Jump to content

Firewheel custom battery pack


esaj

Recommended Posts

Great thread! @esaj best summary I've see of how eWheel motors are controlled. One of the functional aspects I've been uncertain about, is while we have a fairly good understanding of the way PWM pulses input power during commutation, how does the MOSFET bridge work in the reverse direction? With the MOSFETs bridge the voltage is discrete but unvarying (at around 60v), where the length of the pulse simulates desired lower one, dependent on the riders speed; but suppose the rider brakes at 5kph, then the voltage potential of the back EMF is far below the required 67.2v to charge back to the battery pack & it's difficult to see how different length PWM pulses from the back EMF, with any reverse current diode configuration would step-up this voltage.     

Link to comment
Share on other sites

  • Replies 125
  • Created
  • Last Reply

While breaking the EUC does not only take the generated EMF and charges the battery (consume the power). By the bridge they coils of the stator still get powerd, but with a lower frequency as the wheel is turning, so the generated magnetic field from the coils is turning a little bit slower then the wheel itself with the permanent magnets. That is just the opposite technique of accelerating the wheel, where the magnetic field is produced turning a little bit faster then the wheel itself.

Everytime the wheel overtakes the magnetic field (turns faster) the motor goes in the "generator" mode and "produces" EMF. If the EMF is small (low difference in speed and low absolute speed) there is still more power needed to produce the magnetic field with the coils than the motor is generating - so the batteries get discharged.

Just when the EMF is bigger than the energy needed to produce the magnetic field, the batteries are charged.

To control the motor you also have to "types" of pulses. The one main pulse comes everytime with a phase shift of 120° to the three motor wires for the coils (which are placed geometricely with 120° from one to another, too). This generates the turning magnetic field. The frequency of these pulses determine how fast the magnetic field is revolving.

The second "type" of pulses is introduced by the PWM. These first pulses are chopped - the duty cycle of the PWM pulses controll the strength/force of the magnetic field.

So for the Mosfet bridge there is absolutely no difference between the motor and the generator mode.

In one of the articles a couple of post before (about the BLDC controller and the included Break Controller) they describe a break controller, which coult be used to save a fully charged battery from overloading in generator mode. Unfortionately this seems not to be implemented in any EUC so far... But this would be beside the MOSFET Bridge.

The reverse current diodes can just be used to protect the circuit from the reverse current, which comes from the coils when you take them from the voltage. (the topic is explained in detail in the youtube video "active free-wheeling" a couple of posts before)

Link to comment
Share on other sites

But the sinusoidal waveform is just a characteristic of the PWM pulse length right? So there's really only one pulse type but three phases? I still don't understand how the voltage is stepped up to ~67v in regen mode. 

There has to be some feedback logic that's feed from the motor input three phase power wires.   

SinewavePWM.jpg 

Link to comment
Share on other sites

But the sinusoidal waveform is just a characteristic of the PWM pulse length right? So there's really only one pulse type but three phases? I still don't understand how the voltage is stepped up to ~67v in regen mode. 

From what I've understood, it doesn't have to be 67.xV to charge the battery (unless the battery is totally full at 67.x), the charging will always occur as long as the back-EMF voltage of the motor is above the battery voltage during regenerative braking. My voltage meter in the Firewheel also seems to support this, as it goes above the "coasting" (riding at steady speed) voltage when braking and below it when accelerating.

However, if the battery voltage is above the back-emf when braking, there must be something else going on, like changing the PWM-timing to energize the coils so that they try to drive the motor backwards in relation to current direction (using battery power for braking), as from what I've understood, otherwise the motor would start to accelerate again (until reaching the same voltage and coasting)... I suppose that power braking also works by trying to drive the motor in reverse.

Link to comment
Share on other sites

...

However, if the battery voltage is above the back-emf when braking, there must be something else going on, like changing the PWM-timing to energize the coils so that they try to drive the motor backwards in relation to current direction (using battery power for braking), as from what I've understood, otherwise the motor would start to accelerate again (until reaching the same voltage and coasting)... I suppose that power braking also works by trying to drive the motor in reverse.

The coils are always energized while driving - they are always producing a rotating magnetic field. For accelerating the magnetic field rotates a bit faster than the wheel, for breaking it rotates slower than the wheel. Imho that's (more or less) without regard to Back-EMF or not.

Trying to drive the motor in reverse imho wont work. If you have the three motor wires A B and C then for one direction your give the pulses (120° phase shifted) first to A then B and then C. If the motor goes in the other direction the pulse go first to A then to C and then to B. If you change this pulsing scheme while one rides the wheel i assume there won't be a nice outcome ;)

Link to comment
Share on other sites

The coils are always energized while driving - they are always producing a rotating magnetic field. For accelerating the magnetic field rotates a bit faster than the wheel, for breaking it rotates slower than the wheel. Imho that's (more or less) without regard to Back-EMF or not.

Trying to drive the motor in reverse imho wont work. If you have the three motor wires A B and C then for one direction your give the pulses (120° phase shifted) first to A then B and then C. If the motor goes in the other direction the pulse go first to A then to C and then to B. If you change this pulsing scheme while one rides the wheel i assume there won't be a nice outcome ;)

I have no doubt that you understand it a lot better than me  :), but maybe I explained badly, and "reverse" and "backwards" were probably poor choices for words (or I have simply misunderstood the entire concept :P). I meant that the timing (phase-shift) of the PWM-pulses is changed so that the "pull" between the rotor magnet and the coil happen in the opposite direction than it's running, not that the pulses are run in reverse... so if the circuit for the coil is "closed" before the magnet in rotor is in the same position (using the hall-sensor signatures) and that causes the rotor to be attracted towards the coil, in "reverse" it would be "closed" after the rotating part has passed, trying to pull it in the opposite direction? Does that make sense? I still can't draw worth shit, but maybe this will help:

Vyr6Ahx.png

So normally accelerating the motor, the winding will pull the rotor magnet towards itself before it's in the same position, accelerating the motion, whereas braking it will pull the magnet towards itself after it has passed in (in regards to the direction of the turning rotor)? And the current direction is dependent on which part (battery or the motor) has higher voltage at that moment (potential difference), which in turn is affected by the current rotational speed of the motor for back-emf and battery charge/voltage sag/whatever for battery? In "reverse"/"backwards", the attraction will occur on the "other side" of the coil, and the motors' turning direction will gradually be turned to the opposite direction, after it has first slowed down to stop.

I'm having problems wrapping my head around this concept right now, as I'm simultaneously trying to build UI in Android... :D

Link to comment
Share on other sites

Esaj, You have been so far around, that I've lost track of the actual reason for why the modified Wheel no longer handles going Down the hill. Could You wrap that up please?

Like:

1) It is all Down to the 'mystery' wire, coming from the BMS, that is no longer attached in the proper place.

2) The new BMS's are completely different from the default one and doesn't cope well with the control-board and regenerative charging.

3)  The new BMS's has not been shorted on the MOSFET's, since it will violate the guarantee.

4) You have no idea.

5) ?

Regs

Link to comment
Share on other sites

I have no doubt that you understand it a lot better than me  :),

I am not 100% sure - but it seems we both give best to get a thourough understanding of the fundamentals..

but maybe I explained badly, and "reverse" and "backwards" were probably poor choices for words (or I have simply misunderstood the entire concept :P). I meant that the timing (phase-shift) of the PWM-pulses

As far as i understood (nicely to be seen in the "active free-wheeling" youtube video The frequency shift happens in the "main" puls. The pwm pulses are just "chopping" this main pulse and determine the "strength" of the magnetic field.

 is changed so that the "pull" between the rotor magnet and the coil happen in the opposite direction than it's running, not that the pulses are run in reverse... so if the circuit for the coil is "closed" before the magnet in rotor is in the same position (using the hall-sensor signatures) and that causes the rotor to be attracted towards the coil, in "reverse" it would be "closed" after the rotating part has passed, trying to pull it in the opposite direction? Does that make sense? I still can't draw worth shit, but maybe this will help:

...

So normally accelerating the motor, the winding will pull the rotor magnet towards itself before it's in the same position, accelerating the motion, whereas braking it will pull the magnet towards itself after it has passed in (in regards to the direction of the turning rotor)?

Exactly. That's my point of understanding, too (at the moment ;) )

And the current direction is dependent on which part (battery or the motor) has higher voltage at that moment (potential difference), which in turn is affected by the current rotational speed of the motor for back-emf and battery charge/voltage sag/whatever for battery?

With the charging of the battery i still have a little understanding problem. If the Voltage from the motor is higher than the battery voltage the upper MOSFET of the bridge will be operated in reverse - the current will flow through the "Inversdiode/Body-Diode". Is maybe described in detail in the datasheet/application note of the bldc driver i have linked - need to look at it in detail, sometime.

And maybe sometime i'll dig into the EMF / generator topic too a little bit deeper. Imho the moving permanent magnet should induce a voltage in the coils everytime. But in the literature the distinguish between the "Motor mode" (the wheel gets accelerated) and the "generator mode" (braking the wheel). Maybe thats just because the possibility for usable Back-EMF can only happen while breaking (would make sense, because than kinetic energy from the wheel is consumed by the EUC System. Whilst accelerating the EUC System uses battery to generate kinetic energy) - would be a nice roundup to know the theoretical reasons for this, too.

Link to comment
Share on other sites

Esaj, You have been so far around, that I've lost track of the actual reason for why the modified Wheel no longer handles going Down the hill. Could You wrap that up please?

Like:

1) It is all Down to the 'mystery' wire, coming from the BMS, that is no longer attached in the proper place.

2) The new BMS's are completely different from the default one and doesn't cope well with the control-board and regenerative charging.

3)  The new BMS's has not been shorted on the MOSFET's, since it will violate the guarantee.

4) You have no idea.

5) ?

Regs

I now exactly, thanks to 1Rad Werkstatt, but was again asked not to tell the reason publicly. :mellow:  I also now exactly how to fix it in my case, but there's no really fast solution (unless I throw a lot of cash for a faster solution, but am reluctant to do so ;)), so I probably won't do it until over winter.

 

As far as i understood (nicely to be seen in the "active free-wheeling" youtube video The frequency shift happens in the "main" puls. The pwm pulses are just "chopping" this main pulse and determine the "strength" of the magnetic field.

Couldn't you just use a constant DC voltage though? As far as I know, the pulses could be created simply by the PWM, too, as that's how it worked in my fan-driver, although those were just single mosfet per fan, but I used PWM to "chop" separate 12V line to 6-12V depending on wanted speed, and they were "just" normal 12V DC-computer/case/equipment fans with no directional control. And the course was more about embedded programming than electronics anyway, so we never dug more deeper on how the motor actually works, just that the speed can be controlled by adjusting the voltage.

With the charging of the battery i still have a little understanding problem. If the Voltage from the motor is higher than the battery voltage the upper MOSFET of the bridge will be operated in reverse - the current will flow through the "Inversdiode/Body-Diode". Is maybe described in detail in the datasheet/application note of the bldc driver i have linked - need to look at it in detail, sometime.

Hard to say anything about this, I did take a look at those both datasheet, the motor controller and the break controller, note that it really says BREAK controller on the sheets, not BRAKE controller. From what I understood, the break controller cuts the DC-bus by shorting it over a resistor with a (PWM-driven?) switch (it was drawn as a switch in the schematic, not a mosfet), and from what I understood, it does not really handle the "braking" of the motor by itself, but is used to control the regenerative voltage by doing "dynamic" braking to "chop" the voltage coming from the motor? Both datasheets had little to none information on regenerative braking though, although they do explain the PWM-driving of the motor with hall-sensors and the switching patterns.

The examples (mostly with one and two halfbridges) I've seen have shown the current flowing "wrong way" through the mosfet, but never mention how the diodes affect this, there must be of course more to it... maybe the mosfet datasheet has some hints? This is the Firewheel power-mosfet (it's surface-mounted, unlike in most wheels):  http://cdn-reichelt.de/documents/datenblatt/A100/DS_IRF7759.pdf

And maybe sometime i'll dig into the EMF / generator topic too a little bit deeper. Imho the moving permanent magnet should induce a voltage in the coils everytime. But in the literature the distinguish between the "Motor mode" (the wheel gets accelerated) and the "generator mode" (braking the wheel). Maybe thats just because the possibility for usable Back-EMF can only happen while breaking (would make sense, because than kinetic energy from the wheel is consumed by the EUC System. Whilst accelerating the EUC System uses battery to generate kinetic energy) - would be a nice roundup to know the theoretical reasons for this, too.

Many sources have stated that the back-emf voltage is always there when the motor turns, so it seems it is always acting as a generator, even during "motor mode", but the current direction it's just a matter of voltage difference between supply (battery) and the back-emf? Most sources have explained that the back-emf (voltage) of the motor is "pumped" up by disconnecting the motor from the battery (not sure anymore if it's in free-wheel mode or totally disconnected then ;)), so the voltage goes up, and then when the circuit towards the battery is again closed, it "shoots" off some charge into the battery. And this is then repeated fast during the braking.

 

Link to comment
Share on other sites

 

I now exactly, thanks to 1Rad Werkstatt, but was again asked not to tell the reason publicly. :mellow:  I also now exactly how to fix it in my case, but there's no really fast solution (unless I throw a lot of cash for a faster solution, but am reluctant to do so ;)), so I probably won't do it until over winter.

 

 

My guess is, that the really fast solution is to buy complete batterypacks from 1Rad. Probably highest quality, but not exactly cheap. :unsure:

Well, I hope You find a cheaper solution.

Link to comment
Share on other sites

My guess is, that the really fast solution is to buy complete batterypacks from 1Rad. Probably highest quality, but not exactly cheap. :unsure:

Well, I hope You find a cheaper solution.

"Now" = "know" in my earlier post, just noticed seeing you quote me... :D

Anyway, yeah, that would be one (and the most expensive) solution, the slower & cheaper options require some parts to be ordered, and shipping stuff back and forth, and that of course takes time (I haven't even ordered the parts yet), faster shipping is possible, but of course more expensive. I have the fourth pack now also, but haven't installed it into the wheel yet (as that requires yet another tear down), and should help at least somewhat, as there are more packs to soak up the regenerative braking charge... also gives more range & power even with partially charged packs. But I need to work on the app now also to get vee's MCM2s back to him, I'm already late on my schedule with that  ;) 

Link to comment
Share on other sites

Since i did in my breaks inbetween some additional reading in different papers just a little addition/corrections to my previous postings: With the three H Bridges (Mosfets) and the way the coils are conntected by the three motor wires the motor controller can shorten the coils for "Emergency Braking" (with PWM Pulses so the MOSFETS, Coils and connectors will not go up in smoke) and implement Plugging Breaking (reversing the supply voltage of the coils) for breaking. Also in one article a boost up converter was mentioned, which should be easy to implement with the 3 H bridges, so that regenerative breaking is more efficient with low rpms (too low back emf). But i assume for real efficient and battery safe regenerative breaking a DC-DC converter is needed.

The "normal" dynamic breaking (burning the power with some resistors) would need a few additional components. As shown in the data logs the amount of energy is quite much (1-3kW is a normal hot plate!) - but should be possible for quite some seconds to warn the driver and let him stop the e-wheel and/or go uphill again to use up some battery.

Link to comment
Share on other sites

The "normal" dynamic breaking (burning the power with some resistors) would need a few additional components. As shown in the data logs the amount of energy is quite much (1-3kW is a normal hot plate!) - but should be possible for quite some seconds to warn the driver and let him stop the e-wheel and/or go uphill again to use up some battery.

I've only got that high momentary values during hard braking and going down a really steep hill (25 degrees), going down on about 10 degree hill while braking to keep the speed fairly steady, the average regenerative power was around 300-400W (both with the Gotway MCM2s).

Link to comment
Share on other sites

I hope that until my current EUC has blown up the batteries ;) new e-wheels come on the market, which have some high power resistors, use them in a clever combination for plugging-dynamic breaking and have a nice dc-dc converter for regenerative breaking included.

Best with a nice really huge accu pack and low weight :-)

Link to comment
Share on other sites

On the other hand, if a 3kW hotplate was installed on top of your wheel, think of the time savings, you could cook your breakfast while riding to work (Assuming there's enough downhills or braking required on the way) ;)

Link to comment
Share on other sites

Unfortionately to work i just go down ~50-60 Meters the first 6 km. Don't know if this is enough for hot coffee, bacon and scrambled eggs  ;(

Going back home is much better - there i can use 100 m height difference on the first 2,5 km. Should be enough for a good amount of Glühwein (?hot wine punch?) to arrive at home with a nice smile even in wintertime ;)

And once its raining the energy could be waisted with an https://www.kickstarter.com/projects/1243275397/air-umbrella (was imho already posted here once, long time ago...) and some heat guns in wintertime to get rid of the snow and keep the driver cosy and warm :)

Link to comment
Share on other sites

Just went out to test (finally) the battery-warning voltage, first rode it around the hiking paths to get the batteries more empty, and then up and down the hill for a good 20 laps and then used a more steeper gravel-hill, before I got it to trigger. It's pretty near 47V or little below, if my voltage-meter can be trusted, the lowest value I saw on the steep hill was 47.0V, a little before it triggered on a bit steeper bump. What makes it "funny" (or scary, you decide) is that the BMS datasheet says the overdisharge cutoff voltage is 2.9V +- 0.05V per cell, so for 16S-pack, that's 45.6...47.2V. Don't know about the tolerance, if it can change due to temperature or if there's just some variance in the components, but it's cutting pretty close :P  On very empty batteries, a stronger acceleration or hill climb could cause a cut-out instead of a battery warning, especially now that it's getting colder.

On the other hand, I did get the battery-warning twice on my very first try right after I had put the wheel back together, so it could be on the lower side, and probably having multiple packs helps keeping the sag "slower", ie. it won't hit really low values that easy.

Link to comment
Share on other sites

@esaj, with your pack, how much range do you get between he typical low-voltage value of 54v & 47v?

It only hit 47V momentarily during very steep rise (voltage sag), once stationary the voltage resumes to around 54V. It seems the sag becomes much "deeper" the lower the battery is. From around 57.5V to around 54V I rode about 10km, this is still with the 3 packs only, I still haven't installed the 4th pack.

The battery-situation will be fixed over the winter, as it looks like I won't be getting the parts any time soon.

Link to comment
Share on other sites

Now that my flu has mostly gone, I finally got around to take the wheel apart again for adding the last battery pack. The shells had damaged a bit more than what could be seen from the outside. From the outside I could see that one of the screw-posts had come off (the posts have been clearly attached afterwards, and not moulded to the original shell, as it had came off so cleanly), but also two of the posts had cracked in the inside, and the screw from the other has flown off completely. Have to figure a way to replace them, I'd prefer through-bolting to screws on plastic threads... 

Soldered the connectors to the last battery pack, and then forgot it on the charger for a bit too long, went up to 62.8V, whereas the rest of the packs are currently sitting at 60.1V. Either I have to charge the rest closer to the same voltage (as I have nothing to use as load to discharge the 4th pack), or risk some sparking and rushing currents if I just connect them to each other.

I have made some sketches for custom shells (I won't be using the original shells for the mould, as I want to place components differently), but building the shape to make the negative mould from is going to take some time. Also I'll probably combine this project with building a flat-bar steel frame that's attached to the pedal bars (or whatever they're called, the metal parts that come down from the axle and have the pedals attached to them). Don't have the materials yet, but already know where I can order the carbon- and carbon-/aramid-fiber fabrics, release waxes/sprays, gelcoats and high quality epoxy resins for laminating. I remembered I have an old 400m3/h air blower & activated carbon scrubbers in the storage, so working indoors shouldn't be much of a problem (the house also has forced ventilation for both intake and exhaust). Also I can use the top of the fireplace, sauna or baking oven for heat treating, should the resin need that (some do need to be kept at higher temperatures for some time to attain the maximum strength after the initial cure). Not sure when I get around to work on that, preferably I would have just got the packs fixed to prevent the overcharge-problem, but it looks like I'm not going to be getting the needed parts for a while, and the weather report says that the thermal autumn should start next week.

Link to comment
Share on other sites

  • 1 month later...

Minor update: Finally got around to check the parts I ordered (I've actually had them for some weeks already, but been busy with the frame-project and lately with other completely non-unicycle stuff), and... I'll have to wait for at least one replacement before shipping everything forward for the final work. I could do it myself, but it would again void the warranty on the packs, which I'm reluctant to lose after all the money I've poured on them, with my luck, they'll explode or die right at that moment. Looks like I'll have safe packs by Christmas, hopefully... :P At least they aren't totally useless even in their current condition (as long as I don't charge them too much  ;)), so I can keep on working with the custom frame-project, once I get around to that again, been otherwise occupied lately.

Vee was very kind and sent me a replacement tire, haven't actually even unwrapped it yet, but it could maybe be used as a winter tire if I get studs for it. Also on the lookout for battery warming systems, but haven't ordered any parts yet.

Link to comment
Share on other sites

  • 6 months later...

Got the battery packs with the new BMSs, so time to start building connectors for those. Easy job, right? Wrong.

I don't know how, but it took me about 4 hours to make this. Maybe I stepped through a wormhole at some point...

PjpHBeG.jpg

Starting simple, soldering a small piece of 6mm2 (pretty much AWG 9) to the EC5 connector.

AL1jJlE.jpg

Ok, this was a bit tricky, I need to connect the EC5 to four EC3's for the four packs. So I soldered four 2.5mm2 -wires to the 6mm2 -wire. In hindsight, probably should have used a longer piece of 6mm2  -wire and then just split it evenly into four 1.5mm2 -wires (which would have been around AWG15 each, which should have been enough, if you consider the battery cables are usually AWG16)... 

wR6CP0u.jpg

The EC3's I got already had the metallic connectors snapped into the plastic housing. So I added put together both sides of the connector, so the housing wouldn't melt (at least as easily) and the connectors should stay in place.

e93PrzZ.jpg

Soldering "cup" -connectors is relatively easy, although requires a lot of heat so the entire connector is hot enough for the solder to flow properly, which can damage the plastic housing. Heating and partially "filling" the cup and then sticking the pre-tinned wire there usually seems to work pretty nicely. Apply heat for a while after, so the wire gets good attachment, and then keep it in place for a while afterwards while the tin cools down. Do notice that the wire can get pretty hot, like seen above, the shrink wrap higher up the wire has started to shrink, and I needed to cut it down later on ;)

HlugRq8.jpg

Here the positive side is also done, just need to shrink wrap that ugly connection between the  2.5mm2 -wires and the 6mm-wire.

BNmZaf3.jpg

Done! Not pretty, but should work. Now I just pop off the other ends of the connectors, so I can solder them to the actual battery discharge wires.

a2JOyVu.jpg

For fucks sake!!! One of the connectors was upside down... :angry:  So NOT going to do that again / start replacing it! I'll just put a male-connector one of the packs, it's shrouded anyway... :P

 

Link to comment
Share on other sites

Ahh I hate that when it happens.  At least it's pretty minor.  I've had to disassemble an Xbox 360 multiple times after trying to troubleshoot the DVD reader.   The original laser was still working intermittantly so I tried to tweak it to avoid replacing parts.

It ended up to be hopeless so and just a $5 eBay part to swap out.  I spent so much time trying to adjust the alignment distances and play around with it I think I took it apart and reassembled it 10 times.  I could have saved all the trouble by putting in the new $5 part.  It's all a learning process though.  What doesn't kill you makes you smarter.  :D

Link to comment
Share on other sites

Archived

This topic is now archived and is closed to further replies.

×
×
  • Create New...