Bafang/BPM geared-hub specs, teardown, and pics


Staff member
Page 2, Bafang web-catalog
Green Bike Kit
E Life Bike
Bafang 500W-BPM also available from ES member Keywin /

Crossbreak and Miles have calculated that the Bafangs "might" run well up to 2,000-RPMs at the motor (5:1 reduction to the output shell, so 2,000 motor-RPMs = 400 wheel-RPMs), and any RPMs above that: efficiencies and waste-heat both slowly curve to the bad side...

Here is a breakdown of the BPM-II, showing the minor differences.

A short video of "fullthrottle" opening up his BPM to clean, re-grease, and replace the freewheel...

Firstly, the two BPM models [from BMS-Battery] and the new CST cassette-hub model are all rated at 36V. The coding applies to a 36V input and all the values I will be giving you are in unloaded conditions. Let's begin with a code-16 motor and work our way to the code-8 motor.

_8_______378. mind you, all of these specifications are directly from the engineers of Bafang.

Jeremy Harris' and Wurlys' acquired data on Bafang code#/turn-counts

This drawing is reported to be a BPM

Specs per ES member full-throttle:

BPM 500W Code-9
Phase impedance 70mOhm
Phase inductance 275uH
9 turns x 10 strands of d0.56mm (23-AWG)
34 lams 0.50mm
18 teeth: 17mm wide x 10mm deep x 14mm high
Stator diameter 130mm
Air gap 0.50mm
Magnets 8 pairs 18mm wide/deep x 26mm long (along the rim) 2.5mm thick
Flux ring 6mm thick

The motor pictured below was purchased from Green Bike Kit in April of 2013, and may vary slightly from older Bafang-BPM versions. Here is a link with pics of a BPM from 2012 which show some of the internal differences:

Bafang 002.JPG

Bafang 007.JPG

View attachment 10

Wooden stand central hole is 80mm diameter, the cord slot is 18mm wide
The distance between square sides when the stand is flipped over is 5-1/4 inch / 133mm

Bafang4 014.JPG

Bafang 005.JPG

Bafang 006.JPG

Bafang 010.JPG

Bafang 009.JPG

Bafang 011.JPG

Bafang 012.JPG

Bafang 013.JPG

Bafang 015.JPG

Bafang 014.JPG


Staff member
Shaft bearing on the flat sideplate is 15mm ID / 32mm OD / 8mm thick? part number is 6002RS
Shaft bearing on planet-gear side / disc brake side is 20mm / 37mm / 9mm, part number is 6904RS

Large central bearing 25mm ID / 37mm OD / 7mm thick, part number is 6805RS, there are TWO of them side by side.
The three planet-gears each have a bearing, a 6001RS, 12mm X 28mm X 8mm

Steel sun gear, 16mm wide, 21T
Plastic planet gears (three of them), 42T / 14mm thick,
Steel ring gear, 105T, 16mm wide

Stator tooth width 16.80mm~17.20mm
Stator diameter (air gap) 5-1/8-inch / 130mm

Axle-threads are 12mm X 1.25


Staff member
Bafang3 001.JPG

If you look closely, you can see three extra holes on the sun-gear flange that are threaded. I see no purpose for these on the Bafang-BPM, so I assume this sun gear is used for several different models. Once you remove the sun gear and the circlip that is then exposed, you can re-attach the sun-gear, and then make a puller out of three M4 bolts that are longer (to separate the magnet-ring from the stator if you don't have a 3-arm pulley-puller). It's possible these threaded holes are for jacking bolts to help separate the sun gear if it is difficult to remove.

Bafang3 002.JPG

Bafang3 003.JPG

Bafang3 004.JPG

If I were doing this often, I would make an aluminum sleeve that was 17mm ID, 25mm OD, and 42mm+ long. If you slide such a sleeve over the 3-arm puller side of the axle and tighten a nut onto it, this sleeve will slide along the central bearing ID and force the stator and axle to slide out (and slide in on re-assembly) following a straight line, rather than pulling back and forth to the sides.

Bafang3 005.JPG

View attachment 5

Bafang3 007.JPG

View attachment 3

Bafang3 009.JPG

Bafang3 010.JPG

Until someone asked me to check, I assumed the entire magnet bowl was steel. Upon closer inspection, I tested the different parts, and the outer ring is steel (of course), while the bowl face with the six wide spokes is aluminum. This pic is framed to show the color difference at the steel/aluminum joint. The round dimples on the steel ring perimeter have no apparent purpose. Since there are four of them that are evenly spaced out, I suspect they held this part onto a lathe jig.



Staff member
Oil cooling the Bafang-BPM:

This is an experiment, based on the experience of others. I hope to report back after a while with any additions/deletions/or changes necessary to improve this operation. Here is the oil-cooling experiments thread:

I am drilling and tapping threads all the way through on two opposing bolt-holes from the disc brake mount, to allow for future oil-cooling by filling 1/3rd of the motors internal volume with Automatic-Transmission-Fluid (ATF). When the motor is not spinning, the oil level should be just below the axle bearing, since the axle bearing would leak no matter how well the rest of the motor-case is sealed.

The sealant I used here is a small tube of "Permatex High-Temp red RTV Silicone Gasket Maker", which should be cheaply and readily available at most automotive parts supply stores. It is formulated to be resistant to all common automotive fluids, including motor-oil, engine coolant, and automatic transmission fluid (ATF).

The threads are M5/5mm X 0.80 (removed with a 4mm hex-wrench), so I used a tape-wrapped 1/8-inch bit (3.2mm) to drill through-holes from the threaded side. I flipped it over, and the aluminum was so thin at that spot, I hand-turned the drill bit with a screwdriver handle (my small bits have a hex-shaft). I used an 11/64-inch bit (0.172, 4.3mm), which is actually a perfect bit for this. Then, I started tapping the new threads from the threaded side.

First, cover the bearing hole with tape, and remove the two bolts in the holes you wish to drill and tap threads. Holes were originally 8.5mm deep, now open all the way through, and now 11mm deep before screw end will protrude. When using 3mm thick brake disc with no washers or lockwashers, you can now use screws with 14mm of thread.

Bafang6 033.JPG

Bafang6 037.JPG

View attachment 6

Bafang6 040.JPG

View attachment 4

BPMoilCool1 002.JPG

BPMoilCool1 003.JPG

BPMoilCool1 009.JPG

BPMoilCool1 012.JPG


Staff member
This post is about how to double the phase wires copper to allow more amps to be used on this motor.

On the stator with phase wire terminations facing up, going clockwise with the halls at the bottom, the phases wires from left to right are Yellow, Blue (center), Green. The three screws holding the shaft onto the stator core are common M5 thread. Each phase wire is 19 strands.

Hollow shaft ID: 0.300" / 7.6mm

BPM axle flats width: 10mm

Two 0.060" copper-diameter 16Ga wires that have been stripped of insulation and then twisted together equal a roughly 0.090" diameter of copper, which is approximately equal to Hobby King 12Ga.

The stock wire insulation is thin, but is actually quite good. I tried stripping a few inches of the stock insulation on the paired wires, and re-insulating the two of them with a single sleeve of heat-shrink, but the result was slightly bulkier than just the paired wires with the stock insulation still intact. I then tried taking the six phase wires (with the stock insulation intact) and covering that 6-stock-wire bundle with a sleeve of 6mm heat-shrink. This worked well, and even passed through the axle fairly easily with a small amount of lube on the outside of the 6mm heat-shrink sleeve.

The six-wire bundle inserted so easily, that I wondered if I could do that again with three of the original hall wires added to the bundle to be used to connect to a temp-probe. It worked! There was enough air-space between the fatter wires that three of them easily worked. The original sheath over the wire bundle is quite thick, and the 16-Ga wires are easily adequate for 500W, which allows a diameter of bundle that still easily and quickly inserts through the hollow axle on an assembly line with no lube. By using lube, removing the thick-walled sheath, and removing two of the hall wires...the phase wires can easily be doubled.

In this phase-wire upgrade, the stock cable length was cut in half to result in two equal-length sections. This also results in ten shorter hall wires, instead of the stock 5. It might be possible to double the phase wires, retain 5 hall wires for sensored operation, and also add three of the extra hall wires for the temp probe operation, since the hall wires are very thin.

Bafang6 001.JPG

Bafang6 002.JPG

View attachment 12

Bafang6 004.JPG

View attachment 10

Bafang6 006.JPG

Bafang6 009.JPG

Bafang6 007.JPG

Bafang6 010.JPG

Bafang6 013.JPG

Bafang6 014.JPG

Bafang6 017.JPG

Bafang6 020.JPG

View attachment 1

Bafang6 026.JPG