Precision, energy and battery for a beginner


#1

Hello

I am french (but I take care of myself) and I find this project great to avoid the use of stepper motors in CNC machine.
Before I get started I have some questions :

1>>Are there any other demo videos with the latest hardware and software version?
2>>Odrive makes it possible to go fast, ok, but what about the precision of the positioning ? typically for a machining machine?
3>>How can one compare the energy consumption of Odrive compared to a stepper motor with equal torque?
4>>You use a lithium battery to recover energy and to withstand peak consumption; Were not you afraid of destroying it or worse, that it explodes? (strong currents and overloads)

Thanks in advance


#2
  1. not as far as I know. Did you watch the Handcrusher5000 video in the blog?
  2. depends on the precision of your drive or lead screw and the resolution fo the encoder. In theory, something around 0.1µm could be done using Renisaw encoders with 0.02µm resolution. Sadly, they are very expensive.
  3. Depends on how much you the steppers are overdimensioned. There is no ballpark figure. Energy consumption isnt the main interest here but generated heat. The less heat a motor produces, the more load it can take, so a smaller and cheaper motor can be used for the same task.
  4. nope. You try to keep the battery at its nominal voltage. If the battery is large enough, it’s impossible to overcharge it. Which size you need exactly depends on the current demands of your application. Some example calculation:

6s Lipo battery, charged to 22V. It schall not exceed 25V, so there is a 3V margin before the battery could take damage. Say you want to brake with 50A, then you can calc the maximum inner resistance of the battery: V = R*I or R=V/I = 3V/50A = 0.06 Ohms for the battery or 0.01 Ohms per cell. A 30C lipo with 4ah or a 90C lipo with 2ah should be ok. Better use a bit more, as contact resistance, wire resistance etc add up to this figure


#3

Thanks for your answer

1>> I was looking for videos with slow movements, static load tests, precision tests etc.
2>> Which oscillations, number of steps, must be taken into consideration when considering the precision of positioning of motor + encoder assembly? 1 encoder step ? 2 ? 10? 100?
3>> I will see well, but surely better, a stepper motor is a radiator by principle.
4>> I understand nothing at all to your explanation… sorry.
For me a Lipo battery must be charged at constant current and your calculation shows current values very much higher than this acceptable current, even for a battery called “ultra-fast charge”.
The PCM on Lipo batteries is there to block under loads, overloads, voltage and current as well as short circuits. If the motor consumes a constant current, without peaks, the Lipo does not discharge and has time to complete its load; then braking is inevitably an overload. Normally if the PCM lets do it is good but the pcm costs 0.5$ (good security ?) and normally the PCM is not a regulation, it must act only in ultimate safety.

Thanks


#4

Hi @po220, welcome.
Just out of curiosity, what interface do you wish to use for your CNC application, step/direction?

All the demo videos I have are here. I will put your suggestions (slow movements, static load tests, precision tests) in my list of future demos.

This really depends. Of course the stiffness of your mechanical assembly is the most important. Assuming you have a very stiff machine, then there are some other factors:

  • Disturbance force: vibration from the cutting
  • Stiction and other non-linearities.
  • Moving mass (larger mass is better, since it makes smaller accelerations for the same forces)

For a very ideal machine, where stiction is low, and under no disturbance, I think something like 1-2 encoder counts is doable. For something like a CNC machine it’s really hard for me to guess about the stiction and disturbance forces. I think someone just needs to measure it on a representative CNC machine, or we perhaps we can find some data from other servo manufacturers.

The short answer is that ODrive is way more efficient, mostly because at zero load it generates zero losses. Skipping a long discussion: you shouldn’t compare torque directly, since these motors can run much faster, so you should gear for that, and then you can use less torque and run more efficient.

For this reason maybe starter motor lead-acid batteries are more suitable. I haven’t actually tried either, but there was some discussion about trying the lead-acid here: Lead Acid battery as charge buffer.


#5

The larger the battery, the more current it can eat. Chose a large enough battery for your application and all will be fine. Using a BMS is always good but pricey. lead acid batteries that need no BMS/PCB…what ever PCB shall mean, but i think you mean a BMS = Battery Management System


#6

Just out of curiosity, what interface do you wish to use for your CNC application, step/direction?

Today I’m only discovering. Tomorrow I would use Odrive for pick and place and small machining machines. I saw that it was possible to use Odrive with step / direction as a stepper motor driver. I do not know if there is a more appropriate solution for that.

Thanks for the futurs demo. I am part of a French community on the machining diy and many people are wondering about the behavior of Odrive in machining (slow movements with load)

For the battery, I need to look at the code and the doc to understand how to use it. I will look carefully at the subject on the significantly more reassuring lead acid battery.

A BMS is important to properly charge the Lipo battery
A PCM (Protection Circuit Module) is integrated into lithium batteries to limit the risks associated with their use. It cuts off the battery when it is overloaded.


#7

For a very ideal machine, where stiction is low, and under no disturbance, I think something like 1-2 encoder counts is doable. For something like a CNC machine it’s really hard for me to guess about the stiction and disturbance forces. I think someone just needs to measure it on a representative CNC machine, or we perhaps we can find some data from other servo manufacturers.

My worry about this is more about the nature of the brushless motor itself, which is not here originally designed for this kind of slow and high torque movements.
But you are right you should try. I can not work on this until 2018 but afterwards I will do testing if no one has done it before me.
It would be nice if all Odrive users could show what they are doing with!