CA Current Throttle Usefulness/Experience? + throttle damper

[hillzofvalp]

I gave in and bought a CA from methods... but I'm wondering if I ought to try the "current throttle" feature in which the throttle becomes a current control.

Will this just give more consistency to the feel of the throttle? the controller does not all of a sudden give the motor more current at the wrong times, does it? I would hate to be giving the motor heat it didn't need on by accident while traveling between 20-30mph... the current would still taper off at higher speed, correct?

thanks.

[hillzofvalp]

anyone?

[jonescg]

I've never used the throttle input options on my CA's, so I unfortunately can't be of much use.

[lester12483]

The cycle analyst is a great piece of equipment to have for your ebike. It's worth the price.

[MadRhino]

http://endless-sphere.com/forums/viewtopic.php?f=2&t=31518&p=470572&hilit=torque+throttle#p470572

[hillzofvalp]

Idk how people get results with this search function unless maybe they've read the thread previously.. Thanks

[methods]

Idk how people get results with this search function unless maybe they've read the thread previously.. Thanks

Old thread but relevant topic. Use Google. Google crawls and updates this forum constantly. The search engine above sucks.

I found this with Google - not even a site specific search - just a standard key word search.

-methods

[GrayKard]

Heres an excerpt from this thread: PaulD's Old Race Bike

Thanks guys! I really appreciate the comments!!!

Here's the latest chapter of the story, brings us about up to current.

Maybe you're wondering- will it all work? I certainly was wondering that!

The next thing to do was to put it back together and test it. The only remaining major complaint about the bike's performance was the touchy throttle- it was difficult to smoothly lay into the power when accelerating. I had had ideas about possible solutions- potentially trying the CA current-controller throttle. To that effect I got out my spare CA- large screen stand alone version and wired up connectors for the throttle as per the CA manual.

I spent a little time testing it and trying to tune it but was unable to find a good combo of gains to make the throttle response better. It tended really strongly to want to oscillate, and with the gains low enough to prevent that it became extremely sluggish. It may work with more tuning.

But- that may not be necesessary- I changed the half-twist Hall throttle for a Magura pot-based throttle with 10 turn trimpots on the high and low wires to tune the high and low output limits. This allowed me to set the limits of the throttle travel exactly at the points where the controller begins to deliver power and is maxed out. This resulted in a nice linear, sensitive easy to feather throttle. It feels fantastic and I think should work really well.

Hope that helps, Gary

[methods]

Thanks for the link.

I think I can solve this issue. I am building up a CA thrust current setup right now. First thing I am going to do is bring in the voltage thresholds to maximize the voltage resolution from the hall throttle. Then I am going to neck down the current limit until it is "smooth". Then I am going to geek the feedback if needed. I am pretty sure we can get this to work very well.

I also suspect there is some interaction between the controller current limit and the CA set limit.... My hypothesis is that if the CA set limit is significantly lower than the Controller shunt limit the bike will be more likely to go into oscillation.

Maximize throttle rotational resolution
Minimize current limit that rotation maps to
Align controller limit to ca limit

Lets go test and see. I will report back

-methods

[motomech]

Idk how people get results with this search function unless maybe they've read the thread previously.. Thanks

I use the "search within a search" function.

[Skippic]

Thanks for the link.

I think I can solve this issue. I am building up a CA thrust current setup right now. First thing I am going to do is bring in the voltage thresholds to maximize the voltage resolution from the hall throttle. Then I am going to neck down the current limit until it is "smooth". Then I am going to geek the feedback if needed. I am pretty sure we can get this to work very well.

I also suspect there is some interaction between the controller current limit and the CA set limit.... My hypothesis is that if the CA set limit is significantly lower than the Controller shunt limit the bike will be more likely to go into oscillation.

Maximize throttle rotational resolution
Minimize current limit that rotation maps to
Align controller limit to ca limit

Lets go test and see. I will report back

-methods

Working on the same. I'm using the Arduino for programming and a separate current sensor.

http://www.hydroponicsonline.com/store/ ... 22785.html

I think the most important part is to have quick feedback so the automatic throttle can use finer steps to adjust without creating a noticeable delay.

I'm already thinking it might be easier to disassemble the controller and eliminate the speed feedback. At least my controller tries to achieve speed based on the throttle instead of thrust. What we actually need is just to control the PWM, regardless of the speed of the bike. The main issue is with powerful ebikes. At 200W the cruise control setup might be more desirable.

[methods]

I am making some good progress here.

First off, I can say that the system is not at all "tuned" out of the box. Everything is "loose" in order to accommodate a wide range of inputs.

I am going to start a thread - but here is some info.

First measure your exact throttle output. Mine is 0.817V to 3.6V.
Then measure exactly where your controller starts to output - mine is 1.50V
Then measure exactly where your controller switches from PWM to Constant output (you can hear it). Mine is 3.3V

So what we want to do is map the voltage range of 0.817V - 3.6V over to the controller input range of 1.5V to 3.3V

map(throttleMin, throttleMax, controllerMin, controllerMax)

THIS WILL TOTALLY ELIMINATE ALL OF THE SLOP IN YOUR HALL THROTTLE.
I HAVE ZERO SLOP
THE INSTANT I CRACK THE THROTTLE THE MOTOR STARTS TURNING.
I REACH FULL SPEED/CURRENT IN JUST THE VERY LAST FEW DEGREES OF THE THROTTLE.

Now here is how you do it....

I like to set up a motor in a test stand (or turn your bike upside down)
Set the current limit in the CA to 2A (maybe 3 or 4 for yours, my motor is a 12 turn).
Why 2A?
Because there is no load on the motor and we want to have some resolution. Freewheel current on my motor is about 1A so this gives me a testing platform to work with.

First: ELIMINATE SLOP AT BOTTOM OF THROTTLE

This is the easiest... first off, if you are not reading the manual then get the F out of here I am not looking to spoon feed anyone... just looking to help those who are trying to help themselves

8.18 AUX Threshold
This is what sets the "turn on" threshold for your throttle. You want to set this just a tiny bit over what your throttle sits at. Mine sits at 0.817V so I set this threshold at 0.820V. Setting it below 0.817 (for mine) results in the motor turning with no throttle. Setting it any higher than it needs to be results in slop at the bottom of your throttle

Easy eh?

Now we are mapped -> 3.0V range starting at 0.820V and ending at 3.820V. (range is 3.0V no matter what... cant change that)

That is the first part of our mapping.

Now all we have to do is set the output range.
My controller responds to 1.5V to 3.3V
SO....... (RTFMD)

iTermMin = 1.49V
iTermMax = 3.3V

This means that when my throttle first cracks (0.820V) the CA will begin by putting out 1.49V
See?
Now we have mapped 0.820V of input to 1.49V of output.

iTermMax is the highest value the CA will put out, so setting it at 3.3V eliminates the slop at the top of your throttle.

Ah.... ain't that better?

Now there are only 2 more things we care about. I will tune the feedback gain under load as this is totally system dependent.

As a side note....
For those that do not want current throttle but do want full range on their hall throttle you can do this exact same procedure but set it up for Speed instead of Current.

-methods

[hillzofvalp]

Thanks for the input. I will try your tuning method today. I have my current throttle wiring working, but it seems like the oscillations occur very randomly between 25 and 40mph.. probably because the wind is so variably affecting the power requirement..


edit: come to think of it, do you realize how dangerous an oscillating current throttle could be with a motor such as my future cromotor? I wonder if I can work the kinks out with my 9C before I switch over. I would assume a current cap will always help for testing as well.

[hillzofvalp]

I may add a particular issue that you guys may keep in mind: disconnecting the throttle output wire from the CA and leaving the CA throttle output wire connected to controller after having current throttle enabled will cause your bike to go full speed. It apparently need throttle reading to be sane. This could be dangerous, obviously. Luckily I Was on a dusty garage floor or my bike would've been in my face. IMagine a disconnected throttle lead while on the bike.

Is there a safety mechanism we can implement in the event the throttle voltage is removed? or is it just because I haven't tuned my CA? edit: I just followed your procedure and the thing will still take off without the throttle connected to the CA. I'm starting to wonder if the insane throttle protection in the infineon controllers would sense this issue and save the day... ? I have mine off atm

EDIT: METHODS:

bare throttle idle output(no controller connected) is .889V. when output wire is connected to controller it reads .856V at rest. You are talking about the first, correcT? top end voltages are 3.796V bare and 3.714V with the controller attached.

[methods]

Thanks for the input.

Just fyi I was speaking to the public so dont take any comments I made personally (i.e. about rtfm etc)

As far as the exact voltages... tune them in. I had a rude surprise when I switched from my 18S testing station to my 24S bike. My controller was set up in reverse and about 3 seconds after I plugged in the controller the bike exploded backwards

I brought the aux threshold voltage up 100mV and now it is great.

Just keep going down until the bike goes with the throttle off, then come up a tiny bit. Sneak up on it.

So I took a ride to tune in the feedback. I was running only a 5Ah pack and 75A on a 12fet so after only a few minutes everything was heatsoaked... my motor, wires, controller, and pack. (big, hot, puffy lipo ). That aside (soft response) I had to turn the IntAgain down to 40 to get rid of the oscillation.

(IntAgain is the only feedback related setting that affects the current feedback loop. It defaults to 300. I tried it at 900 and it was like riding on rail road tracks. I tried it at 10 and it was like a slow start controller. Even at 40 it was still not perfect and I had already given up quite a bit of throttle response. Now to be fair... 99% of us are just screwing around on these bikes (no serious competition) so a little lag on the throttle is not going to lose the gold metal for anybody... Some slow response actually helps keep motor temps down and saves on current.

RE: Blastoff
I hard-wired my throttle to my CA with an short piece of wire. No worries for me

RE: Poor performance
If someone is setting up for SPEED control it is very important to set the speed limit! IF you leave the speed limit set to 99mph then your throttle resolution will scale to 99mph and even just cracking it open will tell the bike to do something like 30mph... which will result in hitting the current limit instantly. To use speed mode set the speed of the CA to just a hair above your true top speed (not freewheel speed...) to maximize resolution

So... after a quick test I think that the feedback will be able to be tuned to a rideable point but as it stands now it is not going to be tits. Part of the problem is throttle resolution. I think the feedback routine is a little simple for higher power bikes. I think to get it real smooth with our monster bikes we need to take speed into the feedback loop - not just current.

-methods

[hillzofvalp]

I hard wired mine, too, but Im talking about the what-if-it-broke situation. I do appreciate your input very much, as the bike feels a lot better.. NOW I just got to make the twist throttle experience more ergonomical.

I think the cromotor will be great with this throttle, but I wish the throttle curve was more exponential than its current linear state. That way I'm less likely to bump it and fly off the bike. I suppose this could be adjusted with some sort of simple circuit.

I bet Justin will address the tuning in the next CA. Word has it he will even have a watts limit.

[Farfle]

Ill do a writeup on the RC circuit throttle damper circuit I did. A 6k resistor and a100 uf cap made throttle response so much better at low speed.

[Arlo1]

Thanks for this thread guys I will have to go over this and see how it works for me with higher power. I set it up on my brothers bmx and it was not bad...

[methods]

Ill do a writeup on the RC circuit throttle damper circuit I did. A 6k resistor and a100 uf cap made throttle response so much better at low speed.

Bah... we call that "slow start".
We must have spent 1000 hours trying to get the Chinese to *stop* putting this LPF on the controller inputs.

The thing is... yea, it makes the low speed response much better but it totally trashes the high speed response when using speed throttle. If you are going 40mph and you let the throttle go, then wrap it back around again there is like a second delay before the power hits. Drives me absolutely mad.

How long is the delay in your RC?
(yes... I know I could do the math )

So you tuned it such that the tiny wrist variations get filtered out but large changes still hit fast?

Ideally we would just do it digitally like Justin did with the CA but add a second speed factor. Just a slope to account for speed so we can still get a fast throttle response without getting bucked off below 3mph.

-methods

[MadRhino]

Yep, some days I miss the clutch lever that gave so fine control.

[methods]

Maybe I can write an arduino program
"Virtual Clutch"

I am joking... sort of.... that could actually be kind of interesting.

-methods

[MadRhino]

That is a pleasing eventuality. I can imagine a left brake and clutch unique lever, where the clutch action would start on release just after passing the brake bite point.

[hillzofvalp]

Do your write up for the buffer, please! Sounds good for current throttles

[Sacman]

Methods you and I breifly discussed in PM's using the CA's throttle control abilities as a Cruise Control. But you said you never really looked into it. Could you take a look at the Cruise Control possibility now since you're looking at the CA's throttle? I just want to know if it's possible to use a momentary switch to set/release the cruise control speed. Thanks.

[gensem]

humm.. and what do I do when the last 20% of a magura throttle does not vary in voltage? cry?