Paralleling 3 motors on a single H Bridge ?

Thread Starter

RIKRIK

Joined Oct 11, 2019
93
Currently i have been wanting to make iver a delivery robot or a luner rover. So iv keep that thought in the back of my head and instead am focusing on a cheap robotics platform.

My main idea is a rocker boogie design. Made from pvc pipe and 3d printed parts. I will need 3 motors per side. I have ordered a 150rpm motor to do tests with. I decieded on that speed as it would give me with my wheel size a travelling distance of 1km in 14minutes. So 4km a hour. Obviously this may change to a slower rpm motor. Depending on my tests. As stated below: Under load 0.8kg per motor 500ma. Stall 1.3a.

I am wondering as too make things simple if i could parallel all 3 motors together. For tank style steering.

The h bridge design i am using isnt fantastic but i know it works. And it has cost me £5 in parts to make 2 drivers.excluding heatsinks which i already have. 5A NPN TIP120, 5A PNP TIP125, 2N2222A NPN

This may also be a good thing to comment on, the use of extremely inefficient obsolete BJTs.

motor2.pngmotor1.png
 

shortbus

Joined Sep 30, 2009
7,847
I am wondering as too make things simple if i could parallel all 3 motors together. For tank style steering.
Sure, why not? If the H-bridge can handle the current. The only problem would be actual speed control of individual motors connected to the bridge. Small motors aren't known to be very RPM consistent one to another.
 

Irving

Joined Jan 30, 2016
947
The stall current isn't necessarily worst case, its a nominal figure. I have motors that take 20% more than rated depending on motor temperature. Also if motor is being forcibly driven backwards the current can be even higher... Similarly if you reverse a motor at speed it instantly goes electrically into stall and more until it can start spinning in the reverse direction.. in other words your bridge should handle 2 x Istall ideally.
 

Irving

Joined Jan 30, 2016
947
I think the big problem with running motors in parallel is you have no control of the current sharing between them, this will make accurate speed control difficult and starting could be compromised as well
 

Thread Starter

RIKRIK

Joined Oct 11, 2019
93
The stall current isn't necessarily worst case, its a nominal figure. I have motors that take 20% more than rated depending on motor temperature. Also if motor is being forcibly driven backwards the current can be even higher... Similarly if you reverse a motor at speed it instantly goes electrically into stall and more until it can start spinning in the reverse direction.. in other words your bridge should handle 2 x Istall ideally.
Well i did think of the stall current. But not times 2. The tips can handle 5a. So may have to upgrade the bjts. My approach to the stall was brute force. So supplying all the motors anything they needed. Tbh i wasnt two concerned with speed control . As for stearing one set of wheels turning clockwise the other anti clockwise.

My other idea was using induvidual motor drivers made from smaller even cheaper bd677 / bd676, 4A RATED. Its not really that much more work to make 6. They would look alittle bit like a ram stick. One in each leg.

I am wondering if with the current circuit if basic pwm could be done. Nothing super high speed. Also if i did make 6 seperate motor drivers. Would there be any problems i may run into.

Kind regards,
Rick
 

Irving

Joined Jan 30, 2016
947
Any h-bridge can do PWM. Common practice is to keep the bottom mosfet/bjt enabled and switch the top one, but other strategies exist. You will need to revise your top/bottom linkage. You need 4 gpio per bridge. An Arduino might struggle physically with only 19 ports depending on what else is needed. You could use a dedicated i2c addressable PWM board, or 3 Nano one per axle. I'd probably go the latter route.

How have you sized your bridge transistors? Done the thermal calculations yet?

Skid steering is incredibly hard on batteries (and carpets) and very hard to do with multiple wheels each side especially if there's any length to the machine. I doubt you'll have enough torque with those motors. If you have 3 jointed segments you might want to think about wheelchair steering each segment. The turning circle will be large even so. Or go really adventurous and rotate each wheel on a vertical axis at the end of each leg with a servo; could make for some interesting obstacle avoidance strategies.

Have you thought about/sized your batteries yet?
 

Thread Starter

RIKRIK

Joined Oct 11, 2019
93
Thermal management wise. A darlington can get up to 150c on full wack. Im still debating on using heatsinks plus fan for the big one. Or putting them on a piece of copper flat bar, with a insulator inbetween if im going to do the induvidual ones.

I was thinking maybe using a uln2003 as my brains. And incorporating a additional pnp transistor with pull up resistor on the input of the two 2n2222a. So when the uln2003 is activated. It will pull the pnp low.

And because the uln2003 has 7 switching transistors. I can comman 3 inputs for left and 3 for right.

Its not a full rocker boogie. I took my inspritation from this
. Skip to 5mins in to see it working.

Kind regard, rick.
 

Irving

Joined Jan 30, 2016
947
Yes it can get to 150degC but it won't survive long there - it's an absolute maximum not an acceptable operating condition! Unlike MOSFETs BJT suffer from thermal runaway.

Forgive me for asking, but do you know how to size the heatsink or is this just guesswork?
 

shortbus

Joined Sep 30, 2009
7,847
Its not a full rocker boogie
when I first saw this I had to look on Google to see what you were talking about. I thought maybe it was for some kind of "effects pedal" or something from the 1980s. Your spelling BOGIE wrong.

"is that bogie is (rail|british|australia|new zealand|canada) structure with axles and wheels under a railway carriage or locomotive, called railroad truck in us english also used under semitrailers, and lorries with more than one rear axle while boogie is (informal) a piece of solid or semi-solid mucus in or removed ... " from -
https://wikidiff.com/boogie/bogie

 

Thread Starter

RIKRIK

Joined Oct 11, 2019
93
Yes it can get to 150degC but it won't survive long there - it's an absolute maximum not an acceptable operating condition! Unlike MOSFETs BJT suffer from thermal runaway.

Forgive me for asking, but do you know how to size the heatsink or is this just guesswork?
Thats what i ment. 150c. What i will probably do is just make the two motor drivers. Heat dissipation wise. I have a 20mm by 3mm bit of copper flat bar. About 60% more thermally conductive than aluminium.

Obviously i think i may need more heat dissipation. But it will be 4 mica isolated to220s , heat siliconded and all bolted down on to the copper bar. If it gets to hot. I will try doubling up on the 3mm copper flat bar.
 

Irving

Joined Jan 30, 2016
947
OK, but what I meant was do you know how to calculate the junction temperatures?

Yes, copper is a good conductor of heat, but how good a heatsink is a flat bar?
 

MisterBill2

Joined Jan 23, 2018
6,727
I think the big problem with running motors in parallel is you have no control of the current sharing between them, this will make accurate speed control difficult and starting could be compromised as well
You can raise the voltage and put al three motors in series, which will assure that the current is the same, and that they share the load evenly. The second advantage will be longer battery charge life.
 

Irving

Joined Jan 30, 2016
947
You can raise the voltage and put al three motors in series, which will assure that the current is the same, and that they share the load evenly. The second advantage will be longer battery charge life.
Putting them in series also has its problems. The current flowing is the same for each motor and therefore cannot be higher than the least mechanically loaded motor. The result is you get very little torque.

Yes, battery life is improved, but that's because you're doing less work!
 

MisterBill2

Joined Jan 23, 2018
6,727
If the motors are driving on each side and the suspension assures that each has adequate contact then the load should be similar for each. Looking at the picture it seems that the load distribution would be similar between the three motors, at least unless one of them fell into a hole. Just where would this rig be traveling, anyway? If a bit of active suspension was added then a lot could be done. OR, add one more set of wheels and a bit of lateral flex and the skid steering can be a lot more effective, and you can put 4 motors in series on each side. Handling more voltage is simpler than handling more current.
 

Irving

Joined Jan 30, 2016
947
You won't skid steer at all with motors in series... You'll be lucky to skid steer at all even if individually powered...

That video admittedly shows it working. But that's a very lightweight structure with hard plastic tyres on hard surfaces. On the shine floor it turns, on rougher concrete it's starting to struggle. Give it a softer surface like sand or gravel that bulks up against the sideways sliding wheel and it'll get stuck. You can't change the laws of physics! Been there.

And thats with individually driven wheels, not compromised by being all on one controller.
 
Last edited:

MisterBill2

Joined Jan 23, 2018
6,727
You won't skid steer at all with motors in series... You'll be lucky to skid steer at all even if individually powered...
With an adequate amount of lateral flex capability in the structure it will bend as the drives speeds vary, with neither side skidding. But certainly the flex capability will need to be carefully included. Probably it will not look much like that picture.
 
Top