DC Motors - Shunt?

[4Springs]

I'm intending on converting a tractor to electric. Tractors have some specific requirements that are a bit different to cars, so I'm looking at motor options. I'm wondering if a shunt wound DC motor would be suitable.

1. Tractors use a hand throttle, you set it at a particular position and then leave it there. This is so that you can essentially set an RPM to suit what you're doing, for instance driving an implement on the PTO. A ground speed is selected by changing gears, so tractors have lots of gears. You can change gears while keeping the RPM more or less constant.
2. I'm looking to convert a tractor rated at about 40 kW. Unlike cars, tractors are sometimes used at a continuous high power. So either the motor I use should be rated at 40 kW continuous, or I need to de-rate the tractor specifications.
3. The brakes on a tractor are normally not very good. The low gearing means that engine braking is extremely significant, and is normally the only braking required to avoid runaway when driving down hills. However, on some propertiess (e.g. here) there are no hills, so this is only required when driving the tractor on the road once in a blue moon.

Points 1 & 3 rule out a series DC motor. Series DC motors are torque controlled, so if I was driving along with a set throttle and pushed in the clutch, the motor would over-speed.
An AC induction motor would tick all of the boxes, but they are expensive!
My question, then, is how about a parallel (shunt) wound DC motor? The DC motors that I've worked with for EV conversions were provided with the connections so that you can wire them in series or shunt. But I have not used one in shunt mode, and I have some questions about how it would work.

Matching up with my numbered points above:

1. Shunt DC motors apparently are speed controlled. So on the face of it, this looks like a good solution for the way that tractors are driven. Set the throttle lever to a position and the controller will deliver torque until the requested motor speed is reached. Push in the clutch (take the load off) and the motor speed will remain constant.
2. I assume that motor power ratings are the same whether wired as series or shunt.
3. What happens when a shunt DC motor is pushed faster than the requested speed? Does it just 'coast' faster and faster like a series motor? Or is there some kind of engine braking effect? If there is engine braking, where does the power go? If there isn't engine braking, is there another way of causing this effect (e.g. shorting the terminals). This situation is rare enough that regenerative braking (putting power back into the batteries) is not required.

[coulomb]

The DC motors that I've worked with for EV conversions were provided with the connections so that you can wire them in series or shunt.

My understanding is that shunt fields are very different to series fields; shunt are designed for high voltage, low current (so many turns of fairly fine wire), whereas series fields take traction current, so only a few turns of very stout wire or even metal bars.

Shunt DC motors apparently are speed controlled. So on the face of it, this looks like a good solution for the way that tractors are driven.

Agreed.

What happens when a shunt DC motor is pushed faster than the requested speed?

It tries to regenerate power back to the battery. But many simple buck controllers won't be able to handle that, I think.
[ Edit: I was thinking separately excited, not shunt. See Ian Hooper's post here. ]

If there isn't engine braking, is there another way of causing this effect (e.g. shorting the terminals).

Yes, this is called plugging. I don't think you can just short the terminals of a 40 kW motor with a contactor; some controllers will have an IGBT or similar that can connect the armature to a resistive load. I note that that would have to be a prodigious resistor (it would not need to be rated at 40 kW continuous, of course, but still be pretty big to take that sort of power for tens of seconds). Also the battery to drive a 40 kW load for hours on end would have to be quite large (well over 80 kWh for 2 hours of continuous operation). The controller would also have to be rated for continuous operation.

[jonescg]

You could even set up a load bank like dynamic braking on locomotives.

...or pony up and buy a synchronous AC motor and inverter

[francisco.shi]

A shunt wound DC motor will be able to regen energy to the battery as long as you have the right controller. You will need a half bridge instead of a single transistor for the series wound motors.
An AC motor may be more efficient and it could be liquid cooled which would make the motor smaller.
I think a Hyper9 is rated at 35kw continuous. That may be good enough. I don't know how the brushed motors compare pricewise.

[Bryce]

Personally, I'd go AC - the price difference with DC is very little now and you get a whole lot more flexibility through programming if you select the right brand ... and a whole lot more confusion if you get one with a less than helpful computer programming interface ....

Will dig out my notes on DC motors and write more later (that BTW will be AFTER I finish resting from organising the 2020 EV Vision conference .... could be a while - I used to teach the AC and DC motor subjects to apprentices, but I am a bit rusty as I last taught that stuff 5yrs ago
Cheers
Bryce

[francisco.shi]

I think I can answer most questions on motors.
I got an brushless inverter working from scratch.

[4Springs]

Ok, so from the comments above it looks like a DC motor could be used in this application, as long as the controller can handle regen current.

Let's look at prices.
I've found some examples from the EV Works website. All out of stock, but the list prices give an indication.

For instance the Zeva MC1000C controller is $2000. I'd need to ask @zeva if it could handle regen from a DC shunt motor.
The Netgain Warp 11 is $5000. Or perhaps the Impulse9 for $3000.

In AC, the HPEVS AC-51 is $7500, including controller.

I'm finding it difficult to compare these motors, some power outputs are peak, some are continuous, and some graphs are given at a completely different voltage to what the motor is advertised at... But I think the $7500 HPEVS AC-51 is roughly equivalent to the Netgatin Impulse9.
If so, the DC option is about $2500 cheaper.

I think that the HPEVS forces air through itself for cooling, and the Netgains certainly do. It would be of some advantage to have a motor that did not draw air in, since a tractor could be used in a hay paddock with lots of hay and dust floating around. I assume that a brushed DC motor might theoretically be able to ignite dust if it collected around the commutator?
The Netgain website lists a cover that can be mounted onto the motor to allow forced air cooling. The air could be drawn through a dust trap such as the one used on diesel tractors.
I think that some AC motors are sealed, and cooled through fins on the outside of the case?

The graphs I was looking at for the HPEVS motors developed their power at very high RPM. Tractors are generally quite slow, with 2000 RPM being typical.

[zeva]

I'd need to ask @zeva if it could handle regen from a DC shunt motor.

Unfortunately not, the motor controllers I make are a "quarter bridge" design which can't support regen. As far as I know, Zillas and Curtises are the same story.

I think Kelly do some half bridge controllers that can support regen on DC motors, but only for PMDC. SepEx can do regen but need a different sort of controller again. Shunt and Series DC motors can't do regen because there's no permanent or separately excited magnetic field for the rotor to spin in to generate current. But, I'm not so sure regen would be important in a tractor.

As far as speed controlling brushed DC motors, it's possible but you would need some kind of speed sensor and feedback loop, since their speed is highly dependent on load. Could be done with a PID controller dynamically adjusting the throttle on a DC controller, but not trivial. In the case of AC, the controller inherently knows motor RPM, but you would need it to support speed control rather than torque control (more common for traction applications) - but most decent ones should.

[4Springs]

I think Kelly do some half bridge controllers that can support regen on DC motors, but only for PMDC. SepEx can do regen but need a different sort of controller again. Shunt and Series DC motors can't do regen because there's no permanent or separately excited magnetic field for the rotor to spin in to generate current. But, I'm not so sure regen would be important in a tractor.

Indeed, I'm not after regen per se, I'm after a way to stop a shunt DC motor overrunning if I'm going down a hill when my brakes don't work. I'm after engine braking, and the discussion above suggests that regen would be the best way to do it.

The Kelly KDH range say that they support regen, for instance $2000 would get you one of these: https://www.green-scooter.hu/kdh14801e
This is the modern version of the controller I have in my car. The software lets you put it into 'speed' or 'torque' mode. Looks like it will drive a series or Permanent Magnet motor. I understand that regen current is not developed for series DC motors, but the discussion above suggests that it would be for shunt DC motors?

[coulomb]

I understand that regen current is not developed for series DC motors, but the discussion above suggests that it would be for shunt DC motors?

Sorry, I was thinking separately excited DC motors for regeneration. I've just realised how rusty I am on motors, so I'll butt out of this discussion, except to suggest that a separately excited DC motor and controller might be what you need.

[francisco.shi]

A quarter bridge controller will not slow you down but it will hit a maximum speed.
It will not overspend like a series wound motor.

[brendon_m]

Does it have to be a new motor? Could you use a 2nd hand ev/hybrid motor and inverter. I know EV wrecks are hard to come by in Australia but might be an avenue to look down. Might even be worth the freight to import something

[4Springs]

Does it have to be a new motor? Could you use a 2nd hand ev/hybrid motor and inverter. I know EV wrecks are hard to come by in Australia but might be an avenue to look down. Might even be worth the freight to import something

Good point.
A single Outlander motor would probably be a good size. Outlanders have been around for a while now, so there might be some wrecked ones in Australia?
I'm not sure that the throttle response in the existing inverter would be set up correctly for my application though. When driving an Outlander the throttle pedal is a torque control, takng torque from positive to neutral to negative (regen). In the tractor I need speed control. I have no idea how versatile those inverters are...

[brendon_m]

Well, a repairable write off phev went at pickles auction in Queensland yesterday. Last I saw with 12 hours to go it was at $3100.
I don't know if the mitsubishi inverter has been cracked, or if it's been cracked to the level you need but one from a prius or lexus (which are both pretty well sorted) should work with the mitsubishi motor.
You just need to make frankensteins tractor

[francisco.shi]

Speed control is quite simple. I presume all inverters take a torque command (at least the leaf and my inverters do) and will send speed feedback. All you need to do is have a set point for your speed work out the difference between your set point and the speed feedback from the inverter and multiply that by a constant and that is your torque set point for the inverter.
I think the Mitsubishi inverters have already been hacked. I think they work very similar to the leaf. You send a CAN message to a particular address with your torque command and the inverter replies with the current speed and other useful info.

[4Springs]

Speed control is quite simple. I presume all inverters take a torque command (at least the leaf and my inverters do) and will send speed feedback. All you need to do is have a set point for your speed work out the difference between your set point and the speed feedback from the inverter and multiply that by a constant and that is your torque set point for the inverter.
I think the Mitsubishi inverters have already been hacked. I think they work very similar to the leaf. You send a CAN message to a particular address with your torque command and the inverter replies with the current speed and other useful info.

Hmm.., Looks like an interesting project.
Any leads for where to look up information on this stuff? I like the idea of re-using motors/controllers from written off vehicles. I've never worked with CAN, and it looks like it might be a long journey, but it might be worth it.

[francisco.shi]

The velocity loop is very easy to do. Just 3 or 4 lines of code. (Meaning a 5min job)
You need a board that has a CAN bus interface.
A friend is making a body controller and it was meant to be used to run a leaf motor.
We already have the leaf messages.
Grahame from Suzie auto offered to lend his imev so we could hack the messages.
I also found some website somewhere that had some imev can message logs. I didn't look further because I didn't have a motor to test it on.
CAN is not difficult if you have something that talks and listens on the bus.

[francisco.shi]

I am sure my friend will sell you a board. I am using it on my conversion so I could help you with the software.

[HuffnPuff]

Sounds like an interesting project. I’m a long way from an knowledgeable about this stuff, but have you considered a motor for PTO and a motor for drive? The example I’ve seen is an electric ride on. Has several motors because it’s easier to drive the blades independently of the motion systems instead of a bunch of belts trying to drive both off one shaft.

Rather than thinking how to make an electric motor behave like an ICE but think how the end result works most efficiently. Although I understand dealing with an existing gearbox is probably cheaper than trying to reinvent it.

[4Springs]

Sounds like an interesting project. I’m a long way from an knowledgeable about this stuff, but have you considered a motor for PTO and a motor for drive? The example I’ve seen is an electric ride on. Has several motors because it’s easier to drive the blades independently of the motion systems instead of a bunch of belts trying to drive both off one shaft.

Like you I'm also a long way from knowledgeable! But I have thought about your proposal.
The trouble is that the PTO motor would need to be about the same power as the drive motor, so everything gets doubled. I'd need a second motor, a second controller and another set of controls. A tractor gearbox has all of this built in already. They generally have a two stage clutch. The first stage lets you stop or start forward motion while keeping the PTO implement going at it's rated speed. You can select the gear you need for the desired ground speed or direction. The second stage stops motion and the PTO, so it's great in a panic.
My electric ride-on lawnmower has three motors with three controllers. As you say, this is a cheaper way to design from scratch, at least within the low power requirements for a lawnmower. I'm trying to avoid designing from scratch...

[Adverse Effects]

ok thinking out side the box here

have you though about a outrunner BLDC?

i am NOT talking about an inrunner but an outrunner specifically

easy to cool because the only thing moving is a ring of magnets and if you put a small fan blade on the out side they make there own draft and the center is all the electronics and no brushes or sliprings and you can easly install some sort of liquids pipes in to it to cool that so no need for dusty air flow through there, and as we all know the bigger the diameter of the motor the slower it revs and the more tuque it has and that is what you need for a tractor it would also be very easy to mount and the flat face that spins could simply have something like a clutch center plate and splines bolted to it,

it would also give you controllable regen and thus controllable engine braking

GOOGLE IMAGES OF SOME

[4Springs]

have you though about a outrunner BLDC?

No, I have some learning to do.
Looks like they are often used on aeroplane propellers.
This one for instance: here
4.6 kg for 40 kW is light!
85 Nm is ok I guess. I'm not sure what I need though. They give this same figure for the 40 kW, the 50 kW and the 100 kW versions...
28 KV would give up to about 4000 RPM at 150 V.

easy to cool because the only thing moving is a ring of magnets and if you put a small fan blade on the out side they make there own draft and the center is all the electronics and no brushes or sliprings and you can easly install some sort of liquids pipes in to it to cool that so no need for dusty air flow through there, and as we all know the bigger the diameter of the motor the slower it revs and the more tuque it has and that is what you need for a tractor it would also be very easy to mount and the flat face that spins could simply have something like a clutch center plate and splines bolted to it,

Hang on, the heat will be made in the coils, which are on the inside. They look like they're made to be open so that they are cooled by the flow of air around them (on a plane there would be lots of air!). I'll want it to be sealed off from the air. I can't see how they could be liquid cooled, pipes would have to go through the centre somehow.

[Adverse Effects]

outrunners are not like normal motors you can have a 100mm center shaft if you want because it dose not spin

here is just 1 of the motors i am talking about

now just because it says XXX RPM at XXX volts dose not mean you have to run that voltage you can run less volts the motor just wont spin as fast

emrax 188

[Bryce]

As an FYI - the Emrax motor is commonly used in FSAE racing EVs. (https://motorsport.org.au/about/education/formula-sae)
These are the teams I work with in my day job.

Cheers
Bryce