[remy_martian]

from: https://www.automobile-catalog.com/curve/2022/3000230/chevrolet_bolt_euv.html

Preview:

So much for a "flat torque curve from 0 RPM"...

Appears to be some rather crude limiter code in the inverter controller at ~3900RPM

I wonder if it's the measurement system/technique doing all this curve weirdness or if these are actual curves.

These cars are currently scarce in salvage, but that will hopefully change in a year or two - if they don't kill the model off.

 

[kennybobby]

That is not measurement or test data.

It is from a ProfessCar software simulation, written by Pawel Zai, using OEM sales and marketing "performance specs" to guesstimate 0-60 and 1/4 mile times.

Without voltage and current data it is just mathturbation.

Here is a Tesla ludicrous launch mode run data from 2016 showing the "flat" torque command at T=0 which then has a pullback after launch to control wheel spin. Even with all that it takes 1.3 seconds to reach full current draw. i'm sure Chevy and all the OEMs do something similar to limit current if given WOT

 

[Ken S]

Measuring torque in an automotive driveline is not trivial. Can anyone recommend an economical technique to measure torque in the range of 200~250 ft-lb?

 

[brian_]

Measuring torque in an automotive driveline is not trivial. Can anyone recommend an economical technique to measure torque in the range of 200~250 ft-lb?

Do you want to measure the torque output of a separate motor, or the complete powertrain in a car? Dynos are available for both, and 200~250 ft-lb at the motor output shaft is a routine automotive value. It's a service you normally just buy.

If you want to do this in a home garage way, especially for a separate motor, you can drive a brake rotor with a caliper clamping it, and the caliper on a pivoting mount restrained by a lever arm with a load cell (such as a hydraulic cylinder with a pressure gauge) measuring the force required to anchor the arm. Torque equals force multiplied by lever arm length; power equals torque multiplied by power, and the whole thing will only work in brief spurts until the brake pads overheat. This is basically how the data published by NetGain for their motors is obtained; it's also where the term "brake horsepower" originated.

Real electric motor performance testing is done by driving a generator with the motor, varying the load by controlling the generator, and measuring torque in the shaft with an inline torque measurement (typically based on strain gauges in a calibrated shaft section). That's not economical to set up.

 

[Ken S]

Thanks, Brian.

I see the error in my late-night posting (serves me tight) It seems I omitted a key term from my post: "In situ."

I'd like to measure motor torque in an EV while the vehicle in motion, under normal driving conditions.

 

[remy_martian]

In the Bolt you have the instantaneous power usage given to you in the instrument panel in kW. You can derive the torque with some math from that and knowing the SoC to get the current consumption (which is torque). An OBD-II monitor might give you more/better data if you want accuracy.

Easier to go to a tuner and do a dyno run.

 

[brian_]

Thanks, Brian.

I see the error in my late-night posting (serves me tight) It seems I omitted a key term from my post: "In situ."

I'd like to measure motor torque in an EV while the vehicle in motion, under normal driving conditions.

Ah... that makes sense.

I see three approaches:
Power or current and efficiency
remy_martian described this one: the car reports the electrical power used at any moment. That power divided by an assumed efficiency is the power being delivered. Power divided by speed is torque. This will be close except in extreme conditions (very high or low load or speed), where efficiency changes substantially.

If the car reports motor (not battery) current that's even better: the motor has an ideally constant ratio of current to torque, and if you can determine that from any condition you can apply the same ratio with reasonable accuracy even in significantly different conditions.

Apparently some of the Tesla in-car displays include torque, and this is essentially how they are determining it... but they have not just one ratio but a detailed map of motor current to torque built during the development of the motor. And they used a torque meter to measure the motor's output torque during testing to build that map. This level of sophistication is not realistically a DIY solution.

Performance
The acceleration of the vehicle, minus the drag, indicates how hard it is being propelled. You can work this out in terms of power (torque multiplied by shaft speed equals power), or of force (torque to wheels equals force times tire radius), but either way you need to know how much drag there is, and that can be determined by a coastdown test (and is roughly indicated by power consumption at constant speed).

Direct measurement
... or at least relatively direct: turn the axle shafts into torque transducers by fitting the with strain gauges (and a way to read the gauges on a spinning shaft!). This is theoretically sound, technically feasible (it's how real inline torque meters work), but in practice not an easy project or anything that I would expect a DIY enthusiast to do.

 

[Ken S]

I am familiar with adding strain gauges to a drive shaft. I think for now, I will stick to the data I can get from the motor controller. I like the idea of a coastdown test on level ground. I should be able to build a model of total drag (the sum of air resistance & rolling resistance & drivetrain friction & ...) as a function of speed. Then, acceleration tests on level ground. Then...