1
\$\begingroup\$

I have an old Black&Decker battery powered mower that uses a 36 volt SLA battery pack. The original motor failed catastrophically and I now have a 24 VDC motor adapted to fit. I added a PWM buck controller that can be set to drive the motor with the required voltage. I also added a fuse and a switch to disconnect the battery. I have a lot of 16A 250V AC switches that are small enough to fit the enclosure, and at first it seemed OK, but then it would not turn off. I inspected the contacts and they looked OK, but obviously they must have tack welded at some point. I replaced the switch and it happened again, and I have not yet connected the motor.

In a previous post, I sought ideas about limiting surge current into a large capacitor with 240 VAC applied through a bridge rectifier, and one suggestion was a series inductor of 1 to 10 uH. I didn't think it would be very effective, but I thought I would see if it might work for this problem. My simulation showed a current surge of nearly 100 amps, and a 10 uH inductor only reduced that to about 85 amps. I used a freewheeling diode to battery negative to handle turn-off transients. 100 uH reduced the turn-on surge to about 60 amps, but a 100 uH inductor able to handle the expected 15-20 amps into the controller would be prohibitively large and expensive. An NTC thermistor might work, but would generate a lot of heat, and could "starve" the controller when it needs to pull current from the battery pack.

I will probably need to find a suitable DC rated switch, but space limitations will make that difficult. Here is my simulation with 10 uH:

Surge current into capacitor simulation

\$\endgroup\$
7
  • 1
    \$\begingroup\$ I don't think you understand what is (probably) happening. The switch, when you turn it on, is bouncing (this is normal.... all mechanical switches do it). The first time the switch contacts all is well. But then current ramps up and the switch bounces, breaking contact while current is flowing. This causes an arc that melts the two contacts. When it makes again, the molten contacts stick together. Adding more inductance will probably make it worse. The easiest fix probably is to pre-charge the capacitor through a resistor before turning on the main switch. \$\endgroup\$
    – user57037
    Commented Jun 30, 2022 at 5:51
  • 1
    \$\begingroup\$ So you could have two swtiches, one pushbutton pre-charge switch, and the main switch. Never turn off the main switch while the motor is running either or you may well have the same problem. If there is no other way to stop the motor then you have a difficult situation and electronic switching might be your best solution. \$\endgroup\$
    – user57037
    Commented Jun 30, 2022 at 5:52
  • \$\begingroup\$ There are several different strategies related to inrush limiting and if you search for that term you may find some good ideas. \$\endgroup\$
    – user57037
    Commented Jun 30, 2022 at 6:16
  • \$\begingroup\$ One thing I am not clear on is whether the controller automatically just starts spinning the motor as soon as you turn on the switch, or if a separate step is required before the controller starts to drive the motor. \$\endgroup\$
    – user57037
    Commented Jun 30, 2022 at 7:05
  • \$\begingroup\$ That relay seems very much undersized. why are you using such a wimpy relay? \$\endgroup\$ Commented Jun 30, 2022 at 7:13

3 Answers 3

2
\$\begingroup\$

I think a good, practical solution to this problem would be to use a PMOS device with an ON resistance of 20 mOhms or so, and a continuous current rating of 74 amps, such as the IRF4905. They are available for less than $3 each in a TO220 package. The switch would provide gate voltage through high value resistors, and a gate capacitor to limit the peak current during capacitor charging to about 70 amps for 1 mSec or so. The continuous 15 amp current during normal operation would result in about 5 watts which would require a modest heat sink. I have some IRF5210 devices that I could try, although they have 60 mOhm resistance for 15 watts, but at least I can try one to verify the principle at lower output power. The closest device I could find in LTSpice is a RSJ250P10 with 45 mOhms:

Simulation with PMOS device

Please let me know if you think this is the best approach, or if you have a better idea. Thanks for your answers and comments so far.

\$\endgroup\$
5
  • 1
    \$\begingroup\$ If you are going to implement an electronic solution, I think the best approach is to pre-charge the large capacitor through a resistor THEN turn on M1 quickly. A sequenced approach. Once the large capacitor is charged up, there should be no problem with the contacts welding. Turning on M1 slowly may have a tendency to blow-up M1 unless you do it REALLY slowly. \$\endgroup\$
    – user57037
    Commented Jun 30, 2022 at 7:03
  • \$\begingroup\$ Pre-charging the capacitor will power up the controller, which draws some tens of milliamps from the battery pack. I need to be able to remove all significant load from the batteries to avoid draining them when not in use. I could use a three position switch with a precharge position, and make sure the capacitor is charged before turning it fully on, but I don't have room for such a switch. The IRF4905 can take a 240 amp pulse drain current and the estimated 400 mOhm total series resistance at 36 volts is limited to less than 100 amps. \$\endgroup\$
    – PStechPaul
    Commented Jun 30, 2022 at 8:32
  • \$\begingroup\$ This schematic in linear mode can cause FET failure. Why did you add the cap that causes this problem? Let the controller manage the switching properly. \$\endgroup\$ Commented Jun 30, 2022 at 12:27
  • \$\begingroup\$ There appears to be some confusion here. The charging current has nothing to do with the controller switching or the motor. The current surge is from the 36V SLA battery pack through the switch into the large electrolytic capacitor on the input of the controller (and I am only guessing the 1000 uF and 200 mOhm ESR). The MOSFET will be in linear mode only during the initial charging period. The total energy absorbed by the MOSFET will be about the same as that of the charged capacitor, or about 600 mJ, and the IRF4904 can handle 900 mJ avalanche energy. \$\endgroup\$
    – PStechPaul
    Commented Jun 30, 2022 at 19:18
  • 1
    \$\begingroup\$ You could have an electronic circuit that manages the pre-charge. If you don't want to put a little microcontroller in there you can use comparators and RC time delays and whatnot. I do have experience with connecting batteries to large caps. I have designed two different motor controllers for products with user removeable battery packs. You get giant destructive sparks unless you implement some type of pre-charge. Using a FET to turn-on slowly will lead to blown FETs. The FET is a good idea. It is just that you need to move the dissipation out of the FET into a resistor. \$\endgroup\$
    – user57037
    Commented Jun 30, 2022 at 22:46
1
\$\begingroup\$

Perhaps the problem is the input capacitor on your PWM controller. Essentially a capacitive (dis)charge welder. You may be able to replace it with a beefy TVS rated for more than than 36V but well less than the maximum input voltage of the controller.

It might not work or might lead to damage of the controller though. At least I can imagine such situations.

\$\endgroup\$
3
  • \$\begingroup\$ I don't think a TVS will help with reducing the capacitor charging surge current. The problem does not appear to be a voltage transient, although there may be enough inductance to cause arcing and tack welding. There is a considerable arc when I insert the fuse, but 36 volts is enough to weld contacts. \$\endgroup\$
    – PStechPaul
    Commented Jun 30, 2022 at 3:54
  • \$\begingroup\$ @PStechPaul there will be very little capacitor charging surge current if the capacitor is unsoldered and sitting on your bench, and replaced with a TVS. \$\endgroup\$ Commented Jun 30, 2022 at 7:44
  • 1
    \$\begingroup\$ Now I see what you meant. But I don't think it's a good idea to remove the capacitor from a purchased unit, unless defective or poor design. The PWM expects to have a low impedance energy source (capacitor) close to the switching elements, and it will probably be problematic to draw those surges from the battery through wires, fuse, and switch. \$\endgroup\$
    – PStechPaul
    Commented Jun 30, 2022 at 8:05
0
\$\begingroup\$

The low DCR motor inductance is best regulated with a ramp up in frequency of constant pulse width such that the desired I impulse dI= V/(Ldt) where V reduces from rising back EMF RPM. The time to reach max RPM will increase but the peak current can be regulated without exceeding rated current at idle motion of lawnmower.

Added: your (commented links) ESC (electronic speed controller) should do this properly.

my Estimates

Cap surge current= 24V/0.25ohm = 96 A
Motor impedance Rm= V^2/W= 1.64 ohms. DCR ~ 10% of Rm= 0.16 Ohms
Motor max surge/stall current= 3600 A in theory if battery was applied direct to motor, which is why it has a soft start.

Recommendations

  • You should not need anything besides motor and ESC
  • Do not add a 1mF cap . The battery ESR won't be much different but a million times larger C {Farads} than 1 mF.
  • do not use a DC switch to cut off motor power, unless rated for arc voltage or surge current to turn on or surge current to stop blade.
  • do not stop the motor with anything else other than the motor ESC controls

If you do add anything please add your design specs, and calculations in your question.

Your motor and ESC are adequately paired, what's the problem?

Conclusion

It looks like the ESC has the front-end low ESR cap and the "XT-90" is the cheapest soft charge solution. (suggested by @mkeith ) These have a power resistor like 5 Ohms that makes the initial contact on the + side and charges up before you insert fully to 0 resistance.

schematic

simulate this circuit – Schematic created using CircuitLab

I am guessing on the ESC input capacitance but this will suppress the arc to 90A up to 450V (in theory) if the internal R is 5 ohms. (verify req'd) Thus with 24V the current will be <5A then rise as the DCDC converter charges up in the ESC then the current drops towards no load. These could be rated like ICL's for Joules of charge and transient heat rise. \$E=1/2CV^2 [J]\$.

\$\endgroup\$
11
  • 1
    \$\begingroup\$ The problem exists without the motor connected. Once the capacitor is charged, the battery current through the switch into the controller should be no more than about 15 amps, for motor power of about 500 watts. \$\endgroup\$
    – PStechPaul
    Commented Jun 30, 2022 at 4:20
  • \$\begingroup\$ Did you add the Cap? Is it a brush motor or BLDC? Without acceleration control the motor surge current can be 5x to 10x rated current and with PWM the 1mF cap will smoke the driver with 0.25 ohm ESR and the cap. \$\endgroup\$ Commented Jun 30, 2022 at 5:16
  • \$\begingroup\$ The controller is a 5000 watt 10-55V 60A unit ebay.com/itm/… and the motor is a 24V 350W unit similar to ebay.com/itm/… \$\endgroup\$
    – PStechPaul
    Commented Jun 30, 2022 at 5:40
  • \$\begingroup\$ Ideally I would use the ESC controls as you suggest, but I am trying to retrofit a 24 VDC motor in a mower that used a 36 VDC SLA battery pack and a 36 VDC motor. The mower's hand controls are connected from the motor to the battery pack, which I have essentially modified to emulate a 24 VDC source. I could not find a reasonably priced and compatible 36 VDC motor, although that would have been best. Considering the time and money I have already spent, I should have just purchased a 36 VDC motor like: ebay.com/itm/154769201801 for $66. Maybe I will just do that! \$\endgroup\$
    – PStechPaul
    Commented Jun 30, 2022 at 19:39
  • 1
    \$\begingroup\$ @TonyStewartEE75 connecting an ESC to a battery will lead to sparks unless inrush is managed. I believe that is the issue OP is having. Upon connecting the battery to any device with large input capacitors, the inrush has to be controlled. The OP doesn't want to leave them permanently connected because the ESC consumes power. \$\endgroup\$
    – user57037
    Commented Jun 30, 2022 at 22:49

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.