21 pm

davidcoton wrote: ↑

Sat Sep 28, 2019 1:54 pm

A 1K resistor should be enough load to set the voltage, unless the regulation is so poor that the supply is useless anyway. You could use a lower value resistor, but make sure it can dissipate the heat. Don't go below 600R at 12V with a 0.25W resistor, for example.

Interestingly, I found a web site that does Watts/Volts/Amps/Ohms calculations (https://www.rapidtables.com/calc/electr ... lator.html) and if you plug in 2.5A and 5.2V, you get a resistance of just 2.08 ohms. Seems like a pretty light load.

A light load is all that is necessary to tame the PSU output -- probably don't need anything, but it will help discharge the output capacitors. You are not trying to do a full load test, just avoid no-load (which should be alright, but then, why the warning?).
Did you calculate the power at 5.2V and 2.5A? Can your resistor handle that??

Yeah, I used a 1K and adjusted the output voltage without any issues. But now I have another problem.

I set the output voltage to 5.3V and connected the Pi. The output voltage from the converter is still measuring 5.3V, but the Pi is stuck in the (low power) boot loop. The converter I am using is a popular one (https://smile.amazon.com/gp/product/B07 ... UTF8&psc=1) and here are the relevant specs:

Input voltage range:3~35VDC(Note: the input voltage is less than 4V, voltmeter failure)
Input current:9A (max)
Output voltage range:3.5~35VDC(please to ensure that output voltage above on input voltage)
Output current: 6A (max)
Output power: 75W (when the input and output voltages greater than 20V up to 128W)
Conversion efficiency: 96.4% (max)

When I connect the 1K resistor output, the onboard LED displays the in/out voltages correctly and the POWER LED is illuminated, but with the Pi attached to the output, the display and power indicator are both off. Any ideas?

pixelpop wrote: ↑

Sat Sep 28, 2019 5:47 pm

The output voltage from the converter is still measuring 5.3V, but the Pi is stuck in the (low power) boot loop.

Use a good USB cable or wire directly to the Ov-5v test points.
The μUSB connector has got too much losses and most USB cables has ridiculously thin wires.

rin67630 wrote: ↑

Sat Sep 28, 2019 6:49 pm

pixelpop wrote: ↑

Sat Sep 28, 2019 5:47 pm

The output voltage from the converter is still measuring 5.3V, but the Pi is stuck in the (low power) boot loop.

Use a good USB cable or wire directly to the Ov-5v test points.
The μUSB connector has got too much losses and most USB cables has ridiculously thin wires.

I'll give that a try. I am just using breadboard jumper wires now, but I'll beef it up to production (18 gauge) wiring.

Well, I am starting to get depressed. I attached 18 AWG wiring between the boost converter and a micro USB and plugged it into the Pi. I then connected the battery to the input of the converter. Absolutely no change in behavior. Boot looping and no display on the converter. I turned off the backlights on the LCD attached to the Pi and during boot, as soon as there was disk activity, the converter display would go blank and the Pi would boot cycle again. When there was no disk activity during boot, the converter display would pop back on and show 5.5V on the output (exactly what I had set it to).

So I guess I have choice. Either accept the fact that a Pi 3 A+ with an attached 7" LCD cannot be run from a battery or try a much larger battery (9-12V) with gobs of amperage and use a buck converter to get to 5.5V and pray the amperage is sufficient. I had thought that a LiPo battery rated at 3.7V and 3000 uAh would have given the Pi at least 30 minutes of 5.5V at 3A, but I guess not.

So should I try a larger battery with a buck converter or would that just be pouring more time and money into a Pi black hole?

5.5v spells danger to your RPi....
...just set it to between 5.08v and 5.10v (like I did with my RPi4B NAS's DIY power distribution board) and you're within safe operating parameters.

LTolledo wrote: ↑

Sun Sep 29, 2019 5:09 am

5.5v spells danger to your RPi....

I used this setting based on Adafruit setting all their power adapters to 5.25V. "This adapter was specifically designed to provide 5.25V not 5V, but we still call it a 5V USB adapter. We did this on purpose to solve a problem that occurs often with USB-powered gadgets: they draw so much current than the resistance of the cable causes a voltage drop, so instead of 5V, the device sees 4.75V or so."

I also watched a YouTube video where a guy was testing loads on a Pi using a professional adjustable power supply, and he had to go up to 5.5V before the lightning bolt would stay off throughout the boot process. I was watching my Pi boot up while I cranked up the voltage, but when it was still in boot loop at 5.5V I figured that it was time to give up.

You should not go above 5.25V, which is the upper limit of the USB spec. The risk is that, with the Pi idling and drawing less current, the voltage at the Pi will rise and be passed on to USB peripherals (and HDMI monitors) that could be damaged.

If you are genuinely supplying 5V+ through 18SWG/AWG wires then something else is wrong. Does the PSU have a variable current limit? What is it set to?

To get 2 Amps at 5 volts the lithium cell will need to be giving about 3 Amps.

Is the cell well charged?

drgeoff wrote: ↑

Sun Sep 29, 2019 8:30 am

To get 2 Amps at 5 volts the lithium cell will need to be giving about 3 Amps.

Is the cell well charged?

I haven't used it that much since I got it, but I'm going to give it a good charge on a charger tonight.

According to the DigiKey calculator (https://www.digikey.com/en/resources/co ... ttery-life), I should get about 40 minutes (0.7 hours) with 5V and 3A with this battery (3000 uAh):

Battery Life = Battery Capacity in mAh / Load Current in mA * 0.70
*The factor of 0.7 makes allowances for external factors which can affect battery life.

That voltage regulator is probably like the multitude of other cheap Chinese regulators I've seen with the same specs.

One of them was honest enough to state in its datasheet that to get 9A output, you must use a high voltage close to its input limit.

I would not use such a device.


Might I recommend an MP2636, by soldering this to my Pi (bypassing the USB ports) I have managed to power a Pi3A+ (and other Pi before it) with an attached 5 inch screen (powered by the Pi) and run cpu burn A53 with no low voltage issues. YMMV
A single (genuine) 2000mAh 18650 Li-ion battery supplies the power for a compact solution.

pixelpop wrote: ↑

Sun Sep 29, 2019 8:55 am

drgeoff wrote: ↑

Sun Sep 29, 2019 8:30 am

To get 2 Amps at 5 volts the lithium cell will need to be giving about 3 Amps.

Is the cell well charged?

I haven't used it that much since I got it, but I'm going to give it a good charge on a charger tonight.

According to the DigiKey calculator (https://www.digikey.com/en/resources/co ... ttery-life), I should get about 40 minutes (0.7 hours) with 5V and 3A with this battery (3000 uAh):

Battery Life = Battery Capacity in mAh / Load Current in mA * 0.70
*The factor of 0.7 makes allowances for external factors which can affect battery life.

1. Lithium cells are not fully charged at the factory.

2. 3000 uAh is 3 mAh is 0.003 Ah.

Sorry, I mis-typed the battery info. It is, in fact, rated at 3000 mAh (not uAh). And thanks for the pic! I'll will study that further this evening.

I think I'm getting closer to the truth. I had been running with this setup:

5V/2.5A wall wart --> PowerBoost 1000C --> Chinese boost converter --> Pi.

I was using the PowerBoost because I am trying to get the Pi to boot from a battery. So aside from the 5V/2.5A wall wart going into the PowerBoost, I also had the battery connected to it. In this scenario, nothing was working. I would stick a 1K resister on the output of the boost converter and adjust the output to 5.2V (as indicated on the boost converter's display). Then I would remove the resistor and plug in the Pi and every display on the boost converter would go out and the Pi would not boot. I measured the output of the boost converter at 3.7V!

I removed the PowerBoost from the setup and it works! I'm not sure what the PowerBoost was doing, but it was killing my voltage. Just to verify, I removed the boost converter and replaced it with the PowerBoost. The Pi started doing again what it had been before, boot looping and noise on the LCD...both indicative of undervoltage. And that would jive with the 3.7V output I saw before. I couldn't measure it because the LCD had a hat on the GPIO.

So I'm going to get rid of the PowerBoost and try running on just the battery. Have to make a JST connector for that. If that works, the I just need to figure out how to charge the battery when the Pi is not powered.

The power boost 1000C has an output current limit of 1A, it isn't suitable for most Raspberry Pi models.

3.7v sounds like the nominal voltage of a lithium ion battery (cell), how did you wire it?

I just wired the battery to the input of the boost converter via JST, then the output of the boost converter to the Pi via micro USB.

I like the module you are using, although it's a bit more complicated to make adjustments. But I read some other posts about it (and probably most other boosters) and it looks like best results are achieved by bypassing all the connectors and hard-wiring things together.

I'm going to get one of those MP2636 boards and see what I can rig together. I'll report back. Fortunately, I found a local source so I should have it in just a few days. Thanks!

Most of those posts are likely based on my experience with the device

Anyway, I found an older, but slightly clearer photo of my setup that may be easier to see how I wired it compared to the previous photo.

My soldering in this version was not as clean as it is now.

The rough soldering job on the Pi itself is just how mine arrived out of the box.
The switch is required to turn the boost regulator off and stop it from depleting the battery.
https://i.imgur.com/ObogwjW.jpg

Nice job! Does the switch turn off power to the Pi whether it's running on external power or on the battery? Also, will the battery be charging while the switch is off? Here is my desired scenario:

External power applied (plugged in) and Pi turned on - Pi runs on external power (battery trickle charge optional)
External power removed (unplugged) and Pi turned on - Pi runs on battery
External power applied (plugged in) and Pi turned off - battery charges
External power removed (unplugged) and Pi turned off - minimal to no battery drain

Think of it like you cell phone or tablet. That's what I am shooting for anyway. Here's a pic: