conceptual question

[graevus]

Greetings,

Been reading quite a bit here lately. Let me just say right off the bat: good work guys; this is quality engineering stuff all the way, regardless of pedigree. Everything from capacitive soldering design to control system mods, nice. This is an amazing resource.

Not trying to sound greedy, but I believe E-Bikes will be a major growth industry in the next decade or so, and you guys here have the resources ad knowledge to much improve the crappy Chinese lead acid E-bike design which is the current big seller in the word today. IMO, there is no better ecological or financial solution to our country’s current transportation cost/pollution issues.

On that note, I have a few questions if you experts wouldn't mind.

Living near Seattle I am faced with an interesting set of problems, specifically steep hills and water damage.

I understand the whole Ohms law thing, where power is the product of voltage and current, and although my motor theory is weak I know voltage=speed and current=torque. Why could you not then have a switch which would change the battery voltage from say 48V to 24V, thereby doubling the amperage and torque while halving the speed? It seems like it would get you up the hill a little better to me, although it would require an unavailable controller as far as I know. It sounds better than a geared motor, and it could even be an automatic change with proper motor feedback. Pulse width modification might fit right into this idea.

The FAA stipulates a minimum of 8G wire for a max of 45 Amps, what gauge do recommend for the battery to controller and controller to motor connections? I also read that Hall Effect sensors do not function well in extremely wet environments, such as the one I live in, and Anderson connectors, although cheap, are really not made to withstand abuse of this type from my research. Dead shorts are really bad news for batteries and controllers; they tend to fry.

If I had the disposable income, I would buy a lithium Milwaukee battery bank using 2 28V’s in series, and 8 more in parallel to make a 56V (really 48V) 15Ah pack. Run it through a high amp controller and a 500W(?) minimum brushless hub motor on some sort of steel frame cruiser bike w/ torque arm if it is a front motor. Batteries in front if the motor is in the back, or vice versa. Only reasons I would go with the Milwaukee are that it is easier to return than to someone named Ping in China, and it’s beyond the alpha test mode.

Anyway, keep up the good work; this is great stuff here. My hat is off to you.

 

[Anonymous]

.

 

[TylerDurden]

Howdy...

We have at least a couple of O&Os in the Seattle area... Toshi and pwbset come to mind.

Switching series to parallel was the method of choice before electronic controls became abundant. Electronic controllers don't care much for it (LVC, spikes, et al.).

Andersons are fine if you get big ones. Not exactly cheap...

Geared systems have advantages for hilly terrain, not very stealthy tho.

 

[swbluto]

Howdy...

We have at least a couple of O&Os in the Seattle area... Toshi and pwbset come to mind.

I think it's just Toshi. And, soon enough, me. 8)

For the weather-proof conundrum, I think trailer-connectors are going to be a good bet for the continuously detached connections(And weather-proof the controller by coating it and wrapping it in bags; Tape, in a weather-proof manner, open connections; still haven't figured out the throttle, yet. I think I might just craft together my own photo-sensitive potentiometer, and waterproof all connections. Basically, you'd control the amount of current by positioning your thumb over the "photo cell"; Complete cover-up means 100% throttle, no cover means no throttle, and some range in-between depending on the amount of covering-up. An external LED would be provided for the light source so it could work in the dark, and the internal circuitry would be the voltage logic of the device. Maybe I should patent this. :lol: ). You might have to hook a few in parallel for high current: I don't really know what some of the top's current ratings are, but they seem like a good design.

 

[pwbset]

I think I might just craft together my own photo-sensitive potentiometer. Basically, you'd control the amount of current by positioning your thumb over the "photo cell".

Oooh.. that sounds yummy for sure. Keep us posted on your development process. :wink:

PS - I live in northwest Montana, which is basically Seattle's weather 24hrs later and 20deg colder and like 5 people instead of 5 people per square inch. Essentially the same thing. :wink:

 

[graevus]

Howdy...

Basically, you'd control the amount of current by positioning your thumb over the "photo cell"; Complete cover-up means 100% throttle, no cover means no throttle, and some range in-between depending on the amount of covering-up. An external LED would be provided for the light source so it could work in the dark, and the internal circuitry would be the voltage logic of the device. Maybe I should patent this. :lol: ).

You know what might work good for that is a LED driven phototransistor. A phototransistor will allow current to flow from the collector to emmitter depending on the amount of light the base sees. Waterproof the crap out if that side of the circuit, and then drive the phototransistor with a second independant low voltage circuit consisting of a LED and a potentiomenter which would act like a throttle as the LED light varied. That way if the pot shorted all you would loose, in theory, would be a 2 cent LED and not the $200 controller. I wonder what voltage range the output from the throttle uses to tell the controller to speed up or slow down... You would need to design that in to not have to mod the controller. Optical isolation is what we EE's call this type of stuff. You could steal the tiny amount of juice to drive the LED from whatever source you use for the head/tail lights. regulate it down to 5V using something like this:

http://www.allegromicro.com/en/Products/Part_Numbers/8450/8450.pdf

The problem with using your thumb to control the amount of light on something like this is that these devices are very particular about the intensisty, and even the color of the light input. It would work differently on a sunny Vs. a cloudy day, and then change again when dusk/night occured. The trick is to give them inputs which are consistently the same parameters, hence the LED.

Opticaly isolated throttle; i like it.

 

[DahonElectric]

I also live in pretty hilly areas too in Vancouver, BC. Vancouver BC terrain is very similar to Seattle Washington. As for conquering hills. There are 2 ways going about it. You either pedal it up with the motor together or you motor up. If you plan to your use your bike as an electric motorbike, then yes get the biggest and baddest motor you can afford with a high voltage high capacity battery, since this is where it's going to be doing all of the work to push yourself, the bike, groceries and the weight of the ebike kit up any hill. More power, heavier batteries, which means need much more power to combat even heavier weight..

It's easier to decide if you plan to pedal with the motor as a 36v or even a 24v would do just fine with a relatively fit cyclist. In fact, it's always a good thing to be doing some form of low intensity aerobic exercise anyhow and that's what a bike is for. If you want more torque for better hill climbing, you would increase the battery voltage which then increases both the motor speed and torque. That's why my Bionx 24v system gets a big boost if I insert a 36 volt Li-Mn battery in its place. But it comes to a point that you would be putting so much juice into the motor that couldn't deliver anymore speed and any extra energy left would be totally wasted in the form of heat. So if you are looking for an efficient steady state power assist hill climber, you either go for a geared hub motor or a big Crystallite 5 series motor and head for a decent battery pack of your choice.

In terms of waterproofness, there are 3 items that can be damaged by water. The hub, the controller and throttle control. I don't worry too much about the wirings if it's relatively sealed. If you have a cycle analyst, then that's the 4th possibility.

The problem with the ebike industry is the loosely used term of waterproofness. Yeah it's waterproof, but the problem with this term is that, my expectation of waterproofness can be slightly different than the maker's expectation. When it comes to reliability, I expect the user to be made aware of what their ebike equipment can or can not do. However, because the industry is still in its infancy, we are expected to be treated like guinea pigs for awhile, so expect yourself to ask questions.

When we rate the waterproofness of any enclosures that houses the electronic parts, we rate them by the IEC 529 IPX standard. The maker can opt to have their equipment certified. Garmin certified some of their outdoor GPS units with an IPX rating of 7. Why couldn't Bionx, Crystallite or eZee have the same ratings? What good is the assurance of someone who says that the hubs are really well sealed, the controllers can withstand a downpour from the tap and the controller throttle being at least waterproof when you're the one riding the bike? An ebike is an electronic equipment and to be expect to be ridden in the rain. Unless the ebike industry begins an honest certification of their equipment against water and dust, it is up to the user to make the components at least water resistance.

Here's a link to the IPX standard..

http://www.opticsplanet.net/water-proof.html

 

[Link]

I understand the whole Ohms law thing, where power is the product of voltage and current, and although my motor theory is weak I know voltage=speed and current=torque. Why could you not then have a switch which would change the battery voltage from say 48V to 24V, thereby doubling the amperage and torque while halving the speed? It seems like it would get you up the hill a little better to me, although it would require an unavailable controller as far as I know. It sounds better than a geared motor, and it could even be an automatic change with proper motor feedback. Pulse width modification might fit right into this idea.

I considered this earlier on in my build, but decided it wasn't worth it, since 96V is just a bit too much speed on a Golden Motor (~50mph). I don't think that particular motor could take it, to my dismay. :wink:

Anyway, it's a very doable idea with a few suitable relays (quite similar to using a boost pack). As for the controller, you could just rig up a small voltage regulator circuit that worked over a wide voltage range to feed the logic with what it was made to work with. Upgrading the FETs and caps to handle the increased voltage is already a fairly common practice.

However, doubling the voltage won't double the torque (barring all PWM current multiplication effects, anyway) since there's a shunt in the controller that regulates how much amperage flows through it regardless of voltage. It will give you max amperage for a longer time when accelerating, though, since the BEMF from the motor will take longer to catch up to the battery voltage and start limiting the current by itself.

I'm still considering an automatic thing with a future, more complex twin-motor build that switches between series/parallel batteries AND (brushed) motors, for four different power levels. 8) Not yet sure how exactly to go about it at this point, though, since I'm really not that great with digital ICs (yet). :?

The FAA stipulates a minimum of 8G wire for a max of 45 Amps, what gauge do recommend for the battery to controller and controller to motor connections? I also read that Hall Effect sensors do not function well in extremely wet environments, such as the one I live in, and Anderson connectors, although cheap, are really not made to withstand abuse of this type from my research. Dead shorts are really bad news for batteries and controllers; they tend to fry.

Wire ratings are really confusing, IMO, since what they're rated for will depend on how they're being used. For example, we use "undersized" wiring compared to, say, what's used for a house, and the drag-racing Killacycle uses undersized wiring and connectors compared to what we used to save on weight.

I use 12ga for 30A or under, and 10ga for around 55A max. I wouldn't have a problem putting 90A or so through 8AWG.

Hall sensors don't usually get wet (being inside the motor), but covering them and their legs completely with epoxy or whatever should solve that problem.

I'm not actually sure how well PowerPoles handle wet weather, but they're pretty much the standard around here regardless of where you're at. I haven't heard of any problems with them getting wet, actually.

My advice for shorts is keep the fuse close to the battery. I accidentally shorted out a comparatively weak lithium pack by accidentally touching the exposed two lead wires once. Yeah, dead shorts are very bad.


I don't actually have much experience with waterproofing, since if it rains here in So-Cal we run outside and praise Ceiling Cat. :wink:

 

[pwbset]

I don't actually have much experience with waterproofing, since if it rains here in So-Cal we run outside and praise Ceiling Cat. :wink:

I see what you did there.

(sorry.. it's almost obligatory at this point to respond with that isn't it... at least I didn't rickroll you or hasselhof you... haha.. and prairie drama is sooooo last year.. better update that sig :wink: )

 

[graevus]

So the controller is the current limiting device then... If that is the case you would want the wire gauge a little thicker on the battery side, and a little thinner on the motor side to save weight and minimize voltage drop.

The tricky part about wiring is the voltage/current thing. At high voltages the electricity is more prone to jump or arc, causing a short. At high current the tiny amount of resistance in the copper is enough to generate a lot of heat. From my previous job experience in microsolar system design we recomended 48V solar panel arrays over say 12V systems because to get the same power through the wires the current was 1/4 in the 48V system meaning you save a ton of cash in wire cost. The arcing concern is really a non issue with voltages under 100V; if i recall it will jump about a millimeter for every 2000V depending on humidity etc. 100V probably will short twice as easy as 50V, but the difference is really small. Wires for higher voltages should be better insulated. Signal quality is another strange wiring issue, but unless you want to watch DVD's on your e-bike, or have have e-mail too, you probably wont run into it.

thanks for the answers

 

[Link]

I don't actually have much experience with waterproofing, since if it rains here in So-Cal we run outside and praise Ceiling Cat. :wink:

I see what you did there.

(sorry.. it's almost obligatory at this point to respond with that isn't it... at least I didn't rickroll you or hasselhof you... haha.. and prairie drama is sooooo last year.. better update that sig :wink: )

I know, but I'm running out of memes. :? And I don't want to have to resort to ones specific to 4chan...

So the controller is the current limiting device then... If that is the case you would want the wire gauge a little thicker on the battery side, and a little thinner on the motor side to save weight and minimize voltage drop.

In brushed motors, that would probably be helpful, but there's so little wire to run through (mine goes through all of about 10ft round trip), it's not usually worth getting the extra spool worth. And with brushless motors, there's 50% more wire, anyway.

As for arcing, most shouldn't have a problem. Very few run over 100V, 72V being more typical for higher-powered ebikes. Most wires are good for 300V in worst-case scenarios, anyway. But I would suspect a coating over the power traces on the controller's PCB would be a good idea at higher voltages, and infinitely more so if you ever ride in the rain.

 

[swbluto]

You know what might work good for that is a LED driven phototransistor.

I don't really what that is, but I can take a guess! Does the phototransistor actually control the voltage in some variable manner, though?

The current way the throttle works(in the crystalyte system, anyways) is that the controller supplies the throttle a 5V signal. The throttle itself is just a potentiometer, with a bicycle-appropriate way of dividing the voltage on the fly; the voltage sent back to the controller, after division, determines the current output. 5V(or around there) sent back is "full throttle" and less than some small voltage(let's say less than around .5V) is no throttle.

I was thinking that a "photo-resistor" could replace one of the "resistors" in the potentiometer design, and that appropriate values for the photo-resistor and other resistor would be chosen so that it divides the voltage "accurately" in an expected way. The LED would be there for consistency, as you stated, and that could be driven off by a separate supply of power. I was just thinking about drawing off the 5V of power supplied by the controller and thoroughly waterproofing the connections, but the shorting-possibility definitely makes the "optical isolation" idea more appealing.

Yeah, I'm not an EE, but I've been trying to learn from online resources. I'm trying to master radio theory and am still trying to master the demodulation part of FM radioteletype to send digital data. I've read that the Phase-Lock-Loop is one negative-feedback system for demodulation, but I'm still trying to make the connection from "two different frequencies" to "decoding the 1s and 0s". I've seen a frequency to voltage converter in my project book and it seems apparent how you could turn different frequencies into 1's and 0's from that, but I have no idea how the circuit works.

 

[graevus]

No offense intended with your not quite EE status; you are here learning and thats more than can be said for most engineers ive met...IMO its the drive not the degree which really matters.

Your idea to make a voltage divider using a photo reactive resistor on one side is a good one; it might work well with a little fine tuning. It got me to thinking of a project i did a few years ago with phototransistors. A phototransistor is a device which can be configured to act kind of like a dimmer switch controled by the amount of light which hits the sensor. No light the switch is off, some light and some current flows through the junction. Lot of light and its full power. It only works with low voltages and low currents, strictly analog. Its good stuff as a control system input, thats what i used em for last time. Sorry if i seemed to hijack your idea, i just liked the concept.

Frequency modulation is some hairy stuff eh? It essentially hides a signal inside a much longer wavelength signal. In analog the signal could be sound or TV image, but in digital all it has to be is either a one or zero defined by frequency (or amplitude) with nice crisp corners on the squarewave. I did that once in a lab in 03, the circuit was ugly big on the breadboard. Never did it again.

 

[Link]

The current way the throttle works(in the crystalyte system, anyways) is that the controller supplies the throttle a 5V signal. The throttle itself is just a potentiometer, with a bicycle-appropriate way of dividing the voltage on the fly; the voltage sent back to the controller, after division, determines the current output. 5V(or around there) sent back is "full throttle" and less than some small voltage(let's say less than around .5V) is no throttle.

Er...not usually. You're exactly right about pot throttles, but most are hall sensor types, meaning an analogue hall sensor gets fed a voltage and outputs a signal based on the position of a magnet in the moving part of the throttle.


What's this about a phototransistor throttle? I can see why you'd want a water-resistant throttle, but normal hall-type ones can be made completely waterproof by covering the legs of the hall sensor with epoxy. I used hot glue because I'm cheap, lazy, and it pretty much never rains here . The thing would work under sea water, if I was so inclined to try it...

 

[swbluto]

Er...not usually. You're exactly right about pot throttles, but most are hall sensor types, meaning an analogue hall sensor gets fed a voltage and outputs a signal based on the position of a magnet in the moving part of the throttle.


What's this about a phototransistor throttle? I can see why you'd want a water-resistant throttle, but normal hall-type ones can be made completely waterproof by covering the legs of the hall sensor with epoxy. I used hot glue because I'm cheap, lazy, and it pretty much never rains here . The thing would work under sea water, if I was so inclined to try it...

Oh. I assumed mine was a "regular" potentiometer-type of throttle since it clearly operated off some voltage-dividing circuit and its rotary action would've seem to lend itself well to some resistive-type of pot. Mine was a crystalyte thumb-throttle: Is that hall-based?

I think I'm going to head over to the Trouble-shooting page at ebikes.ca to see what more information I can glean.

 

[Link]

Only certain types of controllers even work with pot throttles (mostly the expensive, high-amp ones like Alltrax, Curtis, and the sort). Plus, for some reason, pot throttles tend to be much more expensive (e.g. Maguras are something like $50 each, IIRC). I believe the C'lyte ones are hall-based. Same for all of the ones I have.

 

[fechter]

There's no reason a photo sensor throttle wouldn't work. I think you could use an opaque tube with a narrow v-shaped slot to vary the amount of light that passes as you twist the throttle. I think I saw one somewhere?

The phased locked loop locks onto the signal and you pick the 1s and 0s off the control voltage signal. If you look at the control voltage with an oscilloscope, you'd see sort of a square wave that follows the modulation changes. High is a 1, low is a 0. You use some kind of comparator or threshold detector to trigger the digital output.

 

[safe]

I understand the whole Ohms law thing, where power is the product of voltage and current, and although my motor theory is weak I know voltage=speed and current=torque. Why could you not then have a switch which would change the battery voltage from say 48V to 24V, thereby doubling the amperage and torque while halving the speed?

You are thinking the exact same idea I had when I first got here too. Let me explain what I learned...

The answer is that it's already happening.

What? Yes, it's true, and the reason is due to the bizarre way that a PWM (pulse width modulation) controller works. Because of the weird way that electricity ebbs and flows you get this thing called "inductance". With "inductance" you get varying effectiveness of current flow through a motor depending on how "loaded up" the wire is. It's kind of like driving on a freeway with open traffic verses trying to drive the same route during rush hour... crowded current flows less energetically because there is a negative pressure. (traffic)

Anyway...

For the most part the PWM creates power using the formula:

Power = Voltage * Current

Fictional Numbers: 100 = 10 * 10

...and this means that with a low duty cycle (in other words the pulses are "on" only 50% of the time) you will get:

Fictional Numbers: 100 = (10 * 0.50 ) * (10 * 2)

...so if you've followed me on this the current actually MULTIPLIES because everything needs to balance.

It's been called "current multiplication" though technically that's a lousy name. :wink:

The answer is that current already increases at lowered voltage.

But there is one last thing and that is that the motor itself has limits to how much current it can use for a given voltage. So you are somewhat boxed in for a given gear ratio. To really break out of all this you need to downshift the motor into a lower gear or expand the current limit (on the battery side) so that more power can be delivered. Many people here use hub motors that cannot be downshifted, but can be overvolted which effectively delivers the power you need.

 

[TylerDurden]

Because of the weird way that electricity ebbs and flows you get this thing called "inductance". With "inductance" you get varying effectiveness of current flow through a motor depending on how "loaded up" the wire is. It's kind of like driving on a freeway with open traffic verses trying to drive the same route during rush hour... crowded current flows less energetically because there is a negative pressure. (traffic)

Good thing safe's here to explain inductance to an EE. :lol:

 

[docnjoj]

Because of the weird way that electricity ebbs and flows you get this thing called "inductance". With "inductance" you get varying effectiveness of current flow through a motor depending on how "loaded up" the wire is. It's kind of like driving on a freeway with open traffic verses trying to drive the same route during rush hour... crowded current flows less energetically because there is a negative pressure. (traffic)

Good thing safe's here to explain inductance to an EE. :lol:

Now I ve seen it all!
otherDoc

 

[Link]

Because of the weird way that electricity ebbs and flows you get this thing called "inductance". With "inductance" you get varying effectiveness of current flow through a motor depending on how "loaded up" the wire is. It's kind of like driving on a freeway with open traffic verses trying to drive the same route during rush hour... crowded current flows less energetically because there is a negative pressure. (traffic)

Good thing safe's here to explain inductance to an EE. :lol:

Now I ve seen it all!
otherDoc

I got a minor LOL out of it. :wink:

 

[graevus]

Inductance? Whats that??? Kind of like a radiator? it radiates things...I find you sound the most proffesional when you use words like things and stuff. If they understand what you're saying, you aren't saying it right (typical EE thought process).

The phase angle between the sinusoidal waveforms was pi radians, so of course it caused deconstructive criticism; the output was zero! hehhehheh.

Just kidding Safe, if someone reading it didn't understand the concept, its all good guy. I knew the question was sort of basic, and i figured if it was that easy someone would have done it already. I did not however know why exactly it would not work. Thanks

I do think there is probably room for improvement on controllers though. PWM is great stuff for resistive devices, but my thinking is that there might be a better way to run an inductive (thanks safe) DC motor. Everything with windings is inductive.

Ever hear of an error amplifier? I witnessed one in creation a long time ago. It took a 5v motor and made it so i could not stop it with my fingers. Without the extra circut it had hardly any torque, and it was automatic. Strictly analog stuff, no micro involved and not too many parts either. Been meaning to look into that again. The circuit came out of a book called the op amp bible, although the guy who built it had 20+ years experience in EE. Maybe someone already though of this though..

Now all i need is time and money. ;(

 

[swbluto]

Inductance? Whats that??? Kind of like a radiator? it radiates things...I find you sound the most proffesional when you use words like things and stuff. If they understand what you're saying, you aren't saying it right (typical EE thought process).

The phase angle between the sinusoidal waveforms was pi radians, so of course it caused deconstructive criticism; the output was zero! hehhehheh.

Just kidding Safe, if someone reading it didn't understand the concept, its all good guy. I knew the question was sort of basic, and i figured if it was that easy someone would have done it already. I did not however know why exactly it would not work. Thanks

I do think there is probably room for improvement on controllers though. PWM is great stuff for resistive devices, but my thinking is that there might be a better way to run an inductive (thanks safe) DC motor. Everything with windings is inductive.

Ever hear of an error amplifier? I witnessed one in creation a long time ago. It took a 5v motor and made it so i could not stop it with my fingers. Without the extra circut it had hardly any torque, and it was automatic. Strictly analog stuff, no micro involved and not too many parts either. Been meaning to look into that again. The circuit came out of a book called the op amp bible, although the guy who built it had 20+ years experience in EE. Maybe someone already though of this though..

Now all i need is time and money. ;(

Cool! I was just recently thinking about the general action of the Phase Locked loop and how it sends a voltage signal proportional to the error and I was wondering if there'd be someway of making an electric motor that was controlled in such a way that'd it'd be, nearly, "constant" speed no matter the incline - so a sort of "constant speed source". :lol:

It'd clearly have to increase current somehow to increase the torque necessary, but I have no idea how you'd increase the current such that it could potentially approach infinity(or some really large value) with a fixed voltage source and a ... motor that has little variance in the "on demand" resistance(Or does it?).

I was thinking a mechanical way would be to adjust gearing, but a 300 watt motor just simply won't climb an incline that requires more than 300 watts due to competing "gravitational-based-normal acceleration" and "motor acceleration". I guess it might be possible with a motor that has a huge power range for all practical purposes(I.e., no 40 degree inclines), although I don't know much about motors beyond the one I have.