Adding external power source

I’m dissappointed to say the least to realize that the videos shown on the kickstarter were not the board that comes with it.

To say that it was a prototype is misleading as all the other iterations of hexy were shown before the ‘final’ version. It should’ve been made clear that the servotor board is still beta.

I’m switching to the ssc-32 and I wish all the luck to Arcbotics to get the kinks worked out of the servotor board.

Just some info that I’ve gathered tonight after pulling the servotor board out of the hexy and installing 18 of my own personal servos to the board and doing some testing.

  1. All the servos move properly with all four of my power sources that I have been trying. A 3s Lipo with a ubec set at 5v, a 5 pack of rechargable nimh batteries from Walmart, a 4 pack of regular alkaline batteries, and a 5v 2 amp wall plug.
  2. Shaking occurs still no matter which power source I’m using.
  3. Idle power draw with no movement, just servos activated and holding position is between .1 and .2 amps.
  4. Single to four servo movements are generally around .3 to .7 amps depending on the motions (I tested the stock typing, wave and pointing motions)
  5. Activating the more complex movements like walking forward/backward, rotating, belly flop, get up, dancing all never got higher than 1.9 amps peak, and generally was around 1.4 amps.
  6. With the Lipo I was using (a 1400 mah pack) the current draw appears to give a run time between 2 to 3 hours depending on the amount of activity. This pack is a nice small one and weighs around the same as the 5 nimh AA’s but smaller in size.

All of the above was done controlling it with the latest version of pomoco, and the latest 1.7 firmware.

Useful benchmark!

Did you check to see if the numbers were the same with the Tower Pro servos?

[quote=“LuisSala”]Useful benchmark!

Did you check to see if the numbers were the same with the Tower Pro servos?[/quote]
I would if they still worked, but they stopped working.

[quote=“ncbob”]Just some info that I’ve gathered tonight after pulling the servotor board out of the hexy and installing 18 of my own personal servos to the board and doing some testing.

… bench mart testing here…

All of the above was done controlling it with the latest version of pomoco, and the latest 1.7 firmware.[/quote]

Wow… huge thanks for this testing!
I got the powerex batteries Wed, then noticed I need 5, not 4 (doh!, I skim read more than I should), so got the other 4 pack this afternoon (Amazon prime for the win). They are charging now. Excited to see this thing do more than have a spastic colon tonight!
Will go digging at the “good army” later today for a wall wart too.

I also had issues with one of the body upright panels being a little too wide front-to-back (as it sits in the assembled robot). I also noticed that the hip pieces that sit opposite the “H” seemed to have been cut on an angle on most of the legs; this was most noticeable at the tip where the screw was inserted to secure the servo. Everything did eventually fit together with a bit of elbow grease but I would have expected better tolerances for laser cut parts.

As for power supply, I’ve had good luck with good ole Energizer NiMH rechargeables in the 4x clip.

[quote=“ncbob”]Just some info that I’ve gathered tonight after pulling the servotor board out of the hexy and installing 18 of my own personal servos to the board and doing some testing.

  1. Activating the more complex movements like walking forward/backward, rotating, belly flop, get up, dancing all never got higher than 1.9 amps peak, and generally was around 1.4 amps.
    [/quote]

I haven’t looked at the software yet on how the servo’s are started, but servo’s like that easily pull like 600 or 700 mA when stalled, which also means they will pull something like that if they start to move when there is “load” (=mounted in the robot).

so if you would for example make a move that starts 10 servo’s at once you will pull a (very short) spike of 7amp
You might not see this spike on a multimeter, most are rather slow.

this of course does not mean you need a 7amp power supply, but the power supply you use will need some beefy capacitors to handle this gracefully or the voltage will drop during a very short time. Cheap power supply’s often lack beefy capacitors .

If you use a 5V power supply for both the servo’s and the logic and on complex moves you have erratic behavior it might be an idea to simply add a beefy capacitor ( 10000uF 10v or so) to the servo power rail directly.
They cost almost nothing.

[quote=“guntherv”][quote=“ncbob”]Just some info that I’ve gathered tonight after pulling the servotor board out of the hexy and installing 18 of my own personal servos to the board and doing some testing.

  1. Activating the more complex movements like walking forward/backward, rotating, belly flop, get up, dancing all never got higher than 1.9 amps peak, and generally was around 1.4 amps.
    [/quote]

I haven’t looked at the software yet on how the servo’s are started, but servo’s like that easily pull like 600 or 700 mA when stalled, which also means they will pull something like that if they start to move when there is “load” (=mounted in the robot).

so if you would for example make a move that starts 10 servo’s at once you will pull a (very short) spike of 7amp
You might not see this spike on a multimeter, most are rather slow.

this of course does not mean you need a 7amp power supply, but the power supply you use will need some beefy capacitors to handle this gracefully or the voltage will drop during a very short time. Cheap power supply’s often lack beefy capacitors .

If you use a 5V power supply for both the servo’s and the logic and on complex moves you have erratic behavior it might be an idea to simply add a beefy capacitor ( 10000uF 10v or so) to the servo power rail directly.
They cost almost nothing.[/quote]
This was tested the same way as when the servos were jerking and most of them didn’t move. The firmware was the only difference.

While a stalled servo would pull that much, to have all of them stall at the same time is very unlikely.

I’m not using something as crude as a multimeter to monitor/record the current and spikes.

The supply I use with the Lipo pack is more than capable of handling the current, it. Has a constant rating of 5A and a surge of 7a for 15 seconds. It works quite well powering my blade sr helicopter. Also once I switched the new firmware and my own servos I was able to use just about any power and all the servos worked. When I went back to the older firmware it started displaying the non-functioning behavior even if I just tried the ‘wave’ motion. In all the tests I did the only time I saw a difference was with the old firmware and the Lipo pack. It would shake and vibrate and more of the servos would work but it still wasn’t functional. When I used any other power source they just wouldn’t work, and it would just twitch when powered up.

stall no, start to move yes :slight_smile:

[quote=“ncbob”]
I’m not using something as crude as a multimeter to monitor/record the current and spikes. [/quote]

not everyone better testgear than a multimeter , hence my remark :slight_smile:

[quote=“ncbob”]
The supply I use with the Lipo pack is more than capable of handling the current, it. Has a constant rating of 5A and a surge of 7a for 15 seconds. It works quite well powering my blade sr helicopter. Also once I switched the new firmware and my own servos I was able to use just about any power and all the servos worked. When I went back to the older firmware it started displaying the non-functioning behavior even if I just tried the ‘wave’ motion. In all the tests I did the only time I saw a difference was with the old firmware and the Lipo pack. It would shake and vibrate and more of the servos would work but it still wasn’t functional. When I used any other power source they just wouldn’t work, and it would just twitch when powered up.[/quote]

R/C bec things (i don’t know your model but most are the very alike) are typically not very capable of delivering high current bursts and keep the voltage stable. Note that this is a totally different property than the “constant rating of 5A and a surge of 7a for 15 second”.

Hence why I personally would try to take the power source out of the equation and use a stable power supply, at least for the logic (which accepts up to 10volt IIRC , so you can hook a 2cell lipo directly to it and use the bec for the servo supply). The schema ( github.com/ArcBotics/Servotor32 … otor32.pdf ) does not indicate what voltage regulator is used, might be worth checking the voltage output of that also.
Maybe the regulator on the board does not like noise/spikes put on the Vin rail by the provided servo’s? Or it’s faulty?

Of course the simple fact things improve with your own servo’s and the “jitter” are not explained by above, but it’s a start to make sure the ATMEGA has no intermittent voltage drops. Then there is at least one thing you can cross of your list :slight_smile:

The ubec I’m using has to be able to work on my helicopter which requires very precise and constant control, and if the ubec was not able to hold the voltage, then I would not be able to fly my helicopter very well.

This is why I tried SEVERAL powere sources, the board pretty much behaved the same way whether it was the Lipo, the alkaline, the nimh, the wall source. The wall power supply is the same one that I’ve used on my other Microcontrollers, it’s almost more a standard bench ps rather than a wall supply, and it’s a known good quality power source. But like I’ve said the only difference was with the 1.5 firmware and the Lipo, it almost was able to move all the servos. With the 1.7 firmware the was no difference in the behavior between the different power sources.

Having them all start at the same time really didn’t cause that much of a spike, so that’s not an issue.

The only time I’ve seen issues like this is with problems with software, or cheap hardware that can’t handle the demands placed on it.

I ordered a few current monitoring devices to more precisely power profile the kit. I’ll make sure to post the results online. I’ve got about a dozen different types of batteries I’ll be using in the tests. (I didn’t arrive at the powerex’s first attempt :wink:

I’ve also ordered a number of sample ubecs from manufacturers to find the best price/performance ratio for hexy. I’ve got it working with a more expensive ubec, but I’d prefer to offer something more affordable in the store.

The external power brick I’ve been using from mouser works pretty well, but again, trying to find something more affordable for the store.

Any news on the official testing of the external power supplies? Or any users have good bad experiences with the suggested options? I’ve got an event next month and would like to get stuff ordered.

Thanks, R

If you have a CPU powersupply you cant go wrong Rob, just pull out the 5v power lines and hexy will work like a charm.

That’s a great idea. A cheap 400w silent ATX-sized PSU costs 12 Pounds on Amazon, and would probably supply 18 or 20 amps at 5v. I don’t know about the ripple.

I recycled a laptop-style switching power supply from an old NAS device, it delivers 4.2 Amperes at 5v. Works very well, particularly for development and testing.

This works great without any need to modify… Just plug it in…

amazon.com/gp/product/B00KTQ … UTF8&psc=1

$11 & free shipping… Got mine in about 4 days… :stuck_out_tongue:

Ok, got my Hexy back out again to try one more time. I’ve had much better success with it. I am using this BEC now just to make sure it’s got enough power :slight_smile: hobbyking.com/hobbyking/stor … utput.html

I hooked up the servo and logic plugs, the servos powered up, but the board wouldn’t work unless I plugged in the usb cable. I went ahead and tested everything, and it worked great with this BEC, no twitching and all the preprogrammed moves work great. I went to troubleshoot why the board isn’t working and from what I can tell the jack for the logic power is bad. I’ve hooked up a second servotor board using the same BEC and plugs and it works fine from the BEC.

Even have it working remotely over BT and the ultrasonic sensor working.

Need to get a slightly smaller battery so it fits without sticking out the back, for the test I was using a 3s 5000 man 15c pack, going to switch to a 2000mah compact pack, and then I’m going to try the 5a bec again as it’s less expensive and smaller.

Very excited to get this working finally. I’ve also 3d printed one and given it to a friend to assemble, but she hasn’t got it all put together. I didn’t have enough normal 92g servos so I ordered some ones with the carbon gears so she has 1/2 her servos with the carbon gears.

Once we get that one together I’ll post some videos of it as well.

Here’s the video. youtu.be/7qxB9p5sH24