Is separate power supplies needed for robot?
HoangTran83
Posts: 52
My robot has a 24 volt power source (2 x 12volt SLA batteries in series), is it a good idea to have a separate power source for the microcontroller? or would it be okay to use a voltage regulator and tap the power off of the 24 volt power source and then connect the output to the microcontroller? your inputs are greatly appreciated.
Hoang
Hoang
Comments
I have used separate power supplies while testing just to make sure the motors and servos aren't causing any of the problems I may be experiencing.
However I've used two separate power supplies several times to work around noise problems. I had servos that were really noisy, and kept resetting my microcontroller. Attempts to filter the noise with capacitors didn't work, but running the microcontroller off a separate nine volt battery did the trick. You do need to connect the grounds so the two power supplies are at the same potential.
A. You get better performance out of motors at 24 volts because there is less wasted energy in the H-bridge circuit.
B. You have less problems of noise in the microcontroller and logic circuits because of the isolation provided by a second power supply.
Can you do it with all one supply.
A. Yes, many people have. You can either drop voltage down to provide for the microcontrollers and waste a lot in the conversion OR you can boost voltage up for the motors to make small low voltage batteries work better with a small motor (Pololu does this with their line follower robot).
B. One supply will generally waste more power than two supplies unless you build a setup from multiple cells and tap the low voltage applications from the somewhere in the middle of the cells. Or course, one supply is likely to have less weight that two supplies.. and less bulk.
Tapping the low voltage from the middle of the cells creates problems for charging the batteries unless each cell or group of cells is charged separately. The cells that are tapped to provide the low voltage will be discharged further than those that are only connected to the high voltage load. If the low voltage current is a small percentage of the total this may not be a problem, but if not it may shorten the life of the cells providing the low voltage.
Better to use a switching regulator to provide the low voltage. That way all the cells are discharged to the same level.
Be aware that switching regulators can cause radio interference. I've had several wireless projects (including GPS) not work well when using a switching regulator.
I strongly suspect you are idealizing. In the real world, the internal resistances of individual cells vary a great deal... even when they come from the same production run. SO balanced discharge may simply be mythological. And in many contexts, the proportion of discharge that the control circuits impose is a very small fraction of the whole.
A linear regulator might waste 30%, but a switching regulator still may waste 20% in regulation. Ideally, it would be best to eliminate regulation entirely and work within the operating range of the microcontroller - some with operated from 2.5 to 5 volts nicely.
Do the motors and H-bridge require regulated voltage? Maybe none at all.
Loopy, perhaps I am being a bit over cautious, but I did say "If the low voltage current is a small percentage of the total this may not be a problem" in that post, and you are assuming that the current drawn by the control circuits is a small fraction of the total so we are somewhat in agreement on that point. On the other hand if there are a lot of sensors and other circuits that need a regulated low voltage to operate properly this may not be the case. You really need to look at the total current (amp hours) drawn by the control circuits in relation to that of the motors.
Another problem to consider is the voltage drop across the cells powering the control circuits when a motor starts up. The high peak current will cause a drop in voltage that disrupts the control logic. There have been many posted cases of this happening when the logic has been powered from a robot's 7.2V battery pack via a 3 terminal regulator.
As for the switching regulator vs a linear regulator, it very much depends on the battery voltage. Here is a comparison for a common setup (7.2V battery, 5V logic) and HoangTran83's case (24V battery, 5V logic) where the logic circuits draw 300mA.
For the 7.2V battery the linear regulator would have an efficiency of ((0.3 x 5)/(0.3 x 7.2) = 69.4% and would have to dissipate ((7.2 – 5) x 0.3) = 0.66W. Not much worse than a switching regulator at 80%, and I might use a LDO linear in this situation.
For a 24V battery the linear regulator would have an efficiency of ((0.3 x 5)/(0.3 x 24) = 20.8% and would have to dissipate ((24 – 5) x 0.3) = 5.7W.
The switching regulator efficiency would be 80%, and would also be much less susceptible to voltage dips caused by motors starting up.