0
\$\begingroup\$

I'm trying to figure how to protect batteries from overvoltage on my electric skate. Specific condition is : motor would run above max RPM based on system voltage.

Motor is 85Kv, battery voltage is 33.6V. This gives max RPM of 2856.

Goal is to be able to spin the motor up to 4250RPM, which would bring "motor" voltage up to 50V.

Can I use TVS diodes (starts working at 30V, fully operates past 33.5V) to redirect / suppress exceeding voltage toward a loop of resistors ?

I have a hard time figuring their use. Was thinking of making a combo TVS + Schottky diodes to force one-way direction + a big resistor in between to dissipate exceedding power.

Thanks

\$\endgroup\$
5
  • \$\begingroup\$ If you want to allow regen braking at high speed, you should consider the maximum charge rate of the battery. Usually batteries have much higher discharge rates than charge rates. Regenerative braking without a dynamic brake resistor could lead to battery over-charge (meaning charging too fast). Any scheme that causes a TVS to experience prolonged conduction will not work. The TVS will rapidly overheat and fail. The 'T' in TVS stands for "transient", which means temporary and of short duration. \$\endgroup\$
    – user57037
    Commented Jun 11, 2017 at 16:45
  • \$\begingroup\$ @mkeith Thank you for the explanations. Ok so I should check for a 500-600W resistors setup to get a safety margin (based on max 480W of regen). \$\endgroup\$
    – Vanarian
    Commented Jun 12, 2017 at 10:11
  • \$\begingroup\$ It is a starting point. I don't know what your actual regen current will be. It would be nice to measure the battery charge current during actual regen. I am only saying that 480W is a reasonable amount of power. Braking down from 30 km/h should be doable with 480W. Because of physics, the faster you go, the more power required to brake. If your controller allows you to brake hard at 75 km/h, the regen current will be very large. Something may fail. This is really all I can say without getting into a lot more design detail. Be careful! \$\endgroup\$
    – user57037
    Commented Jun 12, 2017 at 16:04
  • \$\begingroup\$ @mkeith True it also depends on the force of braking ; I'll try to modulate the braking current - hard brake at high speed is dangerous, progressive is better. Just in case I'll overdo the resistors. Will come back with a diagram for a double check. \$\endgroup\$
    – Vanarian
    Commented Jun 13, 2017 at 9:11
  • \$\begingroup\$ You can limit charge by coasting or braking. with/without regeneration. Your choice. \$\endgroup\$ Commented May 10, 2019 at 2:59

2 Answers 2

1
\$\begingroup\$

In some cases the best thing you can do if an overspeed condition exists is to short all three phases together. This does not necessarily result in a lot of braking torque. It depends on the motor. I know it seems like this should create enormous current and torque but I have in fact done this on a dynamometer with a current probe, and with the motor I was testing, there was no drama, no high torque, and no overheating problems.

The key to understanding why is to recognize the role of motor inductance in limiting the motor current. As you overspeed the motor, the voltage grows linearly, but the frequency also grows linearly, so the impedance due to inductance of the motor winding increases linearly, too. So the phase current approaches a steady state value which is, in some cases (depending on the motor) not that high.

\$\endgroup\$
2
  • 1
    \$\begingroup\$ Shorting shouldn't do too much to an induction motor (but it depends on the field state just before shorting, I think?), and will do a pretty tremendous halt on a synchronous/PMAC/BLDC type. Probably not a great idea on a skateboard, lol. So be careful with that. Hm, it should be possible to give a PMAC motor a highly reactive load as well, so that energy circulates through it, dissipating power in winding and core resistance, rather than dumped into the battery. This will require a special controller mode though. \$\endgroup\$ Commented Feb 1, 2023 at 5:45
  • 1
    \$\begingroup\$ @TimWilliams the highly reactive load will produce far more braking than shorting the phases. Actually FOC can be thought of as presenting a highly reactive (or even active) load to the motor in order to produce the desired torque. The best thing to do is use FOC and enable field weakening. This will allow controlled torque over the maximum range of speeds. But there could be scenarios, even on a skateboard, where transitioning to shorted phases is the best and safest bet. \$\endgroup\$
    – user57037
    Commented Feb 1, 2023 at 8:05
0
\$\begingroup\$

I would consider using a relay (possibly a solid state type) that disconnected the motor from the drive electronics when the motor terminal voltage rose above a critical value. when the voltage dropped back to a safe limit the relay would reconnect.

If you need to load the motor (now a generator) in the over voltage state then this is also possible however, I see no immediate reason to make a load if the motor is disconnected via a relay.

\$\endgroup\$
8
  • \$\begingroup\$ Thank you for advice. My concern is to still be able to brake with the motor as long as it remains under the 50V margin. Actually I'd use the relay in a situation where terminal voltage cross ESC limis (limit is 58V) in order to protect as you suggested. Which relay type would you recommend ? I kinda try to make a 2-step protection, first step between 33V & 50V to still be able to control the board ; past 50V to be in "freewheel" mode and not shutdown the system. This gives more chances to slow down back to a speed where system can control the board again. \$\endgroup\$
    – Vanarian
    Commented Jun 11, 2017 at 10:59
  • 3
    \$\begingroup\$ To electronically brake, you need a place to dump the power. With carefully designed circuitry that could be regenerative charging of the battery pack, but only within spec and not when it is full. Look at electric railcars and you'll often see a big bank of air-cooled resistors. If you want to calculate their sizing, consider the amount of hill you want to be able to descend per unit time at constant velocity and the weight of the board and rider, and convert the lost potential energy per second to watts. \$\endgroup\$ Commented Jun 11, 2017 at 11:59
  • 1
    \$\begingroup\$ @Vanarian there isn't enough information provided in your question to make a judgement on a specific relay. Also this isn't a shopping site except for guidance, generalities and engineering specifics. I'm not against giving product recommendations but it requires a lot more work on your part precisely defining things and quite possibly too much of my Sunday. \$\endgroup\$
    – Andy aka
    Commented Jun 11, 2017 at 12:09
  • \$\begingroup\$ With the braking requirement a relay probably isn't going to work anyway. Or at least there would need to be two - one to disconnect the ESC, another to apply the brake. But an all-or-nothing brake may make for a challenging ride, either minimally effective or having the potential to throw you when applied. So then you need to module the brake, which sounds like a job for FETs as found in an ESC, rather than a relay. So what it comes down to is that you need an electronic drive rated for your requirements, rather than the underrated one you want to use. \$\endgroup\$ Commented Jun 11, 2017 at 12:12
  • \$\begingroup\$ Thank you guys for the replies ! About my electronics : I run a custom Vedder ESC based on LM5109 driver, good up to 100V but smaller components on the PCB make it only 58V proof without mods. It runs 6x FETS ifrs7530 air cooled ; Batteries are 8S4P made of LG HB2 ; each cell will be balanced up to max 4.10V at charge. ESC provides regenerative braking, motors are sensorless. I was also reflecting on the dump system, I'm glad I was on the good path. So in the end you'd recommend switching FETs to redirect the exceeding power toward resistors ? Still a learner here. \$\endgroup\$
    – Vanarian
    Commented Jun 11, 2017 at 23:29

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.