TRS300 HELLO. Harley Conversion. ME1616

[TRS300]

Hello everyone. I've been reading from this site for some time now and I want to express my gratitude to all the smart information found here.

I'm working on converting my 2015 CVO street glide to EV. The project has guidelines and parameters one of which is that I can restore the motorcycle back to its original engine if desired.

Highlights of the project are that I retain full integration of the instruments, fob, abs, etc.

I'm using an me1616 that I opened up and modified. New bearings, varnish, new power leads separated to each set of windings. As a sub project I was curious to see if the motor could be converted to hall position sensors and I did achieve that using a suron position sensor with a custom magnet ring. It works well actually. I've run the motor on a simple sinusoidal ebike controller and it works fine.

I do have a Kelley KLS96601-8080NPS controller that I had the firmware upgraded so that it can be more easily integrated into the Harley CAN bus. The Kelley controller is huge however so I'm still on the fence about what controller I'll use. The good news is I can easily use a hall sensor controller now if I prefer to.

I posted this project on a harley forum and got zero interest. At least they did not call me names haha.

Anyway, going forward I have some questions that I hope to get some input on. And if anyone is interested in what I'm doing I hope I can help others on this forum also. Thanks for reading my introduction.

 

[bananu7]

Looks fun! I'm sure you'll get much more interest here. Integrating with the stock harness sounds like a challenge, but at the same time I think we'll be seeing more and more of such projects as the median year for the conversion base increases. Good luck!

 

[harrisonpatm]

Looks pretty cool!

I posted this project on a harley forum and got zero interest. At least they did not call me names haha.

I posted my 89 Kawasaki EX500 conversion on a forum specifically for EX500's and they were very supportive and appreciative. I guess the Harley stereotype lives on!

 

[fechter]

Is that a water cooled motor?
Any idea what batteries you're going to use? A Harley is pretty heavy to start with.

 

[TRS300]

Looks fun! I'm sure you'll get much more interest here...

Thank you. I modified the OEM harness by surgical removal of all the loom related to the ECM. Every cut point is marked and recorded. My custom "EVCM" is based on an Esp32 connected to the CAN bus and other sensors etc. The bike still has the OEM modules (abs, security, gauge, etc.) my new EVCM basically transmits information on the bus to drive gauge displays and let the other modules think the ECM is on the network by transmitting the VIN number. This avoids CELs being thrown by the other modules.

 

[TRS300]

Looks pretty cool!.... guess the Harley stereotype lives on!

Thanks. Honestly, I think it's just beyond the scope of most folks on that forum. Like I said, no negative comments. I think it's just a fish out of water there. Personally I don't think it's a stereotype issue.

 

[TRS300]

Is that a water cooled motor?
Any idea what batteries you're going to use? A Harley is pretty heavy to start with.

Actually batteries is a topic I'm going to need help with. That and a good controller. The bike was a twin cooled model which means it already has water cooling components. The bike is heavy which is why I kept a transmission in the build. At least to begin with. This project is an "around town" runner bike. I don't expect to get significant distance. As for the weight. I'm trying to keep everything heavy below calf level. So even though it's heavy it should feel lighter than it is.
Also, I've recorded the weight of every part removed from the bike. I think in the end it will be a little bit lighter than it was.

 

[amberwolf]

Actually batteries is a topic I'm going to need help with.

Since it's the core of the whole system, and must supply all the power ever needed to do anything you are doing with it, some thoughts:


First, if you haven't already, sit down and determine what job you need this bike to do for you, under the worst case riding conditions you will ever encounter, for how long in time / range, at what speeds, at what acceleration rate, with what total weight (rider, bike, anything you're carrying, or hauling if you ever use a trailer, etc).

Then you can use that information in various calculators / simulators like those at ebikes.ca to guessitmate how much power it will take to do that job under those conditions.

Then you can use those results to guesstimate the minimum capability and capacity for the battery, and the minimum power required for the motor and controller.

Add at least 25% capacity to what you think you need to account for loss during aging. Add at least 25% more to account for unaccounted condtions, winds, detours, etc.

So, "always"*** use a battery that is more capable than you need by as much as you can afford and fit, both in capacity and current capability. That way it will not be stressed even as it ages, and will still have the capacity to do the job you need it to even as that capacity drops over the years.


Use the largest capacity cells you can, vs a bunch of smaller ones, so there are less interconnects to create, manage, and potentially fail. If an interconnect fails, and disconnects a cell or cells in a parallel group, the rest of the cells have to supply the whole load, which stresses them, and could damage them if enough others aren't present to share. It also drops the capacity of the whole pack by the amount those disconnected cells represent. If a pack is made of a single 1p block of large cells, an interconnect failure is obvious becuase there is no power up at all, no operation, so no cell damage can occur. (cell damage can lead to fires).

Also, larger cells are easily found in used-EV-modules from places like Greentecauto, Batteryhookup, etc.
(GTA often has free shipping on some pretty good EV modules). They come in a wide variety of capacities, cell types, capabilities, and voltages. Some are easily reconfigurable, and most come in a fairly sturdy framework if not a complete enclosure. Just make sure the specs provided are better than what you need it to do, and that it is a reconfigurable module if it doesn't already come wired up for the voltage / capacity you're after (but could be).

 

[SlowCo]

Interesting project. Hope you'll document it here with lots of photos. Good luck and enjoy building!

 

[TRS300]

Since it's the core of the whole system,... So, "always"*** use a battery that is more capable than you need...

Thank you for taking the time... I've read many posts from you on this site and appreciate your knowledge.

So here's what I've been thinking. And I admit it doesn't really align with more conventional designs. That said, number one priority for me is safety. Not only in protecting myself and others, but also in protecting the equipment. Performance is second to safety. But of course I want the build to work well.

I used to work in a technical industry where safety in design was always required. But there was also an emphasis on redundancy and often required distribution of resources. I'm retired now and the systems I speak of had nothing to do with EVs.

This being a DIY hobby like project I'm not building something highly engineered. And frankly motorcycle EVs are not very practical for many reasons but... Therein lies the challenge.

OK. So here's the thing... On the subject of batteries there is real benifits IMO to being able to have multiple battery packs distributed in different places on the motorcycle. Both for weight distribution and for redundancy reasons. I understand your comments about managing a safe system in the case of a component failure. And this build will have a custom brain that will be monitoring the overall system. So to try and limit the length of this response, my thought was to implement a monitored system and if a failure were to happen, the vehicle would be placed in a limited duty mode. Similar to a limp-in mode seen on many vehicles.

I guess the point I'm trying to make is I expect to use multiple battery packs that are distributed. And to add safety systems to ensure that a component failure doesn't result in an adverse condition that causes a safety issue or damages hardware.

 

[amberwolf]

On the subject of batteries there is real benifits IMO to being able to have multiple battery packs distributed in different places on the motorcycle. Both for weight distribution and for redundancy reasons.

With or without the limp-mode (below) if the systems are all equally capable of powering the system then it would be redundancy-capable and would still follow the guidelines I posted.

Assuming no limp-mode, if they're not all equally capable of powering the system then it's not redundancy-capable, and the caveat I gave about parts of a group having to take more load than they are capable of applies.

I understand your comments about managing a safe system in the case of a component failure. And this build will have a custom brain that will be monitoring the overall system. So to try and limit the length of this response, my thought was to implement a monitored system and if a failure were to happen, the vehicle would be placed in a limited duty mode. Similar to a limp-in mode seen on many vehicles.

It's not too difficult to do this, in principle. The limp mode could tell how many packs are operating, and decrease current limit of the system by the amount each inop pack would have been able to contribute. To decrease the current limit it would need to be able to communicate with the controller and change this on the fly.


Even using standard parts such as regular individually-BMSed packs that are simply paralled, you can install something that monitors each BMSes FET (or contactor) driver control, and if any of them stops engaging the discharge control, it sends a signal to the EVCM to lower the current limit to acceptable levels for the now-existing situation of however many packs are left. Failsafe is that if a signal stops occuring, it assumes the pack is inop.

There are also BMSes that have data outputs that you can monitor (CAN, RS232, BT, etc) which could even include monitoring per-pack current and cell-voltage-sag, but this is more complex and relies on correct data being available. Unless you can monitor per-cell data (voltage sag under load, the way the BMS does); the only failsafe available would be to monitor the BMS "alive" state with a heartbeat send/receive to make sure everything is still working, periodically during operation, and if it is not repsonding or doesn't provide exactly the requested / expected data, assume the pack(s) inoperative and initiate the limp mode. You wouldn't want to assume that the pack is not telling you there's a problem just because it doesn't have one.



You'll note that I generally don't worry about limiting the length of my responses. (even though it means many people don't read them)

 

[TRS300]

With or without the limp-mode (below) if the systems are all equally capable of powering the system then it would be redundancy-capable and would still follow the guidelines I posted.

Assuming no limp-mode, if they're not all equally capable of powering the system then it's not redundancy-capable, and the caveat I gave about parts of a group having to take more load than they are capable of applies.



It's not too difficult to do this, in principle. The limp mode could tell how many packs are operating, and decrease current limit of the system by the amount each inop pack would have been able to contribute. To decrease the current limit it would need to be able to communicate with the controller and change this on the fly....

I think I misstated when I used the term redundancy. It's not really applicable in this case. I think having multiple batteries being monitored for load should be sufficient to invoke some type of protective response. It would be nice if I could have BMS data transmitted via CAN bus. If I could get voltage, current and temperature for each battery on some type of bus communication I could initiate a system protective response based on the data. I don't really want to get inside the BMS operation, just want to monitor the batteries real-time.

 

[bananu7]

At least on ANT, you can talk to it via a serial port. You could build a simple controller that would periodically poll it for data and expose it on the bus.

 

[TRS300]

At least on ANT, you can talk to it via a serial port. You could build a simple controller that would periodically poll it for data and expose it on the bus.

Sorry. What is ANT? is it a BMS?

 

[j bjork]

Yes, ant is a brand of bms. This thread might be interesting to you too, even if you are not using vesc: Making bms's VESC compatible

 

[TRS300]

Yes, ant is a brand of bms. This thread might be interesting to you too, even if you are not using vesc: Making bms's VESC compatible

Thank you! Actually, I just had this delivered yesterday. I'm curious to learn more about it.