You are here

Protecting PSU with resistive load when braking

9 posts / 0 new
Last post
henxan
Offline
Last seen: 11 months 2 weeks ago
Joined: 2019-11-13 22:52
Posts: 6
Protecting PSU with resistive load when braking

Hi!

I'm doing a DIY-project for Christmas, a gift for my wife. It is an electric winding machine - a machine that winds thread on spools. Example of such a machine, controlled by a pedal, is shown next:

My setup is as follows:

[PSU] ------------<X>---------[VESC]------[BLDC]

Where the VESC is controlled by a modded servo tester with a potentiometer pedal. The "X" in my drawing symbolizes a resistive load circuit.

The winding machine does not require a lot of torque, at least not compared to 100kg (200pound) person on a skateboard. Also different from skateboards, I want the winding machine do decelerate/brake as fast as possible (maybe not recommended with skateboards :) ).

Since I am using a PSU to power the machine, it will not be able to absorb back emf/braking current. So, when I plan on using regenerative braking, I want to dump all reverse current to the resistive load.

I have done a great deal of searching on the net, and I have found several articles and forum posts outlining a solution, however they do not describe the specifics of how to construct such a resistive load. The closest I get is circuit diagrams, but I do not know how i convert the circuit to my specs. I am going to run between 18V and 23V psu, and the circuit need to suit this. Also, I want to brake at max possible power. I have ordered a Flipsky Mini FSESC 4.20, which has voltage rating at 8-60V and amp rating at 50A (150A intermediate). My motor is a Maytech 5055 Hall sensored with 70kV.

Here is image of two such circuits I found on the net:

So, basically, my two questions are:

  • How do I construct a circuit suitable to my need?
  • How should I set up VESC for this situation?

Any help is appreciated! :)

district9prawn
Offline
Last seen: 2 years 2 weeks ago
Joined: 2018-04-26 12:18
Posts: 123

Have you considered attaching a rechargeable battery in parallel with your power supply? Even a small lithium pack should be more than enough to prevent the controller from creating high voltages at the input during braking.

Sorry I know this doesn't directly answer your question but its a practical solution if you want to avoid faffing around with circuits.

henxan
Offline
Last seen: 11 months 2 weeks ago
Joined: 2019-11-13 22:52
Posts: 6

@district9prawn Thanks for the reply!

Yes, I have considered using a battery in parallel, however, I have decided against it. Eventually, the battery will fail - and I don't view it as a permanent solution. I also don't feel comfortable with lithium batteries which could fail and catch fire. When disconnecting the PSU, and putting the device away, I'm reasonably sure it will be safe.

However, I read another thread on this forum, about braking and use of  "duty cycle mode". As I understand this mode, it could be set up such that it functions as "dynamic breaking"? That is, the VESC shorts the three poles of the motor when braking, thus using its own generated power to break. This would consume and dump the braking energy as heat in the motor, and not generate a high reverse emf/current? If this is the case, I will probably utilize this mode. In my case, the torque is small, and I believe I won't expose the motor to any degrading heat. 

district9prawn
Offline
Last seen: 2 years 2 weeks ago
Joined: 2018-04-26 12:18
Posts: 123

Duty cycle mode is still using regen to slow the motor when the commanded speed/duty is below the current motor speed. 

Modulating shorted phases still results in some regen as the inductance of the motor results in voltage spikes. The mosfets can't block these spikes from pushing current back into the battery as their body diodes will rectify the pulses back to your supply.

If you don't want a lithium battery attached to your supply maybe a mechanical brake is your best bet? A mechanical brake would also allow faster stopping and better control.

twikis
Offline
Last seen: 5 years 2 weeks ago
Joined: 2018-01-09 22:04
Posts: 4

Depending on the PSU a (big enough) zener diode in patapara should do the trick.

Also, I'd recommend you to read this, if efficiency and maximum power out of the motor are important to you: http://vedder.se/2014/10/chosing-the-right-bldc-motor-and-battery-setup-...

Your kv is to low for that voltage...

pf26
Offline
Last seen: 3 weeks 5 days ago
VESC FreeVESC Silver
Joined: 2018-12-04 08:44
Posts: 53

Possibly you could power a little lamp from your PSU output, so that regen will never exceed this lamp power.. A 5055 motor should not regen that much if you ramp it down to zero speed slowly enough ? Might be little late for dec 25th though..

henxan
Offline
Last seen: 11 months 2 weeks ago
Joined: 2019-11-13 22:52
Posts: 6

So, I've built a "mock-up" of the electric winding machine - YAY!! :D .. However, when spinning up the motor, it gets choppy when I increase the RPM. Do anyone have an idea of what the problem is?

You can watch a video of this here:

Here is some realtime data I captured of this happening. I have to confess, I do not understand what exactly what I am looking at.

henxan
Offline
Last seen: 11 months 2 weeks ago
Joined: 2019-11-13 22:52
Posts: 6

Okey! Finally, I've got everything working! laugh

I still have one big question, as I have identified my problem:

The VESC switches from sensored to sensorless mode at a given ERPM - apparently at 4000 ERPM. Well, I changed this limit to 3000 ERPM, and the stutter suddenly started at lower RPMs. I tried increasing this to 5000, 7000 and finally 10000 ERPMs - at every increase the stutter appeared at higher and higher RPMs. Finally, at 10000 ERPM sensored cutoff, the BLDC was at max RPM, and no stuttering had appeared. smiley

Question:

  • Why should the sensored mode cut off at certain erpm? It seems to work fine, but if this could cause some problems later on, it would be nice knowing. smiley

BTW, here are two sampled data graphs from the problem before and after I fixed it. A stuttering episode is observed in the first graph:

Stupidspencer
Stupidspencer's picture
Offline
Last seen: 2 days 2 hours ago
Joined: 2019-06-20 00:06
Posts: 32

The reason for switching to sensorless at higher RPM is that motors typically run smoother this way. I think of it kind of like ignition timing advance on a gasoline engine. As the RPMs increase, more and more advance must be used to keep the engine running well because the amount of time it takes for the fuel to combust is constant whereas the amount of time available for it to combust is decreasing with increasing RPM. So to counteract that, the ignition needs to start that process earlier in the cycle. But the hall sensors in a BLDC motor are stationary and cannot advance, so at some point as RPMs increase, it becomes more efficient to use BEMF to get position data.

Your motors should run smoother sensorlessly, and the fact that they don't is certainly due to some setting being detected wrong.  The first setting I would mess with would be "Observer gain X1M". You can't hurt anything playing with this setting, so go ahead and (as the help popup will tell you) try halving or doubling its current value and see how your motor runs. Note that "observer" is the name given to the collection of circuitry and algorithms in a motor controller that monitor rotor position by sampling the BEMF generated by the motor.

If you don't end up getting it to run well sensorlessly, I think the only downside to running higher RPMs on sensored commutation is that the motor may draw more current and thus run warmer. You are not running on batteries, so the small bit of wasted power is not an issue for your application. So as long as the motor isn't overheating, and is performing to your expectations, I don't think there is any problem running on the sensors all the time.

I've seen this thread here for a few weeks, and while I'm responding here, I must ask:  Don't you think a VESC is a bit overkill for a bobbin winding machine? I don't want to belittle your project in any way, I am just curious as to why you chose such a powerful (and relatively expensive) controller for performing such a delicate task as winding thread.