I'm the caretaker of a brand-new (not really) Sebring Commuta-Van which has been repurposed as a runabout for my employer. The original schematics and manual are truly a museum piece, and I've concluded that none truly match the setup in the vehicle itself. By dint of a series of solenoids, the motor is given power at three speeds:
from cell pack A and B in parallel running through a huge resistor
from cell pack A and B in parallel
from cell pack A and B in series
Since it's a series-wound motor it's got a lot of torque, and I've disabled Speeds # 1 and #3, #3 after we took the car for a spin and somehow exploded the negative terminal of pack B. Oddly, there was no sign that we hit the 200A which would have blown the fuse on the system, but the terminal lead was very hot near to the blown terminal, and it was hot enough to melt / explode the lead.
In any case, I've gotten a handle on how I would like to charge the system (individual cheap solar charge controllers, one per cell), but I'd like to be able, ideally, to use the Arduino to monitor each cell's voltage, and serve as a unidirectional speed controller. From what I can find, 30A seems to be the limit of most MOSFET circuits. In a pinch, a relay module would probably get a bit more current-carrying capacity, but this isn't ideal.
What have been toying around with:
Build H-Bridge with truly beefy MOSFETs, or a unidirectional version
Purchase several 30A Arduino current source modules and mount them in parallel
Purchase relay module and retain the current (very primitive) motor control functionality
I like the third option best, but I am not well versed in laying out PCB's and determining the proper trace width for such a current load, but it'd be a good excuse to learn. What does surprise me is that the MOSFET's rated to handle 200A seem to have tiny pins for the amount of current they are supposed to handle. It also seems like there may be more steps to the puzzle, if there are no logic-level power MOSFETS that can handle such a high current, I may need to first convert the logic-level voltage to the 20V required to switch the MOSFET.
The ones I see on Mouser web site use multiple pins in parallel, except for those with screw terminals. All require heat sinks and the SMD devices ALL have belly pads that must be soldered to copper on a printed circuit board. You are not going to do that with your soldering iron.
200A is not suitable for any PCB! You need substantial Cu or Al busbars and high power
DC rated fuses for the high current path.
A motor that takes 200A in normal operation might be pulling a lot more when stalled or at start up - you
first need to measure that with a peak-reading clamp meter to get a feel for the size of the problem.
Hopefully being series-wound the stall current is managable and the 200A fuse is a useful indicator
of the peak current spikes.
200A continuous duty needs some pretty exotic devices with screw terminals bolted to a big heatsink
block, trying to parallel 20 or so ordinary MOSFETs is not a great idea, they need to be
very well matched and once some start to fail the whole lot can pop. Parallel a couple of 200A
rated iso-top (SOT-227) style devices is more reasonable approach. IXYS make some nice parts in
this range that are affordable
High power H-bridge design is fraught issues that have to guarded against - is super easy to blow MOSFETs
when many kilowatts of power are available and your protection circuitry doesn't cover all the bases.
To successfully make your own very high power H-bridge is going to take a lot of resources and care -
testing at low power, checking all the waveforms on a 'scope, testing protection circuitry, various sizes
of dummy load will be needed before going anywhere near the motor itself.
NB Never power a series-would motor without mechanical load - they have no upper speed limit
and can explode or sieze up.
Thank you all for the advice. On that note, I have opted to select a motor control unit such as the Curtis 1204. In any case, I see there is no Schottky diode or capacitor across the motor terminals, just a relay that closes to complete the circuit. I took the car for a test drive and somehow the battery terminal exploded on one of the cells. My question is: Does this indicate resistive heating, or a sudden inrush of current? I had the car coasting in gear (large field built up in the coils of the series-wound motor), and by closing the contactor I may have caused some kind of strange, high-current transient event. Would this sound plausible as a cause/contributing factor? How might I remedy this siuation and ensure it does not happen in the future?
NB: When re-assembling the pack, I made sure to polish all contacts.
Extremely good contacts must be made - at 200A every milliohm of contact resistance means 40W dissipated in
the contact resistance, leading to heating/oxidation. Its very important to prevent oxidation or corrosion
setting in so lead-acid battery contacts should always be thoroughly cleaned to remove the oxide layer,
clamp securely with grease/vaseline to exclude oxygen and moisture. If sanding the lead contact use wet
sandpaper with paraffin so no lead dust can float up and get into your lungs.
The initial failure would be because the contact resistance had risen to a few tens of milliohms, at which point
there's enough heat to start melting stuff.
Wow, thank you very much! I was just using a Scotch-Brite pad with nothing on it, so probably inhaled some Pb (and Cu). Before I run the car again I'l disassemble, re-sand and re-assemble. I have some NOALOX Anti-oxidant joint compound, would it help to use that on the connections, or would something else be recommended? This time I'll check the resistivity carefully before giving it another go.
Here's my understanding of the circuit, ignore the L values as I don't have an AC source to determine those, but W1 is an automotive solenoid running the accessory circuit actuated by a microswitch (not shown). L1, L2 are the field and armature windings, with R1 being that of both. R2 and L3 represent the resistance and inductance the coil for the automotive solenoid W1.
Hi,
Are you using the Curtis Controller?
If so then this diagram of yours.
Is nothing like how Curtis would like you to connect their controller.
Where are the fuses for the batteries?
Are you aware of why you need a pre-charge resistor and what MAIN is?
I have worked with these controllers and they are not to be fooled around with, you need to know what you are doing.
Tom...
PS Can you tell us what any electrical experience you or your have?
No, the schematic in the photo is the most accurate description of the control system (in one of its four possible modes of operation). Down the road I will likely opt to select a Curtis controller to replace the relays. I can take a few photos of what the set-up looks like as it currently stands. Four solenoid relays running off of a separate 12V battery switch between 3 speeds; attached is another schematic which simplifies the layout somewhat. There is (now) a 200A fuse installed in the main motor circuit, when I started work on the car it had been bypassed and a bit of paper-clip was being used as a washer between the bolted-together sections of cable. Note the lack of flyback diode on prints.
I have bypassed the microswitch to activate the first speed (to bypass the series resistor, which is HUGE) and the third speed (shifting the battery pack from parallel to series), leaving the second speed the only speed. It's a binary control system. I may remove the series resistor bypass solenoid, in order to use the 12V signal to actuate the go solenoid, as the present connection is faulty.
I am not familiar with the pre-charge resistor. It appears to be a resistor to moderate the current inflow into the controller, as the resistance value increases with the voltage? MAIN looks like it's the main contactor. I just Googled the former, it's there to keep the caps in the MC topped off. In that case, it seems as if you're just selecting a resistance value to supply half an amp to the unit.
I have some knowledge of electronics, having taken a circuits course six or seven years ago, and if needs be can get into the complex math relevant to passive elements. I've not had much opportunity to use this in a practical context until now, the practicality of the present venture notwithstanding. At this point I have to prove that the current set-up works before I push to purchase a Curtis to mount in its place. I am now leaning much more towards this option, rather than building a controller from MOSFETS - but hey, I'd like to use an Arduino in here somewhere! Perhaps to run other bits down the road. I'd like to know if there's a way to make a six-channel SoC meter or passive battery balancer. I see that there are six analog inputs so I could at least display the voltage on each cell, perhaps use the outputs to control transistors to relay voltage from one battery to another if the potential difference exceeds a certain value.
