Clean work Don’t forget to use red loctite on the kegel adapter pins
you have taken all the guess work out of my build using the Neptune, thank you sir
Please tell us how the Maker-X SV6 ride, I love the way they look and want to make a gokart-type thing with 2 of them.
Where did you buy your motors from? And when?
I really love this build thread, it leaves nothing to the imagination
I ordered from esk8supply on 3/9. You can check my entire Bill of Materials here.
Thanks very much- I enjoy doing documentation as part of my professional life, so to have a chance to apply that skill to a personal project is very rewarding. I’m hoping some of my learnings and mistakes can help other riders in some way.
Update 2020-04-10
Part way through integration with battery pack and BMS completed. Drop through mounting also finished.
Arrived
Janux Quattro 4.25" TB DD Hubs
Yep, I know that I’m using eLofty drives, but word on the street was that the 5 pin eLofty adapters would fit the TB DD hub pattern. In addition I’m leaning heavily toward the 6" Evolve tires and wanted to validate that Evolve tires fit on these hubs. After getting the parts in I’m happy to report success on both fronts.
M5 Hardware
Rather than going with standard #10 hardware for mounting the trucks, I decided to use M5 hardware instead. This should also allow me to replace the stainless bolts on the eBoosted enclosure with black bolts (which fits the board aesthetic better). Having common hardware between the trucks and the enclosure mounts simplifies the parts list slightly.
1/2" Heat shrink
This allows me to complete the positive lead on the battery pack by giving rigidity to the in-line fuse.
Kapton Tape
It ends up that the enclosure is just a bit too narrow in the middle to fit all the N.E.S.E. caps. As such, I’ll use Kapton tape to cover the remaining bolts of the enclosures to minimize the chance of short.
XT30 Pigtails
I wanted to connectorize the charge port and since the charge current is ~4A, XT30 will be sufficient for this.
New
New Button
The Neptune 15 is a very full featured BMS. As such, it has its own power button which is required for proper function (more on this below). The button which is included is small and flimsy, so I’ll need to order a new button (latching not momentary) that is capable of panel mount and install it next to the AntiSpark button at the front of the enclosure. This button will also need to include an in-line connector to maintain N.E.S.E. maintenance access.
Changed
Button Move
Looking at how packed the rear of the enclosure is getting, I decided to relocate the AntiSpark power button to the front of the enclosure. This will require extending the leads between the button and the PCB, for which I’m considering using a simple 3 Pin ATX fan extender cable. Also note that these leads will run across the top of the battery pack, which means they must include a connector between button and PCB in order to maintain the ability to easily service the N.E.S.E. modules.
Complete
BMS Battery Leads
Completed the rebuild of the positive pack lead using the in-line fuse. The picture below is the naked lead. The heat shrink covered lead can be seen farther down.
N.E.S.E. Bolt Coverage
The eBoosed enclosure is just a little too narrow at the middle to fit the N.E.S.E. pack with the bolt covers on. As such, I’ve used some Kapton tape to cover the bolts on the pack edges in order to minimize the chance of short.
30Q Binning, Fish Paper, and Installation
I binned all the 30Q cells, but there wasn’t a single one out of the 50 that was move than 10mV out of alignment. I added fish paper rings the positive end of each 30Q and dropped them all into place in the pack.
Here’s a shot of the pack with the lid securely fastened into place.
Charge Port Soldering
I had a heck of a time getting the solder cups to behave when soldering the charing port. Finally got XT30 pigtail on and heat shrink covered the leads at the charge port:
Next was adding the XT30 pigtail to the load/charge leads of the BMS:
Finally I attached the whole charge group together:
BMS Setup and Test
Next was testing the BMS. Interestingly, the power on test only requires connection of the balance leads but also requires the connection of the power button to the BMS (latching not momentary). I had not taken this into account to this point so I ordered a panel mount button to fulfill this function. After attaching the included power button I was able to test the BMS using the Android App:
Once in the app, I customized the BMS settings. Anything not on this list I did not change from default:
Section | Setting | Value | Comment |
---|---|---|---|
Dashboard | Guard Bar Lowest | 38.4V | 3.2V / cell |
Dashboard | Guard Bar Lowest Alert | 40V | 3.3V / cells |
Dashboard | Guard Bar Highest Alert | 49.8V | 4.15V / cell |
Dashboard | Guard Bar Highest | 50.4V | 4.2V / cell |
Dashboard | Dashboard | Voltage | Displays Pack Voltage |
Dashboard | Cell Bar Lowest Alert Zone | 3.2V | |
Dashboard | Cell Bar Highest Alert Zone | 4.15V | |
Device | Cell Number | 12 | 12s |
Device | Battery Type | Li-ion | |
Device | Cell Voltage High | 4.10 | Increase Cell Longevity |
Device | Cell Voltage Low | 3.20 | Increase Cell Longevity |
Device | Cell 0V Margin | 25mV | Threshold to prevent never-ending charing if calls cannot balance closer than this |
Protection | Charge Current Limit | 6A | Adapter can only supply ~4A |
Protection | Temperature Limit | 65C | Termistor will only be near cells, so temp threshold should be notably lower than 80C |
Protection | FET Temperature Limit | 70C | Unlikely we’ll hit this in the current charge only configuration |
Charge / Balance | Cell Balance Delta (mV) | 50 | How far apart P-groups get before balance charging kicks in |
Charge / Balance | Stop Charging CV Current | 0.10 | Threshold to prevent never-ending charging if calls cannot balance closer than this |
Battery Pack Charging
After customizing all my BMS settings, I verified pack charging using the Android App
Drop Through Mount Test
Using the M5 hardware, I mounted the trucks to the board drop through. Will need to test for motor and wheel bite later.
Anyone used backing frames for drop through mounting before? Considering that the majority of the impact will be absorbed by the pneumatics, I’m leaning toward not using them.
What’s Next
- Finalizing foam placement in the enclosure
- Wire routing for the motor leads and sensor wires
- Testing wheel bite and purchasing tires
- Testing motor bite and (possibly) purchasing harder bushings
- VESC programming
- Remote Programming
- Bench testing of fully integrated system
Research
Nothing in addition to previous posts.
I use Moonbeams on my drop-through setups, but over-tightening my hardware has caused them to crack a bit, so those Riptide frames are probably the best option if you care about protecting your deck in those spots.
Super clean build , can’t wait to see the finished product
Update 2020-04-17
Great progress. Bench tested and ran into a problem I could use some help with (see below)
Arrived
Bones Swiss 6 Bearings
Decided to go with some quality bearings for the build
Evolve 6" Tires
I cranked down the trucks to just before the bushings started to bulge and I was still able to get the 7" tires to impact the board relatively easily. Even though I really like to the look of the 7" tires I decided to go with the 6" Evolve tires instead. These ship from SoCal and arrived quickly. Here’s a shot comparing the 6" and 7" tires on the Janux 4.25" hubs:
Flipsky VX2 Remote
I ordered this about 5 weeks ago from banggood and it finally arrived. I love the feel of this remote in the hand, the bright screen with telemetry, and the rechargeable battery. The thumb wheel feels solid and appears responsive.
Latching Waterproof Button
To enable control of the BMS from outside the enclosure, I needed a latching, waterproof button. I like the quality here but it’s a bit big (hole diameter is ~0.7"):
New
No major additions to the design at this time.
Changed
Motor Spacers
As part of the integration process I began putting together the hubs, adapters, bearings, tires, etc. and ran into an interesting issue. I’m having aluminum spacers made and will need to get longer M5 socket head bolts to mount the adapters to the motors. This should allow a near flush mount of the adapter against the hub.
Complete
Test Wiring Complete
All hardware attached as per the wiring diagram. Power buttons work and LEDs light up.
Remote Programming
Powered on the remote and adjusted the settings.
- Battery Cells: 12
- Pole Pairs: 14 < Can someone verify this please? I cannot seem to find definitive data on the 58kv motors.
- Speed Option: mph
- Wheel Type: HUB motor
- Wheel Diameter: 150
- Throttle Calibration: Completed
- Pairing: Completed
- ESC Type: FSESC
VESC Programming
Watched videos, read, and reread the VESCtool Manuals. Started cautiously.
Auto Connect
- Connection to Primary VESC immediately successful
- Firmware needed updating. Updated to 4.2.
- Disconnected, connected to 2nd VESC. Updated 2nd VESC firmware to 4.2.
- Disconnected, reconnected to primary VESC.
Setup Motors FOC
- Medium Outrunner
- Battery Cells Series: 12
- Battery Capacity: 12.000 Ah
- Direct Drive: Checked
- Wheel Diameter: 150.00 mm
- Run Detection Button- Success! Both motors detected and settings seem reasonable.
Setup Input
- PPM Remote
- Didn’t change anything on first screen
- Didn’t change anything on next screen
- Finish- tested: nothing!
- Checked wiring schematic vs VESC connector and shuffled the pins around.
- Tested again: Remote makes motors spin and I get full telemetry on the VX2!
Bench Test (partial failure)
This is where the problem arose. The remote is super sensitive, just a small flick past center dead zone and the motors attempt to go full speed. Once they get to nearly 100% duty cycle one (or both I can’t tell) appears to vibrate and loses most of its speed before trying to ramp up again. It sounds bad:
Debugging:
- Disconnected CANbus cable
- Connected VESC1 to motor1, ran Setup Motors FOC- success. Spun motor up to full- one drive spins up, no vibration, remote still super sensitive.
- Connected VESC1 to motor2, ran Setup Motors FOC- success. Spun motor up to full- one drive spins up, no vibration, remote still super sensitive.
- Connected VESC2 to motor1, ran Setup Motors FOC- success. Spun motor up to full- one drive spins up, no vibration, remote still super sensitive.
- Connected VESC2 to motor2, ran Setup Motors FOC- success. Spun motor up to full- one drive spins up, no vibration, remote still super sensitive.
- Connected CANbus cable again- both drives spin up and vibration is back
- Changed remote to VESC2, both drives spin up and vibration is still present
- VESCtool / App Settings / PPM / General Tab / Multiple VESCs Over CAN: Changed to False, wrote change to VESC- one drive spins up, no vibration, remote still super sensitive.
- Turned on stream realtime data and stream realtime app data, VESCtool / App Settings / VESC Remote / General Tab / Control Type: Changed to Off, wrote change to VESC.
- Set remote to mode H. Pushed remote thumbwheel full forward: VESCtool reports 99.2% throttle
- Set remote to mode M. Pushed remote thumbwheel full forward: VESCtool reports ~75% throttle
- Set remote to mode L. Pushed remote thumbwheel full forward: VESCtool reports ~50% throttle
- VESCtool / App Settings / VESC Remote / General Tab / Control Type: Changed to Current, wrote change to VESC. Once remote gets past standard 15% deadzone, motor speed goes to full before VESCtool reports remote getting to 25% throttle.
Help Needed
VESC Output for given Remote Input
From the above testing and debugging it appears that the VESC output to the motor is set to ramp at a rate at least 10x that of the remote input. I have looked around the tool, the manuals, and the forums but I’m failing to find a location where this can be manipulated. I have tried adjusting the Throttle Curve but even at -80% Exponential this change only makes the most minuscule difference. I don’t need to change the curve, I need to change the multiplier. For example, I would assume at 50% throttle I would hit 50% motor speed and at 100% I would hit max motor speed. Can someone point me in the right direction here?
eLofty 58kv Pole Pairs
I’m seeing conflicting information on how many pole pairs the 58kv motors use. If someone who uses these drives could verify this info that would be helpful.
What’s Next
- Fix VESC Programming
- Bench testing completed without failures
- Get tubes and tires mounted
- Get spacers installed and adapters mounted
- Mount wheel to the adapters
- Finalize foam placement in the enclosure
- Wire routing for the motor leads
- Wire routing for the sensor wires
- Wire routing for the buttons
- Mount buttons, cable glands, and charge port to enclosure
- Mount enclosure to board
- Test ride
Research
Nothing in addition to previous posts.
20 magnets
Nope. Not how current mode on vescs work. Your remote controls the torque / acceleration of the motors, and without any load on them (i.e. bench testing) it’s only normal that they go up to max speed quite fast.
Those loud sounds from your motors are a bit worrying though. You should ask in the elofty thread if anyone has experienced something like that.
I find it odd that the vibration / noise only happens when both motors are connected over CAN and everything is > 90% duty cycle. With only a single motor there is no problem getting to 100% duty cycle with no vibrations, so I’m hesitant to pin this on the motors.
Good to know about the Current Mode. Is there a different mode which will cap speed?
More testing needed!
Not a practically usable one.
Whats the height of the enclosure? internal and external? Also have you thought of replacing the go-foc’s with neoboxes? I know the go-focs have a height of 19mm, but wouldn’t 22mm fit?
Those measurements are difficult to articulate since the enclosure “flows” with the shape of the board. That said, my best estimate is that external height is ~35.5mm and internal is ~28mm. I have added some additional foam around the perimeter which gives me a few more mm, allowing me to attach foam to both the top and bottom of the internal components, which should provide some protection to vibration.
I have not really looked into the neoboxes after having committed to the go-focs. The layout of the phase wires at 90 degrees to the XT60 connectors was advantageous to my layout.
What type of on/off switch are you going to use for the bms?
I just installed mine and had the same realization that I’ll need one.
Update 2020-05-26
Figured out the wire routing (which was the biggest remaining unknown). Also complete are tires, tubes, Loctite, foam, buttons, more.
New
RJ12 Ends
In looking at how to route the sensor wires and phase wires, it was important that the design allow the enclosure to be fully mounted to the board and then connected to the motors. With the limited cabling length, the wires need to be extended from the VESCs to get these connectors to the outside of the enclosure.
The JST 6 pin 2.0mm connectors (used for the sensor wires) do not easily lend themselves to extending, are not really compact, and have no real strain relief. I looked at many different options for 6 conductor connectors including the Haggy sensor connectors, but eventually landed on using an RJ12 connector. The RJ12 connector offers a small form factor and integrated strain relief in addition to allowing for easy connection to other RJ12 connectors.
RJ45 Coupler
In order to mate the RJ12 connectors between the motors and the VESCs, a simple compact RJ45 coupler will be used. It should be noted that this coupler maintains proper pin order allowing connectors with the same wire order to be connected together. This connector was also chosen both for the metal casing and compact form factor.
4mm Bullet Connector Extenders
Just as the sensor wires required extending to reach the outside of the enclosure, the 4mm bullet connectors on the phase wires also needed extensions
Bluetooth Module
Once the enclosure is bolted to the board, it should only be removed in the case of hardware maintenance. That said, VESC programming was initially completed through a direct USB connection to each VESC. To get around this issue, a compatible Bluetooth module will allow for programming of the VESCs (including firmware upgrades) when used with the VESC mobile app. This module will connect with COMM connector on the secondary VESC.
Cogging with MakerX VESCs
Reading through the site, I came across this reply in the Go-FOC SV6 and Go-FOC SV4 …Maker-X thread. This sounded very similar to the issue I was having with my SV6 VESCs. The next step here is to upgrade firmware to v5 and test again.
Changed
Smaller latching waterproof button
The 0.7" button I was planning on using was just too huge. I found a 12mm option which matches nicely with the button from the AntiSpark.
Complete
Get tubes and tires mounted / Get spacers installed and adapters mounted / Mount wheel to the adapters
This was mostly covered and completed over on the Janux-esk8 Aluminum Hubs to fit Direct Drive thread on post 161.
Additionally, here is a side by side of an uninflated Evolve 6" tire vs an inflated tire at 50PSI:
Also a full shot of the first wheel mounting:
Loctite Janux hubs and adapters
Applied Loctite 242 to the Janux hubs and to the bolts which hold the Kegel adapters to the motor CAN.
Finalize foam placement in the enclosure
The corners of the N.E.S.E. modules required the thicker foam but the sides of the enclosure needed the thinner material in order to squeeze everything in. Additional foam is located on the top of the N.E.S.E. modules and on top of the VESCs in order to firm up objects in the enclosure once the enclosure is mounted to the board.
Wire routing for the motor leads
The motor leads (phase wires) will go through the cable glands as originally planned. Additionally the extensions will allow these phase wires to connect between the VESCs and the motors after the enclosure is mounted to the board. Quick shot of the extensions:
Wire routing for the sensor wires
The choice to use the RJ12 connector was detailed above, but some additional detail is required here.
First, with the extensions made available through use of the RJ12 connectors, the original design of using a single cable gland for all three phase wires plus the sensor wires (per motor) can be achieved.
Second, it should be noted that the wires used for the JST connectors are ~1.2mm, too large for the ~1.05mm channels in the RJ12 connector ends. As a result, each of the sensor wire leads required soldering to CAT5 conductors. Since only 6 conductors are required, the brown pair was removed from the cable length and the blue, orange, and green pairs were used. The sleeve of the CAT5 was kept for the RJ12 strain relief clamp. CAT5 was chosen over flat 6P6C wire due to a) a better fit with the (3) phase wires within the same cable gland and b) the larger gauge of the conductors.
Third, the wire colors of the sensor wires differed from the motors to the cables supplied by the VESC manufacturer (these motors are also missing the temp sensor lead). As a result, special attention needed to be paid to how the connectors were wired.
Lastly, was the consideration of total connector length. When (2) RJ12 connectors are mated using the RJ45 coupler, the entire length of the assembly is 35.6mm including strain relief for each RJ12 connector. This length gives some flexibility in cable placement for the connectors.
Wire routing for the buttons
As mentioned in my update from 2020-04-10, I relocated the buttons to the front of the enclosure. The antispark uses a momentary switch, requiring (3) conductors whereas the BMS uses a latching switch requiring only (2) conductors.
For the antispark, I used a standard 3 pin fan extension cable. For the BMS I used a small two pin connector I had available:
The connectors are both located near the front of the enclosure and the wires are run such that they have the minimum chance of being crimped when the enclosure is mounted to the board.
Regrettably the placement of the buttons in the enclosure did not go quite as expected:
Now all I can think of when I look at the front of the enclosure is Sloth from Goonies.
Mount charge port to enclosure
The charge port (on the other hand) mounted exactly to plan. The XT30 connector kept the installation simple.
Help Needed
Nothing right yet
What’s Next
- Verify functionality of Bluetooth module
- Update firmware to v5
- Bench testing completed without failures (with fall back to BLDC)
- Mount cable glands to enclosure
- Mount enclosure to board
- Route motor cables to avoid motor contact
- Test ride
Research
The helmet thread has lots of food for thought.