I have a Pi 4 that I want to power by 18650 battery @ nominal 4.2v, including charging, but it needs to be above 1Amp output.
My existing boost boards work perfectly from a single 18650 for the Pi Zero 2 W, but obviously that doesnt require as much herbs as the Pi4.
So, I’m looking for a DC-DC boost board that can take <4v and boost it to 5v, but at >1A.
I know that that is a big ask, and this is why I am having issues finding the magical unicorn.
I can run the Pi4 direct from the 18650 fine, but that doesnt work ideally for clean shutdown or regulated power, I would rather a stable 5v output.
If you convert everything to Watts and allow for conversion efficiency you will find that the 4.2V battery will have to supply significantly more than 1Amp… Connecting several 18650 cells (the correct term) in parallel should do but you will still need a converter that has the output current capacity.
I actually have two cells in series, and they are 2500mA, so I should have 5A at my disposal.
Doesn’t work like that unless your terminology is wrong.
2 x 4.2V 2500mAhr cells in series will result in 8.4V @ 2500mAhr.
2 x 4.2V 2500mAhr cells in parallel will result in 4.2V @ 5000mAhr.
Actually you would not have 5A available either. Capacity (mAr) is usually quoted with a load of C/20 or in the case of 2500mAhr 125mA. This capacity reduces as load increases and in the case of lead acid the reduction is quite dramatic at close to capacity loads. I have not studied the curves for Lithium cells but I must do so as these cells are very wisely used now.
I am just pointing out the possibility of a serious reduction in capacity under higher loads and urge you to find out exactly what that reduction is and keep it in mind.
Sorry, I meant parallel, and realise that I was talking about capacity, not discharge current.
Regardless, the 18650s are not the limiting factor, it is the boost board, hence why Im looking for one that can take 3v-5v input and output 5v at a current of multiple Amps.
I am pretty sure Core have something to do your job. I might be one of the adjustable ones but that should not be a problem.
Potentially, but I couldnt find anything over 1A
The Pololu U3V70F5 might be something like what you are after.
We’ve got a handful of power supplies that run on 18650 cells that might fit the bill for what you are after.
The first consideration would be to check the rated current output of your cells as different manufacturers will promise very different continuous current output within the 18650 form factor.
The Pi 4Bs power requirements will also vary considerably based on the number of USB peripherals it is assumed to be supplying, the Raspberry Pi documentation recommends a very conservative 3A supply.
DFRobot make a number of 18650 power supplies.
I’ve used their adjustable 6 way power supply for my own Pi and touchscreen setup, I don’t trust their claims that all 6 channels can output their rated load simultaneously as it’s all being drawn from the same 18650 cells that certainly can’t output that much power. I dismantled my unit and inspected the board and each channel does have a separate regulator tied back to the batteries.
The 4 battery 18650 power supply is also described in a way that suggests it has 11 separate outputs across 2 voltage levels but looking at the back of the board suggests there is only one regulator for each voltage level and mutliple points to connect to that output.
The 2 battery 18650 power supply is a similar story.
We’ll work on rewording the descriptions on those pages to make it clear that connection points and regulator output channels are a different matter, sometimes these things get mixed up in translation from the supplier.
If you already have a battery holder and want to keep things simple I think Bob is on the money with getting a separate regulator like the Pololu U3V70F5 with good documented specs that you know will be able to supply the current you need.
Thanks Trent, that all sounds good.
I was pretty non-descript in my usage, so let me elaborate a little.
This is for my camera dolly controller that is attached to the sliding cart that holds the camera.
We run these of 12v for multi-day time-lapses and that is from a big ol heavy car battery, so there’s no issues there.
There are times where I need the Pi4 to stay running when I unplug the 12v, such as packing up and I dont want to turn off the pi as I am going to plug it back in and turn on in a short period of time, or for testing purposes so that we dont have to keep shutting down and re-powering for each different setup.
The primary reason that I need this currently is that I am shooting overnight at the Australian Compact Telescope Array in Narrabri next month and you cant have any radio devices active due to the radio telescopes.
Of course the Pi has WiFi and Bluetooth and I want to turn that off prior to getting there as I dont want to risk having the radios on. This means that I need to keep the pi powered for about 15 minutes, shich should be all good from the two 18650s as long as I can find the boost controller that will run the Pi4.
I cant use any of the pre-made Pi4 UPS hats as I have my own custom hat in place.
Yep, the Pololu U370F5 seems to look like it will do the job, thanks Robert.
OK, another question for you knowledgable peoples
I found a 4A boost controller whose specs suggested that when supplying it 3.2v it could generate 5v and supply 4A. I have no idea how the magic pixies produce the current, but that’s what the specs say.
I cant power the Pi4 cleanly
So my thoughts are that the issue I have must be the LiPo charge board.
It has terminals for the battery connection and output terminals for the load, that is protecting the battery from under voltage.
It’s a generic chinese board that I got years ago, and I dont have any specs for it, unfortunately.
My question to the group is do these style of LiPo charge controllers limit the current from the battery?
If you can’t power the Pi cleanly from that device then there are two possibilities. One is that the boost device cannot supply the rated current. It’s not unusual for the manufacturer to quote the rated current for the best possible circumstances, which includes correctly rated components and adequate cooling. If the device is overstressed then the voltage would fall off as the current draw rises. The Pi is quite sensitive to voltage, and a small drop might be enough to make it unreliable, or stop working altogether.
The other possibility is that the battery cannot provide the required current. All batteries have an internal resistance which limits the amount of current that they can deliver. This is only loosely related to the battery capacity - you can have a high capacity battery with a limited current capability, and vice versa. You need to look at the specifications for your battery to see if this might be the problem. As the current draw from the battery approaches the battery’s limit the voltage will drop off. The symptom of a battery with limited current capability depends on the booster - for instance, the booster might just shut down when the battery voltage drops below a predetermined level, or it might not be able to maintain the rated voltage. But there is no technical reason that a 3.7V battery couldn’t provide sufficient current for a booster to provide 5V at 4 Amps - it depends on the type of battery.
Further testing is leaving me more in despair.
I have configured this setup to be powered by the mains through a plug pack, to the LiPo charger, to the boost board, to the Pi.
The screen that I am using I have wired specifically to use a modified USB cable so that I am powering it from external 5v power.
The Pi boots and screen displays, then dies.
So I checked the plug pack and it was only 1A, so thats underpowered, so I went and lugged to my lab the AGM battery that is going to power this whole setup. 12v AGM 35kg battery can definitely provide whatever amps this little setup wants, but again, the pi reboots just as the boot process is almost complete, when the screen is plugged in.
This screen is an aliexpress special, HDMI with USB for power and touchscreen. I need HDMI screen as Im using the whole GPIO of the Pi.
Is it possible that the screen is drawing power from the pI over the HDMI?
I started this project with the thought that powering the Pi would be simple.
The same setup works flawlessly with the Pi Zero 2W, but obviously that is a far lower power draw.
Firstly, the charger will not be powering the Pi. But if it is charging the battery and the battery can supply enough current, then that doesn’t matter, other than you have to be sure that the system spends more time charging than it does running the Pi or you will have a flat battery. For the same reason (and with the same restriction) a 1A plugpack is adequate for running the charger.
Secondly, it appears that your testing has shown that the boost device cannot deliver the current that the Pi needs when it is driving the display. You could confirm that by running the Pi from a plugpack that you know is properly rated for the Pi. I have seen boost devices advertised as 4A when all that means is that the chip they use has a rated maximum current capacity of 4A, but the way the device is manufactured, with smaller components and little or no heatsinking, it will never get to 4A without dropping the voltage well below the nominal 5V. I suspect that if you measure the voltage being supplied to the Pi you will see it drop when the display is plugged in.
FWIW I use an old PC power supply for my 5V when testing power-hungry devices.
thanks Jeff, makes sense.
I was looking at a known device and giving it a try, so something like this:
Seems to be able to deliver 2.5A @5v
Yes, but if you look closely at the V-I diagram for the 5V output at the bottom of P16 of the datasheet you will see how the voltage regulation drops (very slightly) as the current increases. It is possible that a sensitive Pi would detect a voltage dropout at something less than the specified current limit. The power supply might be the unglamorous workhorse of the setup, but it still needs to be very carefully considered.
There are no magic Pixies.
You need to do what I said a couple of days ago and work in power (Watts) to get your head around that. I will do it one more time.
5V @ 4A = 20W. Allow for conversion efficiency, suggest 85%. Input power = 23.53W
23.53W @ 3.2V = 7.35A.
That is the current you would need to supply from 3.2V to achieve an output of 5V @ 4A assuming a conversion efficiency of 85%
Yes, all well and good, but it doesnt make a difference in real world terms.
I can power a raspberry pi 4 with an HDMI touchscreen from a 5v / 2A output USB Powerbank.
That is my benchmark for the max current needed out of the boost board.
If I have a boost board that SAYS that it can output 5v @ 2A from a 3.2v input, and it isnt running the pi, then the manufacturer is not being truthful.
The first one that I tried was my mistake as it was only a 600mA output, so that was never going to work anyway, I simply forgot the output was not up to scratch for this project.
Anyway, I have a working solution at the moment, using a USB Powerbank that has passthroough power (Outputs while charging) and doesnt drop the output when disengaging the charge input.
I have two other options on their way to me that say that they will do what I need, so I’ll update this thread with any definite working solutions at that point (with photos).
I think you may find at the end of the day this IS “real world” terms.
Not a very accurate way to measure the desired current but will do for now.
What I have being trying to say here is in this world you get nothing free. The manufacturer is saying you can get 5V @ 2A out of your boost board with a 3.2V input. What they don’t say is the 3.2V source (battery or whatever) has to supply this POWER, 10W at max output plus conversion inefficiencies. This varies between manufacturers but in the example previously I assumed 85% efficiency. This means in this example (5V @ 2A) the 3.2V source has to supply 10W + 1.76W = 11.76W or in rounder figures 11.8W. Now at 3.3V (the source) this equates to 3.687 or 3.7A, significantly higher that the 2A output.
Now THERE IS NO GETTING AROUND THIS. If your 3.2V source cannot supply the required power the booster WILL NOT provide the rated output. The manufacturer may not explain this adequately but we can’t help that. If not satisfied you will have to take it up with them or your supplier.
I am sorry I can’t explain this any simpler. There is no magic way of getting the extra watts out of thin air, they have to come from somewhere.
In support of this while on the subject of power banks you will find that reputable brands actually state that the mAhr capacity is AT THE INTERNAL BATTERY which could be 3.2 or 3.6V. IT IS SUBSTANTIALLY less at the 5V output.
If you are having trouble understanding this maybe someone else can explain better.
I really dont follow why you think that I dont understand any of this. It doesnt matter what I am measuring the output in, Watts or Amps as I know the Volts, and it’s not difficult to multiply it out or know what is needed.
Just to spell it out, I was asking for alternative boost boards as obviously the manufacturers of the gear that I have are lying about the performance, I am powering the boost boards by two Samsung ICR18650-26C Cells with a max discharge output at 5200mA, so there should be no issues with them delivering enough Amps or Watts or whatever you want to measure it in. Specifically as there are two cells sharing the load, and they are fully charged at 4.1v.