I'm building a mobile robot which is quite heavy and needs some serious HP. For this and other reasons I want to mount 4 induction motors to each wheel (for a total output of about 8kW) and power them via batteries. However, those motors need 3-phase current and I obviously can't plug them directly to the batteries. This is why I need a variable frequency drive but without the prepended rectifier (input current is already DC). I have been looking all around but couldn't find any devices that seem to be able to do just this. Can anyone with a bit more oversight maybe lead me into the right direction?

 

What induction motors are you using, or intend on?
I very much doubt you will find a off the shelf VFD for a induction motor that does not take 1ph/3ph.
The other potential problem is synchronizing induction motors.
Most application like this (locomotives etc) have gone to AC synchronous with P.M. fields.
If you were to use off the shelf VFD's, all you need to is remove the 3 phase rectifier on the VFD input.

 

Most of the VFDs I've worked with have the dc bus broken out to terminal strips to daisy chain vfds with one input supply. So if you have a high enough dc bus voltage from you're battery pack you could tie into the dc bus at those terminals (for 480V you're looking 650ish VDC) and bypass the input rectification entirely without having to modify the drives. You'd still have to deal with the issues that @MaxHeadRoom mentioned. The synchronization issue is a biggie. Could use servomotors rather than induction motors and you could synchronize in software.

 

Thanks for your answer! I'm thinking of using a default synchron induction motor that will run off 3ph, so this part should be taken care of. Where do you see the problem with synchronizing the induction motors? I thought about taking a VFD for about 8kW because I then can hook all of the four motors up to the same VFD output. Is this going to be problematic? Why so?

As far as I can tell electromotives use a similar principle they also alter the frequency via VFD from the DC/AC supply. But I will have a deeper look at that!

 

Okay, well in that case synchronization wouldn't really be an issue, unless one of the motors is messed up. I think we were under the impression it would be 4 motors 4 vfds. So yeah I'd look at the vfd you want to buy and make sure you have the dc bus broken out to terminals. Then all you have to figure out is the battery setup to power it.

 

That sounds promising! Another issue is the automatic control of the frequency. Do VFD's usually have some kind of bus to hook them up to a microcrontroller? Because I want to do the regulation of the frequency in my own software

 

Most do in some way. You'd be looking more into the industrial protocols though. EthernetIP, ProfiBus, ProfiNet, Modbus, ModbusTCP, DeviceNet, CanBus, and a few others. Those are your digital controls that might be available. I know sparkfun has a canbus breakout board that could be used.
Alternatively I don't recall ever seeing one that doesn't have an analog control input that can be used. You define that kind of stuff in the parameters of the drive. It would really depend on the model you go with and how you want to control it.
<Edit> also be careful. There are a number of cheapo drives from Amazon and Ebay that have zero documentation. And ones you'll find that say they're rated for what you need but they're obviously not nearly big enough to be rated.

 

VFD's used in machine control, use Sensorless vector versions for speed control, if extra accuracy is required, they can include a encoder feedback option if ordered at time of purchase.
One VFD for several motors is not such a good idea.
Also I believe synchron motors are P.M. field, if so these do not run well on a off the shelf VFD's I have tested P.M. versions with less than stellar results.

 

Thanks @jiggermole for pointing that out. So connecting the DC shouldn't be a problem then when I pick the right device.

@MaxHeadRoom Why not connect several motors to one VFD? Yes synchronous motors are typically permanent magnet driven, but I wouldn't have guessed that this makes a difference to the VFD.

I'm a bit worried about the high DC voltage input required. I need at least 320V to PWM-reconstruct the AC curves which have 320V peak with a mean of ~230V. But as somebody pointed out to me, this could get quickly really dangerous.

Also why not construct the VFD myself? I could throw away the rectifier and directly PWM with my microcontroller of choice. This could save me cost, weight and parts. But it could also mean that I die testing lol x_X.

 

I did contact Hitachi on using a P.M. motor and they said they would have to customize the VFD it to suit.

 

This isn't in a facility or industrial environment is it? If it is than yeah multiple motors on one vfd is discouraged, though not illegal if installed per the manufacturers instructions. In our plant off the top of my head I know we have 3 machines that have one vfd driving 4 50hp motors. For the hobbyist at home it should be fine.

You could try and make a vfd yourself with igbts and the like, but with mains DC voltage its probably not the greatest of ideas, but not the worst. Just don't lick it.
The dc bus voltage is going to be a major challenge, if you want to do it with batteries. With what you want to do you'd need a very good amount of stored energy. A scary amount of it.

Could do it like mining equipment and use a long heavy duty extension cord. SJOOW is rated for extra hard usage and will take quite the beating. Might be easier to do it that way. Use a phase converter to go from 1ph to 3ph and feed the 3ph into the vfd. Than you only have to deal with mains AC and not a DC bus at mains peak potential. The assumption you only have 1ph available, if you have 3ph than nix the phase converter.

Another option would probably be to use smaller dc motors running off of 12v and gear them down a bunch. That might be a lot easier to implement than mains voltage induction motors directly driving wheels. And would eliminate the need to be up at mains voltage. Slower but a lot safer to do. Found a number of smaller dc motor drivers on amazon. I included some links that look like good places to start looking at required torque you'd need to move an object.
https://blog.orientalmotor.com/moto...lculate load torque, multiply,x 0.05 m = 1 Nm.
https://www.researchgate.net/post/H...ve-a-massive-object-by-means-of-gear-assembly

Or rather than re-inventing the wheel you could go with a premade solution like some of these.
https://www.rdergo.com/powered-cart-movers/
or similar. Less fun than playing with high voltage DC but a lot lower chance of crispifying yourself.

 

@MaxHeadRoom Why not connect several motors to one VFD?

The only time I did this was in conjunction with the manuf. tech dept (WEG) and this was a case where the motors were fitted to a double ended lathe, one motor for each chuck, they simultaneously held the same component so the load was identical for each motor.

 

I know Allen Bradley Powerflex drives have a section in the manual for connection of more than one motor to one VFD. These were asynchronous, velocity controlled induction motors (V/Hz ONLY!!, SVC, FOC etc are not possible). I think synchronous induction is also acceptable and gives better speed uniformity. I would question if you can do this with PM motors as commutation is a big deal on those and you'd never synchronize the commutation between motors.

In any case, general practices are:

• Sum FLA of all motors and add 20%. This is the drive output requirement
• All motors require separate overload and short-circuit protection AFTER the VFD. This is because the VFD cannot distinguish a condition where one motor is overloaded but the sum of all motors is below the trip limit of the drive.
• All motors will start and stop at the same time, and run at the same frequency - obviously.

Some drive systems are designed to be only run from a DC bus. So this is nothing new you're trying here. 325V with batteries could be hazardous, but not specifically more so than working on low voltage (<600V) AC drive equipment. Remember the drive has big capacitors inside which store dangerous amounts of energy, even when power is off. They typically discharge within 5 minutes due to bleeder resistors.

Another thought is some small servo drives. DMM technology makes a product called the DYN2 drive which runs on ~70VDC and drives a low voltage AC servo motor with encoder feedback. A DC fed drive, cables, and 400W motor (with high-res encoder) run about $320. Not cheap but these are the kinds of motors used in industrial robotics. DMM is about the most hobby friendly priced servo. Communication can be done with 5V/24V digital inputs OR through modbus/RS485 communication.

https://www.dmm-tech.com/

For a bit more money, Clearpath servos offer an integrated motor and drive, saving on weight and space.

https://www.teknic.com/products/clearpath-brushless-dc-servo-motors/

I'm just a hobby guy who has used these in the past with good success.

 

The vdc servo drives @mcardoso recommended would work great with a forklift battery. Will still have to gear it down a lot. The torque for those wasn't that high, but them being servos you can have them split and by changing velocity you can steer. Definitely a good option. Looks comparable, price wise, to the independent dc motors and dc speed controls option.

 

What is the power source for this 8kW of power?

 

Unless I am misunderstanding this thread you want to power your machine with batteries. At 2kW a motor(8kW total) why not just use DC motors? They will be much lighter than similar output AC induction motors. Even the newer permanent magnet car starter motors are ~3kW each, and they don't suffer over speeding like the older wound field type did. And control of the speed of the starter motor is way less complicated than using a1 or more VFD's.

 

Unless I am misunderstanding this thread you want to power your machine with batteries. At 2kW a motor(8kW total) why not just use DC motors? They will be much lighter than similar output AC induction motors. Even the newer permanent magnet car starter motors are ~3kW each, and they don't suffer over speeding like the older wound field type did. And control of the speed of the starter motor is way less complicated than using a1 or more VFD's.

I'm wondering why Tesla went through all the trouble to go from dc battery power to ac motors. I always thought it was because accurate, varying speed control is tricky with DC particularly when needing electrical braking. I'd like like to know, I've got a similar project.

 

I'm wondering why Tesla went through all the trouble to go from dc battery power to ac motors. I always thought it was because accurate, varying speed control is tricky with DC particularly when needing electrical braking. I'd like like to know, I've got a similar project.

Errr.... because ac is much easier to deliver in long distances? ... and also because it's cheaper and more efficient to generate?

 

Errr.... because ac is much easier to deliver in long distances? ...

Here DC is delivered over many Km's, it is done for cost benefits over long distances , converted at the Gen. stn. and re-converted at the destination.

 

Here DC is delivered over many Km's, it is done for cost benefits over long distances , converted at the Gen. stn. and re-converted at the destination.

Yeah, I know about that ... but rich was wondering about Tesla in particular