Voltron-Evo; jonescg's new electric race bike

Post up a thread for your EV. Progress pics, description and assorted alliteration
Post Reply
Stiive
Groupie
Posts: 295
Joined: Wed, 31 Dec 2008, 08:26
Location: Melbourne

Voltron-Evo; jonescg's new electric race bike

Post by Stiive »

7circle wrote: In fault state it is no longer controlling the 6 pulse bridge for motor current field weakening.


Had this happen to me the other day except in an induction motor Image

Doesn't matter if in the field weakening region or not, you're effectively halting the stator flux linkage and you'll quickly reach full breakdown negative torque (if the field was strong enough). In an IM this sinusoidally reduces to zero as flux is dissipated.

So in a fault condition, no, you cannot just stop controlling the motor (where possible) - but in this case a dodgy connection lost LV power to my controller and I had no choice in the matter Image

Working on a solution so this won't happen again - it's a big safety concern
Rgds,
Stiive
User avatar
Johny
Senior Member
Posts: 3749
Joined: Mon, 23 Jun 2008, 16:26
Real Name: John Wright
Location: Melbourne
Contact:

Voltron-Evo; jonescg's new electric race bike

Post by Johny »

Stiive wrote: Working on a solution so this won't happen again - it's a big safety concern
Full negative torque... That would just about lock up my rear wheels at 100km/h. The only way I can see to stop it would be to have a contactor between the motor and controller.
Stiive
Groupie
Posts: 295
Joined: Wed, 31 Dec 2008, 08:26
Location: Melbourne

Voltron-Evo; jonescg's new electric race bike

Post by Stiive »

Johny wrote:
Stiive wrote: Working on a solution so this won't happen again - it's a big safety concern
Full negative torque... That would just about lock up my rear wheels at 100km/h. The only way I can see to stop it would be to have a contactor between the motor and controller.


Not a good option either. You're still halting the stator vector (if not more than before because now the the currents cant even drain out of the FWDs), and you'll be trying to stop current through an inductor - you're going to get a massive voltage spike which will probably reach break down of the contactor, motor winding insultation, or arc somewhere.

Last edited by Stiive on Thu, 12 Jun 2014, 05:17, edited 1 time in total.
Rgds,
Stiive
User avatar
Johny
Senior Member
Posts: 3749
Joined: Mon, 23 Jun 2008, 16:26
Real Name: John Wright
Location: Melbourne
Contact:

Voltron-Evo; jonescg's new electric race bike

Post by Johny »

Stiive wrote:Not a good option either. You're still halting the stator vector (if not more than before because now the the currents cant even drain out of the FWDs), and you'll be trying to stop current through an inductor - you're going to get a massive voltage spike which will probably reach break down of the contactor, motor winding insultation, or arc somewhere.
Yes it wouldn't be pretty but at least the vehicle would be steerable. I get the feeling that LEAF and iMiev just live with the remote possibility. You are not pulling full motor current when in field weakening so the chances of controller failure are probably lower than , say, starting from stall.
Stiive
Groupie
Posts: 295
Joined: Wed, 31 Dec 2008, 08:26
Location: Melbourne

Voltron-Evo; jonescg's new electric race bike

Post by Stiive »

Johny wrote:
Stiive wrote:Not a good option either. You're still halting the stator vector (if not more than before because now the the currents cant even drain out of the FWDs), and you'll be trying to stop current through an inductor - you're going to get a massive voltage spike which will probably reach break down of the contactor, motor winding insultation, or arc somewhere.
Yes it wouldn't be pretty but at least the vehicle would be steerable. I get the feeling that LEAF and iMiev just live with the remote possibility. You are not pulling full motor current when in field weakening so the chances of controller failure are probably lower than , say, starting from stall.


In an IM a controller failure is probably worst case at nominal speed peak power (full load amps and full field). The field strength is important because the torque depends on it, and the torque will oscillate till it dissipates.

In a PM, probably the higher the RPM the worse the case gets due to the Kv

Jonescg, have you got a transmission w/ clutch intact? That's the biggest safety feature



Rgds,
Stiive
User avatar
Sutho
Groupie
Posts: 122
Joined: Thu, 24 Nov 2011, 15:04
Real Name: Brett Sutherland
Location: Newcastle
Contact:

Voltron-Evo; jonescg's new electric race bike

Post by Sutho »

7circle wrote: Just wondering...
With field weakening of this PM motor...
Say you get the revs up to 5000 rpm and the controller goes to a fault state.
In fault state it is no longer controlling the 6 pulse bridge for motor current field weakening.
So will the Motor Bemf Voltage increase up to meet th Kv relationship for 5000rpm?
And if this occurs then the motor pk-pk voltages would be way over the battery pack voltage.
This V-dif would drive huge currents into the battery pack and huge regen braking torque.

Or am I missing something?
If 3700 is max at 690V then at 5000 you would be 690 x 50/37 so 932V


Some input from RMS: "Definitely a valid point, but if you are mixing the full load break speed with the no load break speed you will come up with a different answer. I calculate 800V at 5000 rpm. We use 900V capacitors and 1200V IGBT module, so no problems."

Our ELMOFO race car has dual-Remy IPM motors and we have encountered high-RPM (in field weakening) fault scenarios a number of times during testing with no damage or issue. We are overdriving the inverters a little, so this was expected.

In our cases, the above scenarios have resulted in the inverter disconnecting the pack when a fault occurred, so no heavy regen or overcharging. Not saying that this would be the case for all scenarios though.
Stiive
Groupie
Posts: 295
Joined: Wed, 31 Dec 2008, 08:26
Location: Melbourne

Voltron-Evo; jonescg's new electric race bike

Post by Stiive »

Sutho wrote:
In our cases, the above scenarios have resulted in the inverter disconnecting the pack when a fault occurred, so no heavy regen or overcharging. Not saying that this would be the case for all scenarios though.


I would say it still isn't relinquishing control though, especially if it's opening the main contactor - voltage will rise quickly (even with a large cap) with nowhere for the power to go. Rather it probably goes into a shutdown procedure where it will ramp the current down as quickly (and safely) as possible.

I'd imagine in a case where you "accidentally" turn off the car with full throttle @ high RPM, effectively opening the contactor and losing control over the motor, results would be different.

Sounds like Chris B needs to make his own AEVA forum account :)
Last edited by Stiive on Thu, 12 Jun 2014, 06:05, edited 1 time in total.
Rgds,
Stiive
User avatar
Sutho
Groupie
Posts: 122
Joined: Thu, 24 Nov 2011, 15:04
Real Name: Brett Sutherland
Location: Newcastle
Contact:

Voltron-Evo; jonescg's new electric race bike

Post by Sutho »

Stiive wrote:Sounds like Chris B needs to make his own AEVA forum account :)


I don't think that would be a good idea at all as he could get caught up in discussions like this in lieu of building stuff we need! Image
GRMarks
Groupie
Posts: 140
Joined: Sat, 13 Feb 2010, 18:14
Real Name: Glenn Marks
Location: Dandenong,Vic

Voltron-Evo; jonescg's new electric race bike

Post by GRMarks »

Chris I think your forum has been hijacked a little.

Chris have you thought about a pressure sensor on the brake line to activate regen at low pressure and if regen is not enough then pulling the brake harder gives regen plus extra normal breaking?

With your motor cooling problem, its hard to see from the front on photo, but is it due to no air flow behind the radiator, or you need a bigger radiator or water pump?

Glenn

PS the bike looks great.
Regards

Glenn Marks
User avatar
jonescg
Senior Member
Posts: 4721
Joined: Thu, 21 Jan 2010, 23:05
Real Name: Chris Jones
Location: Perth, WA.
Contact:

Voltron-Evo; jonescg's new electric race bike

Post by jonescg »

Woah, just got back after an overnight camping trip and the thread exploded...

OK, it's obvious that the resolver isn't tuned properly, which is a result of solving one problem while creating another. I think I have it headed in the right direction now (literally).

I need to swap a pair of phase wires and the son/cos wires, as these were originally swapped in order to get it to spin the right way. Turned out I could have achieved the same thing by tuning the resolver to 180 degrees more or less than my current settings. It's almost as if it will happily spin in either direction, but will only field weaken in one direction.

To test this (without a wheel attached) I changed the resolver Gamma adjust from -7.4 degrees to +173.6 degrees and as expected it spins the opposite direction, however as soon as it hits base speed it goes into a runaway state all the way up to 5200 rpm or more, and throws an error.

The regen thing is a tough one. Obviously it needs to have some dead spot in order to work, but Danny wants no such deadspot. The alternative is to have a clutch lever which gives the sensation of a rear brake, and is hooked up to the regen potentiometer. This requires more thought and more getting used to, and might not be ideal. Perhaps the dead spot need only be 5% of WOT (wide open throttle) in order to engage coasting regen.

Glenn - we have added a second radiator up behind the ram air intake on the fairings and we've moved the original one up a bit and tilted it forwards. Hopefully this allows more air behind it to get out of the way - the battery pack is a fair old brick wall for airflow. The motor sits directly behind it, so it gets as much cooling as your dog in the back of a ute - when his lead is tied to the middle.

Lots to do, lots to do...
AEVA National President, WA branch director.
Coulomb Racing
Groupie
Posts: 87
Joined: Tue, 28 Feb 2012, 15:53
Real Name: James Hill
Location: Perth
Contact:

Voltron-Evo; jonescg's new electric race bike

Post by Coulomb Racing »

Hi Chris,

I finish exams next wednesday, need any help/third hand and I can come over!

James
Please like us on Facebook: Coulomb Motorsport
User avatar
Adverse Effects
Senior Member
Posts: 1172
Joined: Sat, 01 Jan 2011, 03:30
Real Name: Adverse Effects
Location: Brisbane

Voltron-Evo; jonescg's new electric race bike

Post by Adverse Effects »

for the regen why not just add a "thumb throttle" (acting as a regen brake) on the left hand grip that the more you press it the more regen there is?
pottz
Groupie
Posts: 34
Joined: Fri, 23 Aug 2013, 05:29
Real Name: Danny

Voltron-Evo; jonescg's new electric race bike

Post by pottz »

For sure, ideally for me would be to have as little as possible dead zone on the throttle, it would be best if possible to have throttle closed=regen and throttle open=power.
If the options are a large dead zone or no regen, I'll go with no regen and just use the rear brake, I would like it to come in as I close the throttle but really it's no drama.
top speed is way more important, that was hard to deal with after how much power it had. Such a hard cut had me nearly falling off the bike on some corner exits :)

I don't see much point personally on working at regen as a clutch lever, to me it seems like extra work setting it up when in reality I have a perfectly good rear brake that will do the same thing. I'd be better off with a rear brake lever up there for simplicity. (everyone really seems to want a lever there!)

User avatar
Sutho
Groupie
Posts: 122
Joined: Thu, 24 Nov 2011, 15:04
Real Name: Brett Sutherland
Location: Newcastle
Contact:

Voltron-Evo; jonescg's new electric race bike

Post by Sutho »

jonescg wrote:To test this (without a wheel attached) I changed the resolver Gamma adjust from -7.4 degrees to +173.6 degrees and as expected it spins the opposite direction, however as soon as it hits base speed it goes into a runaway state all the way up to 5200 rpm or more, and throws an error.
Tuning the gamma and putting a wheel on it should help this a little....but it's that "thing" I mentioned that the EVO motors seem to do when in the field weakening range. The field weakening current generates some torque when it shouldn't. Apparently biasing the gamma setting in one direction (forward) can help.
jonescg wrote:The regen thing is a tough one. Obviously it needs to have some dead spot in order to work, but Danny wants no such deadspot. Perhaps the dead spot need only be 5% of WOT (wide open throttle) in order to engage coasting regen.
You just need to set all the bottom throttle (pedal) settings close together to minimise any deadband. You could pretty much have Pedal_Lo, Accel_Min and Coast_Lo all within a few counts of each other. It's all possible. See Appendix G of the Software manual. Image
User avatar
jonescg
Senior Member
Posts: 4721
Joined: Thu, 21 Jan 2010, 23:05
Real Name: Chris Jones
Location: Perth, WA.
Contact:

Voltron-Evo; jonescg's new electric race bike

Post by jonescg »

Sutho wrote: ...but it's that "thing" I mentioned that the EVO motors seem to do when in the field weakening range. The field weakening current generates some torque when it shouldn't.
I think it's a little scary, but hopefully with the inertia of the wheel it won't flip out. I'll give it a go tomorrow.
Sutho wrote:You just need to set all the bottom throttle (pedal) settings close together to minimise any deadband. You could pretty much have Pedal_Lo, Accel_Min and Coast_Lo all within a few counts of each other. It's all possible. See Appendix G of the Software manual. Image


Just starting to get my head around it now. Would this interpretive dance / MS Paint image help explain the current situation?
Image

This should mean that the largely regen will work whenever the throttle is chopped. And protect against any scary runaway bike situations...?

[ Edited Coulomb: repaired smiley in preparation for phpBB transfer ]
Last edited by coulomb on Wed, 19 Jul 2017, 07:59, edited 1 time in total.
AEVA National President, WA branch director.
User avatar
weber
Site Admin
Posts: 2936
Joined: Fri, 23 Jan 2009, 17:27
Real Name: Dave Keenan
Location: Brisbane
Contact:

Voltron-Evo; jonescg's new electric race bike

Post by weber »

Here's what you want. Note that "chopped" is Accel_Min, not Pedal_Lo. Pedal_Lo and Pedal_Hi must not be reachable in normal operation. Pedal_Lo should be about halfway between chopped and zero volts. Pedal_Hi should be about halfway between WOT and 5 V. Also note the much wider regen zone, drawn approximately to scale, unlike yours.

Image
jonescg wrote:To test this (without a wheel attached) I changed the resolver Gamma adjust from -7.4 degrees to +173.6 degrees and as expected it spins the opposite direction, however as soon as it hits base speed it goes into a runaway state all the way up to 5200 rpm or more, and throws an error.
I wouldn't worry about that unless it still happens after you get regen sorted.
The regen thing is a tough one. Obviously it needs to have some dead spot in order to work, but Danny wants no such deadspot. The alternative is to have a clutch lever which gives the sensation of a rear brake, and is hooked up to the regen potentiometer. This requires more thought and more getting used to, and might not be ideal. Perhaps the dead spot need only be 5% of WOT (wide open throttle) in order to engage coasting regen.
You seem to be misunderstanding how this works. "Coasting" and "regen" are two very different things. There is no need to have any coasting region of the throttle -- no need for a deadspot between motoring and regen. It can just pass smoothly from motoring to regen and back again. Sure there will be a zero torque point somewhere between motoring and regen, but if it's set up as I describe below, you won't feel it. The requirements for this are:
(a) that Coast_Lo and Coast_Hi are set to the same value, and
(b) this value is at a specific point between Accel_Min and Accel_Max that depends on the ratio of Regen_Torque_Limit to Motor_Torque_Limit as follows:
Coast_Lo&Hi = Accel_Min + (Accel_Max - Accel_Min) * (Regen_Torque_Limit / (Regen_Torque_Limit + Motor_Torque_Limit ))

For example, if Accel_Min = 20 and Accel_Max = 480, Motor_Torque_Limit = 5430 and Regen_Torque_Limit = 1086 (20% of Motor_Torque_Limit) then Coast_Lo and Coast_Hi should be set to 20 + (480 - 20) * 1086 / (1086 + 5430) = 97.

I tried to explain this, but without giving a formula, at the end of my first post on the topic, way back here
viewtopic.php?title=voltronevo-jonescgs ... 553#p51421
and in my post that followed it.

There will be no dead spot or dead zone once the bike is moving, but sure, when starting off, nothing will happen until you move the throttle past the regen zone, but Danny should be able to gently wind that on in advance while waiting for the start signal.

If you instead put Coast_Hi and Coast_Lo within a few counts of Accel_Min then when you back off the throttle for a corner, maximum regen will cut in very suddenly. You won't have smooth control over regen.
One of the fathers of MeXy the electric MX-5, along with Coulomb and Newton (Jeff Owen).
GRMarks
Groupie
Posts: 140
Joined: Sat, 13 Feb 2010, 18:14
Real Name: Glenn Marks
Location: Dandenong,Vic

Voltron-Evo; jonescg's new electric race bike

Post by GRMarks »

if you have ridden a vetrix scooter, they have a throttle that does a simple version of what weber just described except it has a mild fixed regen setting.
when you let the throttle go it returns to off position which is webers coast lo/cost hi position. To engage regen you have to push the throttle (turn it backwards) against spring pressure to activate regen. This way you physically feel when regen will occur.
If you could get your hands a throttle setup like that it is really easy to ride.
No unexpected regen halfway through a corner.   
Last edited by GRMarks on Thu, 12 Jun 2014, 22:39, edited 1 time in total.
Regards

Glenn Marks
User avatar
jonescg
Senior Member
Posts: 4721
Joined: Thu, 21 Jan 2010, 23:05
Real Name: Chris Jones
Location: Perth, WA.
Contact:

Voltron-Evo; jonescg's new electric race bike

Post by jonescg »

weber wrote: Here's what you want.

Ta - got it now. And thanks for drawing it to scale...

The dead spot at the start sounds unavoidable, but that's probably OK as it's only something you need to worry about once. We might yet need to return the throttle back to it's original mechanical setting...

Full regen on this machine is not going to be much - when you're hard on the front brake the rear will have very little rubber on the road, so it will be set to 100 Nm or less at the motor.
AEVA National President, WA branch director.
User avatar
weber
Site Admin
Posts: 2936
Joined: Fri, 23 Jan 2009, 17:27
Real Name: Dave Keenan
Location: Brisbane
Contact:

Voltron-Evo; jonescg's new electric race bike

Post by weber »

Right GRMarks. I have ridden a Vectrix. I converted two to LiFePO4 for a friend, using Antiscab's excellent kit. A Vectrix type throttle that is spring loaded in both directions, either side of the coast point, would be ideal. But it's not essential. With the existing one-way throttle, letting go of it will cause it to command the regen torque limit, so the important thing is to find out by experiment what is the maximum amount of regen torque that will not lock up the rear wheel even under the worst conditions.

I'd start with the throttle setup for 1/6th regen, 5/6th motoring, as I suggested in the previous message. Although this is intended for the case where max regen torque is 20% of max motoring torque there's no need to go straight to 20%. The present setting has regen at about 6%. You can work your way up gradually from that until it starts to feel dangerous then back it off a bit. Then when that's settled, recalculate the ideal Coast_HiLo value for that amount of regen torque according to the formula I gave earlier.

But the first thing you have to do is set up the throttle so the wiper voltages at the two extremes are at least 0.2 V away from 0 V and 5 V. It doesn't matter exactly what they are. Anything between 0.2 V and 0.5 V away would be fine. If you can't do it mechanically, by adjusting stops at both ends, you can do it with resistors as I mentioned earlier. Or maybe you can do it mechanically at one end and need a resistor at the other end. But once you've set them in the right ballpark, you need to know accurately what these extreme values are. Ideally, get the inverter to tell you what it thinks they are. Then set the Accel_Min and Accel_Max values just inside those values, and calculate the other four throttle parameters from those, as described previously.
One of the fathers of MeXy the electric MX-5, along with Coulomb and Newton (Jeff Owen).
Stiive
Groupie
Posts: 295
Joined: Wed, 31 Dec 2008, 08:26
Location: Melbourne

Voltron-Evo; jonescg's new electric race bike

Post by Stiive »

why not have regen kick in quarter/half when 0% throttle, and full when either front or rear brake pedals are pressed (&& 0% throttle)?

Rgds,
Stiive
User avatar
Johny
Senior Member
Posts: 3749
Joined: Mon, 23 Jun 2008, 16:26
Real Name: John Wright
Location: Melbourne
Contact:

Voltron-Evo; jonescg's new electric race bike

Post by Johny »

Keep in mind that regen will contribute a fair bit to motor/inverter heating. Having any substantial amount of regen on the throttle will definitely make the bike handle differently to an ICE. Maybe have regen set to roughly the equivalent amount of ICE braking. I would go for around 10% to 15% throttle opening for the transition.
User avatar
jonescg
Senior Member
Posts: 4721
Joined: Thu, 21 Jan 2010, 23:05
Real Name: Chris Jones
Location: Perth, WA.
Contact:

Voltron-Evo; jonescg's new electric race bike

Post by jonescg »

Johny - that's why I was trying to make the "regen zone" so small, so that whenever the throttle is chopped (as you are hard on the front brake) the rear wheel provides some resistance, steadying the rear of the bike. The force should be no greater than say kicking an ICE bike down a gear. The roughly 20% allocated to non-forward torque command is pretty generous - I was thinking 10% at most.

In any case, today I swapped the phase wires and the sin/cos wires back to their original positions, which makes the wiring sit very nicely around the shock. I ended up changing the ID value back to zero while I did the Gamma adjust process, and ended up with a value of +2476 (247.6 degrees). Now the motor feedback speed is positive, and the Delta_In_Fil value is pretty close to 902 when spinning forwards, and -898 when spinning backwards. These values (representing +90.2 and -89.8 degrees) are roughly close, as the value travels through a range either side of 90. I'm picking the value at roughly 1200 rpm.

Without changing anything else, the motor will spin with the rear wheel attached, up to about 3600 rpm and not go any further, but I haven't added any ID nor have I done anything with the throttle settings. That's for tomorrow...
Last edited by jonescg on Fri, 13 Jun 2014, 14:48, edited 1 time in total.
AEVA National President, WA branch director.
User avatar
jonescg
Senior Member
Posts: 4721
Joined: Thu, 21 Jan 2010, 23:05
Real Name: Chris Jones
Location: Perth, WA.
Contact:

Voltron-Evo; jonescg's new electric race bike

Post by jonescg »

Quick update on the field weakening issues - it's working!

After tuning the motor in the right direction and adding about 100 amps of ID, it now spins up to full speed. We had the opportunity to do a couple of laps at Barbagallos on Sunday during the lunch break as a bit of a demo of the bike. For me it was a chance to try out the new settings, as speeds of 170 km/h+ aren't really testable on the street. Now it pulls well over 200 km/h over the hill, and gets there just as fast. With the added cooling we have installed it's not getting too hot either.

Murphy's law of demonstrating EVs states that it will only throw an error when people are watching, and sure enough at the top of turn 4 it dropped out requiring me to turn it off and on again. The same happened after the main straight too, and I have since worked out it was a throttle error. The Pedal_Hi and Pedal_Lo are probably too close to Accel_Max and Accel Min respectively. I've put a good 15 unit gap between them now and the faults appear to have gone away.

I do know that the electrical environment is very noisy, so it's going to require shielding, especially on the phase wires.

All up we're about 90% there I think. Next step is to get the bike crated up and sent to Queensland!
AEVA National President, WA branch director.
User avatar
weber
Site Admin
Posts: 2936
Joined: Fri, 23 Jan 2009, 17:27
Real Name: Dave Keenan
Location: Brisbane
Contact:

Voltron-Evo; jonescg's new electric race bike

Post by weber »

Yay! Glad to hear you've got field weakening working. Last I saw you had the Id limit set to 300 amps. Are you saying you added about 100 amps to that (about 400 amps total), or that you now have it set to about 100 amps total. I suggested it should be 50 to 100 amps total. Was I way off?

Glad you got Pedal_Hi and Pedal_Lo sorted too. How close were they to Accel_Max and Accel_Min when the dropouts occurred?

Any joy with throttle regen? If so, what settings did you end up using there?

I like to learn from my mistakes. Image
One of the fathers of MeXy the electric MX-5, along with Coulomb and Newton (Jeff Owen).
User avatar
jonescg
Senior Member
Posts: 4721
Joined: Thu, 21 Jan 2010, 23:05
Real Name: Chris Jones
Location: Perth, WA.
Contact:

Voltron-Evo; jonescg's new electric race bike

Post by jonescg »

The current setting is 100 amps of field weakening. Previous setting wasn't doing anything as the back EMF defeating current was happening after the rotor, not before. It's plenty fast for Wanneroo (Danny would make it go much faster) so perhaps 100 is fine.

The regen is going to take some getting used to, but yes it seems to be working well. I have Coast Lo and Hi set to 92 and the regen torque set to 10 Nm. We can bump this up if need be. I wasn't sure if the faults I was getting on the day were due to this or something else, so I erred and made it 10 Nm instead of the 25 it was set to originally.

Next step is getting the inverter to discharge rapidly. It currently takes about 2 minutes for the inverter DC input terminals (which are completely covered and isolated from the chassis) to get below 50 V. The rules state 5 or 6 seconds, so in order to comply I need to make it discharge faster.

I have a few opinions about this rule, but will get back with my plan so far...
AEVA National President, WA branch director.
Post Reply