New Kelly KBL72401E

Actually, come to think of it, the controller can be programmed on my desk - and my antique desktop PC has a an old serial port - I'll probably be able to scrape up an ordinary serial cable somewhere.

This brings up an issue - when I ordered my first (Chinese)e-scooter, I had expected the motor/controller system to be as quiet as the electric vehicles I was already familiar with - in my case metro (subway) cars or streetcars ("trams" for you Ozzies) - they may make a high faint whine pulling away, but the motor is completely silent - just wheels-on-rails.

So, I was disappointed at that crude rough growl the e-max and later, my C124 make. I had expected to only hear tire and wind noise and nothing else. I hope this controller produces the desired result.

When setting the configuration of the new Kelly controller, I encountered something the people at Current told me nothing about. The last settings - after the controller temperature shutoff settings are something called "smooth setting" these settings include:

Smooth Enable/disable, inflection point, low accel, low decel, hgh accel, high decel. The defalut settings are "enable", 10, 100 127, 100, 127.

Matt or anybody, do you know what these settings do? Thanks.

Well, here is the result of the test ride...

I configured the controller settings per CuMoCo's recommendations. I left smooth settings where they were. CuMoCo told me the the scooter would be quieter and have noticeably more starting torque - I noticed neither, in fact it was producing a bit of a "sputter" upon starting off.

I tried enabling the "noise reduction" setting. The result was less starting growl and less of that sputtering, but at the expense of a some starting torque.

I then tried disabling the "smooth setting". Oddly once disabled, re-enabling it did not bring back those five settings on the configuration window.

I then took my third test ride - still no performance improvement - in fact it was probably a bit worse than the old controller. But at least the regen was working properly, and the old controller's annoying bbpppbbbpppbbbpppTT! from the motor upon stopping was fixed.

But then - exactly coincident with hitting a moderate bump in the road, the controller stopped completely, aside from a some sputtering noises from the motor. No error flashes from the controller or BMS. I smelled a slight whiff of something hot or burning. I was able to coast the scooter down hill most of the way home then a hell of a workout pushing the scooter uphill to the house.

So, I guess we can assume that Kelly still has some QC problems - but now I have to deal with a someone in China instead of nearby Ann Arbor, USA. I regret not insisting that Current sell me the controller. I suspect that Kelly, like so many Chinese manufacturers, probably practices lots of QC for products going to OEM's with a nice contract, but they sell crap to retail schmucks like me.

Time to put the malfunctioning - but at lest operable old controller back in. What are people's experience with Kelly's warranty service?

Regarding Sevcon controllers, supposedly John, formerly at Current, was hoping to use Sevcon controllers in their scooters, but they would not return his phone calls or e-mails.

I strongly suspect the "low accel, low decel, hgh accel, high decel" settings will simply limit current to the set values, and the "inflection point" is the point at which the low and high sections converge.

What current would there be for deceleration? I got a response from Fany (male?) It follows:

Hi Paul,

We upgraded the software version to 0404 for all KEB/KBL controllers.The hardware is upgraded also.

We added the smooth control parameters in the user program.
When you enable the smooth funtions,you will see five more parameters.
You can adjust the low speed acceleration and decleration rate,and high speed acceleration and decleration rate in the user program.
The inflection point is at the joint of low speed and high speed.
But this new software version needs the new firmware and hardware to support.
Sorry,it only can be useful on the new controllers which are produced after July.2013.
Your controller can support this new functions.

Usually you don't have to configure it if you think the current setting can drive the motor smoothly.

Thanks,
Fany

Clear as mud. What does this physically mean? Deceleration and acceleration in any kind of motor vehicle is controlled by the human-operated throttle and brakes, what does it mean in the sense used here?

MEroller,

The controller was most definitely bad. The motor has two sets of hall sensors to provide a backup - and it malfunctioned with either set connected. All connections were re-checked and good. I put the old controller back in and it runs - albeit still with the problems I bought the new controller to fix.

Note that the controller drove the motor poorly (sputtering and very noisy at low speed) even before it failed entirely.

I am getting a replacement - I only had to send the old controller back to Steven Li in the Los Angeles area, I would already have had the new controller on Friday had I not forgotten to leave a signed that it is OK to just drop the package at my door. The delivery from China, via DHL Express, is incredibly fast - It goes on a direct nonstop flight from SE China to Cincinnati, Ohio - and then immediately on a truck to Pittsburgh and is ready to deliver to my house 5 hours later.

I'll report how it works.

Time for an update.

The replacement controller, this time -without- waterproofing, is in and working fine. Kelly seems to be balking at refunding the waterproofing option (even though it was the probably cause of the failure) - claiming that they shipped the replacement it for free, so I should not be complaining if they pocket the option. (Chinese business practices are sleazy and enraging)

The regen issues are fixed; however, using the following settings that Current recommended:

Torque mode
77% motor current (308A)
60% pack current (240A or 4C)

yielded little or nothing much in the way of performance improvement. Curiously, changing the motor current setting, even raising it to 100% yields no low-speed torque improvement - even though, presumably, "torque mode" would mean the throttle controls torque - i.e. motor current - so wide open throttle should produce current up to the set limit.

Today I decided to fiddle with the settings some more - and fiddling is the only thing one can do becasue Kelly provides zero documentation on controller configuration. Kelly controllers can be set to operate in torque mode, speed mode, or "balanced" - presumably some programmed blend of torque and speed control. I decided to try "balanced" mode and 100% motor current. Wow! Much better! It had scary amounts of torque especially at low speed - accelerates from a stop up 15-20% up-grades like they are 15-20% down-grades. I was afraid all that torque and current might break something. After some trial and error, I settled for 85% max motor current.

I still don't have a good understanding what the five 'smooth" settings do, and with five settings to permute, it would take a couple days if I have to do it by trial and error. So I will leave them on the default settings.

Lessons:

1. DON'T get the "waterproofing" option. It appears they accomplish by simply filling the case up solid with potting resin (it was as heavy as a brick). I'm sure everyone can imagine the problems with this approach unless the thermal properties of the potting were very carefully chosen. Simply sealing the terminal and connector end-plate intrusions with silicone provided adequate rain proofing for most applications.

2. Don't use "torque" mode if you want access the full controller current at low speed. I'm not sure why, but the Kelly controllers are programmed to be stingy with current at low speed in "torque" mode, but not in the other modes, where they will allow the full max current you set it to until the throttle-determined speed is reached. The down side is that it might be harder to ride for maximum range and economy. I'll determine that in the next few days.

3. Kelly badly needs to provide configuration documentation in understandable English.

What mode was used to make those graphs. Balanced?

This brings up the issue - why did Current, specifically John Harding (who left the company under some kind of presumed legal settlement gag-order on commenting in this forum) choose torque mode over balanced mode - even though balanced mode is clearly better? John seemed to be always worried about exceeding the temperature limits of the Quanshun motors, so he programmed two-step limits on the maximum throttle signal to the controller based on motor and motor lead temperatures). These limits were initially quite restrictive and would trigger a low power mode where the scooter would hardly climb a hill at all after 20 km of riding on a hot day, but were since relaxed considerably as they found out that the motor can tolerate higher temperatures than originally thought. The temperature limits are user-configurable via Currents BCU software but I have not changed any settings. But obviously, changing the controller mode is going to change the effect of the BCU temperature settings. It's been a fairly cool summer here, so I am not hitting the temperature limits to find out.

And on another issue, I always had a suspicion that the test pilot BCU uses an open-source (Arduino or other) micro controller hardware and programming. I have not opened the BCU case to find out. If this is the case, aren't we owners entitled to the source code for the BCU? I'd like to get rid of the way the BCU program shuts of the regen at below 10 mph.

We are also long overdue for wiring diagrams for the scooters, but I'm drifting way off-topic.

On my scooter, the torque above 30 kph in was always pretty good, it was below 30 kph, and especially the first 10 kph that was a problem - especial if having to make a left turn from a stop through a gap in fast traffic while pointing up-hill. It was character building. But now, it will will leap forward from a dead stop and the throttle is a bit twitchy. I definitely needed it to be in torque mode when taking the motorcycle license figure-8 test. There is still a narrow speed-band of annoyingly loud growling and vibration from fast walking speed to 20 kph or so. I have the "noise reduction" activated, and it works well - but only for the first 10 kph at best - it seems to be deliberately designed to "switch-off" right when the noise reduction is most needed.

And all this time, I, and I suspect most other current owners, are thinking: "Well, this electric vehicle technology must not be sufficiently mature yet - I'll just have to put up with the sluggish performance as part of 'doing the right (green) thing'", when all along, it was just a matter of a poorly-documented (in a western language) software setting.

And I just noticed you have a 800 amp Kelly...

Mike,

And while you are at it, it is probably safe to just set the max motor current to 100%. I'll probably do it this weekend. You may recall that I always thought John was being way too over conservative and nanny-like regarding the settings.