Greenworks 80-Volt Battery

The battery is arguably the most important part of any rechargeable product.  It certainly is the most likely thing to fail.

Greenworks rates their battery's watt-hour capacity at 72 volts.  Thus, a 2.0 Ah battery is rated 144 Wh.  This is a reasonable compromise because 72 volts represents being about halfway through its useful capacity.  However, one battery I have is rated 160 Wh (2.0 Ah at 80 volts).  This makes it look like a better battery, but it isn't.

The cells listed below were discovered by disassembling a number of Greenworks 80-volt batteries.  A Torx T10H security bit is needed to open the case.  The standard form-factor batteries all contain 20 cells configured 20S1P.  

I have listed my estimated theoretical peak kilowatt rating (volts x amps) for each battery configuration assuming full voltage.  Is it possible a battery can briefly supply more than the cell manufacturer's rated continuous discharge current?  Yes.  But it's even more likely that the voltage will drop precipitously when that much current is drawn.  And it's a sure bet the battery won't stay fully-charged very long.  My point is to illustrate that regardless of what a motor's rating may be, if the battery can't keep up, the system will fall short of its specification. 

 The various cells found are listed below:

• Samsung INR18650-20R cells.  These have a green wrapper and are rated 2000 mAh (2.0 Ah).  Rated continuous discharge current is 20A.  Estimated theoretical 1.6 kW peak.

• LG Chem ICR18650HE4 cells.  These have a yellow wrapper and are rated 2500 mAh (2.5 Ah).  Rated continuous discharge current is 20A.  Estimated theoretical 1.6 kW peak.

• Tenpower INR21700-40TG cells.  These have a red wrapper and are rated 4000 mAh (4.0 Ah).  Rated continuous discharge current is 35A.  Estimated theoretical 2.8 kW peak.  Note that this is a 21700 cell.

• Unknown TIR6RXA & TIR6RX8 cells.  These 18560 cells have a red wrapper that is of poor quality.  I could find no information on these cells.  They are possibly from a very early or counterfeit battery.  The BMS power board was also different (physically smaller) than the other batteries.

Greenworks batteries in the long form factor contain twice the number of cells.  They are configured 20S2P.

• The use of 40 Samsung INR18650-20R cells provides 4.0 Ah.  Continuous discharge current doubles to 40 A.  Estimated theoretical 3.2 kW peak.

• The use of 40 LG Chem ICR18650HE4 cells provides 5.0 Ah.  Continuous discharge current doubles to 40 A.  Estimated theoretical 3.2 kW peak.

The battery connections are marked:

• P-  Pack Negative (this is the return for the charger and communications as well)

• COM  Communications (a serial bitstream) 

• CRG+  Charger Positive

• P+  Pack Positive

Samsung INR18650-20R 1C Discharge Curve

The graph below shows the discharge curve for a typical cell from an 80-volt Greenworks pack.  The cell is being discharged at its 1C rate (2.0 A continuously).  Note that this is very light duty compared to what it would be subjected to in a chainsaw.  The blue line is cell voltage.  The left side Y-axis graduations are cell voltage.  Multiply that number by 20 to get the total pack voltage.  Observe the rate of decline in voltage over time.  Several takeaways are apparent:

• Initially, the unloaded cell produces about 4.17 volts (83.4 volt pack).

• At the instant current is draw, the cell drops to 4.07 volts (81.4 volt pack).

• At about the halfway point, the cell produces 3.59 volts (71.8 volt pack).

• Just before the knee of the curve the cell is at 3.38 volts (67.5 volt pack).

• Remaining energy is lost quickly, and the test is discontinued at 3.0 volts (60 volt pack).

Battery Management System

The battery management system (BMS) comprises two circuit boards.  One is located in front for cell monitoring, and one is at the rear for power switching and fusing.  The older design shown below provides a wealth of information. 

The front PCB is marked “Globetools” and a search brings up a website in China for Greenworks:  https://greenworkstools.com.cn/en/  I had assumed Greenworks was a US-based company manufacturing in China, but it's not. It appears the company was founded in China in 2007, and they claim to hold more than 66 patents for invention.

The BMS uses four integrated circuits in a 24-pin SSOP package.  The common marking on this IC over multiple batteries has been “8905AM05SN”.  I assume this is an analog front end (AFE) but cannot find a datasheet.  This leads me wonder to if it's a custom IC for Greenworks?

It's clear from a cursory observation that one AFE IC serves a group of 5 cells.  In this way 40, 60, and 80 volt batteries could be built using 2, 3, or 4 of the ICs. 

Transistors (Q1-Q20) for cell balancing are evident, as are 100 ohm load resistors.  This implies a maximum balancing current of ~40 mA (0.16 watts) per cell.

The microcontroller is packaged in a 20-pin plastic LSSOP marked “10268A”.   This indicates it's from the Renesas RL78/G12 family which features an ultra-low power 16-bit core.

The 2.0 Ah batteries use two IR P379Y power MOSFETs on the rear board.  The 2.5 Ah batteries have threes such MOSFETs.

The newer BMS design is quite opaque.  It is shown below for reference.  Manufacturing costs have been reduced through the use of flexible printed circuit boards for cell monitoring where the older design used individual wires.

Serial Communications

The two photos below are oscilloscope captures of the serial data steam observed between P- and COM from a CSB410 chainsaw.

A burst of data is sent every 200 milliseconds while the saw is ready to work.  Each bit has a high time of about 100 uS.  The line swings between 3 volts (the idle state) and 0 volts.  My Siglent oscilloscope has the capability to decode serial data, but I could not determine the content of this message.

The battery does not autonomously emit this data stream.  The motor controller must be connected via the COM wire which therefore probably initiates a query.

Battery Chargers

I am aware of two chargers for the Greenworks 80-volt battery.  Both are specified for a 120 VAC 50/60 Hz input.  Battery chargers of this type are essentially switched-mode power supplies.  This yields a high AC to DC conversion efficiency.

• GWX0800250 is rated 4 amps on the DC side.  The battery sits horizontally in this charger.

• GCH8020 is rated 2 amps on the DC side.  The battery sits vertically in this charger.

Serial communication between the charger and the battery occur via the COM connection.  Without this connection, charging will not commence.

Capacity Testing a 2.0 Ah Battery

My active electronic loads can't handle an 80-volt battery, so I used a simple resistive load and a Chinese PZEM-051 power meter (rated 100 V and 100 A) for this capacity test.  

The load was made from two Dale 20 ohm, 100 watt power resistors connected in series.  This yields 40 ohms with a 200 W rating.  With an 80-volt battery, the current draw is 2.0 A and the worst-case power dissipation is 160 watts.  As the battery discharges, the power dissipation decreases.  The test is concluded at 60 volts (3.0 volts per cell).  At that point the power dissipation is only 90 watts.  This is a less demanding test of the battery than one conducted at a constant current of 2.0 A (the 1C rate). 

I used a small fan to cool the resistors.

Hot off the charger, the battery measured 83.3 volts unloaded.  As soon as the 40-ohm load was connected, the voltage dropped to 81.6 V.

After 27 minutes, the voltage was 72.0 V.  At that point the current draw was 1.8 A.

At the 60-minute mark, the voltage was 62.9 V, drawing a current of 1.31 A.  So far, the battery had provided 131 Wh of energy.

A few minutes later the pack voltage slipped below 60, and 133 Wh had been provided.  Note the small change in energy from the 60-minute mark.  Towards the end, voltage drops rapidly and not much useful energy remains to be withdrawn.  This is comparable to the more rigorous test of a single cell above (which yielded 1916 mAh x 20 cells = 138 Wh).

The battery is rated for 144 Wh.  I am satisfied with this outcome.  The PZEM-051 is not a laboratory-quality instrument and I have never attempted to validate its accuracy.  Furthermore, I've rarely seen a battery provide its rated capacity.  It's possible the rating is made down to a cell voltage of only 2.5 volts, but I feel that's hard on the battery.  There's very little energy remaining when the pack voltage drops below 3.0 volts per cell anyway.

Immediately after removing the load, the voltage recovered to 61.6 volts.

Recharging with the 2-amp charger took less than 90 minutes and consumed 168 Wh of AC power.  I have not measured the input/output efficiency of a Greenworks charger, but 90% would be a reasonable guess.  This means that battery accepted approximately 151 Wh of energy. 

Heavy Load Testing

The following test shows the importance of the battery in high-load applications.  A perfectly healthy 2.0 Ah battery was tested at roughly halfway through its useful state of charge (starting at 73.9 volts, unloaded).  

The load was a series string of six nominally 2-ohm Ohmite L175J power resistors.  These resistors are rated for 175 watts each and can sustain an overload of 10 times their rated wattage for 5 seconds.  Each wire-wound resistor exhibits an inductance of 36 uH.  A Peacefair PZEM-051 power meter was used for instrumentation.  The battery holder was repurposed from a broken charger, and the power wires are fine-strand #8 AWG with Amass AS150 connectors.

This is a challenging test to perform with the simple equipment shown.  Each test was conducted only as long as was needed to record the data (well under 10 seconds).  The negative terminal clamp was moved by hand to the required resistor tap.  This created an arc during the connection and disconnection process.  It was literally a bit like welding.  One resistor fractured during the test.  I had considered testing all the way down to 2 ohms, but decided against it as the trend illustrates my point, and it would be hard on the battery.

 The table below shows the results.  A 2.0 Ah battery would provide very short run-time in a saw rated 3.4 kW.

Maximizing Battery Life

I've written about maintaining batteries for electric vehicles here: https://www.electricmotiontech.com/home/ev-tech-101/battery-care-and-feeding  However, only partially charging a power-tool battery is unlikely to be acceptable from a run-time perspective.  But there are other things the user can do to maximize battery longevity — especially when it's used in a seasonal application.

If you are going to fully charge the battery, it's best to do so shortly before using it.  See next section.

Self-Discharge

The self-discharge rate of lithium-ion cells is quite low, however the BMS consumes some energy even when the battery is not being used.  Greenworks batteries are very good in this regard and lose little of their energy to a “sleeping” BMS.

Nevertheless, it's probably wise to periodically test battery voltage and verify it has not dropped too much during storage.  If I observe the voltage drop below a 30% state-of-charge (see the spreadsheet above) I'd give the battery a partial charge to get it back up to around 50%. 

Greenworks Battery Warranty

Greenworks has a 4-year battery warranty.  See: https://www.greenworkstools.com/pages/battery-form  I've never attempted to warranty a battery, but it would seem the only criterion is that the charger must think the battery is bad. 

There's nothing said about a loss of capacity.  Generally, a battery is considered to have reached the end of its useful life when capacity falls below 80 percent of new.  This would be almost impossible for the average consumer to determine.  However…