Failsafes

Please post all questions and answers in here. This way people can easily see if someone else has the same problem.

Moderators: BeligerAnt, petec, administrator

User avatar
BeligerAnt
Posts: 1872
Joined: Wed May 15, 2002 12:00 am
Location: Brighton
Contact:

Post by BeligerAnt »

Simon, there's no point in checking failsafes for known robots that haven't changed in design since the last time we checked them! This is the reason most robots weren't checked at the last AWS.

A failsafe is only part of the solution. As Pete says, it probably won't help when there are multiple other transmitters close by. Also it won't help if the motor/speed controller takes on a life of its own - like Chimera at RRC.
Gary, Team BeligerAnt

Raizor
Posts: 101
Joined: Fri Apr 12, 2002 12:00 am
Location: Ireland
Contact:

Post by Raizor »

What about having a remote kill on the whole robot. I had a look into it and the circuit that Ian Watts has in his heavyweight could be built using surface mount components entirely. The only ones I could not find in surface mount were the transistors, but perhaps one of you qualified electronic engineers would know whether or not these are available. As a First year student we are only using bread boards and vero boards atm. Maybe its not practical but its something to consider.
The Irish Guy
Website Hosting:
http://www.dghosting.co.uk

paul
Posts: 108
Joined: Wed May 12, 2004 12:00 am
Location: chessington

Post by paul »

ok, i got an idea.

failsafes arent really needed on ants that dont poses a hazard to humans, such as flippers, pushers etc. the only types of ant that really need some sort of failsafes are spinners but, as with chimarara (sorry i cant spell) all you need is a large screwdriver to jam it up with while you disconect the power.

User avatar
peterwaller
Posts: 3193
Joined: Fri Feb 15, 2002 12:00 am
Location: Aylesbury Bucks
Contact:

Post by peterwaller »

A remote kill cant be operated if the robot is not recieving the signals correctly and if it is you don't need it.

administrator
Posts: 1
Joined: Wed Jan 23, 2002 12:00 am
Contact:

Post by administrator »

Solution is much simpler with ants - just take a biscuit tin to the event!! If it goes mad, either leave it in the arena until the battery gives up or drop the tin over it :D

I'm actually more concerned about (a) the number of times I see people sticking their hands inside the enclosure with robots running (b) the number of times people run up disks without the lid on the arena (c) the fact that blades are supposed to be covered at all times, but still we see people carrying ants around with no covers on blades - even through groups of other people. :roll:

Raizor
Posts: 101
Joined: Fri Apr 12, 2002 12:00 am
Location: Ireland
Contact:

Post by Raizor »

Pete if you have a look at ian's design on his website www.teambigbro.co.uk you'll see that not only is it a remote kill (which could come in very handy with spinners by removing power to the spinner motor) its also a failsafe in it's self.
The Irish Guy
Website Hosting:
http://www.dghosting.co.uk

User avatar
peterwaller
Posts: 3193
Joined: Fri Feb 15, 2002 12:00 am
Location: Aylesbury Bucks
Contact:

Post by peterwaller »

Your missing the point by the time your spinner gets out of control it is usually due to loss of control signals which means you can't initiate the complete power down. It is rare just to loose one channel. Also if you get interference that initiates the power down you have just been knocked out of the competition. I think I will pass on that one.
I just had a look on the site and that circuit board is at least 3 inches counting the number of holes in the vero board and with all due respects to to a very good robot builder that circuitry is a little out of date these days.

Raizor
Posts: 101
Joined: Fri Apr 12, 2002 12:00 am
Location: Ireland
Contact:

Post by Raizor »

Pete, on loss of signal it powers down as it goes into Failsafe mode, it is infact a better failsafe than the GWS FS-1 and the ripmax x-tra failsafe.

As for size as I stated earlier all the components with the exception of the transistors are available in surface mount so the board size could be greatly reduced.
The Irish Guy
Website Hosting:
http://www.dghosting.co.uk

User avatar
peterwaller
Posts: 3193
Joined: Fri Feb 15, 2002 12:00 am
Location: Aylesbury Bucks
Contact:

Post by peterwaller »

Loss of signal is not usually a problem most receivers fail safe with no signal the problem is adjacent frequencies. All receivers have AGC (automatic gain control) and this circuit is there to try and maintain the signel level. If your transmitter signel fails the AGC increases the gain until the next largest signal is usable. This is invariably the closest transmitter and the pulses coming from that are within the 1 to 2 ms pulse width and therfore happily pass through the failsafe. I tried my new controllers out in MilitANT at the weekend and they worked well initially except for my autocalibrate function. The software is designed to measure the nuetral postion of the controls on power up to ensure it would work with any Tx. The problem appears to be that due to interference or power supply variations the controller experienced an unexpected reset during the battle. This then recalibrated the zero position so when I released the stick instead of stopping it drove rapidly out of the arena. Thus yet again proving the keep it simple stupid theory.

User avatar
petec
Posts: 555
Joined: Wed Jan 23, 2002 12:00 am
Location: Burgess Hill, West Sussex
Contact:

Post by petec »

Oh, the excuses!! :roll: Actually Pete, Gary and I noted that your flea (which is a design based on the original Militant if I'm correct) seemed to perform better than all your heavily tinkered robots. Perhaps you need to stop designing things :D
------------
RobotWars101.org
RobotWars101.co.uk
Antweights.com
Antweights.co.uk
AntweightWorldSeries.com

Post Reply