What features do you want? Throttelizer/Throttle interface

[www.evlogix.com]

This throttle device is now being sold at http://www.endless-sphere.com/forums/viewtopic.php?f=9&t=11877&p=180070#p180070.

Now, back to the original post:

Hey, my company is currently designing a microprocessor based solution to using a throttle on an RC system. I'm trying to understand what you'd look for in a device like this, so I'll explain what it is first.

It's basically a device that allows one to use a regular hall/pot throttle on a RC ESC. Since it's digital in its design, it could also have a load of features such as data logging (Like recording your speed, distance, time, voltage, current, altitude etc. over the ride), a display of some sort, it could be attached to a security system, advanced current limiting (Like say you limit it to Y amount of amps for so many seconds and then a limit of X amps turns on - useful for high current bursts without killing the battery, esc or motor from overheating), low voltage cutoff, temperature protection and ultimately programmability. With programming, you can decide if you want current limiting, you can specify the advanced current limiting settings, you can program what you'd want to see on the display, etc. It's also a current controlled throttle, so more throttle = more current and ultimately more power, kind of like a motorcycle. It's not like a crystalyte controller where above some throttle amount going up a hill, more throttle means the same amount of current and, conversely, less throttle does not necessarily mean less current and what "real" throttle range exists is pretty small and hard to control. Half throttle would mean 50% of the top current, 25% throttle means 25% of the current limit, 100% means full current. So, it feels more responsive and natural.

It also comes with high quality LOW ESR capacitors that are essential to running e-bikes and other EVs on ESCs, so you don't have to add the capacitors yourself. Of course, you could add your own capacitors if you want.

How is this better than say, hacking your own throttle together? You don't have to buy a BEC, a servo tester, add your own capacitors and basically do a lot of work. You also get fairly important motor and ESC temperature protection (Killing $100-$200 ESCs gets expensive!), customizable throttle ramping, customizable current limiting, advanced current limiting, a more natural throttle response, and you can extend it to do datalogging and you can display critical information using a display. There are other devices that could be developed for it in the future.

So, right now, I'm just trying to figure out what's the most important to have as there's only so many sensors that the processor can work with at the same time.

Right now, I have on it..

-Current limiting
-2 Temperature sensors (Can be used on anything, like one for the ESC and motor or batteries)
-The throttle

With the previous, one could customize a current ramp, advanced current limiting and temperature sensing profile.

The temperature sensor works using a pullback scheme. You program what temperature you want the current limiting to begin, and it decreases the current limit as temperature increases until it completely shuts off at the cutoff temperature you program. Usually, the decrease in the current will result in less heating which means your bike doesn't just shut down, it just slows down. This helps to protect $100+ ESCs and motors from dying from over-heating without forcing you to be stranded but you could also ride without temperature protection - it's upto you.

Data logging would be possible with another device and the same with the display. Right after I get this going, the display will be designed.

There are other possibilities that could replace one or both of the temperature sensors. There could be a programmable low voltage cutoff or a pedelec sensor, or anything else that might seem desirable. Could you guys rate what would be most important for you?

Also, how important is weather-proof-ness? Weather proofing makes it more costly to develop and would increase the price by a noticeable amount and ESCs typically aren't weather-proofed, so I think whatever needed protection for the ESC and batteries could probably be used for this device. But I don't know, how important is weather-proofness? How much extra would you be willing to pay for it?

Here's some screenshots of the settings changing program in its current state. More features will be added and programmed in the future!

 

[fechter]

Excellent.

If you can make them, I won't have to. :wink:

If you have current measurement and you're using a microprocessor, then there's almost no reason not to have CA functions like total amp-hours and pack voltage (possibly via optional display).

Throttle calibration to allow hall or resistor throttles would be good. (edit: I see you already have that covered).

Be warned, implementing current limiting is not so easy. I had to resort to a PID feedback loop (but did not really need the derivative component). Justin did something like this on the CycleAnalyst with programmable parameters.

 

[Sacman]

Any way for this throttle to be controlled by a torque sensor attached to the bicycle pedals?

 

[recumpence]

Oh, Man, I have been wanting this!

Fechter's limiter is awesome! But, he did that as a huge favor to us that are in RC "In" crowd. :wink: Production in any quantity is not feaseable for one person hand soldering things together. Plus, microprocessor control is fantastic!

Depending on the price, I can tell you I will stock a number of these for my own bikes, production E-trikes, and customer projects.

Can I preorder a few?

Matt

 

[www.evlogix.com]

A proto-prototype has been tested and current limiting and the temperature functionality has been verified. Right now, I have to design the pcb around the enclosure I'll eventually use and I should be doing that tomorrow. Once I get the revised prototype tested and verified, I could start sending out the order for the assembled boards. I hope that'll happen within 3 days or so, but things may happen.

If the torque sensor is something that puts out a voltage less than 250 volts, it could be used but the board would have to be designed for that. If it's a normal 5v sensor, it could probably be used directly and it could be used to replace the temperature sensor. Searching for a torque sensor online, I can't really seem to see any, and the ones brought up by google are in the $1000+ range. I think a magnetic sensor type of setup would be easier and cheaper to implement, but what rotational speed to measure is something I haven't determined yet. You could measure pedal speed, but that doesn't account for the change in gearing. If the device could detect what gear it was in, then the proportional amount of current could be determined, but implementing that doesn't seem particularly easy. I suppose a crude "how fast you pedal is how much current is applied" measurement, but that probably wouldn't be as sophisticated as the bion-x. Anyways, I imagine that would probably be an add-on type of feature so it'd probably be added later as that's figured out. The device will come with connecting ports, so you can attach one or more additional devices to it.

With voltage-related measurements, like for LVC or using it to calculate watt-hours, it'd have to give up either a temperature sensor or current sensing. I could possibly have the extra temperature sensor as an add-on device, though.

So,

Which sounds better?

Setup One
-1 temperature sensor
-Current limiting
-Voltage detection (Used for programmable LVC or Watt-hour calculations)

Or setup Two?
-2 temperature sensors
-Current limiting

I think the voltage detector could be used on the display if it's needed for watt-hour calculations or voltage display. This would probably be the best option if full-pack LVC isn't needed by the majority.

And recumpence, I can give dealer discounts. :wink: The price will be determined after the final design is chosen, which might be in a few days.

 

[recumpence]

Sweet.

My Paypal clicking finger is itching. :wink:

Matt

 

[Miles]

Anyways, I imagine that would probably be an add-on type of feature so it'd probably be added later as that's figured out. The device will come with connecting ports, so you can attach one or more additional devices to it.

I think this would be the way to go. Getting a decent pedal controlled system is quite tricky. It really has to be torque based, I think.

 

[www.evlogix.com]

I was thinking, what causes torque? It's the force applied to the pedals methinks, and the crank length determines the corresponding torque. If some how that force could be measured using, say, a force or pressure sensor, then it might not be hard to implement in concept (and not hard on the pocket book, but still greater than 80 dollars but still not quite a thousand), but it might be tough to design something that could attach to the pedals and withstand the rigors of pedaling without causing a noticeable loss in pedal traction.

Yes, it seems an add-on is most appropriate for that as that would probably take some time to investigate, test and develop.

Anyways, tell me what you guys prefer!

Two temperature sensors or

One temperature sensor and LVC?

I think I may be able to design it so that could be an option instead of having every board being of one type. It does seem like LVC would be important to some, but then not so important to others (such as those with LVC on each cell). Also, it seems that two temperature sensors might not be useful to everyone as the ESC seems to be about the only critically heat-vulnerable part of the system, unless you're running an undersized motor or you're overpowering your motor.

 

[Miles]

I was thinking, what causes torque? It's the force applied to the pedals methinks, and the crank length determines the corresponding torque. If some how that force could be measured using, say, a force or pressure sensor, then it might not be hard to implement in concept (and not hard on the pocket book, but still greater than 80 dollars but still not quite a thousand), but it might be tough to design something that could attach to the pedals and withstand the rigors of pedaling without causing a noticeable loss in pedal traction.

There are many places you could measure the force/torque at. It's a subject that has had a lot of attention from the developers of the bicycle power meters used for training.
Ref:
http://www.g-cog.com/G-Cog.htm
http://www.quarq.us/cinqo
http://www.torq.ltd.uk/acatalog/Powertap_hub.html
http://www.thisisant.com/index.php?section=30

There are also different ways to measure at the chain (deflection force; frequency etc.).

Pedal input force is pulsed, so there would need some kind of smoothing before it could be used for control.

The system that I would like to see implemented would use a torque threshold. The motor would only kick in once this threshold had been crossed and would then be controlled (closed loop) to maintain the pedal input to the threshold level, as far as possible. The user would have real-time control over the threshold level via a control on the handlebars.

 

[Miles]

I think I may be able to design it so that could be an option instead of having every board being of one type. It does seem like LVC would be important to some, but then not so important to others (such as those with LVC on each cell). Also, it seems that two temperature sensors might not be useful to everyone......

Choice between different options would be great

 

[liveforphysics]

I don't pedal, and I don't care about torque sensors.

My perfect throttle would be something that feels like a proper motorcycle throttle rather than a plastic toy like the magura and the kelly throttles that I have. Definately something with a full hand grip throttle, no thubs or half grip stuff.

If its going to have current limiting, it needs to be able to be calibrated for an external shunt. I use a 500amp shunt on my bike right now, and I plan on sending more than 500amps in bursts. (With the new controllers and motors).

LVC and 2 temp inputs would be everything I would want.

 

[johnrobholmes]

Two temp sensors here. I think you have the product spot on, just what is needed!

 

[Grinhill]

Super-small is good for me. My present circuit is about 25mm x 20mm x 15mm. However, today I discovered that it ceases to function correctly below 10 deg Celcius. I held it in my hand for a minute and away I went.

 

[fechter]

A proto-prototype has been tested and current limiting and the temperature functionality has been verified. Right now, I have to design the pcb around the enclosure I'll eventually use and I should be doing that tomorrow. Once I get the revised prototype tested and verified, I could start sending out the order for the assembled boards. I hope that'll happen within 3 days or so, but things may happen.

It sounds like I should talk to you.
Where are you getting your boards done?
Are you assembling the prototypes yourself?
Are these surface mount?
What are you using to measure the current?

I have lots of other projects that are kind of bottlenecked by the board assembly part.
I'm building another batch of the through-hole analog RC throttle adapters and it is very time consuming (not fun). I'd much rather get them made by machine. This should also bring the price down, which would also be good.

What kind of micro controller are you using? Good at programming? I can think of several things to do with microcontrollers but I suck at programming. I guess I'll have to learn...

Sorry for all the questions.

 

[dumbluck]

I'm in for the pre order!!! Paypal ready to go!!! I'm staring at a box of rc parts that just showed up and boy would it be nice to just plug and play. I hope you have some good luck getting started soon.

 

[www.evlogix.com]

A proto-prototype has been tested and current limiting and the temperature functionality has been verified. Right now, I have to design the pcb around the enclosure I'll eventually use and I should be doing that tomorrow. Once I get the revised prototype tested and verified, I could start sending out the order for the assembled boards. I hope that'll happen within 3 days or so, but things may happen.

It sounds like I should talk to you.
Where are you getting your boards done?
Are you assembling the prototypes yourself?
Are these surface mount?
What are you using to measure the current?

I have lots of other projects that are kind of bottlenecked by the board assembly part.
I'm building another batch of the through-hole analog RC throttle adapters and it is very time consuming (not fun). I'd much rather get them made by machine. This should also bring the price down, which would also be good.

What kind of micro controller are you using? Good at programming? I can think of several things to do with microcontrollers but I suck at programming. I guess I'll have to learn...

Sorry for all the questions.

I'll be getting a quote for the boards from several places. The prototype board is created in-house.

The prototypes are assembled myself.

The majority of the components are surface mount. There are a few that aren't, such as the needed electrolytic capacitors, voltage regulators and wires, but it's definitely all migrating towards automation given the inherent labor and time advantages.

The current is measured using an excellent temperature-coefficient shunt resistor. The signal is then processed and software is used for calibration.

The chip is an AVR. PICs would be easier, but it's what I know.

 

[www.evlogix.com]

I finally got the prototype built and almost completely soldered and now it just needs to be tested! How quickly I can move from this to production depends on the testing results, but I suspect they'll be good. Hopefully, I can get comprehensive testing completed by tomorrow and I'll get quotes for the boards within a few days. It's going to be soon! (I hope!)

 

[fechter]

Thanks for the replies.
What is the shunt resistor you're using? I'm using the Allegro hall effect current sensor, but they just jacked up the price on those to over $6ea.

I'll be getting a quote for the boards from several places. The prototype board is created in-house.

Your house has better toys than my house

 

[www.evlogix.com]

I'm using between 2-3 of these kinds of sense resistors. With the additional signal processing parts, the extra needed board space and extra amount of needed soldering, it seems like Allegro's solution is pretty close in terms of cost effectiveness if not better. Well, assuming you needed a 5v supply in your circuit for other reasons.

http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=CSNL20.0005FRCT-ND

So far, it seems there was a few unanticipated goofs in the board design. Cut a few traces, jump a few things, and away I went! It seems like I won't get testing completed until possibly tomorrow morning with all these corrections I've had to make, so figuring in about another day's worth of some redesign time and it seems the board design won't get finished until probably tomorrow or the day after. After that, it should be ready to order!

And, with about 300 dollars and little to none DIY time, you can get your own mini-fab going. I use the photolithographic method which is good enough for 10 mil traces and 10 mil spacing. The per board cost is about 6 dollars with all consumables considered for a single small board, so with the extra DIY time it takes to make the board, it might not be worth it over a board manufacturer. But it's true you don't have to wait a couple of days for a prototype which I think definitely helps speed up the development cycle!

 

[fechter]

Thanks for the info on the shunt resistors. If my math is right, a .5milliohm, 2W one can handle about 63 amps.

I do like the Allegro sensors. It eliminates the need for an op amp on the shunt, and they seem to be quite accurate.

 

[www.evlogix.com]

Thanks for the info on the shunt resistors. If my math is right, a .5milliohm, 2W one can handle about 63 amps.

You're 100% correct, which is why upto three is used in parallel to get the desired current rating. Also, as a reminder, that's the continuous rating and 190 amps is pretty high for continuous use - If you're running at 50V, that corresponds to about 10 kW which is definitely greater than 50 mph on the flats and more than 45 mph up relatively steep hills - bursts upto 10-30 seconds could be upto 20 kW or so, which is plenty. But, that's the ordinary shunt resistors. There's are also higher rated shunt resistors available on special request. In the next board design, I think I'll accommodate external shunts that can be programmed.

 

[MitchJi]

Hi,

Hey, my company is currently designing a microprocessor based solution to using a throttle on an RC system. I'm trying to understand what you'd look for in a device like this, so I'll explain what it is first.

It's basically a device that allows one to use a regular hall/pot throttle on a RC ESC. Since it's digital in its design, it could also have a load of features such as data logging (Like recording your speed, distance, time, voltage, current, altitude etc. over the ride), a display of some sort, it could be attached to a security system, advanced current limiting (Like say you limit it to Y amount of amps for so many seconds and then a limit of X amps turns on - useful for high current bursts without killing the battery, esc or motor from overheating), low voltage cutoff, temperature protection and ultimately programmability. With programming, you can decide if you want current limiting, you can specify the advanced current limiting settings, you can program what you'd want to see on the display, etc. It's also a current controlled throttle, so more throttle = more current and ultimately more power, kind of like a motorcycle. It's not like a crystalyte controller where above some throttle amount going up a hill, more throttle means the same amount of current and, conversely, less throttle does not necessarily mean less current and what "real" throttle range exists is pretty small and hard to control. Half throttle would mean 50% of the top current, 25% throttle means 25% of the current limit, 100% means full current. So, it feels more responsive and natural.

Current control sounds excellent!

How about cruise control?

How about a web browser (cross platform) interface?

 

[www.evlogix.com]

Hi,

Hey, my company is currently designing a microprocessor based solution to using a throttle on an RC system. I'm trying to understand what you'd look for in a device like this, so I'll explain what it is first.

It's basically a device that allows one to use a regular hall/pot throttle on a RC ESC. Since it's digital in its design, it could also have a load of features such as data logging (Like recording your speed, distance, time, voltage, current, altitude etc. over the ride), a display of some sort, it could be attached to a security system, advanced current limiting (Like say you limit it to Y amount of amps for so many seconds and then a limit of X amps turns on - useful for high current bursts without killing the battery, esc or motor from overheating), low voltage cutoff, temperature protection and ultimately programmability. With programming, you can decide if you want current limiting, you can specify the advanced current limiting settings, you can program what you'd want to see on the display, etc. It's also a current controlled throttle, so more throttle = more current and ultimately more power, kind of like a motorcycle. It's not like a crystalyte controller where above some throttle amount going up a hill, more throttle means the same amount of current and, conversely, less throttle does not necessarily mean less current and what "real" throttle range exists is pretty small and hard to control. Half throttle would mean 50% of the top current, 25% throttle means 25% of the current limit, 100% means full current. So, it feels more responsive and natural.

Current control sounds excellent!

How about cruise control?

How about a web browser (cross platform) interface?

Cruise control would be possible if it detected speed. It doesn't as of the current version, but a speed detector may be possible through an add on display or possibly an add-on speed detector.

And the current interface is written using a cross-platform library, so it should work on multiple operating systems. However, communicating over the serial port is pretty operating-system specific and can't be done solely through a generic web-interface, so a different version of the software has to be written for each operating system. Hopefully, it'll be pretty easy to port. Currently it's set to work under Windows XP and Windows Vista, and I might port it to Mac, but I don't have a mac available to me sadly enough that's needed for testing. Even the local libraries only carry Windows.

 

[MitchJi]

Hi,

Cruise control would be possible if it detected speed. It doesn't as of the current version, but a speed detector may be possible through an add on display or possibly an add-on speed detector.

Matt said the HV Esc's do a pretty good job of maintaining speed. In any event being able to choose a setting and have it maintained until a brake is applied would be (IMO) a good start. The brake cut-off could be either mechanical or electrical.

And the current interface is written using a cross-platform library, so it should work on multiple operating systems. However, communicating over the serial port is pretty operating-system specific and can't be done solely through a generic web-interface, so a different version of the software has to be written for each operating system. Hopefully, it'll be pretty easy to port. Currently it's set to work under Windows XP and Windows Vista, and I might port it to Mac, but I don't have a mac available to me sadly enough that's needed for testing. Even the local libraries only carry Windows.

I only have a Mac but it would probably be a better solution for Mac users to borrow a PC than for you to do a Mac port .

 

[www.evlogix.com]

Hi,

Cruise control would be possible if it detected speed. It doesn't as of the current version, but a speed detector may be possible through an add on display or possibly an add-on speed detector.

Matt said the HV Esc's do a pretty good job of maintaining speed. In any event being able to choose a setting and have it maintained until a brake is applied would be (IMO) a good start. The brake cut-off could be either mechanical or electrical.

Wait, so the HV Esc's have a speed limiting function so you could limit the motor's RPM to a certain amount? That's cool. The current version already has an e-brake input, so it seems that could be as simple as adding wires for a user-provided cruise control push button. I'm not yet upto the task of designing things to be held onto the handlebars, but hopefully that'll change with development of the display.

Or it could be a cruise control setting that always on unless the e-brake is engaged, but that sounds like it could be dangerous.