Hi there, I am making a quadruped robot with 16 MG946R servos, 4 SG92R servos and 4 MG92B (continuous rotation) servos. Their voltage specs mean that I can provide the same voltage to them all at about 6V I think. I am controlling the servos using an Arduino Nano ESP32 and I plan to power that separately through the VIN pin using a few AA batteries . I have looked primarily at LiPo batteries, and I want to make sure as much as I can that whatever I buy will work as my time and cost I can spend on this project are both limited. For example, would the Turnigy nano-tech 6000mah 2S 25~50C Lipo Pack w/XT-90 work (https://hobbyking.com/en_us/turnigy-battery-nano-tech-6000mah-2s-25-50c-lipo-pack-xt-90.html), if I used a 6V stepdown voltage regulator as well. The total max current draw from the servos will be around 20A for the MG946Rs + 3A for the SG92Rs + 4A for the MG92Bs = 27A. Or would the Turnigy Graphene Panther 3000mAh 3S 75C Battery Pack w/XT90 (https://hobbyking.com/en_us/turnigy-graphene-3000mah-3s-75c-lipo-pack-w-xt90.html) instead be a better choice? I've also seen people using PCA9685 16-channel servo drivers, would you suggest that? What are your thoughts? Please excuse my lack of electrical knowledge.
Sum up the current needed for the maximum number of motors/servoc/steppers active simultanious. The get power supplies that can handle it.
For servos and servo motors, concider the stall current.
How far have you gotten? You can test each step using power supplies from your mains supply. Adding batteries in the beginning means there is so much more complications that will affect your testing.
I have made one of the 4 legs and set it up and tested if each of the servos separately to see if they can each handle their respective loads, which they can, and I have made some other changes to the design of the leg, and I am about to get the other 3 legs 3D printed too. In terms of testing all the servos together just for one leg I haven't done that yet, although I have a Turnigy 2200mAh 20C 2S LiPo battery that I got a while go as well as a 6V 3.3A voltage regulator which I guess I could use to test the one leg, but testing using power from the mains supply sounds a lot easier, although I am unsure of how I would do that. Could I cut the adapter cable and use something like crocodile clips and jumper wires to attach into a breadboard, and put an appropriate voltage regulator in, or is that not right? Thanks a lot for the advice.
You will need heavy wires with soldered connections, and heavy duty power distribution PCBs to handle that much current.
Do NOT use breadboards for any power connections, as the tracks will burn. Breadboards are strictly for temporary experiments with low power logic circuitry and are inherently unreliable.
It is much easier to break the servo power supply into smaller units (say four 6V 6A power supplies), each supplying a small group of servos.
Be sure to connect all the grounds.
Ok thanks very much I'll be sure to follow that. For the separate power supplies, if I were powering it through the mains supply, do you mean I should power the robot from 4 separate power outlets? I apologise for my lack of knowledge.
One power outlet or one battery will work. The biggest problem is finding voltage regulators that can handle large currents.
A popular approach is to have several switching voltage regulators connected to one higher voltage power supply, each regulator powering a small subset of the load.
Pololu has the best selection of very efficient, well designed and well supported voltage regulator modules.
Ok I will do what you have mentioned, thank you so much this is really helpful.
For the power supply, should I go for a 6V AC-DC power supply, or a 12V one, as I will be using 6V voltage regulators either way? I plan to eventually make this battery powered as well using a LiPo, so similarly would I get 2S or 3S? I've seen in some places that the extra volts come out as heat, but then elsewhere I've seen that switching regulators don't heat up like that. What would you suggest?
This topic was automatically closed 180 days after the last reply. New replies are no longer allowed.