5kW ebike motor / controller testing

Show off your Ebike creation here.
Post Reply
User avatar
owhite   100 W

100 W
Posts: 113
Joined: Aug 03 2020 11:22am

5kW ebike motor / controller testing

Post by owhite » Sep 20 2020 8:50pm

Hello folks I'm posting this so I'll have a chronicle of some testing I'm doing for an eventual final build. This is going to involve a bunch of prototyping before I get to the final setup, so here goes. Right now I'm starting with a Rattan Fat Bear as a test platform:

https://imgur.com/McRF5dY.jpg

The bike has four inch 20 inch diameter fat tires. The bike comes stock with a 48v, 500w motor on it and maxes out at 32mph. I purchased a motor from ncycle on aliexpress. The motor appears to be a 5000w 100KM/H max speed and was discussed on this thread:
viewtopic.php?f=3&t=107820

Another pic of the motor: https://imgur.com/2EzSoxx.jpg

I will be running a 24 FET MOSFET 84-132V Freeway Speed Controller from Lyen.

I'm shooting for a 24S1P system using four 16000mAh packs in series. Each pack is a liperior 6S 12C 22.2V used from rcbattery.com. A pic is here: https://imgur.com/YKcpZjo.jpg. I'm just going to pretend that it cant deliver 380amps at peak and keep at a far more sensible level of amperage for this build.

I didnt want to take apart the fat bear to start testing, so I built a 2x4 frame to help with location of the parts and give a general idea of the length of all the final wiring. This is the 2x4 frame compared to bike:
https://i.imgur.com/ojFCKDT.jpg

The final build is probably going to mount everything along the top tube of the Fat Bear, so I'm prototyping a box that will be in the same position as a motorcycle gas tank. These are laser cut stainless steel flats to build the battery box.
Image

The steel portion of the battery box, folded:
Image

The battery box involves a combination of laser cut stainless steel and wood. Here's a test fit on the bike:
Image

Connecting high amperage stuff kind of sucks, so I'm trying out this idea with XT90 connectors, and a 120 amp fuse, 8ga copper wire, and some butt connectors with allen screws:
Image

Then I'm plunking the connectors in a printed box:
https://imgur.com/RfMQDUt.jpg
The whole system gets filled with with two part epoxy insulating potting compound from amazon.

Here is the assembled battery box prototype on the frame, populated with controller and connectors in potting compound:
Image

Another view is here: https://imgur.com/8S0IRk2.jpg. Notice the fabulous fake handle bars, cycle analyst on front, and motor mounted on the back.

And here's a front view of a combination 3d printed box, high amperage lugs, for the motor phase wires:
Image

That's all for now.
Last edited by owhite on Nov 10 2020 12:31pm, edited 3 times in total.

Dr.mckenney@gmail.com   1 µW

1 µW
Posts: 1
Joined: Sep 23 2020 10:35am

Re: 5kW Rattan Fat Bear Conversion

Post by Dr.mckenney@gmail.com » Sep 23 2020 10:40am

Holy smoke Owen!
You are one creative wizard.
I appreciate your prototyping.
You have way too much power to handle in your battery set.
Is this set for a drag strip Speedtest or a new cannonball ebike record attempt?

Tommm   100 kW

100 kW
Posts: 1262
Joined: Apr 03 2018 2:32am

Re: 5kW Rattan Fat Bear Conversion

Post by Tommm » Sep 23 2020 11:11am

Nice deathtrap build... What's the point of pouring time and money into it when the bike is junkyard material meant for light cruising in the first place?

Balmorhea   1 MW

1 MW
Posts: 1575
Joined: Oct 01 2019 8:30pm
Location: Austin

Re: 5kW Rattan Fat Bear Conversion

Post by Balmorhea » Sep 23 2020 1:33pm

Tommm wrote:
Sep 23 2020 11:11am
Nice deathtrap build... What's the point of pouring time and money into it when the bike is junkyard material meant for light cruising in the first place?
I was thinking it.

Also, there's no way that thing goes 32mph with a 500W rated motor system. 32 km/h maybe, with a tailwind. Or if you drop it off a high bridge.

This project is something a resourceful teenager would get into because he doesn't know any better.
___________
Finish Reconstruction.

Tommm   100 kW

100 kW
Posts: 1262
Joined: Apr 03 2018 2:32am

Re: 5kW Rattan Fat Bear Conversion

Post by Tommm » Sep 23 2020 1:43pm

Balmorhea wrote:
Sep 23 2020 1:33pm
Tommm wrote:
Sep 23 2020 11:11am
Nice deathtrap build... What's the point of pouring time and money into it when the bike is junkyard material meant for light cruising in the first place?
I was thinking it.

Also, there's no way that thing goes 32mph with a 500W rated motor system. 32 km/h maybe, with a tailwind. Or if you drop it off a high bridge.

This project is something a resourceful teenager would get into because he doesn't know any better.
He wants to replace the motor and electronics... the real question is does anyone want to go well above 32mph with a loaded up china folding ebike built with knockoff components.

User avatar
owhite   100 W

100 W
Posts: 113
Joined: Aug 03 2020 11:22am

Re: 5kW Rattan Fat Bear Conversion

Post by owhite » Sep 23 2020 5:38pm

Thanks for the comments guys. It's refreshing for an uninitiated person to be able come to a site and get constructive information about a build.

Balmorhea   1 MW

1 MW
Posts: 1575
Joined: Oct 01 2019 8:30pm
Location: Austin

Re: 5kW Rattan Fat Bear Conversion

Post by Balmorhea » Sep 23 2020 8:39pm

owhite wrote:
Sep 23 2020 5:38pm
Thanks for the comments guys. It's refreshing for an uninitiated person to be able come to a site and get constructive information about a build.
If it causes you not to do something dumb and hurt yourself, then it's valuable information.
___________
Finish Reconstruction.

User avatar
owhite   100 W

100 W
Posts: 113
Joined: Aug 03 2020 11:22am

Re: 5kW Rattan Fat Bear Conversion

Post by owhite » Oct 03 2020 10:45am

Another installment on setting up my build.

This topic is "So your system isnt working and youre trying to figure out if it's the motor or the controller or something else." or "Beginner's notes on popping a MOSFET". There's lots of posts about this on ES, but maybe this will be useful to someone to have all this info in one post.

So for my build wiring seemed to be going fine and I sure AF dont want to talk about how to program a controller - but then I experienced a setback - the motor died. I'm not sure where things went wrong. First, I was testing the motor and controller and was getting sporadic results:
https://www.youtube.com/watch?v=aJTSRYaJqho

The frame rate of the video makes it hard to see but shows a kind of skipping behavior when the throttle was at 0-90%, and then ran at relatively slow rpm at 90-100% throttle around 20s of the video. Then, at one point when tried switching the phase wires and the motor stopped turning entirely. I just hit the throttle and boop, it was dead. This post is of some things I learned to try to fix - the text assumes you know how to run a voltohmmeter and know a few things about electronics. The term "continuity" means if you connect the voltmeter and it measures that you have a complete circuit, the term "short" means there is a connection that you dont want - like, if stand in a bathtub of water and stick your finger in an outlet, that's a short. Avoid.

Throttle testing.

One issue is you throttle may not be compatible with your controller. So I tried two of them, one came with the motor, the other I made using a potentiometer using this circuit:
Image

in both cases I got voltage changing from 0-4vdc on the signal wire of the throttle so that seemed okay. Note: most commercial throttles are based on hall sensors, dont expect that it behaves like a normal potentiometer, you have to have power going to the throttle to see voltage changes on the signal wire.

Hall sensors

Hall sensors inside the motor are set up to detect rotation of the magnets on the stator as they rotate past the stationary stator. My controller has a five wire connector - the red and black wires are for ground and +5v the three other wires correspond to each hall sensor. When the hall connector is powered, if you put the voltmeter on gnd and each one of the hall wires, you should see it vary from 0-4vdc or so when you spin the motor. In my case they were all working fine.
My connector:
https://i.imgur.com/wwUwhVc.jpg?1
View of halls inside the motor:
https://i.imgur.com/vAB2GIb.jpg

Motor tests

A QS205 is a simple animal. Check the hall sensors, they're okay. Other things to test:
  • to find a short, put connect the continuity tester on each of the phase wires one at a time, and an outside chassis of the motor, you should not get a short
  • connect the continuity tester between the plus wire and an outside chassis of the motor, you should not get a short
  • connect the continuity tester between the negative wire and an outside chassis of the motor, you should not get a short
  • connect the continuity tester from one phase wire and another the phase wrie, you should get continuity, or very low resistance
Finally, you want to know your motor will generate some electricity. Motors are basically generators. Turn the motor you should get some electrons out. You can test this by putting your voltmeter on each phase wire, or I have a light tree from back in the day when i used to discharge smaller lipo batteries. So, connect up lights on two phase wires, turn the motor, see if this thing can generate current:

Image

and away it goes.

Then there's the slam dunk. Another way to rule out the motor - if you got it, why not try another controller? In my case I was also lucky because when i ordered the motor it came with a 70amp controller. At first I didnt bother with it, but since the motor may have been toasted i went through the programming steps, and eventually got the motor working. The motor turned. Fine, so ruled out motor problems.

Controller aint working? Check your MOSFETs.

Skipping over a lot of detail, MOSFETs are basically valves that when switched on allow monster amounts of current to flow. In the case of my infineon controller, it has an IRFB4110 MOSFET
Image
https://www.infineon.com/dgdl/irfb4110p ... 15a9571e0b

When the controller sends logical signals to the gate, that gate dump all kinds of current. The MOSTFETs get massive current from the plus and minus sides of the input into your controller, and the logic in the controller drives each phase wire of your motor. Well, that's when they are working the way that you want them to. The problem is that these chips have limits, and if they are exceeded what they usually do is short out, and then they dont operate like a valve or switch.

Typically, when they short, it means that you can easily detect problems by finding a short on either the gate dumping something to the plus or minus side of your controllers input, and in my case i also had a short going to the phase wire as well.

One quick way to test your controller is to leave it connected to the motor and to turn the motor by hand. The motor should turn as smoothly as it does when it is not connected. In my case it didnt turn smoothly.

Another simple test can be done with your voltmeter and is described here:
https://www.ebikes.ca/documents/BlownMosfets.pdf

This involves checking the resistance between the negative line going into your controller, and each of the phase wires. Two phase wires should showed around 10k ohms, one of the phase wires was around 0.18ohm. Next check the resistance beween the positive line and each phase. In my case two of the phase wires were 10k, and the same bad line read 0ohms. A short between one of your phase wires and either of the negative or positive line is bad news.

Another step you can take if youre interested in trying to repair your controller is open it up and get access to the MOSFETs.

I picked this up from this link:
https://www.youtube.com/watch?v=lTssCn5YyTw

When I went into my controller there's a large cooling block to serve as a heat sink for the fets, and the easiest way to test them was from the bottom of the controller. See:

https://i.imgur.com/9aYXCkP.jpg

There is a set of MOSFETs that connect to each phase, and to either the positive or negative side of the controller's pos/neg lines in. In my case three fets serve as gates between each phase, and positive (or negative) line in.

You should be able to find a big fat trace that connects either the positive or negative line in. In my case, I found a short between the positive line in and the fets controlling one of my phases. Boom - those are the fets that I would need to replace.

The controller was beefed up to take more amperage by adding a metric fuckton of additional solder to the traces, and even with a hotair rework station at 500F I'm not really able to melt the solder. A similar sized blob of the solder that I normally use melts no problem - so I'm not sure but one guess is maybe the previous guy used plumbing solder or something. So the bad news in my case is i'm not totally sure I'll be able to swap out the MOSFETs. Anyway, hope this helps someone in a similar situation.

User avatar
owhite   100 W

100 W
Posts: 113
Joined: Aug 03 2020 11:22am

Re: 5kW ebike motor / controller testing

Post by owhite » Oct 29 2020 9:07am

More notes just to have a record of progress:

I grabbed up this 2007 frame on ebay:

Image

It wont fit my fat tire wheel so I designed parts to get cut by water jet
Image

That I will weld partly with a shit welder in my basement, and then take to a professional to lay down good quality beads.

The frame is nice in that it lets you tuck in the batteries in relatively discrete places, shown in yellow

https://i.imgur.com/LInmtmd.png

I’m currently designing the battery box housing that will also be a combination of waterjet cut aluminum and welding.

Still challenged with a decent controller. First purchase was from Lyen. It didnt work on arrival, and determined it was a fet problem.

viewtopic.php?f=6&t=108357#p1587800

Ordered another controller from powervelocity. The site said it would ship in two days. They gave no tracking number, no response at all after repeated requests for a tracking number so good old paypal helped with getting a refund. Definitely can not recommend powervelocity for their customer service.

Now employing the Mad Rhino strategy of getting a 72150 and swapping out the MOSFETS. Fet removal went smoothly and waiting on the replacement parts.

Out of curiousity i also ordered a custom printed circuit board a galp BESC

viewtopic.php?f=30&t=97765

which if it works will be fantastic but board debugging can take months of work.

User avatar
Skaiwerd   100 W

100 W
Posts: 266
Joined: Jun 16 2016 4:55am
Location: CT

Re: 5kW ebike motor / controller testing

Post by Skaiwerd » Oct 29 2020 4:17pm

owhite

Glad you ditched the folding frame.

I like that swingarm, I’ve considered a similar design approach myself, perhaps for the frame also. What thickness would you use. 1/4” seems the standard but to me seems a bit thicker would help in the dropout area. Have you considered weight? I believe a square foot of quarter inch is 10lbs so it adds up quickly and puts you back to thinking tubing or solid rod girder style for that matter. Things to think about.

User avatar
owhite   100 W

100 W
Posts: 113
Joined: Aug 03 2020 11:22am

Re: 5kW ebike motor / controller testing

Post by owhite » Oct 29 2020 9:18pm

Thank you for your comments.

I calculate the weight of the swing arm to be 7 pounds. The frame (with old swing arm) is 6. I did a ballpark estimate and all up weight of the bike (motor, batteries, wheels, forks) to be in the 75 pound range.

Thanks for advice on weight costs for building solid parts - I should say I have two-ish design constraints. The main constraint is I'm mostly focussed on "wtf can a guy do with cheapish ebay parts and basement manufacture?". And, I'm with you, I think it would be interesting to put together a whole frame with waterjet parts.

The other constraint is, this is just meant to be a street cruiser. It's not a downhill bike, it's not a trials bike, it's not a high performance sport bike for leaning into corners. I just want to push 50+ on the street, and not sit in the right lane to get passed.

Swing arm. I'm not optimizing for low weight, i'm just convincing myself that a swing arm will be possible using waterjet and welding. I assume that by being strategic it would be possible to knock a bunch of the .375 parts to thinner - quarter inch should be totally doable.

i'll get this thing welded up, and share results soon.

thanks.

User avatar
owhite   100 W

100 W
Posts: 113
Joined: Aug 03 2020 11:22am

Re: 5kW ebike motor / controller testing

Post by owhite » Nov 05 2020 11:06am

A post about upgrading a sabvoton motor controller to a 100v rating.

Tools. Before starting let me recommend some incredibly helpful tools that I dont think would have made this possible. Item 1 can be found by searching on "claw pen" and helps with grabbing small parts. Item 2 is a desoldering tool. There are a lot of versions of this and the bad news I dont think I could have done this without this, which is quite pricey:
https://i.imgur.com/qe0sDeo.jpg

Search on "HAKKO FR301-03/P Desoldering Tool".

Item 3 is a voltmeter, preferably one that uses an audible beep when it tests for continuity. Youre also going to need a soldering iron, solder, soldering wick and other common bench tools. You may also want to order some non-conductive thermal grease.

Opening the sabvoton. To open up your controller, there are a lot of pics and videos on the net if you search for "sabvoton teardown" and similar terms. Some of them are included here to help, but one issue is there are so many controllers out there, and the internal configurations are different.

In my case I one similar to one shown in this video:
https://youtu.be/2EfIhOmRm5g?t=398

First remove the allen head screws on the bottom of the controller. The front and back of the sabvoton has two plastic caps and a pile of allen head screws to keep them in place. Remove screws, and yank the cap on the end that does not have the screw terminals for connecting wires.

Inside the controller the board is mounted to a large aluminum plate:
https://i.imgur.com/PiLHEns.png

That plate has a thermal heat sink compound and sticks to the out controller housing - lift that up a bit inside the controller, then the whole unit will slide out on the side where the terminal connectors are. Grab some saran wrap or a plastic bag, and lay your greasy thermal plate on that to reduce contaminating your hole bench.

The parts to be replaced are attached to that plate by a series of screws that are accessed through holes in the top of the PCB:
https://i.imgur.com/TnIGBZ8.png

All of these screws have to come out. Loosen, and then use the pen claw tool to lift out. The screws will have a lot of thermal grease on them so keep 'em inside some kind of container. There are also screws that hold the PCB on the aluminum block with plastic spacers on each corner - remove those.

This then exposes the underside of the PCB:
https://i.imgur.com/dOJyj8I.jpg

And what youre looking at is a bunch of MOSFETs, here is a pic from a spec sheet:
https://i.imgur.com/rIde3ll.png

That I've also discussed here:
viewtopic.php?f=6&t=108357#p1587800

Thermal insulation. There are a couple of incredibly important elements to the MOSFETs and how they are mounted. One issue is that the drain (the big flat plate on the back of the FET) can not have electrical conductivity to the aluminum plate. The other is that the FET needs to be thermally conductive to the plate. The way this is dealt with is to coat the relevant surfaces with non-conductive thermal grease, use some kind of insulation tape between the FET and the aluminum plate, and to use insulating grommets on the screws holding down the FET.

Here's a pic of screws with grommets:
https://i.imgur.com/fr8i1QH.jpg

Here's a pic of hardware / mechanical drawing:
https://i.imgur.com/UHTQqR9.png

note that picture shows that nuts are used to with the screws to mount the FET, but in my case the screws went into threaded holes on the aluminum plate.

FET removal / desoldering. I found it easist to cut the through-hole wires on each FET, and then heat the remaining wire tab to extract it out of the PCB. Then next comes the drudgery of removing the solder plugging the hole of the PCB. This is going to involve a combination of heating with soldering iron, hitting with the desoldering tool, and possibly using wire braid to soak up the solder. It's miserable work, and this is where the Hakko desoldering tool really saves the day. These have a heating element at the end of a steel tube, and then a built in pump to remove the solder. These tool work extremely well, other than requiring occasional cleaning to unplug solder from inside the tool.

So which FETs? I decided to go with IRFB4110:
https://www.infineon.com/dgdl/irfb4110p ... 15a9571e0b

I'd be interested in suggestions from other people on possible components. IRFB4110s are rated right up to 100v, and since I'm operated at 24s I am gonna guess that sometimes these will blow at that level. I did see some other FETs that might be able to go to a higher voltage but I copped out on doing anything radical to keep the costs down.

Notice also that other components, such as the biggest capacitors on the board are also rated to 100v - I also assume that 24s lipos could blow one of these guys some day but I didnt replace.

Installing FETs. There is a VERY important element to placing the new FETs on the board. Notice that the mounting hole of the FET has to be perfectly aligned with their respective mounting hole on the aluminum plate. If it is not, they you are much more likely to experience conductivity shorts when you screw in the mounting screws. Use pliers and bend the terminals on the FETs as precisely as possible - then solder them in place. When bending and soldering - do everything you can to check position of the FETs. If I do it again I'll probably try to use some kind of template.

Once the FETs are soldered, get the insulating tape ready on the aluminum plate, and add some thermal grease to FETs or plate if needed. Now you have to carefully reattach the FETs to the plate with insulating grommets on the screws. The super critical element here is you need to make sure you're not shorting the FET plate to the aluminum heat sink plate. To assure yourself of this, clip your multimeter to the aluminum plate, and check for continuity between it and the drain wire on the FET. (note, there is a diagram above, youre looking for the FET terminal marked "D"). A audible continuity checker is really helpful at this point. If you get any continuity, back the screw out and recheck your alignment.

Once done, also check for any continuity issues between the tabs of each FET and nearby blobs of solder. That side of the board is really messy so a lot of potential problems can arise.

After convincing yourself that everything is fine - reassemble the controller.

This post,
viewtopic.php?f=6&t=108357#p1587800

and many others discuss how to test your controller. Good luck.

User avatar
owhite   100 W

100 W
Posts: 113
Joined: Aug 03 2020 11:22am

Re: 5kW ebike motor / controller testing

Post by owhite » Nov 28 2020 3:17pm

Swing arm build.

Image

I am starting with this frame and need to replace the back swingarm to fit my QS205 motor:
https://i.imgur.com/yyEgZCn.jpg

I did a bunch of CAD work for 3/8ths thick aluminum, set up with slots and tabs to help the whole thing fit together:
https://i.imgur.com/oYWNBs2.png

The square box on the right was set off to get cut by water jet:
https://i.imgur.com/ixgRrk6.png

I used big blue saw and total cost was around $300. The parts were cut in 3 weeks or so:
https://i.imgur.com/djFUvNC.jpg

When designing for waterjet, be sure to account for the kerf width of the cut, which in my case was around 1.25mm. They were fine for the most part, but the "3/8ths" thick aluminum was not 0.375 inches - it was actually thicker - probably because most plate is manufactured in mm these days. So in some cases tabs did not fit perfectly, and I was able to buzz the tabs to correct thickness using a router table with a carbide bit.

Prior to welding I did a lot of test fits. The swing arm has an inner core. The bolt you see there will hold the shock to the bike:
https://i.imgur.com/OjSWrjy.jpg, https://i.imgur.com/aEQViwk.jpg

And then the outer portion of the swing arm was added:
https://i.imgur.com/KYQ2fvp.jpg

Whole assembly prior to going to the welding shop, some parts from the old swing arm were used to make sure spacing was correct on the inner core. The dropouts that will fit the axle came great, as did the dropout distance of 185mm:
https://i.imgur.com/gxYm4PK.jpg

I hit every surface with an abrasive wheel to remove paint, oxidation, and grease on the parts. Then I hit every edge that was going to be welded with a scotchbrite pad, and washed everything down with acetone for a final degreasing. I was very surprized by how much residue washed off the aluminium so this was an important step. Parts were then taken to the welder:
https://i.imgur.com/jhIQ0xo.jpg

This guy was incredible. Instead of chasing the customer out of the shop, he was great about letting me work with him. This was really helpful because the parts had to be welded in place sequentially, and I needed to check things for alignment as he went. He was also very interested in the ebike build and we had a lot to chat about. To be honest he was more of a structural welder than a bike frame builder so he laid down thick, heavy beads. He had a gigantic miller syncrowave and used 1/8th TIG rod. Definitely the loudest sounding squarewave buzz I've heard coming from a welder. Some welds:
https://i.imgur.com/vzHmOae.jpg

Not too pretty but strong AF. No regrets. I'm sure everything will clean up nicely with a powder coat. Total weight: 3.2kg.

That's all for now. Also, making great progress on a custom controller build, have a look at that over here:
viewtopic.php?f=2&t=109266

User avatar
owhite   100 W

100 W
Posts: 113
Joined: Aug 03 2020 11:22am

Re: 5kW ebike motor / controller testing

Post by owhite » Dec 26 2020 8:23pm

Next installment of going from here to here:

Image

Larger pic of design:
https://i.imgur.com/phPwDpM.png
Larger pic of progress:
https://i.imgur.com/ZWRHSph.jpg

This process took a lot of work to figure out the lines of the frame and then get them into CAD. Lesson learned. There is no way you can take a single pic of a bike frame and trace the lines into a CAD program. Think about getting a picture of yardstick. If you set up the picture so you the inches in the 24inch range are actually an inch, the inches at 1 inch, and 36 inches are NOT going to be one inch in length. Anyway. What I did was a lot of testing with laser cutting pieces, seeing how they fit, adjusting based on what results I got, placing some of the big outer plates of the battery on the bike with clamps, and learning from the results.

The really sweet thing about having a laser cutter was I could cut up 1/4 plywoood, and make plates that fit into various areas to see how they fit. My laser also cuts sheet metal which I used to make band clamps to hold into place. Once the dimensionss were finalized I started working out the plans for making two side frames of the battery box. Construction was shown here:

Image

that's a combination of cut sheet metal angle pieces, weld nuts, rivets and aluminum angle. Riveting is super fun by the way it's very easy. You basically just drill a hole and mushroom over the rivet. This an example of an assembled set of aluminum angle:

Image

Cutting aluminum angle. This is also really easy. Turns out you can cut it with a miter saw, a jigsaw, a table saw, or a circular saw. Just get you a non-ferrous metal cutting blade and away we go. So one trick I needed was to have very precise angle cuts, and after screwin up a bunch of angle, used this trick of clamping two angle pieces together, and clamping to the miter saw to cut two pieces to the exact same length:
Image

see, the problem was if I was going to make two frames

Image they have to be very precise in length of all all the edges. The clamp em together trick helped a lot with that.

Once the frames were made I made a bunch of outer box pieces with plywood, used countersunk bolts to put the box together:

Image

another pic of the interior:
Image

One issue, with this construction it will never be water proof. The shock/swing arm combination has to penetrate into the battery box, so no chance of handling that opening.

Image

99t4   100 W

100 W
Posts: 238
Joined: May 10 2020 10:31pm
Location: PNW

Re: 5kW ebike motor / controller testing

Post by 99t4 » Dec 29 2020 12:50am

owhite wrote:
Dec 26 2020 8:23pm
One issue, with this construction it will never be water proof. The shock/swing arm combination has to penetrate into the battery box, so no chance of handling that opening.
Yes that one of the major tradeoffs with a FS bike. The shock placement on yours looks to be especially challenging. That's why I'm on the lookout for a Specialized Epic type of design for a FS project-- more of an open triangle.

What is the width of your battery compartment there?

User avatar
owhite   100 W

100 W
Posts: 113
Joined: Aug 03 2020 11:22am

Re: 5kW ebike motor / controller testing

Post by owhite » Dec 30 2020 5:49pm

each batter pack is 2 1/2 inches wide, and in addition to the central frame tube, the whole thing is close to seven inches.

Image

John in CR   100 GW

100 GW
Posts: 15225
Joined: May 20 2008 12:58am
Location: Paradise

Re: 5kW ebike motor / controller testing

Post by John in CR » Dec 31 2020 12:25pm

A lot of great work so far. Before final assembly be sure to get all of the frame measurements so that once you get that bike rolling you can do it right and build an AL frame to match your swingarm that incorporates the battery box for much more battery room with a smaller exterior form.

I'd be very leery of high speed on those fat bike tires.

User avatar
owhite   100 W

100 W
Posts: 113
Joined: Aug 03 2020 11:22am

Re: 5kW ebike motor / controller testing

Post by owhite » Jan 04 2021 4:03pm

Thanks John in CR,

I appreciate the suggestion about the frame. I will be replacing the wood with laser cut AL, it's just easier to use that because I have a laser that cuts plywood in my shop. Most of all my main goal is to just get on the road to see if the motor, battery and controller combination are even worth using for a final bike.

Regarding tires - I had a PM conversation with Mad Rhino where he said the same thing about tires. He also suggested using 120tpi tires. I will be using those for this build. Again, once I get this thing going and learn how well it works I may move to narrower tires -- I just didnt want to do a full rebuild of the wheels yet.

Thanks again. :)

User avatar
owhite   100 W

100 W
Posts: 113
Joined: Aug 03 2020 11:22am

Re: 5kW ebike motor / controller testing

Post by owhite » Jan 18 2021 6:06pm

Power distribution system for the bike. These were made with two printed circuit boards with built up bus bars mounted and epoxied into 3D printed boxes.
Image

They were constructed from custom circuit boards which solder into XT90 connectors
https://i.imgur.com/ZJF79OH.jpg

2mm thick copper was soldered on to the PCBs to serve as heavy duty busbars:
https://i.imgur.com/1GJuX7M.jpg

Another angle of the same boards:
https://i.imgur.com/dUcFWmi.jpg

The parts are then epoxied and mounted into 3D parts:
https://i.imgur.com/fq7TKD2.jpg

The idea is the lower board will be in the inside of the ebike, it will have plugs that face downward to plug into the batteries, and also have plugs facing upward that mate into the top set of plugs that serves to put all the batteries in serial:
https://i.imgur.com/jyUFbCl.png

When that top plug connector isnt plugged in, the upward facing plugs will also be accessible to the charger.

Here, the power distribution is getting mounted to my plywood battery box, along with a large fuse and a touchscreen display:
Image

The touchscreen system is described here:
viewtopic.php?f=2&t=109266&p=1611894#p1611894

Battery connectors facing out:
https://i.imgur.com/QpVFBsn.jpg

The arming plug attached to battery connectors:
https://i.imgur.com/MKDmtLl.jpg

Note: the plywood is just to help make the prototype, I'm planning on replacing it with aluminum.

User avatar
owhite   100 W

100 W
Posts: 113
Joined: Aug 03 2020 11:22am

Re: 5kW ebike motor / controller testing

Post by owhite » Jan 23 2021 6:46pm

an interesting discussion about the gap between my swingarm and the disc brake

viewtopic.php?f=3&t=110116&p=1613438#p1613438

User avatar
owhite   100 W

100 W
Posts: 113
Joined: Aug 03 2020 11:22am

Re: 5kW ebike motor / controller testing

Post by owhite » Feb 13 2021 8:16pm

Additional work on the interior of the battery box. I designed and 3D printed holders:
Image

A pic of the back holder:
Image

The holders in the battery box, with batteries:
Image

Velcro straps will be used to keep the boxes in place.

Another closeup of the battery holders here: https://i.imgur.com/KYm4uma.jpg

User avatar
owhite   100 W

100 W
Posts: 113
Joined: Aug 03 2020 11:22am

Re: 5kW ebike motor / controller testing

Post by owhite » Feb 24 2021 6:30pm

Starting to get the bike to run:
https://youtu.be/VCvrm-DAZ54

That's a motorcycle tire on the hubmotor -- which I dont think is a good idea but gonna go with it now because it is roughly the same diameter of the front 26 wheel.

User avatar
owhite   100 W

100 W
Posts: 113
Joined: Aug 03 2020 11:22am

Re: 5kW ebike motor / controller testing

Post by owhite » Mar 07 2021 5:23pm

Yah the tire in that last video is simply too large.

Purchased a 19x inch rim (36mm, inner width).

https://www.treatland.tv/semi-gloss-sat ... ch-rim.htm

Working on the calculation for a new wheel:
Image

User avatar
owhite   100 W

100 W
Posts: 113
Joined: Aug 03 2020 11:22am

Re: 5kW ebike motor / controller testing

Post by owhite » Mar 18 2021 8:25pm

Image

it's coming along. still doing a lot of work on the controller.

User avatar
owhite   100 W

100 W
Posts: 113
Joined: Aug 03 2020 11:22am

Re: 5kW ebike motor / controller testing

Post by owhite » Apr 01 2021 8:25pm

Now with the battery box closed up:
Image

Some specs -- these are mostly based on getting readings off of the BESC controller
Amps: 80
Batteries: 24s
MPH: 51
Weight inventory:
Image

Post Reply