"peak power" - Trek 4300 20" Magic Pie build.

[o00scorpion00o]

How come the Mac overheat at 20 or 25 mph average? I thought it can handle a lot of watts? Is it the mac 350W?

Maybe you just put too many amps on it? You can still have high speed but with less torque and less amps.

The Mac is a monster of a motor and will climb anything. The problem is it's inability to shed heat quickly.

I have the 320 rpm mac and I have ran 3500 watts through it, damaged my clutch, but it's such fun!

I've never had a problem with heat, Eve at 35-40 mph runs, but long steep hill climbs could kill it unless you back off the power, even at 600 watts you would be surprised how well the mac climbs!

I run 60 volts nominal (16s LiPo) through it, so helping to keep the current lower, that's about 58 amps on acceleration, and about 35 amps to cruise at 35-40 mph on level ground.

Now go to 48-50 volts and that 3500 watts on acceleration becomes 70 amps, and already that's a significant increase in current. Interestingly enough, cruise speed will be less because your top speed will now be 30-31 mph.

So by using 10 less volts your current increases by about 12 amps and that can be enough to get phase cables too hot, especially on hills, so the more voltage you got the, the less amps you pull and the faster you get up the hill taking less time to get things hot.

[Alan B]

Changing battery voltage won't change the motor phase current for the same motor operating conditions. Motor operating conditions (load, rpm, temperature) determine motor voltage and phase current. Battery voltage gets converted to this by controller. Changing battery voltage will change battery current to feed the same power. But if motor is operating at the same speed and load the motor voltages and currents will not be different.

The problem in climbing this big hill is power dissipation. Gearmotors can dissipate about half what direct drive motors can dissipate. Gearmotors are operating at a more efficient RPM but it becomes a trading game of efficiency versus dissipation. In the various modelling I did the only way to win was dual motors to double the dissipation. Boring holes and forcing air through the motor could be important also. Avoiding hard acceleration and other heat drivers can also help. Liquid cooling would really change the game. Then one small motor could do the job. Put coolant through the frame tubes for a distributed radiator.

[neptronix]

Nailed it..

And yep, liquid cooling would change the game, big time. Hot copper = high resistance = wasting more power and/or getting less and less torque out of the motor as she heats up.. the efficiency level is going to be terrible and copper that is gradually getting hotter sure isn't going to help things.

For everyone's reference, here is the thread where i decided i was going to take this task on. After a few pages, i settle on the magic pie motor per Luke's suggestion & also from watching oooscorpionooo dump 5-6kw into his, doing a fairly major hill climb, then the worst thing that happened to him was that the halls busted.. lol.

http://www.endless-sphere.com/forums/viewtopic.php?f=3&t=31490&hilit=pikes+peak+magicpie&start=75

[o00scorpion00o]

Nailed it..

And yep, liquid cooling would change the game, big time. Hot copper = high resistance = wasting more power and/or getting less and less torque out of the motor as she heats up.. the efficiency level is going to be terrible and copper that is gradually getting hotter sure isn't going to help things.

For everyone's reference, here is the thread where i decided i was going to take this task on. After a few pages, i settle on the magic pie motor per Luke's suggestion & also from watching oooscorpionooo dump 5-6kw into his, doing a fairly major hill climb, then the worst thing that happened to him was that the halls busted.. lol.

http://www.endless-sphere.com/forums/viewtopic.php?f=3&t=31490&hilit=pikes+peak+magicpie&start=75

HAHA funny!

Don't give up Dave, not at least until you try it, that motor will eat hills like the mac, though not nearly as fast, but I can only imagine the torque in a 20"

But I would go the high voltage less amps route and let speed take you up, giving less time for heat, just be careful with the regen, I would not use it again after a long climb! That pie takes a very, very long time to cool off!!!

I rember smelling it in the back of the car all the way home and my Girlfriend saying "how much is that gone up in smoke ?"
It's still getting me around in sensorless, but it's definitely lacking without halls!

Just watch the heat and you know yourself when it's too hot, then it's time to drill covers.

How much it bogs down is another matter that will depend on how much it heats and that I can't tell you in a 20" wheel, in a 26" wheel at 13 mph on 60 volts, 3500 watts 10-15 mins is the max run time on 16% grade, and frock me is that spooky when the front wheel keeps flipping and you risk falling over the edge, but hey it's all part of the E-Bike fun!!! yeah

[o00scorpion00o]

Changing battery voltage won't change the motor phase current for the same motor operating conditions. Motor operating conditions (load, rpm, temperature) determine motor voltage and phase current. Battery voltage gets converted to this by controller. Changing battery voltage will change battery current to feed the same power. But if motor is operating at the same speed and load the motor voltages and currents will not be different.

The problem in climbing this big hill is power dissipation. Gearmotors can dissipate about half what direct drive motors can dissipate. Gearmotors are operating at a more efficient RPM but it becomes a trading game of efficiency versus dissipation. In the various modelling I did the only way to win was dual motors to double the dissipation. Boring holes and forcing air through the motor could be important also. Avoiding hard acceleration and other heat drivers can also help. Liquid cooling would really change the game. Then one small motor could do the job. Put coolant through the frame tubes for a distributed radiator.

Hi Alan,

I see what you mean, but the thing I noticed is that for a lesser grade of hill a 48 volt battery will get you up fine, but give it another few % and the motor slows down enough to be pulling a lot of amps and the motor is not nearly efficient creating a lot of heat. Volt it up and it spins faster more in the efficiency zone pulling less amps, making less heat and getting you up faster not allowing more time for heat. Does that make sense ? And I'm not nearly an expert, just my own experience.

[Alan B]

Changing the operating point does change things. Clearly the motor can be overheated by going too slow, and totally destroyed by applying too much power. So in between there must be a speed which minimizes net heating. Finding that speed is the difficulty, and then you need enough voltage and current to achieve that speed. Even at this optimal speed the motor may or may not melt before reaching the top.

It would be interesting if we could equip the bike with enough instrumentation to immediately read out the power input to the motor (easy) and the power output (not easy). Then we could immediately tell if this is sustainable (providing we already had some data for dissipation). Measuring motor temperature is an easier way to tell, but the data rate is very low so it might take many runs to determine the optimal speed. Essentially if the motor temperature is rising above some level then it is probably not sustainable for a long climb.

[o00scorpion00o]

What we need is water cooled motors, but I think that will be a long time.....if ever!

Even water cooling looses efficiency because of the pumping of coolant, it would save the motor for sure but wouldn't that still mean inefficiency, because heat means inefficiency, so water cooling only transfers the heat to the coolant? meaning transferring the inefficiency from the motor to the coolant ?

I got sooooo much to learn about motors!

[Alan B]

Operating the motor at lower temperatures improves efficiency by reducing copper losses.

Let's say the motor is 75% efficient and is doing 1500W of work climbing the hill, so we are putting in 2000W and it is dissipating 500W. If this is a gearmotor that can only dissipate 400W it will heat up and melt.

If we put in liquid cooling and bring that 500W of heat out of the motor into a radiator we can avoid melting the motor.

So cooling helps efficiency and allows survival.

I guess we don't want to put holes in a gearmotor for cooling, but holes in a DD motor can help. Perhaps even misting in some cooling water would be worthwhile. They hand out water on the way up Pike's Peak during the race, wonder if they would mind if you wanted a couple extra bottles. Then you wouldn't have to carry all that water from the bottom.

[neptronix]

It would be interesting if we could equip the bike with enough instrumentation to immediately read out the power input to the motor (easy) and the power output (not easy). Then we could immediately tell if this sustainable (providing we already had some data for dissipation). Measuring motor temperature is an easier way to tell, but the data rate is very low so it might take many runs to determine the optimal speed. Essentially if the motor temperature is rising above some level then it is probably not sustainable for a long climb.

I was just thinking about that today on my walk. I need a temp sensor, for sure.
BBQ thermometer with an extended wire up to the handlebars

Or can someone suggest a more elegant solution?

Also, it is highly likely that the covers will be drilled.
A misting system is a good idea. I could hide the tank in the rear topeak rack. Only have to fit 25AH/12S into the rear.. i could fit a few computer duster sized cans of something in there.

[veloman]

What size are your phase wires Nep? Did you upgrade them? The ones on my older Mac look really thin, maybe 16ga. Do you know what battery current is the limit for those? (I think phase currents will vary depending on winding and hub speed too though).

[neptronix]

What size are your phase wires Nep? Did you upgrade them? The ones on my older Mac look really thin, maybe 16ga. Do you know what battery current is the limit for those? (I think phase currents will vary depending on winding and hub speed too though).

Yup, they're upgraded on the pie. 12 gauge, and half the stock length. However, it's graded after the axle, not inside.

The stockers were definitely 16ga.
On my MAC, they were more like 14-15ga. and upgrading them to dual 14ga. didn't help too much.

Not sure about your other q..

[neptronix]

GOT GOOD NEWS...

Battery: 20S/5AH charged to 81v.
Hill: 0.32 mile, 8.6% grade average, ~12% peak grade.
Conditions: 37 deg F. & no wind

Speed at the start of the hill ( ~3% grade ): 28mph.
Lowest speed at 12%: 25mph

Cruising speed: 30mph, at 120%, 33mph.

So yeah, i dropped an entire 3mph at the worst part of the hill. The motor was cold to the touch even after ripping around the street a few times in utter childlike glee at the 120% setting.

This motor is the real deal, boys.... I may actually have to dial down the amps because it has this nasty habit of wanting to throw you off. I doubt it is using the max 56 amps at all as the 5AH of year old turnigy 20C would have gotten hot at anything over ~33A for very long.

I do not have a cycle analyst to measure the amps/watts/etc just yet. I have one currently en-route from methods.
More later, including video once it stops snowing.

[neptronix]

Here is a pic of it so far.



( no, my window is not busted out. That's just window wrap since we live in a drafty ol' house.. )

Needs the following things:

1) Shorter suspension fork, OR a 20in. wheel in front to improve the angle of the bike for better aero.
2) Thinking about an aero fairing. A couple mph would help here.
3) Enable regen.. i have no back brake right now. Waiting on cable from cell_man.
4) Another 12S / 25AH array on top of my
5) Rear rack for the addl. ~16lbs of 12S/25AH ( 10 packs ).
6) External controller mounting to make room in the falconEV bag.
7) Depending on how the next tests go, vented hub covers, internal temp sensor and a cooling system i've been dreaming up for a while..
Dial the amps down from about 56A to maybe 50A. We'll see what the cycle analyst reports. Hopefully they can be dialed down a little to prevent a LiveForPhysics-hands-the-bike-to-a-newb type of situation.
9) shorter crank arms as these 165mm ones have about 3 in. of clearance from the ground.

Lastly, thank you all for cheering me on. I had no effing clue that a DD hub motor could have performance like this at all. I am still shocked at how little mph it dropped on the hill. And i am ebike grinning from ear to ear still

[Alan B]

Good job.

You might set the phase amps lower while keeping the battery amps up, that would not affect top speed but would reduce the kick from zero and the heating that it causes.

If you lower the front are you going to be able to pedal?

[neptronix]

Good job.

You might set the phase amps lower while keeping the battery amps up, that would not affect top speed but would reduce the kick from zero and the heating that it causes.

If you lower the front are you going to be able to pedal?

Thanks for the phase amp tip.. i will experiment with it once i get the programming cable in from cell_man.

I updated the list.. added 9) ..
Yeah, i've got 3 inches of clearance with the pedals O_O.. I will check my local bike shops to see if i can get something smaller than my 165mm pedals.

Will hit up the used bike store on monday.

[jmygann]

cassette -speeds ? disc rotor ?

[o00scorpion00o]

GOT GOOD NEWS...

Battery: 20S/5AH charged to 81v.
Hill: 0.32 mile, 8.6% grade average, ~12% peak grade.
Conditions: 37 deg F. & no wind

Speed at the start of the hill ( ~3% grade ): 28mph.
Lowest speed at 12%: 25mph

Cruising speed: 30mph, at 120%, 33mph.

So yeah, i dropped an entire 3mph at the worst part of the hill. The motor was cold to the touch even after ripping around the street a few times in utter childlike glee at the 120% setting.

This motor is the real deal, boys.... I may actually have to dial down the amps because it has this nasty habit of wanting to throw you off. I doubt it is using the max 56 amps at all as the 5AH of year old turnigy 20C would have gotten hot at anything over ~33A for very long.

I do not have a cycle analyst to measure the amps/watts/etc just yet. I have one currently en-route from methods.
More later, including video once it stops snowing.

Brilliant, See I told you the Pie was a monster hill eater didn't I ?

So going from 26" to 20" really makes a major difference to the load on the motor, that's really cool!

It would be interesting to see how hot it would be on a longer climb, so it did 25 on 12% so it should do 20 on the 16% I went up, but the main thing is it shouldn't get too hot!

33 mph is not too shabby, but on 80 volts it's a lot of battery to carry around, but that was just with 5ah right? I was getting 33-35 mph max on 60 volts in a 26" wheel with 120% setting.

You would need about 100 volts to get to about 40 mph, but shit that would be scary with so much torque, on the + side though you could reduce the current even further if you wanted. 100 volts 10ah would be 1000 w/hrs for about 15 ish miles at 35-40 mph, not bad at about 55 ish whr/mi. That's the same wh/mi as a 50 volt 20ah pack for the same range, or more considering you would be going slower! I think you are into electric shock territory there ? maybe nothing serious just a little bite ?

It would have been better if GM had to make a faster wind for the smaller diameter wheels, I think the pie is too slow a wind, it has so much torque to start with it didn't need such a slow wind, but since they designed it to work with smaller batteries they wanted to maximise the amount of torque for the batteries and controllers they sell. They are not interested in making fast bicycle motors!

I would like to see how the HS3540 performs in a 20" wheel at the same 80 volts. According to the ebikes.ca simulator the HS3540 will go up a 12% grade at 24.9 mph, 50 amp controller, overheat after 5.8 mins. 115 lbs of torque. Strangely enough in a 16" wheel it says it will go up at 26 mph with 140 lbs of torque. That would nearly pull the bike apart!!!

So why does it say it will go up slightly faster in a 16", is it because of the extra torque ?

The new Volkswagen Golf 1.2tsi turbo petrol manual 122hp has 144 lb/ft and the HS3540 has 140 lbs in a 16" wheel, holy shit that would out accelerate it something crazy, to about 20 mph lol still, that's some power!



Dave I say to hell and go 100 volts, if you can afford to spend more on a new controller that is, I don't think the 4110's would take that?

2 x 20" wheels and you will be a lot lower, less wind resistance and it would be a lot steadier. Now if you only had a full suspension bike!

[heathyoung]

Re: crank length, you can get 150mm ones (BMX) - might give you some more ground clearance.

[neptronix]

scorpion; you are the one who helped this build happen majorly. Nobody else spoke so highly of the magic pie even though Luke kept recommending it etc.. I would have forgone this motor if it wasn't for your tests.

Your pie wind is definitely a faster one than mine. Like i said, i hit about 30mph on 5ah of 20S, but that's at a full charge, it goes 27mph on the nominal voltage ( 3.85v ), so that's what i have to plan around..

22S will get me a mph or two, that's it.. Mind you, 22S is 92.4V fully charged, that's about as close to 100V as i want to get since the limit of the components in my 12FET is 100v - cutting it close there..

By the way, comparing car torque to bike wheel torque is a bit complicated because of the gearing & final drive of a car. Even an electric car is going to have a final drive / gear reduction, so the torque will be more or less what the motor is outputting.

But yes, this thing feels like a V8 or V12 taking off. I really want to dyno it.

[neptronix]

Note: i finally got the Cycle Analyst in the mail today so i'm ready to collect some data ..

But life has came between me and my build lately. I've been revamping a decade-old SQL design at work and it's been giving me fits all week so i've not had as much time to focus on the important things in life ( ebikes ).

I noticed that gearing for the pedaling is going to be a severe problem on the 20in. wheel. I need to be pedaling around 25mph to keep the speed up & battery draw / motor heat low on the really nasty sections of the hill. The motor wants a 6 speed freewheel, as a 7 speed would rub on the frame.

The lowest gear available on a 6 speed freewheel is about 14T, and that's a major, major problem..

I hope that i can find a 56-60T front chainring somewhere O_O..

I may have the axle machined.. or insert some spacers on the freewheel side and bend the frame a tad.. not sure which way to go yet.

Heath: i will check that out. 150mm would probably be perfect.

[neptronix]

I found a solution..

http://goldenmotor.com/SMF/index.php?topic=2788.0

http://cgi.ebay.co.uk/ws/eBayISAPI.dll?ViewItem&item=270661455664#ht_2672wt_944

spacers is!
I have a 7 speed freewheel with an 11t.. and another with a 12t.. ( old sugino freewheel ). I think that plus a monster size front 130mm BCD will get me into that 25mph zone.

[jmygann]

I'm looking to replace my brushed 20" hub motor with a brushless , on 48 volt . quick and efficient to 25 mph.

Is the magic pie the best choice ?

[neptronix]

I'm looking to replace my brushed 20" hub motor with a brushless , on 48 volt . quick and efficient to 25 mph.

Is the magic pie the best choice ?

Quick, yes.
Efficient, probably not.

It will need about 66V nominal to do 25mph.

If you want quick AND efficient.. i'd point you in the direction of a MAC/BMC motor in a 20" wheel.. on the MAC side, an 8T would be really quick on 48V as long as you feed it the amps it needs on that motor / wheel size ( ~30A )

PM me if you have more questions about your own build though, i'd lke to keep this thread un-cluttered and on topic as much as possible

[DSX]

Another option for a compact 11 tooth freewheel 5 speed (11T 6 speed may be an option too) is make one like Dr Bass did with epoxy.....this is really slick....


viewtopic.php?f=2&t=27108&hilit=freewheel

[chroot]

My 9x7 9C motor gives me 37-42mph (20s LiPo) on flat road. I assume your magic pie is high torque type motor similar 9C 6x10 or HT35 motor?

Is it because Luke recommend you get this Magic Pie due larger stator therefore more cooling due heat soak into larger stator? Mac 8T is very nice as long as you keep maintain 1500 watts. You will be fine long endurance run without tear up the gear box. ilia brouk (BMC guy in SF), He abused his BMC geared motors hardly break the gearbox and I have see his action. His wheel spin like MAD!

I was gonna get MAC 8T for my Genesis V2100 with extended swing arm and I realized MAC geared hub motor is limited at 1500 watt for longevity last long without break the gearbox. I am in middle of the fence between RC motor setup using recumpence's V4 motor kit or Dual cyclone 1200 watts or stick with Direct Drive motor. arggh hard make a decision.

Please make more vids action and I would love see how the magic pie handle the hilly (10% grade avg). Does your MAC 8T broke down yet? I mean are you not happy with the MAC motor?

Your pie wind is definitely a faster one than mine. Like i said, i hit about 30mph on 5ah of 20S, but that's at a full charge, it goes 27mph on the nominal voltage ( 3.85v ), so that's what i have to plan around..