# need help with a buffer circuit (or different solution)

hey,
i am building a robot (hexapod) and decided to do it from scratch, mainly to learn new stuff instead of copy pasting designs and code...
maybe not the best idea

Arne511:
i am building a robot (hexapod) and decided to do it from scratch, mainly to learn new stuff instead of copy pasting designs and code...

That's good to hear but it gives no idea about what you want help with.

...R

What would a "buffer circuit" do in the robot?

when the servos move (4or more servos at a time) the power supplied is not enough. the time for each movement is only about 50ms. i would like to use the capacitors (that are charged in between movements) to supply the extra juice needed to be able to move 9 servos at a time. battery is still supplying power but needs a bit of help.

at least thats my idea. if you have another one or better one, i am more than willing to give it a shot.

ps:buffer, to compensate for the delta between demand and supply.

daaamn. sorry. now i see the problem... 3/4 of my first post is missing. sorry.

so here is the missing part:

setup:
2x 18650 (3500mAh) supply power to arduino nano and ps2 rx
2x 18650 power a 24ch servo controller (with 6v via an UBEC) and 18 rds 3218 via the controller.

servo specs:
5mA in idle
950 at stall torque.

i would like to move 9 ( perfectly all 18) simultaneously.
right now 4+ servos let the controller starve and no more motion happens.

so i basically want to use capacitors for all 18 servos in combination with resistors to supply the extra power.

i dont know how to figure out the correct size of the capacitors.

the bot weighs 2.5kg. based on the leg length each servo should be able to carry 1.4kg. so i am not hitting stall torque when 3 legs (9 servos) carry the whole weight.

Forget about using capacitors and 18650 batteries. You need a much heavier duty power supply.

Successful robot builders budget 1 Ampere per moving servo for light duty servos, more for heavy duty servos. A sensible choice would be a rechargeable battery pack capable of easily delivering 15 to 20 Amperes. RC hobby shops carry them.

so i am not hitting stall torque

Servos and brushed DC motors draw the full stall torque every time they start to move.

jremington:
capable of easily delivering 15 to 20 Amperes.

Can't an 18650 LiPo do that?

@Arne511, please make a simple pencil drawing showing how you have everything connected and post a photo of the drawing. Please DO NOT post a Fritzing diagram.

...R

18650 is just a case size. It tells you almost nothing about the battery. A typical Lithium-Ion 18650 battery is only rated at 2C so a 3000mAh battery can deliver no more than 6A.

But these days there are quite a few 18650 cells with Continuous Discharge Ratings (CDR) of 20A-30A (around 10C). Typically the high CDR cells have lower capacities, 2000-2600mAh e.g. Sony VTC5A. But if you buy cheap high capacity 18650s then expect 1C-2C only.

Steve

jremington:
Forget about using capacitors and 18650 batteries. You need a much heavier duty power supply.

Successful robot builders budget 1 Ampere per moving servo for light duty servos, more for heavy duty servos. A sensible choice would be a rechargeable battery pack capable of easily delivering 15 to 20 Amperes. RC hobby shops carry them.
Servos and brushed DC motors draw the full stall torque every time they start to move.

jremington:
Forget about using capacitors and 18650 batteries. You need a much heavier duty power supply.

Successful robot builders budget 1 Ampere per moving servo for light duty servos, more for heavy duty servos. A sensible choice would be a rechargeable battery pack capable of easily delivering 15 to 20 Amperes. RC hobby shops carry them.
Servos and brushed DC motors draw the full stall torque every time they start to move.

ok. thats more or less what i feared.. sh*t..
i didnt know that they draw that much when the servos start to move... i thought they only reach those high demands when i am either exceeding the force required to move or hit a stop. but that explains why its killing the controller...
bottom line is, that i need to change the whole setup by using a more sophisticated power supply for the servos. capacitors would have been an easier solution

btw: those are the batteries i use.
datasheet
as far as i understand it, it should work...

One big problem is that typical battery holders aren't rated for massive currents,
so cylindrical cells aren't suitable unless all welded together with copper strips.

You probably need a decent RC LiPo pack with heavy duty gold-plated bullet connectors
and decent thick leads to handle 20A pulses without dropping out. And of course suitable
fusing to protect the wiring harnesses.

Those look like decent cells with a specified CDR of 20A so you should be o.k. for quite a few servos. Trying 18 simultaneous may be pushing your luck. If you check the graph on page 3 you'll see how much the capacity is reduced at high currents.

Steve

slipstick:
Those look like decent cells with a specified CDR of 20A so you should be o.k. for quite a few servos. Trying 18 simultaneous may be pushing your luck. If you check the graph on page 3 you'll see how much the capacity is reduced at high currents.

Steve

All those high current graphs will have been done with a decent test-fixture - most standard battery holders
cannot handle such high currents at all, and simply cook themselves, melting the plastic.

There are not many "standard battery holders" available for 18650 cells and most that do exist are of rather better quality than the typical AA cell holders.

Steve

This topic was automatically closed 120 days after the last reply. New replies are no longer allowed.