Kelly controller into a Z-20b
Anybody put a Kelly controller into a Z-20 or similar scooter yet? I'm looking to do this on my Z-20b.
Thanks in advance if you have info on this.
Someone put an expensive controller $500+ (from what I remember) into a Z20 at Electric Motorsport in Oakland, CA and supposedly the bike gets considerable more range and generally has better performance. I'm not interested in doing that myself, so I didn't dig any deeper.
I have a Kelly KEB72601 ($399) sitting in my basement that I'm planning to install this weekend, time permitting. I still have to get the parts and do final engineering on the circuit to adapt the standard brake input (+12V active) to the Kelly brake input (pull +5v signal to ground when the brake is active), but that shouldn't be too hard.
I emailed Kelly controllers before I purchased and asked what controller they would recommend, and if they knew of anyone using one of their controllers in the Z-20. Here's what Steven from Kelly wrote back:
"For 2500W 60V motor, KEB72601 is the best. Actually EVT America Z-20 is converting to our controllers. They tested KEB72450 and KEB72600, and locked on KEB72600 now. The test drive with 450kg (1000 lb) can get 50mph and climbing 3 hills in a row." (This was not confirmed by anyone at EVT America, despite my asking).
Considering the best GPS confirmed speed I can get from my Z-20 is about 32MPH (I'm 6'7" & 260 lbs, so considerable mass and wind resistance), that would be hard to imagine. The KEB72601 should be far bigger than I need for this application (133A continuous battery current rating, 6KW peak power), but I wanted to make sure it would be adequate if I ever change the motor/battery technology/system voltage, etc. From what I've read on this forum and at endless-sphere, the regen capability probably won't buy me much, but I'd like to try it.
I was hoping to get a cycle analyst from Electric Motorsport before I did the conversion so that I could compare power usage, but it doesn't look like I'll get that in a timely fashion.
I'll post my results when appropriate. Incidentally, I got the idea for the Kelly Controller from a post on this board by Astar. I contacted him, and he said he was hoping to install his Kelly in June; perhaps he'll be posting info as well.
I talked to Andy this week. (My new, replacement controller seems to have died). He confirmed they are going to a Kelly controller in the near future (within the month, he said). I asked him why this wasn't a problem before on the ones they had tested before production. The problem with the controller is that it's custom made for EVTA and the guy who makes them "just keeps tinkering with the darn think. He can't help himself". The designer keeps coming up with new tweaks and improvements. This has led to many of the failures. It's been driving Andy crazy. So he's changing to a supplier who can promise to produce a quality product. He's also planning on finding (or might have already found) a new supplier for their DC-DC converter.
Yes... eventually. You know how it is with Andy and "time". Here's how I understand things:
First, he's getting a shipment of controllers from China that have better mosfets (sp?). These are due to arrive any day now - this week, he said. Everyone who is waiting for a controller will get one of these.
But he doesn't believe this will be the end to his controller problems. So he's working with Kelly on a version of the controller that's tested to work with a z20. This controller will have a heat sensor that will speed-limit the bike if the controller gets too hot (a great idea, in my opinion). It also comes with regen capabilities but he's having them turn that off. He's got a version that he's testing now which he believes meets that standard. These are not in production yet, but he expects them to be "in a month".
But there's the issue I have with him about the perception of time. When he says something will be ready this week, I just don't believe him anymore. No matter what it is, it always takes longer than he thinks it will - usually two or three times longer. So when he says the batch of replacement controllers will be in "soon hopefully, this week in fact" I assume that really means two, maybe three weeks. And when he says the Kelly controllers will go in production in a month, I don't think that could be any sooner than the end of next month (July) - more likely the end of the summer.
We'll see. I'm going on vacation next week. When I get back, I do not expect to find a replacement controller waiting for me. Something is bound to happen to delay it.
Unfortunately, as far as the brake shutoff logic, the "Kelly standard" is "the" standard - coming from the larger e-bicycle community - while the street-equipped Chinese scooter "community" logically uses the 12V brake tail light signal as the brake shutoff.
I never much understood the importance of the brake-shutoff as a safety consideration. Are cars or motorcycles required to have such things? Who applies throttle and brake at the same time? Even in the event of a stuck throttle, the brakes can still stop the motor with a bit more effort.
Well, I guess I am still pretty lucky. I have a Z20a with now over 1100 real miles on it (over 2100 Z
distance units). My biggest vexation is still the speedometer. Twice I thought my controller was dead because I got no joy after plugging in the charger when the main breaker was still closed. The situation "automatically" resolved itself by removing the charger, opening the main breaker, letting
the bike sit for a couple days, closing the breaker, waiting for the be-de-beep signal, and then all was OK. Since then I have been totally religious about opening the breaker before connecting the
charger. I taped warning labels on every connection, switch, wire, etc. that has to do with charging. I also have an amateur band radio that loses consciousness for two or three days whenever it gets a static shock. It has happened many dozens of times and the radio has always spontaneously recovered. Who knows where computer brains are involved?
I do have one real problem now though, the horn no longer works. Tested the switch, it is OK. The
noise device still works because the controller beeps and the alarm beeps still work. Any ideas???
Also, I still occasionally get snagged by the fast launch controller shutdown "feature". Real pain.
I used to think it was a current limiter issue, but it occurs much less easily in warm temps than in
cold temps. Current limiting would be the other way around. I now think it is due to an undervoltage limit cutout (low battery protection) that is too aggressive. That low voltage needs
to be time averaged for a few tenths of second or maybe a full second or two before cutout is
invoked. I bet the Kelly controllers will get this right.
I am still using the factory Soneil charger, not bank charging or equalizing. Yeah, I know, but HarborFreight hasn't had a sale on round tuits lately, and I need a whole bunch of them.
I weigh 210 lbs., I get a top speed of just about 40 mph. In 70 F weather, I can get 20 miles range without stressing the batteries (gauge still reads above half, which probably means about 30%).
This bike is my daily driver and has displaced well over 50 gallons of gas now, and created oodles of plain old fashioned fun, not counting stupid drivers though. I really hope the brothers can stick it out and get these things refined enough to be a commercial success. When they are working, they are the best thing currently available to avoid learning to clean camel stables. This hub motor is really a sweet running machine, pure genius I think, if it can hold up for 100,000 miles.
So Andy does not expect the new controllers to solve the problems that Z/R owners are having, yet he's going to send them out anyway?
That's not what I meant. I think Andy expects this next batch of controllers to work much better than the ones before. But he's tired of the variance in quality that he's had with that company. So he's going to Kelly to get a reliable product from now on. I got the impression that he's been working with Kelly for some time. I imagine he didn't want people waiting indefinitely for that to pan out.
If you call him and tell him you'd like to wait for the Kelly version when it becomes available, he'll probably work with you.
I've successfully installed my Kelly controller into my Z-20a. I was delayed a week as I did receive my Cycle Analyst the previous Friday, so I spent the weekend of the 14th & 15th installing that. My Z-20 OEM controller (my second) failed on the following Thursday after getting two days of data on the old controller, so I had even more reason to install the Kelly.
I haven't gone more than a few miles with the Kelly, but so far I like it. Here are my impressions to date:
- Physically it's significantly smaller than the OEM controller. So much so, that I mounted it on top of a 1.5"x6"x7" piece of aluminum rectangular tubing to gain some height and a solid piece of metal for the heat sink (my understanding is that most of the heat is dissipated through the bottom of the Kelly controllers).
- Electrically, it took some work and a tiny bit of engineering to get it installed. The controls are wired through a single 14-pin connector, so I had to de-solder the connectors from the OEM controller, and wire it into the new connector. I also had to use a simple transistor circuit to adapt the 12v brake signal to essentially ground out the brake input on the Kelly (there are three switch inputs on the Kelly, all active low: brake, reverse and throttle enable). I didn't wire in the reverse switch or the throttle enable, and I didn't bother with an analog input for the regenerative braking for now. I simply set it to regen at 50% whenever the brake is activated (50% is the max you can set it to unless you use the analog input, which would require another hall sensor or potentiometer).
- Speed wise, the Kelly is capable of driving the Z-20 a bit faster when I didn't limit the current. Set to allow 100% current, I was able to get to 37 MPH, but I peaked at 113A current draw (yes, I installed the high-current version of the Cycle Analyst). Set to limit current to 50%, I hit a sustained speed of 35MPH at just above 50A continuous, and peaked at 83A during the ride (with the controller set for "Control Mode = Torque").
- As to current draw, with the old controller I saw a peak of 112A once, and >70A draw wasn't unusual, although 50-60 was more typical if I wanted to maintain top speed (about 31MPH). As stated above, with the Kelly set to limit current to 50%, I peaked at 83A on a four mile ride.
- There is no way that the Kelly will get my Z-20 to 50MPH at 60v. The terminal no-load speed (rear wheel off the ground, speed reported by the Cycle Analyst) with both the OEM and Kelly controller is 45MPH.
- The Kelly may be slightly more efficient than the OEM; the CA was reporting 74.4 wh/mi with the OEM controller on a 12 mile commute to work, and 71.1 wh/mi on my 4 mile test run with the Kelly. I'll know more tomorrow after I use the scooter for my commute.
- Regen gains me 1-2% of my energy back. It doesn't provide a lot of braking, at least with my bulk. I may get better regen as I learn where I can best use it other than for full stops (on a hill near my house, I was able to apply the brakes so that the regen was functioning, but I maintained a speed of over 30MPH).
- I really like the flexibility and programmability of the controller, although it can cause problems. I started with an overly aggressive low voltage cutout, and it limited itself to about 12 amps, which made for a tedious first test ride. I did about 5 test runs before I got to settings I was happy with.
As I learn more, I'll share whatever I think is important. Below are two pictures of the install; there are more, but I thought these provided a good overview.
Here you can see the aluminum tubing I mounted it on top of:
Jstibal - Thanks! Great info! I've been off line for months now on this forum, and am just getting back to it. I've had my Kelly KEB72451 since April ($337 including shipping and RS-232 cable), but have not installed it since I've been too busy. I'm just starting to work on that now.
I have not been riding my Z20 except VERY short trips because I don't trust the controller. I'm a bit of a newbie to this stuff, and I might need some advice. I was planning on using a relay to tie in the brake activation and another relay for the throttle enable (tied to the kick stand switch), but it sounds like you did it with a transistor. Here is a picture of what I'm thinking of:
Here is a picture of the naked front of the Kelly controller:
How do I connect to power cable to the blade connectors and still provide some insulation to avoid having a short (B+, B-, A/B/C)?
Which phase wires go to A, B & C? The OEM phases are color coded, and I'm not sure how to map them.
I just got a message from Steven at Kelly controllers - they been pretty good at replying in a timely fashion. He's talking about the controller that EVTA is going to use.
"The controller was settling down on KEB72600 (non-regen).
It works well. It drives 450kg (nearly 1000 lb) gross weight
to 75km/h, with good climbing capability."
jstibal, A few more questions:
- What kind of launch characteristics are you getting with the Kelly installed? I hated the open throttle/stall quality of the OEM controller.
- Did you use any of the Kelly amp limits? With the cycle analist, you probably don't have to, but I don't have a CA and I don't want to fry anything. I e-mailed Kelly to see if they have a recommendation.
- Did you upgrade the wiring on your Z20? What size?
For wiring the phases, I used A=yellow, B=green and C=blue. I determined this by putting the scooter up on its stand, disconnecting the 3-phase power wires to the motor, turning on the scooter then watching the voltages on the hall sensors as I rotated the tires. If you do this, you'll get a sequence of high/low voltages something like
Yellow Green Blue
4.26v 4.26v 0.28v
0.28v 4.26v 4.26v
4.26v 0.28v 4.26v
4.26v 4.26v 0.28v
(This is all based on a post by fechter out at endless-sphere.com)
There are half steps in there as well where only one hall sensor has a high reading, but you get the point. You're OEM controller and motor should be wired the same, but who knows, you may want to check it. Below is a picture of my KEB27601 with electrical tape for color coding the phases (incidentally, yours looks more like a KBL model than a KEB since it has two of the 14-pin connectors; maybe because it was an early model?)
For insulating the power and phase connectors, I just taped the B & C phase connections once it was installed. Those are the only two that aren't obstructed by wiring or the bike itself, and I use taped or insulated tools just in case.
- What kind of launch characteristics are you getting with the Kelly installed? I hated the open throttle/stall quality of the OEM controller.
- Did you use any of the Kelly amp limits?
I never experienced the open throttle/stall condition on the OEM controller, but I was always easy on the throttle to save battery. I've got the Kelly set to limit current to 50% (I don't use the limiting feature of the CA), and take off characteristics seem similar to the OEM controller with a peak current of 83A for the Kelly. The OEM peaked at over 100A. (the Kelly peaked at 113A when I had the current limit set to 100%).
- Did you upgrade the wiring on your Z20? What size?
I upgraded all of the high-current DC wiring to 6 gauge with soldered on lugs. For the motor phase wiring from the Kelly to the wiring block I used 8 gauge since the motor wiring is 8 gauge, the runs were short and I needed more flexibility.
I was planning on using a relay to tie in the brake activation and another relay for the throttle enable (tied to the kick stand switch), but it sounds like you did it with a transistor.
Yes, I used a very simple transistor circuit for the brake input. Here's a schematic of the circuit:
The resistor is small enough to solder directly to the base lead on the transistor and fit the resistor and the lead in a single piece of very-small diameter heat-shrink tubing. The other two leads on the transistor are individually protected by heat-shrink tubing, then the entire assembly was put in larger heat-shrink tubing as well, then slipped inside a piece of the cable ducting:
It makes for a nice small package that should be extremely immune to bouncing and vibrations.
(Incidentally, the resistor probably isn't technically necessary, but I put it there to limit current from the OEM brake circuit in case the transistor should ever short out for some reason.)
I haven't done anything with the kick-stand switch yet, but what I'm considering doing is to tie it to the throttle input. The switch is closed (circuit tied to ground) when the kick-stand is down, whereas the Kelly expects it's throttle control input to be tied to ground when the bike is ready to go (kickstand up). Rather than engineer something to reverse that, if the kick-stand switch is tied to the 5v throttle input (through a suitable current limiting resistor), then when the kick-stand is down, the throttle input gets tied to ground and the bike doesn't move. Much simpler than anything else I could engineer.
Disclaimer: I am not a professional engineer, nor do I have previous experience with scooters, so please don't consider what I've done or am considering doing as having any express or implied guarantee of any kind.
(Updated at 15:23 2008-06-24. I corrected the circuit diagram, which erroneously stated that I was using the analog brake input. I am using the brake switch input, not the analog input.)
Yes, Mine is an early model. After I ordered, they were still making changes and were not ready to ship it for a few more days. They told me it would not have J1, but then it shows up with J1. It will be unused for me - nothing on it I want to use anyway.
I'm going ahead with using relays for the brake and kickstand. I'm also going to use a pot to control the regen level. I plan on adjusting the level until it's where I want it - I won't have easy access to the pot. I'll let you know how that works out. I hope to be done in a few days.
Thanks for your help.
Just to throw this out there I have found I use about 160w per (real) mile to recharge. That is using bank charging and running about 6 miles round trip. It has been pretty consistent at 1000w for 6 miles give or take about 50 watts. I do plan to take a couple trips more economically, as in don’t take off as fast and glide to the stop sign to see how much it makes a difference.
I just finished programming the controller. My computer is upstairs, so I took two batteries out of the scooter and used them to power the controller while I had it connected to my computer. Here are the settings I used. Let me know if anything does not sound right. I'm not really sure what everything is; the ones in bold are my bigger questions. I'm also emailing Kelly also to see what they say.
Throttle sensor: 0 – 5v
Throttle Low dead zone: 20%
Throttle High dead zone: 80%
Startup time: 0.5 sec.
Hall sensor type: 120 degree
Control Mode: Torque
Under Voltage: 50v
Over Voltage: 75 v
Throttle up/down rate: 3 (range is 1 – 20 where 1=fast and 20=slow)
High Pedal disable: Enable (what is this?)
Top speed: 100%
Motor Poles: 8
Regen Brake Switch: Enable
Regen Current by Brake Switch: 20%
Max Regen Current: 100%
Brake sensor type: 0 – 5v
Dead zone low: 10%
Dead zone high: 90%
Motor Temp sensor: Disable
Now to create an interface box for the kickstand, brake & regen.
UPDATE: Kelly replied quickly, and confirmed above settings, but had me lower the under voltage setting from 55v to 50v (already edited into above values).
Wow, another fast reply from Steven at Kelly Controllers. I just asked him minutes ago about the number of motor poles, and also what the "high pedal disable" is. Here is his reply:
Actually the motor pole number doesn't affect operation. It only affects calculated speed, if somebody got RS232 real time communication and watch the data...
"High Pedal Disable": The controller takes it as fault if power up with throttle applied. Driver can just release the throttle to clear the fault. So motor will run if throttle reapplied. It is a safety feature. You know, it won't be fun if the motor takes off once you turn key on...For your scooter you may enable it.
Wow, this is a great post with lots of useful information.
I had my controller fail about 5 weeks ago at about 350 miles on a 90 degree day. Fortunately, I was only about a mile from home when it happened.
I have sent 3 messages to Support at EVT and still haven't gotten a reply. So, this post tells me a lot more about what is going on than I'm getting from EVT. Never the less, I love the bike but want to be able to ride it again.
The controller is installed. I tested on the center stand and everthing checked out. I haven't yet connected up the kickstand switch to the ignition, so that will be another project, and I want to get the fans connected up too. I took it for a short shakedown ride and so far I like it much better than the EVTA controller. It seems to have better acceleration and a smoother throttle response. I have not checked out the top speed yet. Regen is set for 60%, which seems about right; it makes some noise when doing regen which made me think the brake interlock was not working at first.
You can see the black interface box below that houses the relays and pot for changing the regen level. The clunky EVTA controller is in the background.
I am just about to do the same with my E-Moto (e-max clone). I have had two controllers blow up on me in a year.
Thanks for all the great info.
[Readers note: there is some install info on the Kelly controller in one of the E-Max column posts]
I am going to call Steve tomorrow to order.
How easy was it to hook up the regen? I am trying to keep everything as simple as possible but would consider the regen if it is easy to hook up and actually impacts the range in any significant way.
Regen is really easy to hook up as long as you have a brake signal going to the controller. The brake switch input will activate regen. I used a 1K resistor to be able to tune the amount of regen to my liking - also pretty easy to do but I think you can get up to 50% regen programmed into the controller without any hardware connection. No special wires or connections to the batteries - it just used the power leads. Not sure how much benefit regen is but based on other posts it's minimal. Good luck and let us know how it goes.
Went on a speed run with fully charged batteries and got to an indicated 49 mph. I'm guessing this is actually 38 mph or so. The scooter feels a lot better to me with the new controller, and I now have some confidence that I won't end up stranded because of a controller failure. I think the top speed with the EVTA controller was around 43 mph indicated, so a clear improvement. Even more important to me is the improved throttle response. At 20 mph indicated it will snap forward with acceleration, whereas with the EVTA controller it just wanted to wind up. This is how I expected the scooter to respond out of the box (only a little faster top end).
One problem I had with my installation is that the kickstand being down did not disable the throttle. I tested the signal, and my circuit is grounding the throttle enable only when the stand is up, which is correct. So I contacted Kelly, and they told me I needed to set "Footswitch = Enabled" in order to get the behavior I want. Mildly irritating since I had them check all of the settings before I installed it. Now I have to remove the controller and two batteries in order to re-program it. Overall, I'm still really happy with Kelly's customer service though.
Question: Would there be any interest in a topic in the EVT forum with detailed information on how I installed my Kelly controller? I've been having a private conversation with one forum member who would like more detailed information on what I did and how I did it, and if others are interested I could try to break down all the necessary modifications I did in greater detail. There's lots of good info in the posts in this and other topics, but what I'm proposing is a new, distinct topic where I detail step by step what I had to go through (maybe astar and others could do something similar). Because of the format of these forums where topics cannot be edited but replies can (at least to the best of my knowledge), I'd do a new topic with an brief initial overview, then post separate replies for each major subcomponent/subcategory (disassembly, contactor installation, battery interconnects, controller diagnostics, controller wiring, controller installation, etc.). Then I could correct mistakes within a section so that readers wouldn't have to read the entire topic to find relevant information that was posted at a later date.
Would that be worthwhile, or a waste of space? Or is there a better way to communicate this information?
I think you could add a manual in the EV Download Library section of VifV. They use pdf format I think. I would be happy to review and add comments. I don't want to have a separate document. Then you could have a thread to discuss and give feedback on the document, and incorporate changes as appropriate. You are welcome to use anything I documented in the "Z20b teardown" thread. I think we have done similar things to our scooters, although we approached the throttle/brake signals going to the Kelly differently. You might want to contact a forum moderator to see how to proceed. I'm sure there are a lot of Z/R20 owners that would benefit from the documentation. Possible topics:
- Kelly controller replacement
- EVTA controller replacement
- Taking off the covers
- Rewiring with larger gauge wire
- Diagnosing throttle problems
- Replacing the throttle
- Adding a contactor
- Adding bank charging
- Installing a Cycle Analyst
- Installing a PakTrakr
- Silencing/quieting the turn signal
- Fixing the rear turn signal attachment
- Fixing the kickstand sensor.
- Diagnosing controller problems
- Diagnosing DC-DC converter problems
- Bypassing the alarm system.
- Bypassing the controller thermocouple and running the fans off of the ignition.
I've done most of the above. I've probably forgotten a few things. What else?
Does anyone want my old EVTA controller? It's a replacement controller - no thermocouple. It has not burned out (yet). I don't want any money for it, just pay UPS shipping charges. Double your money back gaurantee (shipping excluded) }:)
I was contacted via PM by someone local and so it's already spoken for. I will contact you if it does not happen for some reason.
Does anyone know what is failing on the OEM supplied controller for the Z/R-20's? Is it just the MOSFETs burning out? If so, how difficult is it to have these replaced (hopefully with higher quality components)?
Has anyone opened up their burned out controllers?
I'm still waiting for a reply from EVT (5 messages and over a month now with no reply) but I'm losing hope they are going to help, so either a repair or Kelly replacement seems to be the only thing that's going to get my bike moving again.
If anyone has successfully repaired their controller, I'd like to know. If so, perhaps a new thread would be in order.