EVT "asian" controller

Thanks Patrick

The current shunt I was using was overheating (homemade - it seemed to be discolored after riding awhile) -- that would make it over-read.

I am looking to get a 10% increase in speed just to keep up with the traffic in the midsized city where I live. Therefore, I'm looking at the pros and cons of either a 60 V mod or a controller bypass. If it ever warms up in this bloody country (it was -12° here this morning) I will get out and measure the output from the controller at various throttle settings. If the duty cycle is around 80% at WOT as previously mentioned, I can get a 20% increase in average voltage just with an override. That might be my best bet then.

I would probably just splice in a DPDT 120 amp relay that would be activated by a microswitch mounted inside the controller that would only come on at full throttle. The motor would be fed either controlled (default) or direct 48 V from the battery.

Question -- would it harm the controller to have the output suddenly go from full load to open circuit? If the relay works the way I plan, It would just switch the motor from the controller output directly to the battery output and leave the controller output open-ended during "turbo mode". Would this harm anything?

Thanks, Rick

Jerome:

The harness taps in place between the red anderson connectors, and supplies a 48v lead for the logic side of the bike. There are two long cables that go to a 5th battery (same size as stock) under the seat. It's that 48 v that runs the bike and DC/DC etc. Yes, the battery gauge still works (though of course it is only reporting on 4 of the 5 batteries). It's completely connectorized so that you can turn the bike back into stock just by removing the harness; it's all plug and play. No soldering or heavy crimping required. You will need to crimp on two small bullet connectors in your bike. We sell the harness for $60, an EB-50 for $105. You can call Electric Motorsport at +510.839.9376. I would recommend the Soneil 1214S charger for the 5th battery. It's a 7 amp unit, and so matches the stock Chroma charger. We currently don't have these in stock, but they will also fit under the seat with the extra batt. If you elect to mount it there, do watch the cooling. You may need to leave the seat up, depending on ambiet temp. etc.

Crusher:
I may not be in the rest of the week, but your harness and EB-50 are ready to go. You can pick them up, or give a call if you want them shipped. For the rest of the folks reading this - I made an extra, it's waiting on the shelf if you want one.

Rick:
I'd be surprised if the duty cycle was only 80% - but then, I've been suprised before; you've got a good idea. Wish I'd read your email earlier - we've got a 4000 torn down for service, and it would be easier measure that way. Of course it's up on the stand, so wouldn't be the same as under load, but I would think the duty cycle would be the same. Given all the speculation on this, I'd like to see it on a scope. I'll let you know as soon as I have done that.

I would be very cautious using just a relay. You'll be switching at full current (which could be more than 120 amps), which can cause the relay contacts to weld shut at full throttle. As you'll be bypassing the controller, neither key nor kill switch will shut off the bike - it could be the start of a very bad day. I would use a contactor with magnetic blow-outs and a separate kill switch for safety. You'd also have to make sure you didn't accidentally hit the switch while at less than full speed.

I don't know if a sudden open would hurt the power output stage. Since there is no regen, I'm guessing it's just a single sided switch, and that it would probably be OK. But I am completely guessing.

Patrick

Well, of course you guys were right. The U.S. model of the controller peaks at 82% duty cycle. It switches at full rail (i.e. battery) voltage, and what you are measuring is the average DC value. It switches at 15.67 kHz. Here's a simple shot of the controller output:

The 42 V is the average voltage. You can see that (at 20 volts per division) the battery votlage is about 50 volts.

Then I ran the motor while the charger was connected and on; you can the rail votage go to about 60, while the duty cycle remains the same.

You can also see the inductive kicks at the beginning and end of each cycle. For an even beter view of that:

Notice that the kicks have the same amplitude both turning on and off.

Ouch! When that last image scaled, it got pretty fuzzy. I'll see if I can't clean that up.

Patrick

Before I had a chance to read the forum this morning, I made this same measurements on my USB-based data acquisition board. Although the acquisition rate was a bit slower, I got almost exactly the same results as you. The only difference is I measured about a 2 kHz pulse frequency. I wonder if I was getting some aliasing.

Anyway, it is interesting to note that by bypassing the controller on these limited US models, you gain about 9 V -- almost as much as an extra battery but without the added weight. I think for myself with this limited controller, a bypass is the best way for me.

Thanks for the tip about the relay -- that might be the only time I've ever wanted the scooter to go slower. I am not familiar with a contactor. Is it similar to a relay? Do you have a recommended supplier or can I purchase one from you? I wonder if you could get away with just switching the positive side of the battery past the controller to the motor input (i.e. using an SPST type relay or contact to just to dump 48 V onto the output line to the motor without disconnecting the controller output to the motor). Any ideas? What is the normal way to bypass the controller? I assume somebody out there must've done this?

Thanks, Rick Retzlaff, Saskatoon

PS -- I just learned something about posting on this forum -- don't do it live in the window -- I just lost all my work and had to reenter it.

Craig Uyeda, Deafscooter is Here...
I read your Speed Controller it not help to booster

Deafscooter
have own asian speed controller on my EVT S.E.
with asian speed controller on my scooter ....

and stay use 48 volts asian speed controller on it

I wanted put picture of my asian speed controller
i dont sure if the topic can easier to post the pix...

Here is an interesting page I came across of some Asian EVT’s with lithium batteries. According to a rough translation they say their scooters are getting speeds of 50mph and ranges up to 120 miles per charge.

http://www.ibike.com.hk/01_bike_report/local/others/evt_scooter/evt.htm

Keith
2006 E-Max sport @ 60v

I've looked at a lotof motor designs and raraely see any inductive elements in the output stages. That is important because if there were inductors there and some leakage which allowed current to flow then you have to find a way to dump the charge.

From the photos above it seems there is little else but the Fet's, and possibly a shunt (?) in the output circuit. Technically speaking, with a little bit of past experience , switching the Fet output of most speed controllers does not harm them.

Without completely reverse engineering the output stages of the Asian controller it would be difficult to say for sure but the reasons to use inductors in the outputs are quite specialised and not applicable to this sort of application, and this design does not seem to have any major ly inductive components on the PCB

It would be best to isolate the controller output from the motor completely when bypassing it, rather than shorting over the outputs. A good high current relay/contactor for this application could be a split charge relay rated for about 200Amps, with a 12V coil, or even a starter motor contactor/relay. The only downside is that these use quite a lot of current in their coils, but you may find that you use less current overall to travel the same distance at the higher speed.

Another possibility would be to turn off the motor controller input supply completley so that the full battery voltage can be put ont the motor with the speed controller left in circuit.

Bypass of an output is common on higher current drives, such as Curtis speed controllers used in golf carts and so on, and to my knowledge there were some version which still gave a PWM output signal to the output FET even when the bypass relay had shorted both sides of the fet togeteher, with no ill effects.

I mentioned relays first I think, and suggest that anyone tinkering with this idea should use 'contactors' instead, as suggested by someone else. I only used 'relay' as a generic term.

Good luck,
Chris Barron

http://journals.aol.co.uk/fixitsan/electric-car-building/

Craig,

I wanted put picture of my asian speed controller i dont sure if the topic can easier to post the pix...

How can we help you? V is for Voltage Forums Help/FAQ. Post questions about the operation of V is for Voltage Forums Site.

Also check out our Help Tutorial for help posting images. What Browser do you use?
If you require any additional help please contact support [at] visforvoltage.org.

Stleride
Moderator Team Captain

Moderators are community volunteers who help keep V is for Voltage Forums running smoothly, and provide forum support.

Some general observations:

The stock US EVT will do about 30 mph. The Asian model will do about 33 mph. A US model at 60 volts will do about 38 mph. Using an Alltrax at 72 volts will get you up to 55 mph (for a little while, anyway). Deafscooter will tell you that his SE model will do 55 mph at 48 volts - but it's just something he lucked out buying.

The inductor in this case is the motor. This is part of PWM design - to use the motors inductance to integrate the current; if you look at the motor current it's about a sine wave.

A contactor is a very heavy duty relay, made to switch large amounts of current. The contact throw might be 1/4 inch (my apologies to those who use units that actually make sense) or more, and the surface area is large. These are usually SPST designs. Drive coils are available in the 12 to 48 V range. A 180 amp unit might be $50 - $90.

Opening the outputs on the FET stage is probably OK. The reverse (changing from full batt back to the controller) may not be true. You'll be running the motor at battery voltage, at whatever current is required given your terrain. I'd like to measure with pure DC, and get Kv and Kt for the EVT motor - but it could be a substantial amount of current.

Now, you're going to switch that back to the controller. As soon as you disconnect (contacts open) there is going to be a large inductive spike generated by the motor (V = L * dI/dT); that is going to presented to the FET stage when the contacts close. I don't know how it's designed - it may or may not take that spike with no ill effects. I suppose you could put a free wheeling diode in to absorb that load - put it on the motor side instead of (or actually in addition to) the controller side.

I wouldn't go the bypass route. It won't be a simple splice job - you'll be dealing with 6 gauge cable, probably more than one contactor, a fair amount of wiring, and still have the safety concerns. Not worth it, IMHO.

I did look for that spare EVT controller. Hah! Probably take two guys the better part of a day to shift all the stuff. I will ask Todd - maybe he stashed it somewhere. Just looking at pictures - it's a 28 pin chip? There are two things here - which chip family is it, and do they use copy protection. I checked into that, and Atmel has had very good copy protection for 10 years. If they used it, we're SOL. If not, then we can read the chip. I know a guy whose good with the ABR series - we can find others for the other families. I wouldn't try reverse engineering - that can be tons of work. A simple copy should suffice to get the increased duty cycle.

I'd like to know what the Asian controller measures on avg V to the motor at full throttle - you out there, Lemmiwinks? They may not go to 100% duty either.

Rick - yes, you need to increase sample speed. Either that or you've got some big L/C loads in the measurement gear.

Patrick

A big 'Thank You' to Patrick for meeting with me yesterday when I picked up the 60 volt mod harness and extra EB 50 for my EVT 4000e. We went over the install procedure and I took some pictures of a 4000e sans body panels on the test stand. It looks like this mod can be done without removing the body plastics. If so, it should be easier than installing PowerCheqs. I still need to pick up a tie down strap for the extra battery and I have a 7 amp Soneil 1214 charger on order, so I'll wait to do the install until I have all components. I'll document the process and post to this forum.

Cons:
* Loose under seat storage on the 4000e
* Add fifth battery to an already heavy bike (thereby raising the C.G. even more)
* Requires additional charger
* Possible accelerated brush wear (Electric Motorsport has replacement brushes)

Pros:
* Cheaper than an ungoverned 'Asian' replacement controller
* Twice the speed boost of an 'Asian' controller
* Simple, elegant solution
* Fully reversible
* Available now
* Looks like a 'dead simple' install
* Proven in the field

-Crusher300
Silver EVT 4000e San Mateo, CA

Patrick,

this may seem like a naive question, and you guys probably thought of this already: Have you all at Electric Motorsport, as an authorized EVT dealer, ever asked the folks at the factory in Taiwan if they would sell you the PROM chips for the 'Asian' controller? Would it be possible to buy them as a service part? Sometimes the most direct route is the easiest. Just a thought.

-Crusher300
Silver EVT 4000e San Mateo, CA

I agree with the points about switching back from the bypassed state to the controlled state, and did make some assumptions, importantly here that the controller doesn't have regen.

Not knowing anything abotu the motor makes the situation guesswork, it may be that as soon as you step back from bypass to pwm control that you introduce a large braking action due to the back EMF now flowing.

Bypass is a commonly used technique in a lot of controllers.

There may not be 100% pwm on an unlimited model. From the design viewpoint just by limiting your maximum PWM to 90% you can de-rate your output fets significantly, and hence produce a cheaper alternative for not much of a reduction in performance.

So to the controller IC, If the power pins are known, as well as any connections to crystals or resonators it would be easier to identify who made it.

If it were even just possible to identify what the various connections to the microcontroller consist of then duplicating it with new code, possibly even new features becomes straightforward. There would be no need to know the code on the chip in the first place, just throw it away and start again from scratch !

Chris Barron

Hi All

Interesting notes about bypassing the controller and the problems there might be. I agree there is some significant issues about switching 100 amps in and around inductances. However, here's a random thought that came to my brain: until somebody could get the IC backward engineered and re programed, this chip has to be sending out a TTL level PWM pulse stream. How about just finding this pin (should be pretty easy to find by just snooping around the board a bit with a scope) and overriding the control signal rather than overriding the output. It seems like an awful shame to ignore a perfectly good controller when it exists. Rather than the brute force of bypassing the output, what about bypassing the input -- feeding the FETs a continuous high level in "turbo mode".

Anyway, thoughts on this?
Rick

I stopped myself from suggesting that because at the moment it is unknown if maximum speed comes as a result of 100% PWM, a DC signal in other words.

The FETs have to dissipate power, and with, say, a 90% maximum duty cycle the fets can be over-rated because they will never have to withstand the continuous current. If, and only if, this controller is designed that way, driving them with continuous dc could cause them to fail by letting the smoke out.

If someone could confirm that the non-regulated controller does in fact run at 100% reliably then the best solution so far would be to drive the gate continuously as you suggested.

Chris Barron

So i know I have a non soldered atmel chip sitting right in front of me, it came off a "48V" panterra electric freedom controller that came with the scooter. burnt to a crisp I managed to get AT89 series chipset numbers and started digging, turns out that their is a way I can make a reverse heat sink with one of their chips and still program it blank, they supply the printouts too haha' anyways, I'll get back to you with more info, but im no engineer but i'm hella bored and will come up with something. You guy's are damn resourceful and i'm proud to be a member

I wonder how many of those guys still have the scooters. But I love it when such threads come back to life so I see some of this.

So how does one go about becoming more elecronically included so we can attempt some of this?