ME1616 motor with sevcon gen4 size 6 80V low power??

[laapmetot]

Hi guys,

I did build a E-kart with a ME1616 motor with Sevcon gen4 size 6 80V. I have 5kWh of watercooled batteries at 86.4V nominal and we can use 500A easily.

I am surprised how amazingly user UNfriendly the sevcon software is. I was warned, and they did not exaggerate..... Anyway I got it to work, can access the helper, configure the controller in a way the basic functions are working and the motor is aligned and spinning. It revs up properly, no weird sounds and more or less ok.

However, quite disappointed by the low power output. It feels slow, acceleration is modest at best. I see phase amps of 450A, and battery amps up to 200A. At 90V that would mean we get roughly 18kW of the battery pack. At a dyno we measure peaks of 60Nm only which drops fast at higher revs, and see around 12kW of peak power. So lack of power is confirmed, but fiddling with the controller settings didnt really give any improvement.

I read a bunch of stuff here, but it didnt solve the problem. What I did:
- checking if full throttle is achieved. it is.
- Kick out all power reductions for temp-voltage to avoid some unseen reduction
- Motor current setting is 120Nm, it doesnt reach that. changing this doesnt help
- tried both speed control and torque control
- played with "tuning" torque-rpm settings. It calculates the settings by itself, input them manually doesnt change anything
- raised up the max current to 650A, didnt see a higher phase current nor power change. Did lower it back to 500A
- played with the kP value. Its 4.0 now, raising it results in weird noise but no power change, lowering no result either
- try to find dcf files from others to compare and learn, unfortunately we couldnt find relevant ones.

If anybody has pointers or a useful dcf I can build on, I would really appreciate it! If its helpful I can post screenshots and/or current dcf.

 

[Rotule]

Hello mate!

have you take a look to the power limit table?

edit : after reading again, it must be what you were speaking for "tuning torque-rpm"

It's hard for me to remember because it's been a long time since I used the Gen4 but there is multiple places where I needed to "rise the settings" for increasing the torque.

It start in Motor Nameplate then maximum current and maybe enven drive profiles if you are using that.

 

[Robinev]

Sevcon is sevcon I don’t think we can ask for more then a reliable vehicle operation the propriety software and how to obtain licensed access says a lot about how restrictive they are and limited we can tune all day but to me it feels like some lame way make the customers feel like you can tune race specs add more power functionality. It’s just a app that allow us to sync motor and turning regen and Miscellaneous options on/off. We need more amps more power faster and more voltage here’s my hack maybe we switch gate drivers from mosfets to higher gate make amps like igbt


Sent from my iPhone using Tapatalk

 

[laapmetot]

Thanks @rotule for the files in PM! Unfortunately it didnt help me in the right direction. See attached the settings and a current graph, maybe this gives some more insight of where I went wrong?

 

[laapmetot]

I did send this to my distributor to see if they can get Sevcon to help us out. If anybody already has answers to some questions dont be shy Find attached my dcf, if this gets solved I will post the relevant information and updated dcf as well. (disclaimer: use dcf on your own risk blabla it can brick blabla can malfunction blabla can break something else blabla)


"
We would like support in solving our issues:

Setup information
Application: Electric kart
Motor: motenergy ME1616
Controller: Sevcon gen4 size 6 80V with fluid coolplate, see pics for HW and SW versions
Battery: powerbattery 24S2P 60Ah fluid cooled continuous 480A - peak 720A nominal 86.5V (5,2kWh)
BMS: EMUS celmodules and canbus module
Contactors: 1x at BMS with economizer and 1x at controller without economizer
Cables / busbars: 50mm2
Fusing: 500A on controller and 500A in battery pack
Dcf: find attached

Issues we have:
• The UI software is, to say it politely, unstable and not very user friendly. The manual doesn’t help much in understanding it all: it focusses on setting up the basics to wire it up and get a motor to spin but the advanced settings are not covered. When we try google, we find that we are not alone in struggling with this.
• Low power. After tweaking we get around 18 kW at DC bus amperage of 320A at 90V with field weakening. This means that 11kW disappears since we put 29kW into the controller. And even 29kW peak output would be fairly low. Without FW we have 11kW at 240A DC bus. Again we put in way more power then we get out: we put in 22kW which is double of what me measure at the dyno. See settings in the dcf and screenshot of the currents with and without FW. Driving it also confirms: not really fast. Geared for 130kmu as we expected double the power, not a strange assumption imo since a Zero motorcycle has 34kW, can reach that speed and has a similar setup with a sevcon, 100V and a motenergy alike motor. As what we have is a very common combination, we would expect a better baseline setup and a power output more in line with the specs.
• Motor temperature not ok, the “overall DSP maximum motor T”. Goes up very quickly over 135’C, and goes back fairly quickly towards 100’C. 2 hours cooling down period after testing it still shows 80 ‘C while the outside of the motor is cold. A day later the motor is still 77’C which is of course not the case. We don’t even know if this represents the sensor value or is a calculation based on current, an excellent example of documentation being insufficient and software adding to the confusion. In any case we can’t figure out how to set this up but currently this is not ok, and we have no idea if we push the motor over the limit.
• We already had 3 times a failed inline fuse from contactor towards the controller. It is supposed to be 7A according to the manual, which we have. We measured current on that wire and see around 1A for 2sec during precharging, and after that 0.5A during operation. Seems normal to us, still it keeps failing during operation at times while nothing weird is going on. We are sure there are no shorts in the wire. We definitely did experiment with wrong settings and had cutouts and errors and spikey current graphs but that didn’t trip the fuse at least not at those moments. Any idea’s here? We can’t keep changing those and raising up the fuse value without knowing what is going on, is not the best practice.
• When we let go of the throttle, up to 250A was send back to the batteries while the limit was set at 90A. It seems that limit is not a control setting, just a warning level which gets ignored. When making current graphs we can see that the target Iq is indeed that level so the controller does so intentionally. It could even be higher since a dyno is not the same as driving. This is damaging for the batteries and a potential fire hazard. We reduced this regen current by setting lower rpm/s values for decelerating. Is this the intended way?!? This way setting up useful regeneration is virtually impossible. So we need to know the best way to limit battery currents and how to setup regen. Ideally both are just current settings but we don’t get that to work. Of course we guard current in the BMS, but that will open the contactor under high current every time we slow down so that is not an option.

What we tried to get the maximum power out of the controller:
• Checking the sincos alignment. Seems to be spot on. In any case we tried a couple of degrees difference in either way, but have similar performance as before. When going 10 degrees of in either way we loose quite some power. So this seems to be ok. (what we don’t understand is why there are 2 values which can be changed…… We only changed the top one at 31 degrees and the bottom one of 2 degrees we didn’t. Could that be a sincos sensor calibration value to compensate for angle offset between sin and cos? And how is that determined?)
• We raised up the maximum current towards 650. That only helps with very low rpm torque and does nothing for max power. So we set that back.
• We raised up the nominal current from 300 to 400 and 500, doesn’t make a difference. Maybe this is a warning level but seems not a setting. (would be a good improvement in the software to make a distinction between “settings that influence actual control” and “safety limits which will stop the controller” and “warning limits which will only gives messages”)
• We checked if the throttle signal results in only half throttle or something. Seems ok as max throttle is set at 10V which we can reach, the high current at low rpm also indicates full throttle anyway.
• We tried current mode and speed mode. Apart from throttle behavior no real power changes.
• We took off all safety’s and limits we could find to be sure there is no hidden cutback active. No change in power output.
• When increasing Kp we see torque increase especially at lower rpm, however a strange shrieking noise comes out of the motor and we have large spikes in the current graph. That doesn’t look/sound right so we did set it back to 4 which is the highest setting without those issues.
• Played with field weakening, however from run to run the current graph is completely different: sometimes we have FW from almost zero rpm and sometimes from halfway the rpm, sometimes not at all. Most of the time the transition into FW is very obvious as suddenly torque and thus rpm increases faster when FW gets active. Using the macro “field weakening” it sets the Id to -229A which we guess it calculates based on existing motor parameters. We played with this value but higher gives no power increase nor any noticeable change, and lower gives a decrease. In any case not clear how this needs to be setup to work properly for max power, and for smoother operation.
• We used the macro “max power” which apparently sets only the rpm/torque table to max torque settings. No power increase, only current failures in combination with FW especially on low resistance runs where rpm goes up quickly. So how this is supposed to help get more power is unclear.
• Try other users to see if they can help us. Unfortunately sharing dcf files with others is tricky since many controller hardware and dld versions are available. We tried to learn from setups from others but to no avail. We didn’t dare to flash the dld as we fear to brick the controller doing so, especially since we are not always sure it the dld we get from others matches our hardware.
• Getting desperate, we tried varying settings which we don’t fully understand, like the motor specific setup and gains. No real changes when varying that, except less smooth operation and a bit less power. We did only 1 step up and 1 step down but since nothing made an increase in performance we didn’t try extremer settings.
"

 

[Rotule]

You could try contacting motenergy also, I've worked with john and he is very helpfull. they has a configuration for all sevcon controllers.

 

[laapmetot]

ok will try! Tomorrow I have a call with Sevcon to see if we can improve, will post the results!

 

[laapmetot]

Actually call wasnt with Sevcon but with the distributor here. Helpful dude. Summary:

- Result of 28kW going into the controller and getting 18kW out with field weakening is what can be expected. If voltage is raised up you can reduce/eliminate fw and the power at the axle will go up at similar amps due to efficiency gain. With this controller not possible, we are already at the voltage limit.
- Regenerative current very high: reduce with "neutral braking %" which was 100% and now set to 20%. Basically the battery limit doesn't do anything.
- motor T is an interpretation based on amps vs time. Ignore it, just look at the PTC. The motor T cutback however looks at that temp. Disabling the I2T cutback should let the motor T cutback only to look at the PTC.
- 7A fuse that blows occasionally: expected to be a wiring problem. Doubt that, but will check.

Anyway we can iron out the small stuff, and unfortunately we are stuck with 18kW. Anyway good learning exercise and we will work during the winter on something more entertaining

 

[Upperfoot]

To get around the field weakening issue you can remain within its optimum RPM (about 3800rpm) and gear it appropriately based on the RPM for the speeds you want to achieve.

 

[laapmetot]

exactly what we did, it helps with efficiency but overall power remains very low.

we are going to try the curtis 96V which is much easier to setup and which has apparently good results with this motor. Will post the results.

 

[laapmetot]

Unfortunately we couldnt get the curtis to work. Endless faults and issues and our distributor is not really helpful, so we gave up on that and are stuck with the sevcon.

Now we swapped the motenergy for a 75-7 66V zero motor. Feels better already but compared to the kart from nuxland its definitely slower. Will put it on the dyno next week.

 

[nuxland]

Zero motor is also only peaks around 50kw. If you want to race 10 or more laps, you have to dial it down to 25kw max with your closed zero motor

 

[j bjork]

It will be very interesting to see how it turns out for you

 

[laapmetot]

Nuxland is already a big help! First lets see if we get the 50kW out, then we can see how much we need to dial back. I also like the alternative cooling system, definitely going to look into that, but first things first.

 

[j bjork]

What was the problem that limited the power for you?

 

[laapmetot]

What was the problem that limited the power for you?

No idea! In the process of finding that out.

 

[j bjork]

Okay, I misunderstood and thought Nuxland had been able to help you with that already

Very curious about what you find out, as I soon will start trying to figure out a sevcon too.
A size 6 and a qs 180 90h, I hope to end up in the 40kw range. Around 18 would be very disappointing

A problem for me is that I have practically no motor data, and I am probably the only one with that motor/controller combo so far.

 

[laapmetot]

this is the 3 separate dyno graphs, and one combined. Remember these are actual measurements at the wheel, not the battery.

ME1616
18kw it basically falls flat on its face at higher rpm

zero 75-7 66V based on drbass dcf
33kW while drawing 50kW from the battery. Lower end is weak. Superflat torque, never saw that before!

zero 75-5 66V based on nuxland dcf
22kW and errattic behavior at topend so stopped the run. Lower end is better.

I made screenshots of all screens to compare values, i think I use the 33kw dcf as a baseline and try to get the lower rpm towards what the nuxland version has. Going to post this later. Already attached the 3x dcf of which are frankensteined from my own file and those of nuxland and drbass (thanks guys!) . All based on the following:

- 0-10V throttle using potentiometer
- 12V contactor without economizer
- sevcon gen4 size 6 80V (max 116V)
- battery 24S 86V
- check sincos voltage and alignment
- definitely not all safeties and cutbacks properly set






View attachment config file 25032021 18kw me1616.zip
View attachment config file 23jun21 zero drbass.zip
View attachment config file 23jun21 zero nuxland.zip

 

[laapmetot]

Did the comparisons. I didnt do all screens since I dont think fi contactor data does anything for power. That works, or it doesnt.

what I found:

Peak torque 120 vs 165Nm. Will increase the value.
Rated stator current 550 vs 420A Will increase the value.
Proportional gain 0.0625 vs 0.03125. No idea what that does. Will try.
Drive torque voltage cutback factor. 0.39 vs 0.99. Does this limit the torque to 39%?!?!? Trying to find out. Seems no voltage cutback set.........

I also see no reason why the high torque version has such a bad topend.

Dont be shy, feel free to ask, suggest, educate!

 

[laapmetot]

well, I managed to combine the best of both settings: the higher lowend torque and the higher topend power.

Turned out there was no voltage cutback (or anyway not this time when I looked so I increased the peak torque from 120Nm to 165Nm. This resulted in high pitched whine and errors. I cut the Kp and Ki in half and now it runs smooth to the ear. However when looking at current control this is still erratic when field weakening starts. Anybody an idea how to get this better?

Didnt touch the rated stator current, will try. If somebody explains this more to me that would help.

Also I think the field weakening should start more gradually. I cant find any settings to influence field weakening however. Info welcome!

Will also work on cutbacks and protections.

Anyway a good day: almost doubled the power compared to the ME1616! Find attached dcf again



View attachment config file 24jun21 zero frankenstein drbass nuxland.zip

 

[j bjork]

Arlo1 on the forum seems to get very good results with sevcons and zero motors, maybe you can get a dcf file from him and compare?

I dont remember the outputs, but I think it was higher. But maybe just because of higher voltage?
There should be some videos of dyno runs somewhere.

Edit, here is a thread with a dyno run:

https://endless-sphere.com/forums/viewtopic.php?f=1&t=109896

He is using feet of torque and hp, but 58hp should be almost 43kw

 

[laapmetot]

Hi Bjork I asked him on PM, but as far as I can see he has a dual sevcon setup which is not comparable I'm afraid.

 

[c70r]

I think your main issue is simply with the motor choices rather than controller (sevcon), and even voltage use.

Unofficially the ME motors require high voltages to make decent power, like the best wheel dyno result I have seen for an ME1507 or 1616 was using the Gen 4 Size 4 150V controller and achieved only 35kW iirc. There are many who were claiming 40kW+ but this was just controller demand/battery output rather than what the motor was outputting.

As for the Zero motor you have, isn't this the 2012 4 turn model? If so you really need big amps to make use of it. Iirc both Arlo and Drbass used a dual Sevcon setup to get good numbers, 33kW from the one controller is about right, I don't think you can get past 40kW with this motor and controller combo.

The ultimate solution for high demand (go kart racing) in my opinion is to throw a 75-7 IPM into a ME1616 shell or to at least air cool it like Nuxland. The motor is more likely to give in to heat than the Size 6 controller to be honest.

Or even a 2013+ normal 75-7 SPM is better to be used with a single controller than the 2012 hybrid 4 turn model.

Edit:
To better explain this.

The Zero motor you have was designed to be used in bikes that had the Agni motor/Alltrax controller combo. Those bikes used a 52V battery. Trying to run a motor at twice the voltage isn't going to double it's power, that's not really how it works. It will theoretically double it's max rpm but it's already designed to take full advantage of it's max safe rpm.

The Zero motors whether the 2012 hybrid or spm or ipm are used at their maximum safe rotor rpm which is 6000 I believe.
For numbers sake lets say your motor is achieving 33kW at 6000. The only way to increase your power is to either increase current, because this motor has a low tq/amp because of its winding or to increase max rpm which you can't safely do as the stock rotor will not handle it. Theoretically to achieve the 70kW that Arlo and Drbass achieved you would need to let the motor run at 12000 rpm or 6000 rpm with double the current using two controllers like they did.

Or to simplify it.
Voltage increases RPM. Current increases Torque. You need both for Horsepower.
75-7 2012 Hybrid requires either very high RPM which is not possible with stock rotor or very high current, theoretically twice the current of the normal 75-7 to achieve the same power because it's designed to work with half the voltage.

Nuxland motor is a normal 75-7, therefore it has different motor inductance. If you are using his config without making any changes it's because it's a different motor, you must change the appropriate parameters.

Drbass config at 33kW is quite good, if you can use the full current capability of the Size 6 you should achieve nearly 40kW, I really doubt you will find more power using a single Size 6 or any controller that won't take at least 800 Batt amps.

Second edit:
The best way people and even Zero with their race bike achieved even more power from their motor was to use a rotor that could safely tolerate a higher rpm.

Trying to find an article of a Zero race bike that explained how they managed to pull a lot more power from the stock bike/motor by modifying/upgrading the rotor to achieve a greater rpm.

 

[laapmetot]

Thanks for your input c70r! Indeed this is the 4turn motor for the 66V zero at a single gen4size6. I was thinking my 86Vnom is fairly low so this motor would need less field weakening so runs less hot.

For motor parameters there was not as single thing different between the dcf from nuxland and drbass. So I'm not sure if it is correct now, I dont have the parameters for any of those 2 motors. If you have the correct ones for the 4 turn motor that would be appreciated, or some pointers in what direction to move to find an optimum.

If I can reach the 40 rearwheel kW I would be quite happy actually. I think its best first to correct motor parameters, and to solve the oscillation during fw before trying to push it more.

 

[laapmetot]

Did manage to improve a little bit. I got some info from Arle1 but unfortunately this didnt result in more power. So it seems this is it, around 32kW at the axle. Less then I hoped for, more then I had.

Now going to the track and experience the result! Thank you all for the support, especially nuxland, drbass and Arle1

See graph: red is with me1616 motor and blue with the zero 75-7 66V

Find attached dcf (use at your own risk yada yada

View attachment config file 29jun21 zero 66V gen4size6 86,4V semifinal.zip