#### Equip your Genius

Welcome to the Parallax Discussion Forums, sign-up to participate.

# Cable length limitation for motor power?

Posts: 4,031
I just came up with an RS485 line driver and receiver for 2 cat5 cables to handle hall sensors and encoders which gives very long cable options when needed. Typically the cables I use for 24v brushless DC is 3 conductor 18 stranded tinned copper. I have run up to 50’ before with the standard system but am trying to see if there are any known limitations for cable lengths vs just hooking up a 150’ length and testing. Would 12/14/16 be less loss to make a different at 150’?

## Comments

• Posts: 3,994
You're talking Motor limitations, right? Because RS485 is good for 4000 feet at 115k, eg see
https://www.analog.com/media/en/technical-documentation/product-selector-card/rs485fe.pdf

For the motors, its going to depend on the motor. Do the motors ever stall?
• Posts: 1,981
For power you just need to know the resistivity of copper, the cross-sectional area of the wires, their
length and the max working current - from that you calculate the max voltage loss - losing a volt or two from 24V
isn't a show stopper, losing 10V would be.

R = r L / A, where r = resistivity of copper 1.68e-8, L = length of wire in metres (counting both directions),
A is the cross-sectional area of copper in square meters.

Its more convenient to use 1.68e-2 and cross section in sq mm.

Put another way 1m of 1sq mm wire has a resistance of about 17 milliohms.

The other check you can do is that the stall current isn't going to set fire to the wiring (there are many
online tables of both wire resistances per unit length and max safe current either enclosed or not).

Another approach to figuring these things is that for the same load, doubling the length of wire requires
its area to be doubled too to maintain the same voltage loss.

You'll soon discover if you do a few calculations that long cable runs work much better with higher voltage
and lower current.
• Posts: 4,031
I'm talking about motor power, not RS485. 24VDC, BLDC. I'm trying to understand how the length affects performance.
• Posts: 3,187
Well, you might need more then 24 V on one end to get 24 out on the other end depending on current. To thin wire might heat up and create danger, thicker wire is better.

I think the length of the wire does not affect the performance of a motor at all, it just has some resistance so you have voltage loss.

Enjoy!

Mike
• Posts: 9,341
Always run a higher voltage than you need and then switch mode it down or in the case of a motor the PWM can take care of that for many motors, especially steppers. If your motor is 24VDC then try to run 48V if you can since the motor start/stall current can be very high and cause the end voltage to drop close to zero at times. You can also feed the control electronics from an isolation diode and "beefy" capacitor arrangement so that momentary voltage drops will not immediately affect the supply for the control logic. The other thing to consider is contact resistance so you are better off with screw terminals or plug-in screw terminals than you would be with RJ45 connectors which certainly aren't normally designed for this (which it looks like you are using).
• Posts: 4,031
Thanks guys. The motor power is 3 phases ABC from the brushless driver board and connected via 18guage with Molex MiniFit JR, not RJ. The BDLC driver must be located remotely from the motor, approximately 150' so I am looking into ways to minimize the losses. I can explore a higher voltage on the motor driver which is doable.
• Posts: 3,994
edited 2019-01-11 - 03:21:10
What voltage does the BDLC driver operate from? Is it the same 24v DC or higher?
• Posts: 4,031
The driver can accept up to 30V. The motor is rated at 24V.
• Posts: 825
You will always have a voltage drop when using D.C over long cable lengths. I would use a heavier gauge wire over 150 ft. Run the system under load with a 20 ft cable length and note the current. Then hook up your 150 ft roll of cable. Run the system again and note the current.
You should be able to calculate the wire gauge by the current differential.
• Posts: 1,665
You could take the max expected current for a single phase and calculate the voltage drop for a size and length of wire, subtract that from the source and see if it is still high enough to meet the motor requirements. Also consider a derating factor for heating in the wire in a given ambient temperature range if heating would significant.
• Posts: 1,981
If the driver is rated at 30V, 24V is the highest you should ever consider to give it - given the voltage
spikes and transients encountered with motors, its already close to the limit. Make sure you've got
good decoupling on the driver.

50m is a very long run for high current - so what is the current rating?
• Posts: 4,031
edited 2019-01-11 - 14:18:47
I plan to use a 10amp supply and expect no more than 5 amps max under load. There are 6 IRF540 n channel mosfets. The voltage limit 30v is the BLDC controller IC and mosfet driver, the mosfets are not stressed here.
• Posts: 19,738
edited 2019-01-11 - 15:34:43
Barely related, I just saw these guys at CES Wednesday. Wireless power transfer: aerial charging of a drone. Talk about in-flight entertainment! And possibly cancer if you spend much time inside the huge transmitting antenna power vortex of death.

http://getcorp.com/technology-overview/#tab-542

• Posts: 8,565
T Chap wrote: »
I plan to use a 10amp supply and expect no more than 5 amps max under load. There are 6 IRF540 n channel mosfets. The voltage limit 30v is the BLDC controller IC and mosfet driver, the mosfets are not stressed here.

If your power supply can provide 10 amps then it is better to have the wire rated for 10 amps. As a minimum I would suggest at least 16gauge, and if that were my project I would use 14guage since it is available for use in heavy duty extension cords.
• Posts: 825
Usually D.C. brushless systems have a cable length of 25Ft. of shielded cable. One good source is Anaheim automation.
• Posts: 4,031
edited 2019-01-11 - 20:12:02
I wonder if there is any benefit of using shielded? Maybe less interference on other devices? Here is my metal fab for the latest gadget.
• Posts: 1,681
Awhile ago, when you stated that you weren't using line-drivers/receivers on the encoder signals, I was wondering about your cable runs and the fact that you are PWM-ing the motors.

I realise that this is out of the question now but I would definitely have gone with a distributed approach. Were it not for the fact that I often have to utilise existing servo-motors, I would be using motors with integrated drives.

Another interesting approach would be to use the ODRIVE with a Prop, located near the motor. These are available in 24v or 48v versions.
• Posts: 4,031
Well in most cases I may have very little access to the system once installed, sometimes not at all without a lot of expense to get to the motor. So the electronics are usually the issue to update or repair and I have to have the bare minimum electronics at the motor. I will never wear out a motor or motor assembly, but electronics do need updating or repair on some rare occasions.
• Posts: 1,681
Forgot to mention that there are lots of Android apps, probably the same for iOS, for cable sizing.
• Posts: 4,031
edited 2019-01-11 - 21:19:23
Good call on finding an app! Big difference between 18 and 12.
• Posts: 1,681
erco wrote: »
Barely related, I just saw these guys at CES Wednesday. Wireless power transfer: aerial charging of a drone. Talk about in-flight entertainment! And possibly cancer if you spend much time inside the huge transmitting antenna power vortex of death.

http://getcorp.com/technology-overview/#tab-542

Very, very cool.
I was working at Polaris in WI, a while back and I would go in to the factory at all kinds of weird hours when production was quiet. I would have to wait at the gate, snow up to my eyeballs FOREVER because I always seemed to arrive when the security guards were doing their rounds. I felt bad for those guys and had the idea...why not have a few constantly patrolling drones with cameras. This charging system would be perfect.

I bet Trump would be better off with this system instead of a wall/fence that will only be tunnelled-under/climbed-over/cut-through anyway
• Posts: 5,959
T Chap
Here is my metal fab for the latest gadget.

Very nice work.
• Posts: 1,681
T Chap wrote: »
Good call on finding an app! Big difference between 18 and 12.

Heck yeah.

As I think I mentioned in another thread; look at any machine-tool, open collector from a proximity sensor, driving an opto-coupler....16 gauge conductors (occasionally 18 gauge)...milliamps! And there can be dozens if not hundreds of these signal wires. Drives me nuts having to work with this stuff. This is why I want distributed wireless nodes.
• Posts: 1,681
idbruce wrote: »
T Chap
Here is my metal fab for the latest gadget.

Very nice work.

+1
👍
• Posts: 1,681
@erco

"GET has developed the world’s first hardware that can wirelessly charges heavy drones over long distances (dozens of meters). Our technology provides best efficiency by distance and can deliver kilowatts of power wirelessly."

Wait....kilowatts? I can build a totally wireless machine?
• Posts: 1,665
For wiring questions (at least in the US), you can look in a copy of the NFPA's National Elictric Code, NEC xyearx for just about anything you would want to know for any wiring job. Section 3, Wiring Methods and Materials is good stuff. You will likely have to purchase it. Think the last I got it was when I ended up with way too many different answers to the same question when asking google etc..... Best \$100.00USD I had spent in a while and it allowed me to complete a hot tub installation with an easy session with the city inspector to get the permits done and have no reinspects through the projects. NEC can be had from NFPA in paper and/or searchable PDF.
• Posts: 4,031
Thanks for the input fellas
• Posts: 1,981
edited 2019-01-12 - 17:42:22
A quick bit of analysis shows that for a load power P, a supply voltage V, a cable length l, conductor total mass m,
percent loss is proportional to P * l^2 / (m * V^2) [ to a first approximation for modest losses ]

So doubling the length requires 4 times the copper, doubling the supply requires 1/4 the copper,
doubling the power doubles the loss percentage (losses go up by a factor of 4), etc etc.

That l-squared term means long cable runs are not usually practical unless the voltage is increased, a not
unsurprising result.
• Posts: 19,738
Mickster wrote: »
I can build a totally wireless machine?

Anything can be wireless.

Dwight:
As of this morning, we are completely wireless here on Schrute Farms. As soon as I find out where Mose hid all the wires, we'll have that power back on.

https://getyarn.io/yarn-clip/1f3157db-5100-41a2-845d-345fe8560727

Sign In or Register to comment.