Improving Throttle Response with Headline BLDC Controllers

I've been using a Headline (a.k.a. Cyclone) brushless motor controller
and Headline motor with a cheap, generic Hall thumb throttle (due to
limited handlebar space) and have been unhappy with the throttle
response that is close to that of an on/off switch. I don't blame the
throttle for this as this same throttle gives nice, gradual response when
used with a controller made by a different manufacturer.

So, I set about to seeing what I could do to improve the throttle
response to be more gradual, so that it was fully "on" when adjusted to
be so, fully "off" with no motor creep when at rest, and gave nice,
gradual response in between.

The Headline controller shows a curious schematic on its box:

http://tinyurl.com/lwl5eu

suggesting that the controller is designed to use a 5 kOhm
potentiometer throttle and not a Hall throttle. This is in spite of
the fact that Cyclone motor kits come with a Hall throttle that seems
to work well enough (again after some tweaking that as I recall
involves placing about 150 Ohms in series with the GND lead).

I performed two experiments and arrived at conclusions after much trial
and error:

1) I wired up a 5 kOhm linear potentiometer (pot.) and then proceeded
to add resistance to each of its various leads.

The 5 kOhm pot. by itself was as poor as that of the Hall throttle. No
response for the first 2/5 range, then full response by 3/5 range.
But, when I added 20 kOhms in series with the GND lead I was able to
use the first 2/5 of the range, making the pot. effective from nearly 0
to 3/5 range or perhaps a bit more. Unfortunately, the upper 2/5 of
the 5 kOhm pot. throttle range was still not usable, equivalent to
"full on".

2) After placing resistors of varying size on various Hall throttle
leads, I settled on the following:

I inserted 1.134 kOhms in series with the Adjustment lead and inserted
420 Ohms in series with the GND lead.

Adding resistance to the Adjustment lead extended the upper range of
the throttle, and adding resistance to the GND lead extended the lower
range. I was now getting graduated throttle response between roughly
1/3 and 4/5 of the throw. I tried capturing more at the lower end, but
any increase in resistance on the GND lead caused motor creep when
throttle was at rest. It appears that the first 1/3 of the throttle
throw is dead space.

In fact, I started with 452 Ohms on the GND lead and initially observed
no motor creep when the throttle was at rest. I went riding. After
climbing a long hill, the motor and controller were both warm, and the
motor was creeping when the throttle was at rest. I figured that as
the temperature of the controller warmed, the resistance of the circuit
increased just enough to cause it to pass the "creep threshold".

With the modified throttles, I do experience one minor side-effect:
When the controller is powered up or powered down, the motor will
"jerk" forward briefly, but not dangerously so.

Aside from providing this information to others using these controllers
who may be experiencing poor throttle response, would any of the
experts in this group care to speculate on why these controllers
respond poorly to an unmodified throttle, Hall or 5 kOhm?

Thanks.

Great info, but I worry about that power up & down jerk. Just the other day I bumped the rubber end of my half twist throttle. sticking the throttle at partial throttle. The next ride I got stupid and turned the key before getting on. To bad no video, because as I caught it by the grip, it accelerated to full throttle and there I am with handlebar firmly in one hand with a bike bouncing on a rear wheel that's spinning at 60-70mph. Luckily I was able to catch it on the 3rd bounce by the saddle with my free hand in an extremely uncomfortable position, and recruit my 5 year old to come turn the key off since my hands were quite full with the bike. That could have been much worse, and 0 throttle is absolutely critical for safety on power-up.

John

John in CR said:

Great info, but I worry about that power up & down jerk.
John

Click to expand...

The jerk occurs when the controller supply voltage is near its low-voltage cutoff zone. For the 24-volt controller, this is a supply voltage of about 18 volts, not normally seen but for an instant as the supply caps drain or charge.

This was a question I posed to BIll some time back as I too was having an interestingly odd throttle response from my Headline controller and throttle. My only difference was that I was using the twist throttle but was still getting the very noticeable long dead space at the beginning of the throw. The inline resistor does help the issue, but it still has some room for improvement.

So I pose to the multitude of Cyclone users out there: did you folks notice a weird throttle response and did you do any monkeying around to get it to respond more fluidly? Did you use a different controller/throttle other than the one supplied by Cyclone (which is the Headline products)? Why would there be such a dead space of no voltage being passed onto the controller, or that the controller is not designed to be sensitive to this initial input?

Tom

the cyclone throttle has been bugging me too. i'm tempted to try a magura 0-5k throttle but it's $50 and it may or may not work.

With a resistor pot, you need to add resistance in series with the ground wire and the 5v power wire to correct the travel. Let the signal wire connect directly to the controller. By adjusting the values of the two fixed resistors, you should be able to get the proper response over the entire range. For a normal controller, these values are around 1k to 1.2k on both ends of a 5k pot. If you use trimmer pots, you can dial it in, but leave some dead space at the low end to prevent creep.

oh sweet! i will pick up some pot's today and test it out. any idea what kind of wattage is going through those wires? i have some 1/4 watt resistors laying around but i'm wondering if i need higher rated parts.

no problem on current, you start with 5K ohm and you have 5V, so less than 1 mA to begin with and then add resistance on each end.

you could test each side with a multiturn bourne cermet type potentiometer, or even just a single turn trimpot, then adjust it until you find your sweet spot setting, then measure the resistance on your trim pot at that point and then find a replacement fixed resistor, even 1/16W watt would be ok, and just solder the fixed value in line to replace the trim pot and it is permanent. maybe even find a spot inside the throttle where you could put a small surface mount resistor inline.

is this idea also applicable to the hall effect sensor throttles too?

could you put a resistor in the ground lead to extend the throttle range on the low end?

if you used a 5V three terminal regulator and adjust the feedback on the regulator to lower the voltage to a lower spot for the high rail or red wire to the throttle, say 4.2V, would that also extend the range on the upside?

dnmun said:

is this idea also applicable to the hall effect sensor throttles too?

could you put a resistor in the ground lead to extend the throttle range on the low end?

if you used a 5V three terminal regulator and adjust the feedback on the regulator to lower the voltage to a lower spot for the high rail or red wire to the throttle, say 4.2V, would that also extend the range on the upside?

Click to expand...

Yes, you can do something similar to hall effect throttles. On the ground lead, adding resistance will reduce the travel needed to start output. I used around 120 ohms for this.

On the high side, I don't recommend adding resistance or lowering the supply voltage, since 5v is close to the minimum required to power the hall sensor. Instead, you can add resistance to the signal wire or put a voltage divider on the signal wire. This will adjust the high end. Keep in mind that the two adjustments will be interactive.
It would look like this. If you only want low end adjustment, then only use the 200 ohm trimmer and run the signal straight through.

Also keep in mind that the controllers and throttles are temperature sensitive, so always leave a little dead travel at the low end to avoid motor creep when it warms up.