Checking out a 5Ah Ryobi clone battery

I've been contacted by designers who were amused at my reverse engineering of their products.

Big Clive

Do you ever wonder if your reverse engineering / teardown videos make their way back to the manufactures of these products?

c

The pack is actually quite well designed. The balancing circuitry doesn't need to dissipate more than a couple of mW since each cell voltage only differs from the others by some mV max. That's why your calculation showing a 50V drop across the resistor is wrong, in practice only needs to compensate for manufacturing tolerances between the cells. The circuit that enables the µC to measure cell voltages uses a smart level shifter to get each voltage. The right hand side uses the n-channel FET in a voltage divider circuit to get a voltage relative to the voltage of the cell measured + the voltage of the cells below it. The p-channel is used to have a current path for the other cells, otherwise the current of one voltage divider would need to flow through the other and influence it's reading. By loading all the cells with a reasonable current, you negate this effect.

Paul Paulson

I wonder if the batteries have been condemned and deliberately discharged to zero for disposal.

Big Clive

A common failure with Ryobi batteries is when they have been left in storage while flat or left plugged into something like a work light. If one cell goes low it will often knock the whole pack out. Sometimes trickle charging the low cell back to around 3V will reactivate the pack. If you do that it might be worth treating the battery as risky until you have charged it until all cells are at equal voltage. (If they make it there). After that, leave the battery in a safe place for a day or two and then check to see if any cells have dropped significant voltage. If they have, they may be failing internally and should be discharged and recycled. The good cells can be salvaged.

Big Clive

@devttyUSB0 I'm not really expecting them to. I don't have any Ryobi tools, but they are new in packaging, and probably from a Ryobi employee. Date is 2020, so I'm mostly curious as to how TF it went so wrong.

Rasmus

Yes I'd like to see a comparison of the knock-off to original, too!

Jeremy Impson

Reminds me of the Forrest Mims voltage bar indicator circuit https://www.crowdsupply.com/img/8eb9/bargraph-voltage-blue_jpg_project-body.jpg

Jeremy Impson

Definitely second this. Would love to see a tear down comparison.

Muxer Baker

In my experience, such cells recovered from being undercharged, do not last very long even though it seems like the cells are back in business. Curious if yours do!

devttyUSB0

Interesting. I'm charging them without the battery management system.

Rasmus

I have fixed these batteries before by a quick high current charge on each cell (supply set to 4.2V). You usually have to get all the cells up to the same voltage before the charge IC will actually allow you to charge them (Aim for around 3.6V on each cell).

By pure coincidence two days ago I found 10 original Ryobi 4 Ah batteries at a local recycling station. One of the neighbours work for Ryobi Denmark, so I was intrigued. There is no reaction when pressing the charge state button. We took two apart and measured across. The cells had less than 1 mV across. We are trying to charge one, using low current on a bench power supply. Would you like me to send you one or more of them?

Rasmus

https://youtu.be/I_Gkwfprzio?t=223 looking at this one - it certainly looks like its going to be just as complex on the inside - that board looks to be full of fets and resistors

Russell Peake

Hi Clive have you taken apart a genuine Royobi pack to compare it ?

Private Private

if they'd added in opamps, or dedicated voltmeter hardware - they probably could have done it with less interdependancy, but i suspect relative voltage readings from cell 1 to cell 4 is probably still relatively viable - so long as the voltage dividers are reliable at 'relative voltages' - or, able to handle/detect a dead cell and therefore the voltages are not useful

Russell Peake

I think that is the intent. But the voltages are all relative to the other cells with respect to the zero volt rail, so it will have to read the cells from that end of the stack and then compensate for the others. I'm tempted to hotwire the bare battery block to a charger to monitor current and voltages.

Big Clive

could those values be designed that the voltage dividers all return a value between 0, and 3.3v even though the actual votages are closer to 3.7, 7.4, 11.1, 15.8 etc?

Russell Peake