Here is a video illustrating why it’s a bad idea to use plastic battery holders for servos… Please read the description below.
youtube.com/watch?v=VKBH22Mgogs
This is a video to illustrate how cheap plastic battery holders and “AA” batteries do not have the ability to source even a small amount of inrush current. Test shows the SSC-32 powered by a single 4 x “AA” battery holder and 4 HS-645 servos. This is using Alkaline cells which provide about 6.2vdc to the VS terminal and the VS=VL jumper is installed. If rechargeable batteries were used the voltage would drop to around 4.8vdc and the problem will be much worse.
LynxTerm has been started and the SSC-32 has been attached. The Green LED is lit because the SSC-32 has not received any commands yet. The first test is to type VER in the window, then return. You will notice the Green LED goes out, then blinks with every keystroke it receives. This is normal operation. The SSC-32 replies back with SSC32-V2.03XE. So far so good!
Then I click All=1500 to turn on all four HS-645 servos. The first indication that something is wrong is the Green LED is on again. This means the SSC-32 has reset, because the batteries were not able to provide the required inrush current. Clicking again only puts weird characters on the terminal window, as the SSC-32 crashes with every attempt.
To prove the SSC-32 is still functional after resetting I hit return to clear anything from the buffer, then type VER and return. The SSC-32 replies with SSC32-V2.03XE again.
Lessons learned:
Cheap plastic battery holders use 30gage wire that can’t carry much current. The springs do a poor job of making good contact with the cells, so they also contribute to the problem.
Use a separate 9vdc battery for the VL terminal to isolate VS (servo voltage) from VL (logic voltage).
Use a 6.0vdc NiCd, NiMH, or other chemistry battery that can deliver the required current if you must power everything from one battery pack. These sort of batteries will use heavy 12-18 gage wire.