PDA

View Full Version : 2009 Scoring System Setup



lnxfergy
05-31-2009, 02:57 PM
See the 2010 Scoring System Info here.
(http://forums.trossenrobotics.com/showthread.php?p=37395#post37395)
-Fergs

mannyr7
05-31-2009, 04:53 PM
Gary, have any more pics of your target plate setup and construction?

lnxfergy
05-31-2009, 04:57 PM
Gary, have any more pics of your target plate setup and construction?

Take a look in his gallery: http://forums.trossenrobotics.com/gallery/browseimages.php?c=3&userid=2033

-Fergs

societyofrobots
07-12-2009, 09:03 AM
Out of curiosity, how much current in mA did it consume? (for my bot, anything above 100mA is non-negligible)

Any chance will you use an LDO regulator for next year? (to reduce required battery size to run it)

Keeping the same profile or going smaller? Using smd next year to reduce profile? Whats the height and hole dimensions?

Trying to budget power in my bot for next year . . . and I'm assuming there won't be any significant changes from the users point of view . . .

lnxfergy
07-12-2009, 07:04 PM
Out of curiosity, how much current in mA did it consume? (for my bot, anything above 100mA is non-negligible)

Any chance will you use an LDO regulator for next year? (to reduce required battery size to run it)

Keeping the same profile or going smaller? Using smd next year to reduce profile? Whats the height and hole dimensions?

Trying to budget power in my bot for next year . . . and I'm assuming there won't be any significant changes from the users point of view . . .

The XBEE is supposed to be about 50mA (but may use up to 300mA during bursts for output). The AVR should only be about 5-10mA, so total should be average under 100mA. The regulator is an LDO (LM2940)

Board size is 2.4" sq, holes in each corner 0.15" in from an edge. Only changes to future revs are a better silkscreen, since we had some issues with the last run. The max height is an upright TO-220 package. In theory you could replace the power regulator with an acceptable 5V source from elsewhere, but you have to verify that it won't dip out and reset the board (which will count as hit against you).

-Fergs

societyofrobots
07-12-2009, 09:03 PM
The XBEE is supposed to be about 50mA (but may use up to 300mA during bursts for output). The AVR should only be about 5-10mA, so total should be average under 100mA. The regulator is an LDO (LM2940)

Board size is 2.4" sq, holes in each corner 0.15" in from an edge. Only changes to future revs are a better silkscreen, since we had some issues with the last run. The max height is an upright TO-220 package. In theory you could replace the power regulator with an acceptable 5V source from elsewhere, but you have to verify that it won't dip out and reset the board (which will count as hit against you).

thanks! =)

Probably better/easier to leave as is. But if you ever find yourself redesigning it, at only 350mA peak, a small smd LDO regulator will be much smaller and can easily handle it.

Also, any chance I can get mine without that big blue wire connector?

lnxfergy
07-12-2009, 09:24 PM
thanks! =)

Probably better/easier to leave as is. But if you ever find yourself redesigning it, at only 350mA peak, a small smd LDO regulator will be much smaller and can easily handle it.

Also, any chance I can get mine without that big blue wire connector?

The board was not designed specifically for Mech Warfare, it was my general purpose Arduino-with-motor-driver-and-easy-IO board that we happened to choose to use for Mech Warfare, hence the larger regulator to run sensors/etc. You could desolder the blue terminal blocks.

-Fergs

Sienna
07-13-2009, 04:00 PM
Can you post a schematic and maybe a volunteer can turn it into a SMD version for those wishing for a smaller profile?

lnxfergy
07-13-2009, 05:20 PM
Can you post a schematic and maybe a volunteer can turn it into a SMD version for those wishing for a smaller profile?

The schematic is already out there.. but, you have to look at costs here. Running a batch of the boards is no more expensive, but assembly for small quantities is highly expensive. Tyb basically assembled the boards last year at no cost, you can't do that with SMD, most of us don't have a reflow setup just sitting on our desk. Do we want to tell competitors "here, why don't you assemble this itsy bitsy SMD board"...

I'm not sure you can make it much smaller either, sure you can hack off the motor drivers and power circuit, but you have to keep those 1x3 headers for each of your sensors, there is gonna have to be a header for the ISP, switching to a SMD AVR is a fun time too, since now you're responsible for having an ISP to make sure that the chip works before we put the final firmware on it at the show.

Is size really that big of a deal? The friggin target plates are 3x3", I'm thinking only a very serious mechEng is gonna be able to get a super tiny mech to walk and still carry all the stuff it needs (camera/guns weigh quite a bit). Weight of the PCB is nearly negligible, you might be able to shave a whopping 5-10g off by dumping the regulator/unneeded headers, maybe another 1/2g if you were to shrink the size of the board. Of course, you better carry a spare, cause we won't have one, and if you scoring system craps out you're back using a bigger board. I'd guess all told, you'd probably have $100-200 into your "mini scoring 1-off" board and support equipment.

-Fergs

Sienna
07-14-2009, 05:22 AM
The schematic is already out there..
You are right, I apologize, I didn't look in the manual you made.

I wasn't asking so much to "compete" with the through hole solution. I was more asking for my own benefit.

You adapted a generic AVR - Xbee board you made into the target system. I was thinking about making a similar generic AVR - Xbee board, only a little smaller. Not because I am saying your solution is evil and bad, but because I want more SMD design experience really. If I can make a generic AVR-Xbee board that happens to be compatible with the scoring system firmware, I think that could only be a good thing. If needed, I could assemble a few units for competitors at cost (as long as they don't mind their resistors being slightly crooked on the pads :P)

----

Now, I don't know where the 2010 target board discussion is supposed to happen, but, some questions:

1) Is it possible to get the serial interface that was talked about for 2009 in place for 2010? It would be beneficial for mech pilots I think to know not only have they been hit, but from which side. (Or if not serial, three pins for a binary output?)

2) Has anyone looked at these piezos (http://www.sparkfun.com/commerce/product_info.php?products_id=9198) from SFE? I bought a couple, haven't been able to play with them yet, but they look promising for one axis sensing.

ScuD
07-14-2009, 06:45 AM
I must have missed it somewhere, but what exactly was the problem with the piezo sensors?
Were they triggered by the walking itself? Or legs bumping into the plates?

I ask because these vibration sensors Sienna links to are basically nothing more than the piezo sensors used before, with a weight attached to them.
Wouldn't those pose the same issues?

Again, I only ask because I do not know the actual reason they failed.

@Sienna, there's something like "hygroscopical" soldering tips available, I don't know the true name and for the life of me can't find a link to it, I have one at work for my Weller station.
It's something like a plain conical tip, but one end has a semi-spherical cutout, which sucks up excess solder.

Works like a charm for anything down to 0402, lqfp-144 works a treat.

lnxfergy
07-14-2009, 07:13 AM
You guessed it Scud, the problem was false hits when mechs walked. We also had a really hard time trying to do any sort of 'calibration', the pulse out pretty much hit the peaks everytime, meaning we couldn't do anything to adjust sensitivity. We are looking at using an FSR in the future, but nothing is entirely nailed down.

-Fergs

ScuD
07-14-2009, 08:07 AM
Argh, too bad, you don't run across a lot of sensors that are "too sensitive" :) I'm sorry I didn't check this, but then I didn't have a mech to test it on anyway..

Looking forward to other experiences with different types of sensors, I'll put on my thinking cap again too.

lnxfergy
07-14-2009, 08:12 AM
but then I didn't have a mech to test it on anyway..

That was exactly my problem too! When I was building the system, I had no mech to test on.

-Fergs

ScuD
07-14-2009, 08:28 AM
Guess that's what year one was all about- figuring out what works and what doesn't.

I'm guessing years 2-3-4 will mostly fall along that line as well ;)

lnxfergy
07-14-2009, 08:48 AM
Guess that's what year one was all about- figuring out what works and what doesn't.

I'm guessing years 2-3-4 will mostly fall along that line as well ;)

I think that's pretty typical of any competition though. As a competitor, it takes you 2-3 tries before you've got a solid handle on what is necessary to be competitive. As competition organizers, it's even harder, since nobody knows what's necessary.

I think we've learned enough that we'll have an excellent scoring system for 2010, mainly I just didn't have time/equipment to thoroughly develop and test the system.

-Fergs

darrellt
09-26-2009, 04:40 PM
Hey guys,

I am experimenting with a hit sensor design using "velostat" conductive plastic pressing against some un-etched pcb. I got the idea from here:

h (http://www.instructables.com/id/Conductive_Thread_Pressure_Sensor/)http://www.instructables.com/id/Conductive_Thread_Pressure_Sensor/

The advantage of this design is that its dirt cheep and not sensitive to vibration. I still have more testing to do before I can make a recommendation however.

Anyway I now have three square yards of this stuff as that was the minimum order. So if anybody wants a sample to experiment with then PM your address and I will put some in an envelope and send it your way. Free stuff! Yay!

Adrenalynn
09-26-2009, 05:22 PM
Just as a heads-up - the targets are going to be standardized next year - no user-constructed targets.

lnxfergy
09-26-2009, 06:45 PM
We (tyberius and myself), are actively developing the new target plates. They will work on a similar principle - Force Sensing Resistor (FSR), however the FSR's are premade, and won't require sewing.... (the sensor is also rigid, avoidng the noise that was clearly evident in the video near the end of the page you linked)

-Fergs

darrellt
09-26-2009, 07:10 PM
The system I am testing is a rigid board. No sewing. It is effectively a 3inch square FSR, I could not find any of that size that were that size available commercially.

The velostat is 3mil conductive plastic that is taped over a 3" 3"copper plate (un-etched PCB). The edges of the copper board have tape on them so that the velostat is suspended a hairs width above the copper. The copper is powered and the plastic is grounded. When pressure is applied to the plastic it presses against the copper which lowers the resistance from the mega-ohms range to the kilo-ohms range. The resistance of the target can be checked by using the microcontroller alone, but I may have to add a latch or a Monostable multivibrator to make the signal duration longer, since the bb does not stay in contact with the board for very long. These ICs can be mounted on the back of the PCB if needed.

I will post some pics later.

darrellt
09-26-2009, 09:16 PM
Video of the velostat hit sensor prototype here: http://www.youtube.com/watch?v=0urooe1QJ5I


Like I said, if anybody wants a sample to play with, I can mail you some. I have plenty of this stuff.

lnxfergy
09-26-2009, 09:40 PM
Darrel,

While I appreciate your interest in developing stuff, one of our primary aims of the new target plates is repeatability... that means using FSRs that are commercially produced in a repeatable way. Please understand that any competitor for 2010 will have to use the specified target boards (more details to come). The "build your own boards" was only for year 1, since we didn't have enough time to go farther in development.

-Fergs

darrellt
09-26-2009, 10:00 PM
Fergs,

I agree with you 100% about repeatability. I was not proposing this as a design each person builds himself. I was more thinking of a design that could be cheep and simple to mass produce by one person. If there is a commercial force sensing resistor of the size needed I agree that would be a better route.

Advantages: dirt cheap, dead simple, can detect nerf hits, back of PCB is available to mount electronics
Disadvantages: repeatability, not transparent, durability?

DresnerRobotics
09-26-2009, 10:32 PM
We have a pretty affordable solution in mind, which effectively creates a 3x3" FSR. More details in the near future ;)

darrellt
09-27-2009, 01:28 AM
Tyb and Fergs,

Thanks for all the work you guys have put in on this scoring system stuff. I am quite excited about next years competition. For some un-logical reason the mere fact that fighting mech robots exist reaffirms my faith in mankind.

Sienna
09-27-2009, 09:31 AM
Are modifications to the scoring transmit board allowed? (As in, can I cut out the voltage regulator?)

lnxfergy
09-27-2009, 10:46 AM
I'm ok with removing the voltage regulator and terminal blocks (as yes, they are quite large), as long as you can verify that your power supply is sufficient, and that the board still operates as designed).

-Fergs