Throttle interface for sensitive high power setups

[zombiess]

No more feedback on my part.

So will that hall sensor upgrade work for sensorless setups? I mean my controller works out the motor, and the halls are not being used. So can your upgrade work off the unused hall sensors to do its "thing"?

It will only work if your controller outputs around a 5 volt speed pulse (doesn't matter if you have halls or not, only a speed signal matters). If you have a cycle analyst and it can read your speed, than this can work by connecting it to the interface.

The plan is to have the interface automatically detect if a speed signal is available and automatically change modes.

 

[snellemin]

No more feedback on my part.

So will that hall sensor upgrade work for sensorless setups? I mean my controller works out the motor, and the halls are not being used. So can your upgrade work off the unused hall sensors to do its "thing"?

It will only work if your controller outputs around a 5 volt speed pulse (doesn't matter if you have halls or not, only a speed signal matters). If you have a cycle analyst and it can read your speed, than this can work by connecting it to the interface.

The plan is to have the interface automatically detect if a speed signal is available and automatically change modes.

Cool. Sounds good.

 

[zombiess]

I'm now working on the code for having the interface use a hall or speed pulse to help control the acceleration. I should be able to get the throttle response to be very smooth with zero lag when letting off the throttle and then getting back on it while at speed. Goal is to turn the throttle into a real accelerator vs a speed control.

I just sent of a test batch of 3 boards to OSHPark as well. Got the size of the board down a little so when I get this code done I can have some people test it too.

 

[Stealth_Rider]

The new interface using a hall sensor should really help out a lot on providing a smooth throttle response. Looking forward to try it.

It took some experimentation on the buffer and delay pots but I finalized my Stealth Bomber settings for all 4 pots and found a small plastic enclosure to house the circuit board.
http://www.digikey.com/scripts/DkSe...ng=en&keywords=377-1938-ND &x=13&y=13&cur=USD

The Bud enclosure is smaller than the one shown here and by notching the Bud box for the exit wires you get nice and compact protection for the interface.
http://www.fareinc.com/65/Throttle_Interface/content/IMG_2769_large.html

The enclosure and board now reside inside the Bomber frame and are working fine.

-Jim

 

[ZOMGVTEK]

I managed to dump the bike at something like 20 a few days ago. Apparently one hand on the bars is not terribly sufficient with a bit of unintended throttle. I'm mostly fine and the bikes fine. Just burned a hole in jeans and a shirt, a little road rash and an ankle injury. I'll be fine in a few days.

I'm not putting blame on the board, it didn't really have much to do with it, I don't think. However, I probably will just pull the board off for now. I'm up for testing out any future updates, but something is definitely not working out that well on my setup.

 

[zombiess]

ZOMGVTEK, I sent you a PM about an update if you want to give it a try. I am really curious to see if it's the board or something else in your setup that we are missing that could be causing your delay.

 

[zombiess]

Any more feed back? I'm not seeing anyone posting on this any more. Did everyone find settings that worked well?

 

[snellemin]

No complains so far. I just would like another one for my sensorless DD bike.

 

[zombiess]

Development is on a short pause for a week or so while I finish designing a PCB for another project. Good news is I flashed the V2 beta into my interface and went on a 6 mile ride with my wife. Our average speed was 7 MPH and the interface made the ride absolutely effortless with beautifully smooth low speed control, especially when I have the 3 speed switch set on the lowest power. Response is still fast enough to still pop the front end up to climb curbs so I have my settings perfect and they only took me about an hour of tuning to find. They are the same ones I shipped all the controllers with, but of course they are tuned to my specific bike so I knew everyone would need to adjust theirs differently.

I'm not sure about everyone else, but for me once I have adjusted to the way the throttle works with the interface there is no way I would ever go back to a controller without one.

Detecting acceleration is coming up next. I already have the interface reading the pulses from the halls / or controller so it's just a matter of writing the algorithm.

I also received a my test run of 3 boards from OSH Park today, they are beautiful and smaller than the beta units. I feel once I have the wheel speed function working it will be much easier to adjust the acceleration setting for almost everyone. Stay tuned, more updates to come. I'm hoping to be in production by the end of June at the latest.

 

[Scott]

Detecting acceleration sounds like a good idea and may be similar to what I’m thinking of. If it is possible to reprogram controllers like the two that you sell. Limiting the current with speed would seem like the perfect way to make a high power bike more ride able. In my case I would want to limit the amps so I can lift the front wheel at all speeds without being thrown off the bike. I do not know the exact current to limit this with but just for an example: 0mph allows 30 amps max, 10 mph allows 40 amps max, and 11mph allows 41 amps max, and so on. Then if this setting could be set by the user by the controller programming interface, or if you could possibly make a small device that connects to the controller to do these task that would be ideal. It’s just an idea but let me know what you think.

 

[zombiess]

I'm back writing code again. I have a few problems to resolve but hope to have a test version on my bike by the end of this month which takes wheel speed into account and turns the throttle into an accelerator vs a speed control.

 

[zombiess]

This is proving to be a bit tricky to come up with a good method for control. It's easy for me to get the wheel speed and make the calculations and that code is done. I'm debating on how to deal with resuming the throttle. I can have the interface learn wheel speed vs throttle output and save a table that periodically updates. I've done this before on other projects, it basically just auto tunes itself. The only issue with this is when you start using a 3 speed switch because it needs to sense this occurred, otherwise the throttle output voltage vs wheel speed will be wrong. I think I'm going to write some code that automatically detects if this is the case and switches between tables or a scaler that automatically adjusts the table once it figured out what speed mode it's in. The other solution is to rapidly ramp up the throttle over a very short time period detect when the bike is accelerates and clamp the throttle to this point to continue on in acceleration mode. Not sure if I can pull this off smoothly or not (I don't want it to upset the bike). I have a 20mS window and my code is good at detecting the acceleration. This would make the interface agnostic to what ever speed mode the end user is in if I can get it to work well. I think I'm going to try this method first because my goal was to make this device as simple as possible.

I'm thinking I might do a small production run of the interface as it is now for those who want to use them as I have had some pretty good feedback. I put one on my wife's 9c 2806 20" 2kW setup yesterday and she could handle the bike fairly smoothly even in high power after about 20 mins of quick test rides adjusting the interface. I feel this is a pretty good metric as she is not familiar bicycles and ebikes. I wish I could simplify the adjustment process but every bike is unique. For reference, her bike can not pull the front end off the ground even from a dead stop the way I have the controller tuned, but she still complained the throttle was too jerky while riding part throttle. Riding a bike is not natural for her yet and she can't even shift gears without looking at what she is doing so I'm trying to make everything as easy as possible and adding the interface has really helped. When I rode it, I thought it was perfect and very smooth with no interface needed, but I have years of experience and great throttle control.

I'm going to make a post in the for sale section to see how many people are interested in a V1 of the interface so if you are interested, head over to that thread.

 

[snellemin]

Yesterday I was riding the critical mass and my throttle didn't work. I kept losing throttle and had to keep resetting the controller to gain throttle. When I let go of the throttle for a bit, it would react anymore. I took it out and had to ride with the throttle directly connected.

 

[snellemin]

Yesterday I was riding the critical mass and my throttle didn't work. I kept losing throttle and had to keep resetting the controller to gain throttle. When I let go of the throttle for a bit, it would react anymore. I took it out and had to ride with the throttle directly connected.

 

[zombiess]

Yesterday I was riding the critical mass and my throttle didn't work. I kept losing throttle and had to keep resetting the controller to gain throttle. When I let go of the throttle for a bit, it would react anymore. I took it out and had to ride with the throttle directly connected.

That means it was detecting a fault condition of some sort as I designed it to latch to 0v under any faults until power is reset. The problem should reside between the throttle and the interface. I've had the same thing happen to me when I re-installed my interface on my bike a few weeks ago. Turns out I had a bad ground in my connector that was caused by frequent removal. 10 mins later and a dab of solder and all was good again. Had me puzzled as well because I could hook the throttle up and it worked perfect. Installed the interface, it worked, wiggle cable, no throttle... power cycle, good to go... wiggle cable... no power.

The other time the throttle signal pin popped out of my connector on the controller side, nothing worked, not even the stock throttle because I have the bar protect mode enabled in my controller to prevent a runaway throttle condition.

I think I'm no up to 5 or 6 failed throttle connections in my ebike experience all caused by frequent use of the connectors. Not sure there are many ebike failures left to experience except maybe a LiPo fireball... hope I never get that one. The run away throttle incident was interesting about 1.5 years ago. Made me very thankful I had a kill switch. Ground wire failed from throttle causing a WOT condition. Bar protect was NOT enabled in the controller program. Once I figured out what it did I always make sure to program it and I always have a kill switch setup.

 

[snellemin]

I'll check and replace my connectors when I get some time. I'll look for some different connectors and see how it goes.

I'm right now framing and replacing my rear sliding door. Damn termites got to it. Another unwanted cost. There goes the cromotor and controller money

 

[zombiess]

I'll check and replace my connectors when I get some time. I'll look for some different connectors and see how it goes.

I'm right now framing and replacing my rear sliding door. Damn termites got to it. Another unwanted cost. There goes the cromotor and controller money

That's how it always happens. I spent years trying to buy something I wanted and every time I had the $2k saved up... something else come up making me broke again.

 

[zombiess]

Here is an update for everyone interested. I just received the PCBs today and they fit perfect. Going to build one up tonight and after I verify it works well I will be placing the order for 70 more PCBs and enclosures. I already own the components to make them all so its just a waiting game for the PCBs. Sorry for the delays but I want these to be protected and look good.

They will be shipping in single output only unless dual throttle output is specifically requested. You can easily add the 2 components needed yourself if you ever need to use this functionality.

Some pictures of what is to come.

 

[Supertux1]

I have been toying with the idea of adding a 1 axis gyroscope to one of the chain stays near the axle and 3 axis accelerometer somewhere around the head tube.

Using inputs from those devices and perhaps a speed / current signal from the controller, an Arduino could probably regulate the throttle signal in a closed loop fashion, much the way the CAv3 does speed/current/power/amps, but with angular and momentum data instead.

Depending on how responsive the motor control system is, I was thinking that this would provide enough feedback for an anti-wheelie throttle. It would be able to detect a potential bike flipping/burnout situation and compensate for that by dynamically turning down the throttle until the situation corrects itself.

OR it could give you a nice wheelie mode by doing the opposite, increasing throttle when it detects freefall, but also shutting off when it detects any side to side tilt.

Maybe a combination of both, anti-wheelie/anti-burnout if not moving, wheelie mode if accelerating and detection of a angular event in the range of 30-45 degrees/second. (eg. Yank up on the handlebars.) In wheelie mode the throttle could then operate as a gain sense which would determine the angle to maintain.

 

[zombiess]

I have been toying with the idea of adding a 1 axis gyroscope to one of the chain stays near the axle and 3 axis accelerometer somewhere around the head tube.

Using inputs from those devices and perhaps a speed / current signal from the controller, an Arduino could probably regulate the throttle signal in a closed loop fashion, much the way the CAv3 does speed/current/power/amps, but with angular and momentum data instead.

Depending on how responsive the motor control system is, I was thinking that this would provide enough feedback for an anti-wheelie throttle. It would be able to detect a potential bike flipping/burnout situation and compensate for that by dynamically turning down the throttle until the situation corrects itself.

OR it could give you a nice wheelie mode by doing the opposite, increasing throttle when it detects freefall, but also shutting off when it detects any side to side tilt.

Maybe a combination of both, anti-wheelie/anti-burnout if not moving, wheelie mode if accelerating and detection of a angular event in the range of 30-45 degrees/second. (eg. Yank up on the handlebars.) In wheelie mode the throttle could then operate as a gain sense which would determine the angle to maintain.

I say go for it, but be forewarned, what seems so easy in your head or on paper can become a nightmare in reality. I'm not even using any kind of feedback loop in the version of the throttle interface I'm starting to sell and it was quite challenging to get working as well as it is. It took many attempts because most of my control ideas I had failed when I finally went to put them actual hardware that controlled the throttle. Another thing to think about is safety of adding any device in between the throttle and the controller. There are numerous failure modes that you should account for even if this device is a one off only for your own use.

 

[Ecyclist]

Zombiess, the answer is probably no, but I will ask anyway - is that interface able to work with hobby type ESC like Castle Talon and without CA?
Is something out there that can do it? I'm hoping for a streamlined set-up with RC motor. Alien Power System has something like that, but they are always out of stock.

 

[amberwolf]

The Throttle Tamer is an analog control; RC controllers are generally PWM inputs, so they won't directly work with each other. You'd have to use an analog throttle into the Tamer, then at the Tamer's output convert analog to PWM (not impossible, just another step and bit of electronics).


But the big problem is the Tamer isn't made anymore, so whatever you can find in the for sale sections used is about all there is.

 

[Ecyclist]

The Throttle Tamer is an analog control; RC controllers are generally PWM inputs, so they won't directly work with each other. You'd have to use an analog throttle into the Tamer, then at the Tamer's output convert analog to PWM (not impossible, just another step and bit of electronics).


But the big problem is the Tamer isn't made anymore, so whatever you can find in the for sale sections used is about all there is.

Thank you amberwolf, but I'm looking for a source so I can build the second or third system with the same set-up once I know that my set-up is working well.
Like I said, Alien would work, but they don't have in stock what I need at this time.
Actually, they have interface for about $10, but they don't have a controller that will work with it.
http://alienpowersystem.com/shop/rc-ev-accessories/electronics/alien-throttle-interface-0-5v/
Saying all that, there must be a smart guy on the forum who can put that interface with Castle controller and make it work.
Unfortunately, I can only do mechanical stuff.

 

[amberwolf]

Thank you amberwolf, but I'm looking for a source so I can build the second or third system with the same set-up once I know that my set-up is working well.

Then the Tamer isn't what you're after, cuz it's not readily available.

Actually, they have interface for about $10, but they don't have a controller that will work with it.
http://alienpowersystem.com/shop/rc-ev-accessories/electronics/alien-throttle-interface-0-5v/

Unless I totally misunderstand what that page says and implies

With this interface you are able to connect any electric throttle 0-5 volt to an RC ESC.

then it should work with every RC controller, Castle, Alien, generic, etc., because it is converting a 0-5v input signal (like from a potentiometer throttle) to a PWM signal (for any RC ESC). (assuming that all the different RC controllers respond to the PWM input on their throttle input the same way)

So unless the site is out of all of their RC ESCs, they do have controllers that will work with it. Those controllers might not run the motors you want but I don't know enough about that part to help much.

 

[Ecyclist]

So unless the site is out of all of their RC ESCs, they do have controllers that will work with it. Those controllers might not run the motors you want but I don't know enough about that part to help much.

Well, you hit the nail on the head. At this point, they don't have in stock RC ESCs controllers that will work with the good size motor and the 13s battery.
I'm not an electric or electronic guru, but I think that Alien controllers are modified so they can connect with Alien interface.
Maybe I'm wrong, and I hope I am, but any other controller will need some kind of magic to put into it to make it work with Alien interface.
If anyone proves me wrong, I'll be ecstatic.
If anyone can do the magic with the Castle Talon controller and make it work with Alien interface, I will be super ecstatic.