Cyclone/Headline controller shunt resistor values

[mrbill]

I'm trying to figure out the values for these resistors in a Cyclone/Headline 24-48 volt controller.

There are 3 5-band resistors in parallel:

(2) black-green-silver-gold-black
(1) black-orange-yellow-silver-gold

My color chart says that 5-band color codes gives numbers for the first three bands, multiplier for the 4th band, and tolerance for the 5th band. It has trouble with "silver" as the third band or "black" as the 5th band. I don't have the instruments to measure them accurately. Can someone confirm these values?
View attachment 1

Here's the color chart I am trying to use:
View attachment resistor_color_code.gif

You may notice that the black-orange-yellow-silver-gold resistor is cut out (on the left). Paco of Cyclone did this to limit the current to 50A at 48 volts, so that the one-way clutch used in the "double freewheel" system is not overloaded.

The problem with this is that it limits the current to 35A at 24 volts (nominal), which is a bit low. I plan to use this controller at either 24 volts or 48 volts. So, to maintain the 50A limit at either 24 or 48 volts I'm thinking of wiring up a direct-connect cable for a CycleAnalyst, but I need to know the shunt resistor values accurately first.

Thanks.

 

[heathyoung]

2) 0.05 Ohms
1) 30K - note it is not in parallel with the others.

Two in parallel is 0.025Ohms. This is 25 miliohms, way off a direct plug in for a CA - you would need to have a 1:10 voltage divider, and set the CA to 2.5 miliohms.

 

[mrbill]

2) 0.05 Ohms
1) 30K - note it is not in parallel with the others.

Two in parallel is 0.025Ohms. This is 25 miliohms, way off a direct plug in for a CA - you would need to have a 1:10 voltage divider, and set the CA to 2.5 miliohms.

Thanks for your reply.

The last resistor is normally in parallel with the others, but it was cut out of the circuit, presumably to protect the one-way clutch that comes with the double-freewheel versions of the Cyclone kit. You can see the stub on the left-hand solder pad next to the other shunt resistors. I'm sure it's less than 30k Ohms, otherwise it would have essentially zero effect on the shunt value.

A straight reading of the 5-band color code (black-orange-yellow-silver-gold)* gives 0.34 Ohms (340 mOhm), but even that would have only a small effect in parallel with 25 mOhm, which is why I asked about it, whether I'm reading the code correctly. Resistors measured in milliohms are usually labeled with text, in my experience. I will try to measure these accurately sometime in the next few days.

*0 3 4 times 0.01, 5% tolerance

As for use with a CycleAnalysts, it looks like using a voltage divider (carefully-measured) is the only way to get this to work.

Thanks for your help.

--Bill

 

[heathyoung]

Oops - yeah that would be 0.34 Ohms - dyslexic. Orange black yellow is 30K.

You can make a low ohms multimeter adaptor with a 100mA current source - http://users.tpg.com.au/pschamb/lom.html - makes life a LOT easier.

 

[boostjuice]

The internal BLDC motor controller chip within 'Headline' manufactured (not 'Headway' - they make batteries ) motor controllers, limits peak current based off a 0.825V threshold accross the parallel group of shunt resistors. For your controller version, the parallel equivalent of the three shunt resistors (50milliohm//50milliohm//340milliohm) equates to 23.288milliohm. http://www.1728.com/resistrs.htm

So the peak current limit is calculated as follows;


EDIT: I have since discovered that this estimate of a 0.825V shunt threshold is wrong! I have now researched the BLDC controller chip datasheet to discover that it is either 0.5V or 1.23V depending on chip option used. Cyclone appear to use both options in different models of controllers. Either Hitachi ECN3031F or ECN3030F. The following calculations should be reworked with knowledge of chip option used

I = V/R (peak controller current = microcontroller peak shunt voltage threshold / total shunt resistance)
I = 0.825V/0.023288 Ohm = 35.42A

With the 340milliohm resistor removed (50milliohm//50milliohm) total shunt resistance is 25milliohm and so the controllers peak current is detuned to

I = V/R (peak controller current = microcontroller peak shunt voltage threshold / total shunt resistance)
I = 0.825V/0.025 Ohm = 33A

Harldy much a of a detune. If this is Paco's trivial effort to stop the tide of warranty claims against failed one way bearings within the double freewheel motors then it's pretty pathetic. If he was serious about it he would need to replace one of the 50milliohm resistors with another of higher value, but this would require board dissassembly to unsolder and resolder the shunts that are annoyingly inaccessible under the PCB. This is something that would be prohibitively time consuming compared to just snipping one end of a resistor already soldered in place so i suspect that is why he does this.
However, the fact is that the one way bearing used in the double freewheel motors is simply underrated for the torque requirements at over about 800W with these planetary motors. Another larger more robust unit can't be used because there isn't the space for it as it must have a 20mm bore to fit on to the planetary gearbox output shaft, and the OD must fit inside a threaded freewheel adaptor

The one way needle bearing used is an FC-20/HFL-2026 which is rated for 28.5Nm of holding torque

The Planetary equipped headline/cyclone motors @ 24V exhibit a loaded speed of around 2500RPM. After the 9.33:1 reduction of the planetary gearbox the output shaft torque is is ~9.33 times greater and the speed is ~9.33 times slower (2500/9.33 = 268RPM)
At 268RPM, 28.5Nm equates to a 799.848 Watt limit (http://www.magtrol.com/support/motorpower_calc.htm), and this disregards shock loads and higher torque at low RPM which electric motors all exhibit . No wonder these double freewheel one way bearings keep failing, especially with these planetary motors driven @ 36/48V.

 

[mrbill]

Thanks for the information.

The next size larger clutch (HFL2530) handles 66 Nm. But this would require machining a special race to fit over the keyed gearbox shaft. Then the OD would only work for 16t freewheel and larger, and even then the material for the freewheel adaptor might be too thin. Looks like I'm stuck with a 33A limit if I want to avoid failure. (I actually observed short peaks of 35A on the CycleAnalyst.) Although I should be used to it by now, I'm slightly annoyed that Paco advertised this current limit as "50A", when in reality it is 33-35A. (He also claims that the stock external controller is limited to 75A.)

How does this part fail? Suddenly and without warning, or does it grind or skip like a ratcheting freewheel nearing the end of its life? Since the assembly can be easily slipped off of the motor shaft for inspection, is there anything visible one can look for to warn of impending failure?

I tested this controller/motor at 24, 36, and 48 volts to check efficiency of the system. The curves are slightly lower than I was getting with my earlier tests due to my having changed the calibration of my CycleAnalyst slightly (the net effect of which was to show increased power consumed of about 1.5%) and due to my using a different method of collecting power data by averaging over time (~1 minute) and comparing the watt-hour reading on the CycleAnalyst to the kJoule reading on the PowerTap.

http://bit.ly/6NCBDG

Below are more photos of the double freewheel assembly. I'm using a 14t freewheel to the cranks instead of the usual 16t so that I don't need unusually large chainrings up-front to synchronize pedaling cadence to motor RPM.

http://bit.ly/6J5MdT

 

[mrbill]

Oops - yeah that would be 0.34 Ohms - dyslexic. Orange black yellow is 30K.

You can make a low ohms multimeter adaptor with a 100mA current source - http://users.tpg.com.au/pschamb/lom.html - makes life a LOT easier.

I cobbled together a measurement apparatus using an 8A battery charger feeding a dummy resistor load, measuring current with a CycleAnalyst and voltage across the shunt with a DVM.

The shunt as it is in the circuit gives 26.4 mOhm. I fixed my DVM leads to the far ends of the bus, not to one of the shunt resistor leads.

The cut resistor measured 37.5 mOhm, not 340 mOhm or anything close to it. I'm not sure why it's labeled as 340 mOhm. Maybe that fourth band is "platinum", a multiplier of 0.001...