The High Voltage guys posted a video of my Evant conversion on their youtube channel.
So, yes, this happened.. :/
After running the evant for two charges the controller died. It starts, I can connect to it using the USB dongle but as soon as I give throttle input I get an overcurrent error.
Measuring the phases it turns out that one phase is shorted to B-, typical mosfet failure. QSmotor, that I bought the motor and controller from, was reluctant to send me a new controller even though they state there’s a 1year and 3 month warranty period. Somehow they say I damaged the controller but it’s virtually impossible to damage the controller in this way except when assembling it. In the end they offered for me to send the controller back to be repaired, or pay half the price for a new controller.. I chose to send it back, which it turned out was not a viable option at all.. so..
The problem with buying a new controller at half price is that Shanghai is shut down due to Covid and the factory is closed.. so QS couldn’t really tell me when I could get a new controller, so the Evant project halted..
I’ve already started planning my next project. It’s going to be a couple of awesome dirtbike replacements.. powerful LightningRod bigblock motors and matching ASI BAC controllers. More on this later. The thing is, I’ve already bought and got the motors and controllers. So.. I started out connecting my 50A VESC to the QSmotor, just to try it out, and sure – it runs. Wierd that the 800A APT96800 controller got damaged from pushing the same amount of power that a 50A VESC can handle.. well well..
For my build I’m going to use the ASI BAC4000 controller. It can handle 400+ phase amps but is limited to 72V max nominal voltage on the batteries. Since I’ve already built the 48V packs that’s going to be utilized in future projects as well, I’m running the Evant on two 48V packs in parallell yielding 60Ah and 2,9kWh. After discussing the matter with mr HV on the HV discord I connected the BAC4000 to the QSmotor and made connectors to precharge and connect the two packs in parallell.
One problem with this setup is that the QS motor uses an encoder to keep track of the rotor position while the ASI BAC only supports hall sensors.
But it turns out the BAC4000 actually handles the motor better in sensorless mode than the APT96800 did using the encoder. 😀
With the APT the motor sometimes stalled, loosing sync, requiring me to turn the machine off and on again for it to regain sync. The BAC4000 hasn’t lost sync once during the hours of use I’ve put it through.
Well, now that the machine runs again it’s time to finish the project. Until now I’ve been driving the Evant with an android tablet in one hand keeping track of the battery status..
The ANT BMS is super with it’s bluetooth connection and app showing all the data you need, but it’s a bit cumbersome driving around holding a tablet all the time.
I got a display that connects to the BMS in the battery..
Printed a sealed backside and TPU gaskets for it..
And viola, the last piece of the conversion is in place! The display shows the voltage of the battery, the temperature of all four temp sensors in the pack, the highest and lowest cell voltages and the current draw. Since the display connects to one BMS in one of the two batteries the actual current going to the controller is actually double what the display shows though. But now I can keep track of the battery status while driving the machine with two hands!
Today I’ve been moving heavy wet snow from our lawn just to test the Evant out. Turns out I can drive it for a little more than 50 minutes on a full charge..
.. and then it’s time for charging. I got two of these Eltek Flatpack 1500 from a friend and they’re able to push a little more than 30A each into the packs.
Unfortunately one of them had a problem with the voltage regulation and died after charging one pack but with these chargers I can charge one set of batteries in approx 45 minutes. This means that with 2 sets of batteries I should be able to run the machine back to back assuming the motor or controller doesn’t overheat.
So now the build is more or less complete.
I’ve got a few panels to remount that I need to repair first but the machine is definitely ready for use again.
Testing the limits of the battery I ran out of juice in the middle of the driveway..
So I had to get the charger and charge it for 10 minutes to be able to drive it back into the garage. =)
When stuff goes sideways like this it truly helps having good people to turn to. The Evant is running again thanks to the awesome help from Captain Codswallop and mr HV on the HV discord. They helped me tune the BAC4000 for my setup and have been super helpful in making the right decisions and making a wireloom that fits my needs. If you want to build a high power ebike or just make any kind of electric conversion I really recommend joining the High Voltage Light Electric Vehicles discord server.
Also, check out the HV channel on youtube:
Well, that’s that. I’ll post further updates if the controller or motor fails, but for now this project is more or less finished! Hope you’ve found it interesting!
Started yesterdays work with welding the new controller mount. I decided that it was better to use the existing mounting holes in the motor than to build something around the motor, especially since fitting the battery packs and all leaves very little space around the motor. So I made a bracket that mounts on top of the motor. To fit the M12x1.5 screws I had to mill the heads down on them to be low profile enough to not protrude from the mounting holes.
This mounts rigidly to the motor and with the threaded M6 holes for the controller to mount to everything is snugly and firmly in place.
One advantage to mounting the controller like this is that I can shorten the phase cables quite a lot leading to less losses.. even though the cables are thick enough to withstand all the current the evant should need in the future – less is still more in this case.
Time for the next piece of the puzzle. Since I don’t have all the parameters for the plasma cutter perfectly dialed in yet there is quite some slag buildup on the parts, requiring some grinding and sanding to get them ready for welding..
This is where the battery box is going to sit that’ll house one of the two battery packs in the evant..
.. and this is what it’ll look like from behind. Quite awesome if you ask me.. Still everything needs primer and paint so the warning symbol won’t be as visible, but I’ll try to remedy that somehow..
And this is how the battery sits inside the machine..
I want the pack to sit as far forward as possible to make sure it gets as little abuse as possible from the wheel and whatever debris that might throw at it. I’m going to make a side plate for the battery compartent to protect it further..
The second battery pack is going to go the other way in the machine, sitting as far forward as possible in front of the motor.
Something like that. I’ll add padding to the cover to push the packs firmly in place and to cushion it somewhat – even though the batteries themselves should be able to withstand quite some abuse.. (hopefully)
The only problem now is that the packs actually won’t fit under the “hood” – which is also the mount for the seat, so..
Getting the grinder out and squaring up the problematic part. I just love working with steel, it’s such a forgiving material. If something’s wrong, just cut it up, make a new part and weld it all together, and it’s strong as ever! Add some paint and it looks like new..
So with that modification the battery fits in the machine. Now it’s just a matter of making the pieces I need to cover all holes I’ve made and welding it all together.. After that I need to take care of the electrical wiring, adding the 12V DC-DC converter and installing the BMS displays (whenever I get those from China).. So, quite a lot more to do but the hardest part is done I guess..
So far it’s been a super easy and super fun project.. Let’s just hope it works when all is complete.
So, when everything was installed and fitted I realised there was only 200mm where the batteries needed to go, and the batteries are 230mm wide.. Hmm.. Well..
I started with removing the controller and the controller bracket but still the batteries won’t fit side by side.. So I tried a few placements and some could fit the battery pack but parts of it would interfere with the rear wheels when turning, other placements wouldn’t fit at all until I found one that fit..
To make sure everything fits height-wise I put the seat mount back on the machine. There was some interference so I had to chop it up a bit..
Cutting and grinding discs on the angle grinder has a way of setting things on fire.. mostly me, but today all that caught fire was my gloves.. =)
Some steel removed from the inner side of the seat mount and now it all fits with margin to spare!
Since I got rid of the mount for the controller and since I need some brackets for the batteries I rigged the plasma cutter to cut some steel.
After a bit of fiddling due to the steel sheet bending (I need a THC) it finally made a puzzle!
Tomorrows task will be cleaning these up and welding them to something that looks kind of like this..
After that the batteries should fit and the controller should be rigidly mounted to the top of the motor.. If that happens all that’s left is to build the final version of the throttle, remove all spare wires from the wiring harness and route everything to the dash.. Oh, yes, I need to mount the DC-DC converter to get 12V from my 110V battery too, and wire in a main switch.. Well..
And there it is:
It’s finally moving, powered by the electric motor!
I connected two of my emergency get-back-home ebike batteries of 48V 4Ah with a maximum current output of ~15A. That’s way to little to drive the QS-motor that’s in the Evant but still it runs. Don’t want to run it for too long though since I haven’t calibrated the valve for forward/backward motion and have to keep the P-brake engaged for the machine not to roll off the lift it’s on – which makes for quite some resistance in the hydraulic system.
The twist trottle is quite unusable in the Evant so I’ve made a quick hack for a system to control the RPM of the motor:
It’s just a 10k potentiometer and a switch. The plan is to set the RPM with the potentiometer and use the switch as an on/off switch to turn the motor off whenever it doesn’t need to run. That way I won’t need to adjust the RPM of the motor to turn it off, which I’ll want to do every time I’m standing still and not working with the machine.
Just made a quick and dirty connection to try it out and it works just as expected. So the next step is to make some kind of enclosure for it, put a better cable on it and a proper connector.
So, when test fitting the motor I realised the bottom side of the motor isn’t flat so it needed a spacer to sit flat on the motor mount.
Fortunately, now that I own a CNC plasmacutter, making a spacer like this is just a couple of minutes of work.
This fits perfectly on the motor. Welded the spacer to the motor mount and used the plasma to cut the inner hole out of the mount as well, and after some paint..
.. it was time to mount the motor in the Evant again.
I decided it was way to much work trying to mount and unmount the motor when bolted with the M12 bolts from underneath so I welded the holes I had drilled and drilled and tapped holes for M8 bolts instead. This way I can remove the motor without the hassle of fighting with the hydraulic pump all the time making life much easier for me! 4x M8 should be strong enough to hold the motor in place easy.
With the new, shorter belt the position of the motor is perfect. I can adjust tension in both ways and get a properly tensioned belt. So, with this confirmed it was time to mount the controller.
On goes the controller mount using M6 screws.
And on goes the controller. The wires between the motor and controller are way too long and I’ll shorten them once I’ve test run the motor.
When measuring I’ve got about 20mm too little space to fit my batteries, so that’ll have to be remedied somehow.
The next step is to figure the wiring out. Most of those aren’t going anywhere on my machine so I’ll disconnect them from the connector and make a new wiring harness with only the leads I need. All I really need is the mains switch, the speed input and the hall sensor and thermistor input signal wires.
So, hopefully I’ll be able to spin the motor up tomorrow. I’ll just have to make a wiring harness to get 96V into the controller.. and a better throttle..
When all the copper plates were welded to the cells and all the BMS wires were added..
.. it was time to finish the battery build. I printed the box and the TPU divider and put the cell pack in it’s place.
I’ve bought a couple of ANT 300A BMSes to protect the packs and they come with screw terminals mounted.. so I pressed the same type of terminal on the negative leads of the pack..
After that it was just a matter of adding the BMS.
.. connecting all the wires ..
.. and pushing the positive and negative power feeds through the TPU grommet..
To finish the packs I soldered EC5 connectors to the power leads and printed a “connector” to house the 3 connectors to make it easier to connect the battery pack to the Evant and impossible to connect the wrong leads shorting a battery pack out..
So.. now I’ve got one complete battery pack of 26s10p which gives me a 96V pattery with 30A capacity yielding a whopping 2,9kWh. This is what I’m going to use for testing. I’ve got enough cells to make one more complete setup so that I can charge one pack while using the other. Hopefully I’ll be able to run the machine for at least half an hour on each pack.. Doing the math :
Avant original motor power: 10kW
Estimated mean power output: 33%
Estimated mean power output: 3.3kW
That should give me almost 55 minutes of run time on my pack, but to have some margin let’s say 45 minutes. Charging the pack at 30A (1C) would take more or less an hour to charge the pack fully, so it might be possible to get back to back running.. otherwise I can build one more pack to have two packs charging while I’m running the third. That should give me continous driving with pretty good margins but it’d cost me approx $700 to build.. We’ll see how it works, only testing will tell..
So, at first I finished building the CNC plasma by adding all the electronics and controllers..
I printed a box for the controller and raspberry pi, added all the cables, a relay for the plasma trigger and closed it all off.
I had to do some work with Klipper to get it to work for the plasma cutter. It’s designed for driving 3D printers and making a config that works the way I wanted to was a bit tricky, but as I suspected it was totally possible.
So, after configuring Klipper and writing a custom POST processor for Fusion 360 it was finally time for the first cut..
I placed the CNC plasma on 4040 profiles to get it off the floor, filled a frying pan with water and put a sheet of 1.5mm stainless on the pan.. and off we go!
It actually worked surprisingly well!
Without tweaking it actually cuts pretty cleanly but there is much room for improvement! So, time to put the plasma to work!
I made a shelf for the motor controller for the Evant in CAD, did a quick CAM processing increasing the feed to 2000mm/min and uploaded it to the plasma.
And of course, when I took the video the controller failed. I did some more runs and it seems the X stepper controller has a tendency to reset every now and then.. And then I realised I’m using a 60W 12V PSU.. That means I’ve got a total of 5A to run all the steppers, the controller and the pi.. which was a bit too little. After switching to a 200W PSU everything worked fine!
I made two side walls..
And of course when it’s CNC controlled there’s room for.. well.. flare. 🙂
After tweaking the feed and amps of the plasma I’ve got it to where the cut parts fall off the metal sheet. Sure, there’s room for improvement as I get quite a buildup of steel on the back of the part but a quick runover with the grinder solves that.
I just had to tack weld the parts together to see how well it fits.
It’s not as precise as the mill but MUCH faster, and cuts 3mm steel sheet in seconds!
I’m going to turn the motor 180 degrees to get the cables poking out the back and the controller shelf is going to be turned 90 degrees to mount at the left and front sides of the motor.
Something like this..
We’re getting closer to testing the Evant. Today I got the new belt which I hope will fit the new pulley and motor location.. But this weekend I’m going to focus on building the kitchen so we’ll see how far I get with the Evant. Considering the snowfall here the past week or so it’d be really good to get it running soon..
Ok, so the snow has started pouring down outside and I really feel the need for the Evant to run.. so this week I’ve been tinkering with the battery..
Since the motor/controller has a peak consumption of 200+ amps and the cells I’m building the battery from are able to push 35A continously giving a whooping 350A for the pack – I need something better than nickel strips to carry the current in this pack.. so I got the tip from a guy on a discord server to use copper sheets..
Copper conducts current 10-15 times better than the nickel strips I’ve been using, probably more since they’re more often than not just nickel plated steel strips.
The thing with copper though is that it’s virtually impossible to weld with hobbyist equipment. That’s where RioCole showed me a method to get the copper sheet welded to the LiIon-cells using nickel plated steel strips as a conduit.
First of all I needed to cut the copper sheet to shape. It’s easy to do with a pair of scissors but I wanted better looks for my connecting copper. The design I make for the cell holders also make the poles of the cells sit recessed about 0.2mm, which requires that the copper sheet is able to bend down to weld properly.. to do that without distorting the entire sheet I need tabs that bend down which are hard to make manually..
Then I realized that my vinyl cutter should be able to at least scratch the cu sheet.. At first I had trouble with the cutter not gripping the copper sheet properly which made it twist, but after taping it down to a piece of vinyl it works flawlessly!
I started out with some test pieces..
.. and then went for a design with a cut Z-slot..
At first I thought that putting nickel strips in series would give me better current capabilities, but it was so much easier to get good welds with short pieces of nickel – so that’s what I did..
Starting to get the hang of welding the copper sheets to the cells. All connections are good and tested. The problem now is that the ANT BMS I got for this pack was too big to fit the casing I printed earlier.. so I had to redesign the battery box..
This beast is currently on the printer so an update will follow when it’s done printing and the battery is fully welded and ready for finishing..
Well, since the mechanics were done a while ago what’s left is really just waiting for the new controller board to arrive to start making the electronics for the cutter. I’ve still made some changes to the Z-axis though to make it easier to adjust the belt tension..
These are the bolts that hold the belt on the Z..
This was the intended idler holder where the belt tension was to be adjusted.. But as you can see, and what I didn’t think of, is that the idler obstructs the lower screw making adjusting belt tension hard using this..
So instead I skipped the CNC:ed part and just attached the idler directly to a nut in the extrusion..
After attaching and tensioning the belt all three axis are driven by the belts and stepper motors.. So, if I only had some stepper drivers we could make the cutter move..