18650 equaliser

Brian Drury

Brian Drury

Member
Joined
Apr 25, 2017
Messages
31
I built a cell equaliser for 18650's. This is not a passive balancer.

The battery has 8 series connected cells. 7 are fully charged and 1 is fully discharged.

I ran the equaliser for 7 hours.

How does it compare with what you guys are doing?


image_qjaemz.jpg
 
What is shown on the x axis? I thought it might be time, but it can't be seconds or minutes. Is this supposed to be mAh? If so, what is the meaning in this case and how is ist calculated?

Looks like your cells are still out of balance though, between 4.15V and 4.0V. Not trying to judge your effort or make it look bad, but since you've asked, I can balance the cells within millivolts of each other, if I want. Obviously that is overkill for most applications and I haven't built it myself, I just bought it.
What exactly have you built there, how does it work?
 
DarkRaven said:
The Y axis is Voltage and the X axis is time. The ticks are 10 Seconds.

The equaliser has not finished its work. I shall leave it until tomorrow to see where it gets to.

A photo of my equaliser is below. It is based on the flying capacitor idea but it uses 18650 flying cells instead.

image_qbuipl.jpg
Click to expand...
 
By looking at the graph, it looks to me that your board charges only the lower voltage cell first then starts working on the other cells once the low cell catches up. Is that how it works? If so, I'd put that in the "active charger" domain of things. It would definitely keep good cells from "overcharging" and the need to bleed off those cells so the others could catch up.
 
I need to make this clear:

This is not a charger.

There is no external power source.

The equaliser is connected to the 8 cell battery. Nothing more.

Energy is being transferred from cell to cell to equalize the distribution of existing charge. No energy is added to the pack.
 
Ohhhh, it's a smart active balancer, then. Balancer would be more informative than equalizer ;) But I gotcha now. Nice work, non-the-less

How much power is used by the board to transfer the energy?
 
Korishan said:
Ohhhh, it's a smart active balancer, then. Balancer would be more informative than equalizer ;) But I gotcha now. Nice work, non-the-less

How much power is used by the board to transfer the energy?
Click to expand...

It is definitely not smart. ;) There is no processor, just a tiny bit of clockwork logic.

I chose to call it an equaliser to avoid confusion with balancers which either use selective charging or selective discharge. This does neither of those things so I chose to call it something else.

The power requirement of the board is micro watts. Some energy is lost due to the energy conversion efficiency of the cells but this will depend upon the particular cells in use.
 
Yes, it is my design. 70mm X 85mm

Specified for 8 series connected 18650 cells as the 'static' cells and 7 X 18650's as the 'flying' cells.

It uses powerPAK MOSFET's with rds on 0.005 Ohm.

Cell switching is 0.5 Hz fully asynchronous.

Designed to operate 24/7

My approach has a big advantage. The flying cells are effectively part of the battery. The switching time is nanoseconds so they are connected 99.99999% of the time

If the battery consists of 8 cells each having a capacity of say 8Wh then the battery capacity will be 15 X 8 = 120Wh
 
And the Kickstarter is when? ;P

So you say this is similar to a flying capacitor design, but using a 18650 pack as flying instead? The mechanics are the flying pack is put into parallel with each series pack for "x"( 0.5Hz - 2 seconds ) amount of time and just repeatedly cycles through the series, grabbing charge from higher packs, and discharging into lower packs as it cycles through?
 
Final plot after about 15 hours.

The levels are closer together using a DVM. The arduino monitor is still being improved so I expect it to get a bit more accurate.

My problem is each voltage measurement = the last cell measurement plus this cell measurement so errors are cumulative.


image_injrfj.jpg
 
I *** REALLY *** like this method. Not so much in its "dumb" form, as it is now, but adding a layer of intelligence to the "flight plan", intelligence such as:

- No flights when pack is not in use or in balance (need to define "in balance").
- "Layover time" on packs calculated based upon voltage distribution and some TBD "flight planner" theorem (need to define the theorem).

I will be following your project VERY closely.

What are the pack voltages measured by multimeter? This where an ESP8266 (like discussed here) measuring voltage, current and temperature only per pack may be more advantageous and cost-effectiveas you can be isolated andget goodvoltage resolution. Have it send data back to a "mission control" for it to figure out the "flight plan" ... Hmmm, got me thinking ...

Nice work!
 
It certainly is interesting, but I still have a hard time figuring out what this can be used for in a real world application.
 
DarkRaven said:
It certainly is interesting, but I still have a hard time figuring out what this can be used for in a real world application.
Click to expand...

Not understanding, no application for an active BMS? I have to bemissing something here...
 
Well, sure, active BMS, but the current is super small, isn't it? After 15 hours the cells still aren't balanced. Imagine how long it would take on a big battery. And if your battery has balancing issues then the charge current will probably bring it out of balance faster than this unit can rebalance them. A bigger sized version of this device might be more appropriate.
 
The MOSFET's are rated at 130A pulse, 20A continuous on a small PCB. RDS on 0.005 Ohm

This project is not a final solution it is a proof of concept. You need to remember that the equaliser will be connected 24/7 and when all cells are equal it should keep things that way.

The equaliser appears to work very well but I am having some problems with the software for the arduino based charger and BMS. It's not a show stopper but will take a few days to sort out.

The bottom line is this will be very low cost.

How useful? Well, give me the specification of what you want it to do and I can tell you how well it performs.
 
I see. But forgive me my scepticism, that's just how my brain works :D

For a start, what do you think is the accuracy when balancing cells? In your test there is still a 0.1V difference between the cells after 15h, will the difference eventually get smaller or is accuracy to 0.1V what has to be expected?
 
No problem :)

I came here for a review because this is where the guys with real 18650 experience hang out. If my design has flaws or limitations I need to know.

The first bench lash-up used 2 fully charged cells (4.2V) and one fully discharged. The flying cells were in a random state. After 24 hours all 3 cells were 3.650V. My multimeter is limited to 1mV. Measurements were manual.

Now I have a PCB equaliser and the measurements are made using my the arduino charger/monitor I built. The voltage measuring accuracy is not as good as the meter but I am working on it. I see no reason for this to be worse than the lash-up so within 1mV should be possible.

All the cells I am using are unmatched dodgy reclaims. I figured this would be a good place to start. Later I shall purchase some decent new cells.
 
BrianDrury said:
Korishan said:
How much power is used by the board to transfer the energy?
Click to expand...

The power requirement of the board is micro watts. Some energy is lost due to the energy conversion efficiency of the cells but this will depend upon the particular cells in use.
Click to expand...

Do you have an actual figure (micro-watts) ? vs watt is transferred (pun intended).
 
You must log in or register to reply here.
Back
Top