Totally Offensive 2006 Build Report (Robots)

Robots:

Design & Build Report for Totally Offensive 2006

This report (and the robot) evolved considerably over several months; you can skip down to the Version 6.0 (04/21/2006) redesign, Version 6.1 (04/22/2006), Version 6.1a (04/23/2006), Version 6.1b (04/24/2006), Version 6.1c (04/25/2006) tweaks, Visible Robot (05/27/2006) build, followed by Version 6.1e (05/30/2006) and Version 6.1f (06/05/2006) tweaks. Then Fitting & Filing (06/18/2006), Battery Box Mods (06/23/2006), Frame Arrival (07/10/2006), and Frame Assembly Day 2 (07/11/2006), We Have a Robot (07/20/2006), and Final Weekend Tweaking (07/23/2006). Then we had Teething Pains at Nationals (08/12/2006), which lead to Version 7 Initial Thoughts (8/16/2006).

After

the impressive debut of Dick Stuplich's F-Bomb, it was obvious that Team Mad Overlord could not rest on its laurels; Totally Offensive would have to be Totally Redesigned if it was not to be Totally Outdated.

With this in mind, I started with a blank sheet of paper, taking into account everything I'd learned from building the previous TO incarnations. The only preconditions were that I wanted to continue using the Whyachi TWM-3RS weapon gearbox, and I wanted a Titanium frame. Helping me in this process, providing valuable feedback (and some of the best ideas), were Peter Smith of Team Rolling Thunder, and Nick Martin of Team Overkill.

F-Bomb had some great innovations, in particular the friction drive. I didn't want to copy that, but instead, I wanted to see how I could take advantage of some of the limitations that it places on Dick's design in order to defuse it.

It became clear that the new frame was going to be heavier than the old one, in particular if I wanted to stay with 0.100 titanium. But if I went to a Ti blade, I could get the same moment of inertia, KE and spinup time with less weight in the blade itself, and thus could devote more to the frame. TO hits quite hard enough, and I didn't feel a need to make it hit harder. A faster spinup time would mean it was less vulnerable to the ever-more-frequent box rushes.

The other major basic decision I made is that this would be the first bot I would totally model in Solidworks, and that it would depend on waterjet cutting and welding for most of the construction. With those in mind, I went wild in Solidworks.

Reflecting upon the “metagame”, it was clear to me that Dick beat TO at Motorama before he even got into the box, by designing F-Bomb so it could adjust its blade height. So in order to beat him at his own game, TO2006 had to be able to do the same, but even better. I came up with the idea of having the rear spar be invertable, so that I could rotate it 180° at the last minute, and run with the blade on top, adjusting the blade height with a UHMW slip plate on the new underside of the robot. At this point, I was still planning on using one of TO's signature UHMW self-righting hoops.

But Peter, looking at one of the early rough prototypes, made a crucial suggestion. He pointed out that for not much more weight, I could replace the hoop with a second set of wheels, drive them via belts, and get total invertability without any reconfiguration. This appealed to me because I had come to the conclusion that blade-on-top, with the shaft not touching the ground, would make the bot much more driveable.

It also played well with my decision to move from the Whyachi T-Boxes (now out of production, and with my stock running out due to combat losses) to the TWR-15 gearbox (which is also a lot cheaper). And even better, it suddenly struck me that this provided a third driving configuration, which I'm calling “Dutch Divine Wind” mode (“Divine Wind” in Japanese is “Kamikaze”). It may be possible to start the bot standing straight upright, spin up carefully, and then tip it over on top of the other bot! That'll put a hurt on that Victor on the top of F-Bomb!

With that decision made, the robot quickly evolved to the design you see here. The main chassis will consist of a welded tub with a top panel (itself made of 3 plates welded together) that bolts onto it. Inside this chassis is the weapon gearbox (now fully protected by Ti), the MagMotor (with a modified endbell that admits 3 support bolts to lock it into the tub, battery compartment sized for ringed battlepacks (though the new Lithium-Manganese packs will fit), and electronics above the battery compartment. The MagMotor will be driven by a Victor with a fan, and that fan will also provide some cooling for the stack of 2 victors that run the drive motors. There is space for a full-sized radio and gyro. Oh, and as a cute extra, I can mount a PC-fan blade assembly on the MagMotor shaft to provide a gentle breeze to the batteries!

All the plates are designed with interlocking tabs. This is to make positioning easier for those plates that are welded together, and to provide stress relief on the bolts that attach the top plate (the bolts won't be in shear). The top plate also has a tab that goes into a slot in the rear plate, then the top plate can be hinged down into the tabs on the side plates before bolting.

Everything is designed for ease of access. 10 bolts and the top lid comes off, then a few more and the electronics comes out and the batteries can be swapped. There are small loops on the top and bottom plates that form a simple blade-locking mechanism that can be engaged no matter what side of the bot is up. There's also some cooling slots in the top plate to admit air.

The battery box is foam lined (I got a little sloppy, and didn't trim back the foam in the models in a few places, so there are some overlaps) and has cooling slots in it. In the closeup of the battery assembly, you can see some of the new “3D” nutstrip I came up with, in the vertical sections. This lets me efficiently attach the top of the battery box.

The rear spar is shock-isolated from the main frame (some ideas are too good, and too simple, not to steal). The spar has two stiffening struts (in blue) and reinforced guard plates protecting the wheels. An adjustable belt tensioning system is designed in, though Peter thinks I won't need it. The sides of the strut are actually on a slight angle, so the bot will tend to self-right with the blade on top if it comes down on the side. There is a provision to add side-hoops of UHMW between the wheel-protectors for extra shock absorbtion

4/21/06: Version 6 Design

The trip

to Battle Beach 3 was very useful. Discussions with several builders lead me to make significant changes to the TO2006 design.

The most obvious changes are moving to a tubular spar (as Peter is always telling me, tubes are strong), and the front "skid plate" that can be quickly adjusted to raise and lower the blade height when running blade up (not shown in these renders is the self-righting hoop). The skid-plate is designed so for the lowest blade heights, I can just switch to small wheels mounted directly to the frame. The wheels themselves are just discs of UHMW, the fact that they roll is a bonus, but they are really designed to skid.

Although this gives up the 4-wheel drive system, we can still self-right using the TO-Hoop™, and it has the advantage of being mechanically a lot simpler. KISS.

There are changes inside the bot as well. I decided to take a technological risk and design for the new LiMn batteries, which are lighter and more compact that NiMH. The battery compartment is sized to run a 9-cell pack (33 volts!) but I will initially run 7-cells. The battery/electronics box is designed with interlock tabs to guides assembly, and can be lifted right out of the bot after a few mount screws are removed. It sandwiches between the magmotor and the MS-1 switch for added support. All the various electronics parts will be connected with powerpoles so as to make repair and replacement easier; the Magmotor and gearbox just drop in, for example.

Getting to the Victor speed controllers requires removing the battery/electronics compartment, but this should not be a huge hardship. The weapon-controller has a fan on it, and can be replaced with a contactor if need be. There is plenty of room in the electronics bay for whatever other toys need to go in it.

Note also the small "step" in the frame that gives more blade clearance; this is an attempt to reduce any self-strike problems with springy Titanium blades. I also changed the way the top plate mounts to the side plates; there are now hinge slots on one side plate instead of the rear plate. This lets me avoid having nutstrip on the battery side of the robot.

Another improvement is cooling. In addition to the fan on the weapon Victor, there are two fans in the front section of the robot, above the MagMotor. These suck air through the slots at the rear, down over the victors, around the MagMotor, and out of the robot. As fans are notoriously fragile, I figured it would be a good idea to have some backup for the main Victor fan.

The rear spar slides through two hoop extensions in the side plate and bolts directly to the rear of the chassis. At the ends of the spar are the motor assemblies; as you can see, I've gone back to my trusty T-Boxes. The motor power wires run through the tube (entering through a central hole, out the end, then run back to the motor; there is a slight exposure but it can be avoided with a little tricky machining on the tubes.

The nice thing about this design is that the tubes only require a few holes in them, and all these holes are in the same plane. The robot can be converted from blade-top to blade-underneath by simply unbolting and rotating the motor mounts 180-degrees! The mounts have slots in them to let me strap down the motors as in the current TO design, and protect the motor and gearbox on multiple sides, much better than TO2005. Also, the T-Box bolts in using all of its mount holes and the two end holes, so it's more securely attached.

The big wheel-protection deflector assemblies are actually separate from the motor mounts, and bolt on using the T-Box mount holes; they also incorporate a dual interference tab lock to take stress off the bolts. When they are on the bot, the self-righting hoop can mount to the outside of the assemblies.

AFAICT, with proper design of the hoop, it should still be possible to do "Dutch Divine Wind" mode.

Weight is a bit tight. Based on the current TO, which is a few ounces under 30lbs, and changing to the speed controller from the contactor, I have about 6.1lbs to use for the frame. Currently, the chassis, skid, tube (medium 0.0350-wall Ti), and 2 mounts runs 6.26 lbs, plus there'll be some welding weight gain. The deflector plates are 0.35 lbs each so I won't be able to use them unless I go to a lighter Ti blade. A thinner or shorter (currently 24") tube is also a possibility; TO2005 has a 23" spar width.

I am not counting the weight savings from going to the LiMn batteries, because I may have to back off that design decision, and that's my spare weight in any case. But it goes without saying that I'd like you to speak up if you see any places where I can shave a bit of structure.

4/22/06: Version 6.1 Revisions

Based

on feedback from Nick and Peter, I made the following revisions:

* Added a small access hole that lets me get at the receiver power plug to turn it on and off; it is covered by a swinging tab that bolts into place.

* Mounted the Victors upside down to the bottom of the battery/electronics compartment. This will permit the whole assembly to be removed from the robot by disconnecting the motor drive wires (4), the weapon power feed to its Victor (2), and the PWM cable to the weapon Victor (the main power light and fans will feed from that controller's supply).

* Thickened the motor spar supports and made thickening panels for them to increase their rigidity.

* Made provision for a shock-mount table for the weapon victor if space permits, by putting extra mount holes into the chassis base.

4/23/06: Version 6.1a Tweaks

Just

a couple of minor tweaks today. On the rear plate, I added a third bolt; the real purpose here is to either run 2 bolts and use the third hole for the wires, or run one and use one or both of the extra holes for wires. My feeling is that since the purpose of the bolt(s) is to prevent tube axial and rotary motion, they don't have to be tightened all that much, and so we can probably get away with 1 bolt, and not have to worry about tube crush.

The other mod is to make the pair of tube ring plates slightly different in profile to give more area for a sold weld.

At the other end of the robot, I decided to get a little artistic on the skid plate...

4/24/06: Version 6.1b Tweaks

Based

on suggestions from Nick, I beefed up the mounting points for the front skid plate, made provision for adding some UHMW backup skids to it, and created a smaller version of the plate for use with lower elevations to bring the forward contact point further ahead of the robot center-of-gravity. I also moved the skid plate rollers inside the plate to better protect them, and reinforced the corners a bit.

4/25/06: Version 6.1c Tweaks

Continued

dissatisfaction with the design of the front skid plate lead to a rethink and radical simplification. Now it's just a flat plate with a lot of mounting holes. The minimum height available is 1.33", and just about any height can be created by coupling together several of the plates (I will also make a shorter extender plate). A new set of tabs in the chassis provide an extra mount point, and some access holes in the front plate will form guides for drilling and tapping an extra set of 5/16-18 holes into the gearboxes. The simplicity of the design will hopefully make it more resistant to battle damage, or if not, at least easy to repair.

The mounting holes for the previous version have been retained to provide a fallback to the older design if so desired. Alas, no room for my pretty logo! Oh, and in practice, the mount blocks and roller will be behind the plate; they are in front for purposes of clarity.

5/27/06: The Visible Robot

With

the design finalized, I turned to Metfab to jetcut the acrylic for the test-fit version of the bot (aka "The Visible Robot"). There are many great places to get robot parts cut by waterjet, but at BattleBeach, the Team O-Town boys sang the praises of Metfab, and so I went with them.

While the cut parts were wending their way to me, I busied myself with the task of making up a bunch of nutstrip in the various sizes the robot would need. As I plan to display the test version as an exhibit at events, to show people what's inside our toys, I made twice as much as was needed, plus spares. I also made up the small aluminum blocks that mount the small roller to the vertical height adjustment plate used when the robot fights with the blade on top.

When

the parts arrived, I got to work assembling them, starting with the garolite battery box. I found the tolerances to be quite good, though a little filing was needed in order to get things to fit; the is apparently a consequence of the interaction of the waterjet with thin and light materials like garolite and acrylic. While this is going to be less of an issue with the Titanium, it is something to keep in mind.

The assembly of the main chassis was relatively straightforward, using a plastic weld to join the various parts. It did, however, reveal some minor design mistakes:

The beautiful fan cover outlines I'd laid out were for the wrong size fan (I grabbed the wrong part in Solidworks!
The tub that the motor and gearbox fits into was sized exactly to fit. However, the actual Magmotor is 0.07" longer than the technical drawings say! So it didn't fit. Whether or not I just have a slightly stretched Magmotor, it's clear that sizing the tub to exactly fit the length is a bad idea, so I'm going to extend the tub by 0.200.
The vertical tabs in the front chassis plate that help secure the vertical height adjustment plate do not provide sufficient clearance for the 5/16-18 bolts that lock down the top plate to the gearbox, so they will have to be redesigned.
The slots in one of the side plates into which the rear top plate (that gives access to battery and electronics) tabs are inserted are a little tight and need to be made bigger.
The rings into which the rear spar fits are currently exactly sized to fit the spar; this required a fair amount of filing to make them fit, so I will probably oversize them slightly.
The two holes in one of the small support plates in the drive wheel assemblies that match the motor mount screw access holes (which are tapped 5/16-18 on my T-boxes) were not in the right place, due to an error in the model.

An interesting problem that came up during the build was the issue of how to drill the holes in the tubular rear spar. All the holes have to be precisely aligned, so that the T-Boxes aren't tilted forwards or backwards. A milling machine is perfect for this, except that my mill has a 9 inch travel, and the tube is 24" long. A minor mismatch!

The solution I came up with was to drill the three central holes in the spar, and then use them to mount the spar on top of a 12" long strip of small C-channel. The channel is then used to align the tube on the bed of the mill (using some dowel pins) and the whole contraption is clamped down. I actually had to extend the slots in the clamps in order to get a good grip. Once that was done, it was easy to use a Center Finder (part number 22025A11) to line up and drill away.

A final

design change that I am considering is the position of the weapon speed controller. In the original design, it is flat on the floor of the chassis, and uses a fan to draw air down over the FETs, which then passes over the other two speed controllers and is eventually exhausted by the two fans above the MagMotor. However, looking at the final result, I think the fit is a little too tight; note that the fan is partially hidden under the battery box.

A potentially better solution would be to mount the speed controller against the rear chassis plate, and change the battery box design so that it mounts a fan or two that blows air against the FETs; this would mean that the two fans above the MagMotor are intake fans.

A side-effect of this change is that I could probably increase the size of the "step" in the chassis by about 0.25", increasing the blade clearance. BTW, with the current 24" spar, the bot could swing a 34" blade, though the MOI would be such that it would spin up very, very slowly. Maybe when I'm running at 36v using "Mr Fusion" as the power supply...

5/30/06: Version 6.1e Tweaks

Based

on feedback from Nick and Peter, as well as a few semi-bright ideas of my own, the following “improvements” have been made to the design:

The vertical height adjustment plate has been redesigned to use slots and a locking tab instead of extra bolts to lock it in place and prevent front/back movement.
Careful consideration of structural issues indicates that the front top plate (the top plates were the ones that were removable) is not adding much to the rigidity of the final assembly, given the presence of the big honking gearbox. So while I'm still going to make it (now with the cute Team Mad Overlord logo stencilled in), it probably won't be used in the final robot.
The top middle plate (the one on an angle) has been split up into two plates. The smaller one is removable and will probably be made out of garolite (I'll make a Ti one as well). The larger one (that mounts the fans, resized to handle 40mm fans) will be welded into plate for extra rigidity. With the smaller plate removed and the battery/electronics bay out of the robot, the entire MagMotor/Gearbox unit can be slipped in and out of the robot. Since the battery/electronics bay comes out as a single unit, this isn't a big deal.
The rear top plate has been redesigned with some cutouts to eliminate some weight. Under it is a new sheet of garolite that forms a top plate to the battery/electronics bay. The combination of the eliminate of the front top plate and the modifications to the middle and rear top plates saves about 1/3 of a pound.
The weapon Victor has been repositioned to mount (using a spacer plate) to the rear of the chassis. The battery/electronics bay has been reconfigured to mount two 30mm fans that blow on the weapon victor. The idea is that the pairs of 30mm and 40mm fans give some redundancy, since they both push air through a plenum that contains the Magmotor and all the Victors.
The “step” in the chassis has been increased in size, giving more blade flex clearance.
Some of the nutstrip has been reconfigured to make it smaller; every gram counts!

6/05/06: Version 6.1f Tweaks

In final

preparation for cutting the parts, I tweaked the vertical height adjustment plate locking mechanism based on some observations by Nick, and finalized the battery box design with a pressure plate that sits right on top of the battery, and the top plate that sits right under to Ti top plate (covering the cutouts). The area between the pressure plate and the top plate will be filled with foam (I may also have one underneath the battery to try and “float” it a bit. The electronics area will also be padded with foam. Another tiny tweak is that the weapon Victor now has a shock isolation plate modelled for it that connects it to the rear of the chassis.

The final weight for the chassis comes in at just under 6.1lbs, which is right on my weight allowance as computed with NiMH batteries. The LiMn packs will save me some weight, which will give me some leeway to increase the blade weight, reinforce the rear spar, etc.

The fact that the weight was coming in just right for the NiMH packs got me thinking - if I have the weight for the NiMH packs, will they fit into the robot. And as it turns out, with a very minor tweak to the battery box, they will! I'm currently waiting for the LiMn packs to arrive for a final fit test, and when that is done, I'll tweak the box so I can run NiMH as a backup.

Based on advice from Jack Price at Duralite, I am having my LiMn battery built as a 4- and 3- cell pair of packs. That way, each pack can be balanced individually (balancers only handle 2-6 cells), and if I decide to run at a higher voltage, I can just buy a 5-cell pack and either run 3+5 or 4+5.

6/18/06: Fitting and Filing

With

the design finally set, I sent off the cutting plan to MetFab, and within a week, I'd gotten the jetcut parts back. When the UPS Guy gave me the flat package, with all the loose parts tinkling around, he told me “I think your mirror is broken...” I bade him fear not, and hustled up to the workshop to start the assembly process.

As with the Acrylic version, there was some fitting and filing to be done. While the overall dimensions of the parts are very accurate, the waterjet can wander a bit, and because the jet spreads as it cuts, the cuts are not quite at right angles to the surface. So a few happy hours were whiled away getting everything to fit together just so. The toughest part was getting the inside of the rear spar rings smooth and just the right diameter (and getting the reinforcement rings to line up; I had to make them a loose fit that will be locked into place during welding). However, some work with a dremel sanding drum proved up to the challenge.

The interlocking tabs proved to be a great idea - while they required some extra filing to get the fit perfect, when the rear tube and nutstrip are installed, the tabs form interference fits and hold everything together. This will make it easier to transport and will serve as a guide for welding.

Getting the motor modules assembled was a bit of a pain, because of the little tabs that interlocked, but I found a particular dremel bit that was just the right size, and after that it was easy. I also milled an indent around the lower plate shaft hole (so that the bearing won't scrape on the lower plate). I had to get a bit crazy on the fixturing, since my rotary table wasn't big enough to let me mount the plate centered on the table and clamp it on all sides. So I clamped it on one side, and used the mill vise as a weight to hold down the other side. Tolerances weren't a big issue, so it worked fine. I also drilled the motor mount holes in one of the Ti spars so I could mount the motor modules and see the finished result.

Finally, the Duraflite LiMn battery came in. I had it built as 4-cell and 3-cell packs, so that each pack can be balanced (balancers only handle up to 6 cell packs). The wires made it a bit bigger than I expected, but nothing that cannot be accomodated.

6/23/06: Battery Box Mods

When the

LiMn packs arrived, I did a test fit using the existing battery box design. While it fits OK, it's a little tight due to the routing of the wiring inside the packs. To deal with this issue, and also make it easier to reconfigure to use the old NiMh packs, I redesigned the electronics bay to incorporate an adjustable sidewall, changed some of the nutstrip supports to make them stronger, and rethought the wiring via cutouts. This was complicated by the fact that it had to be compatible with the existing chassis; the only modification needed will be some extra holes in the rear top plate.

7/10/06: The Frame Arrives

Today

the frame arrived from Team Whyachi. As you can see, Terry did a great job welding it up. The only issues that came up in the initial assembly were that I had to do some minor dremelling to clean up the rear spar support rings, and I had to slim the gearbox by about 0.015" in order to get it to slide into the rear of the chassis. While the gearbox fit in its final position was perfect, I didn't think to tell Terry to use a gearbox as a spacer for welding the rear of the chassis as well, and the result was that it was very slightly less than 3" wide at the point where the gearbox slides in (probably due to the clamping he did when welding). But a few minutes on the mill with a fly cutter fixed that.

I am currently waiting for the new battery box and blade jetcuts to arrive from Metfab, but in the meantime, I'm going to work on the fans and the motor wiring.

7/11/06: Frame Assembly, Day 2

Lots

of little tweaks today. I modified the TWM-3RS gearboxes so they had extra tapped holes needed for the vertical height adjuster plates, and tweaked those plates so that everything fit. Also did some filing and fitting to get some of the other plates to fit (the welding caused the fit to change by a few thousandths of an inch). Finally, I did some motor testing, and put the robot together in inverted mode to test the blade height at maximum extension.

7/20/06: We have a Robot

Well,

it took a lot of work, but the robot is finally mostly together. I faced a number of time-consuming problems along the way.

First, after getting the blade jetcut parts from Metfab, I had to bore out the center holes of the blade, plus prepare the mounts to take the tips that Team Whyachi was making for me. Getting centered on the existing bore holes was a real pain before I decided to use the mill holder (which had a convenient bevel) to do the work for me. The final blade is slightly larger than the original TO blade but has less MOI because of the titanium construction. It also has an optional stiffening blade that bolts on to increase the stiffness. The tips are hardened tool steel with two different tip bevel profiles, hopefully giving me some options against different opponent profiles.

I also found that I'd very much underestimated how much space I needed for wiring, which was truly vexing. There was no way the wires were going to fit in the space that I had. I tried a lot of different configurations before I came up with the final solution, which routes the wires through the radio compartment. But even this was too tight a fit, and I ended up having to remove the bulkhead between the batteries and the radio in order to get it to work. Even so, I'm not particularly happy with the result, because everything's crammed in tightly.

I hereby coin “Robert's Robotic Rule”, as follows:

Wiring takes up 3 times as much space as you think it will.

Still, it drives, and the blade spins up (but chicken that I am, I only spun it up about 25%). On the to-do list, installing the main chassis fans, installing the antenna mount, making and testing the self-righting hoop, and preparing the second electronics bay. Oh, and seeing how it drives with the blade on top...

7/23/06: Final Weekend Tweaking

Spent

the weekend doing a lot of finishing touches; I built the TO self-righting hoop and tested out a new front roll point for it (older versions of the bot had a tab that extended forward, but this one uses a disc to create, with the rear hoop, the conical section that the bot rotates around). I also tested the driveability with the blade on top -- as expected, much better than with the blade on the bottom.

And with that, the bot is ready for Nationals. Likely it will have some teething troubles, but that is par for the course. Lord Hill of BattlePack came through and made me some special NiMH packs that fit into the space that I have, much alleviating my wiring woes, though a complete redesign of the battery and electronics module is going to be my #1 priority after the event -- assuming there is a robot left to modify, of course!

8/12/06: Teething Pains at Nationals

We

went to Nationals expecting to work out some kinks in the robot, but as usual, they were in unexpected places (which make sense, since if we expected them, we'd fix them before we went to the event!).

The big problem was the blade teeth - they shattered like glass. Obviously, the design is flawed. Once they shattered, the ends of the Ti blades got mangled. So they'll have to be reworked, and new tips designed

The other big issue was the current draw for the Mag when we used the top secret "long reach" blades was enough to melt the 3200 NiMh packs I decided to use. In addition to my basic blade, I'd made up two longer ones, to give me some reach over the opposition.

The first fight, with Dick Stuplich's Buggy Brick, went ok; we got some lucky hits and totalled him out of the tournament, but in the process lost 4 blade tips and partially munched the end of the short blade.

Second fight was Peter's Xhilarating Impax. Went the distance; I went with 2 tips on the medium blade, they shattered, then we lost the ends of the blades at the notch where they fit. We lost most power towards the end, and found out why in the pits - the battery pack had melted to the point that it came out of the bot as a necklace of cells! The glue had melted!

The 3rd fight was against Relic (F-Bomb). We decided to go with the long blade, set high to go after his electronics panniers. Glancing first hit, nothing big, and suddenly I could barely drive. He got behind and munched my drive assemblies, and that was all she wrote. I think a glancing hit from the rebound jammed up a gearbox.

On the bright side, the frame held up pretty well, only a few minor dings that will require a bit of tweaking. A rear tube and motor module were lost when Relic whomped them directly, but held up perfectly to anything short of that. Best of all, the tweaks to lock down the endbell of the magmotor seem to have worked, the Mag is still working sweetly. I'll have to tear it down to be sure, but it's promising.

There is lots of work to do, but I think this version of the bot can evolve into a contender in the current environment. But the competition is stiff; Relic, of course, but also the latest Killabyte, which was really, really impressive, and Xhilarating Impax shows that it's possible to build a wedge that can swim with the spinner sharks.

In addition to the fighting, we got to test out the BrightSaber lighted blade, which worked quite well.

Kudos to Team Rolling Thunder for their double - 1st and 2nd in 12's and 1st in 30's! It certainly was their day!

8/16/06: Version 7 Initial Thoughts

Here are

my initial thoughts on how to improve the robot and correct the issues that came to light at Nationals.

The first problem is one of space. The robot was simply too tightly packed, due to my error in not providing enough wiring space. The first potential modification is to move the weapon victor and mount it in one of the fan bays above the MagMotor. As it happens, with a 10mm fan, it fits perfectly, with room for a thin lower support plate. In fact, the current nutstrip support holes can be reused -- they are in just the right places!

Secondly, I can move the drive victors back to their traditional placements, out on the rear spar next to the t-boxes. This will require new rear motor modules to be constructed. While this makes them more exposed, I've yet to lose a Victor when they've been hanging out there before, usually tucked under a UHMW bonnet.

A more radical modification would be to eliminate the self-righting hoop and go to a 2-wheel motor module, with one wheel driven by a timing belt. I've quickly modelled it using a second, empty T-Box. While considerably heavier, this would have the advantage of not requiring a decision about which way to fight the robot until it goes into the box, and it would permit “Dutch Divine Wind” mode. I won't know until I get the robot back and can do some careful weighing if it is practical.

One tweak not displayed in the modelling will be some extra through holes for spar support bolts, so that they are further apart. This should reduce stress on the spar during big bounces (though it held up perfectly until Relic directly hit it).

Moving on to the blade, the problem here is how to salvage the current blades, the tips of which were mangled when the original hammers broke. My thinking is to roughly cut off the current tips and add a very simple, much larger hammer that fits over the current blade profile. Since the blades taper, the hammers will slide into position and allow me to use them as templates to both drill the mounting holes, and to trim off the ends of the blades. As the taper of each of the blades is different, I'll have to make complete sets for each.

The new hammers will likely be hardened Chromoly, which should be less likely to shatter. However, I'm concerned about three things: first, the 90-degree inner angle where the hammer meets the blade - perhaps it should be radiused, even though this will require extra shaping of the blade profile. Secondly, the fact that there is no "lock" that keeps the hammers in position, other than the bolts - is there going to be enough radial shear stress during impacts to cause problems (as modelled, it uses 4 5/16-18 mounting bolts). Finally, the bottom part of the hammer that forms the bed into which the blade rests is currently only 0.25" thick - is this enough? Right now, the entire hammer is 0.75" thick, so it has a slight overlap, and I could eliminate this and make the bed a bit thicker, but it can't get thicker than 0.385 (the thickness of the Ti blade) without causing some space problems.

The reason for this very simple design is that I want something that I can easily work with in my current workshop, which does not have a large or fancy milling machine. I'll have the hammers made, and then fit them myself.

As always, your comments and suggestions are most appreciated.

Robots: