Servo problems.

All things antweight

Moderators: BeligerAnt, petec, administrator

Post Reply
Paulmchurd
Posts: 227
Joined: Sun Dec 10, 2017 4:39 pm

Servo problems.

Post by Paulmchurd »

Hello,

I am new to building robots. I've done a lot of back ground reading but seem to have got a little stuck.

I've started with a simple set up card board chassis glue gunned together. Then Using two continuous servos connected into a receiver and battery.

I bought a cheap 4 channel transmitter.

My issue I'm having is after driving the robot a around for 30 seconds or so. One of the servos will start moving on its own. I can trim it back to being still then after 30 seconds it will start to move on its own again. Eventually the trimming won't even stop it and the robot spins in circles .

This is a new problem. It didn't do this for the first week of use. I rebuilt the robot to try incorporate a flipper. 3rd servo added.

How can I fix this?

Is this a common problem with continuous servos?

Is it my cheap transmitter? Turnigy i4x

Thanks for any help
User avatar
Shakey
Posts: 1119
Joined: Sat Mar 13, 2010 8:38 pm
Location: Reading

Re: Servo problems.

Post by Shakey »

This is an inherent problem in nearly all ESC's but servos are very susceptible for it. Essentially the way the servo reads the signal depends on the temperature. As it changes temperature (heating up over use for example) it may see 'Stop' as something else other than when it was turned on. There's no real way to correct for this as it happens without just constantly trimming the robot or making sure the ESC/Servo stays at the same temperature as much as you can.

Expanded explanation:
The RC signal from the receiver is known as a PWM signal. In RC stuff this can mean (Numbers pulled out of air for illustrative purposes) that 0 power is turning the signal on for 0.1 second every second. Full power could be 0.5 seconds every second. The processor times this by watching for when the signal turns on and then waiting for it to turn off again and counts how many cycles have passed, the cycles being how many times it ticks a second (It's clock rate, what you normally see listed in Hz). In small things like this the processor is often left to make its own clock internally and this is affected by temperature. As it heats up it's version of a second that was programmed in by the programmer can become longer or shorter meaning it starts measuring a signal as something different than it did when first turned on. And so the stick being centred means different things depending on the temperature.
Nuts And Bots - For all your components and ready built antweights!

Alex Shakespeare - Team Shakey / Nuts And Bots / Team Nuts:
AWS 44, 45, 49, 51 & 55 Winner - Far too many robots!
Paulmchurd
Posts: 227
Joined: Sun Dec 10, 2017 4:39 pm

Re: Servo problems.

Post by Paulmchurd »

Ah thanks shakey. Ultimately very annoying this issue.

So having a thick layer of glue under the servo is not going to help things.

When people are battling their bots are they constantly trimming during the fight? Or do none servo ESC take longer to heat up and cause less problems?

Thanks again for the response
User avatar
peterwaller
Posts: 3213
Joined: Fri Feb 15, 2002 12:00 am
Location: Aylesbury Bucks
Contact:

Re: Servo problems.

Post by peterwaller »

I tend to build my own escapes and build in a dead band in the middle which tends to overcome this problem.
Standard controllers like Rory's seem far less susceptible to drift so he probably does the same.
User avatar
Shakey
Posts: 1119
Joined: Sat Mar 13, 2010 8:38 pm
Location: Reading

Re: Servo problems.

Post by Shakey »

Normal ESC's handle battle better so don't heat as much and (As Peter says) they normally have a dead band meaning that rather than (say the signal was 0 REVERSE - 10 FORWARD) 5 being STOP they say anything between 6 and 7 is STOP. So the trims have to wander far more before the ESC stops seeing it as STOP.

So normal ESC ants only need trimming every so often, certainly not during a fight.
Nuts And Bots - For all your components and ready built antweights!

Alex Shakespeare - Team Shakey / Nuts And Bots / Team Nuts:
AWS 44, 45, 49, 51 & 55 Winner - Far too many robots!
Paulmchurd
Posts: 227
Joined: Sun Dec 10, 2017 4:39 pm

Re: Servo problems.

Post by Paulmchurd »

Potentially I may have damaged my persistent turning servo with over heating. Now it over heats very quickly. Thus causing it to need trimming constantly.

Are cheap servos a fast turn over?

Thanks
Paul
Paulmchurd
Posts: 227
Joined: Sun Dec 10, 2017 4:39 pm

Re: Servo problems.

Post by Paulmchurd »

https://www.robotshop.com/uk/9g-continu ... gJ7HvD_BwE

These servo have built in dead space. Have you used these before?
Ant Ipodean
Posts: 43
Joined: Mon Apr 24, 2017 3:31 pm
Location: Auckland, New Zealand

Re: Servo problems.

Post by Ant Ipodean »

Thanks Alex for the explanation about temperature effecting the off position for servos. Explains something that had me curious.

A question:
Are all the Robots I see on AWS events that "creep" when not driven (usually the pink ones :)), servo powered, and if so, are they micro or larger servos?

A suggestion:
I have bought micro servos (9g) from AliExpress that have 360 degree operation, though they really still need speed hacking.
They have a custom ESC and specifically incorporate a wider "dead zone" to ensure that stationary is really stationary.
Note that I have bought some and they test out fine, but have not used them in combat yet.
Search for: DM-S0090D servo. They are not expensive.

Mike.
Paulmchurd
Posts: 227
Joined: Sun Dec 10, 2017 4:39 pm

Re: Servo problems.

Post by Paulmchurd »

Super cheap as well. £1 a servo with postage. If I'm not wrong?

Description mentions about massive dead space.
Pity I will have to wait 20 days for it to arrive in the post.

Thanks for the help everyone.
Rapidrory
Posts: 1160
Joined: Sun Nov 24, 2013 9:54 pm

Re: Servo problems.

Post by Rapidrory »

Most ESCs in the higher weight classes and many smaller brushless ESCs use an external crystal oscillator to avoid this drift, but small brushed ESCs tend to save the space and cost of a crystal, and just use the internal oscillator built into the chip, which is very susceptible to temperature changes.
Rory Mangles - Team Nuts

Robots: Nuts 2 and many more...

NanoTwo Motor Controllers: https://nutsandbots.co.uk/product/nanotwodualesc
Post Reply