COTS Parts made just for FIRST Robots

[Andy A.]

I'd like to see a large source of DeWalt transmissions, clamshells and properly bored out pinion gears that are available only to FIRST teams. As the popularity of this transmission grows, it will be harder and harder to find them come January (it was already difficult this year!). Also, having a supply of transmissions already or 90% modified to accept motors would be nice. It's not that it's difficult to do your self, but I'd be willing to spend a little extra to save some time and have it done professionally.

I really can't give a good quote on what that would cost, having no real idea where to start.

Ideally, we would just get DeWalt to include some in the kit.

-Andy A.

[Paul Copioli]

When dealing with the wheels that many teams make from solid 2 inch thick by 4, 6, 8 inch diameters and machine down we may want to think about castings. Quantity is your friend when dealing with castings, but you need the quantity to get the price right. Below is the example for the kit transmission development:

The kit gearbox housing is made from die cast aluminum. The prototypes were machined from solid. The gearbox design mimics team 217's design strategy for the past 5 years: enclosed, machined housing. The main problem was we would only build 4 to 6 per year so castings were out of the question. We had everything machined from solid by one of our sponsors. This is actually quite expensive (around $300 to $400 a pop) due to machining time. The kit gearbox was going to be ordered 3600 at one time so castings were the way to go. There are several types of aluminum casting processes to choose from but two are the most popular: sand casting and die casting. Which one you choose depends mostly on quantity, but wall thickness and surface quality also come into play. Sand castings are good for quantities of 25 to about 500 and die castings are usually more cost competitive for the higher quantities (>500).

I received quotes for both machined from solid and die casting for the housing. The machined from solid was $52 per piece for quantities of >2,000. The price we paid for the die casting was $11 per piece (a factor of 5 difference). Now, this doesn't come for free because we had to pay an up front tooling charge of $11,000 for the die. Amortize the cost of the die over the initial 3,600 units and you get an adder of about $3.06 per unit so the price this year for the boxes was $14.05.

Now, sand castings have about 1/4 to 1/2 the cost of die castings for the tooling, but the piece price is slightly higher and the high quantities can't be achieved by a single tool (so many tools would have to be purchased in order to make the higher quantity). Sand castings seem to be the better fit for a company willing to sell aluminum wheels due to the lower initial cost.

I think wheels of 4", 5", 6", 7", and 8" could be made from sand cast aluminum machined with a stock bore for about $25 to $35. If key ways or hex broaching is required, the price could jump another $10 or so.

[Billfred]

Here's a few things I wouldn't mind...

1) The modified gears for the XRP transmission. The kids can assemble the rest of the deal, but it would be nice to get the heavy machining out of the way.

2) Some form of switch panel for on the robot (for setting starting position, autonomous modes, etc). As long as it can plug straight-up to the RC and is somewhat mountable, I think it'd work well.

I'll think of more eventually.

[Not2B]

I would like to see parts that smaller teams have trouble with. So I asked myself "If you don't have a lathe and mill, what is hard to do?" and "what do all teams need, and thus lends it's self to mass production?"

And I came up with 2 items.
1) Transmission - which we seem to have 2 good options already, namely the kitbot (affordable - THANK YOU) and AM shifter (bells and whistles).

2) Drive wheels - either a good keyed hub on a wheel, or a hub that allows the easy connection of a sprocket. And cheap. Now there are already several products on the market like this - the battle bots universe has several options, as does the go-kart world. So I don't know if there is a market for these... but if you make them cheap enough... or give us flexible mounting options...

Most other "yearly" parts (Frame, mounts, etc...) seem easy enough for a resource-limited team to handle.

[Steve W]

Originally Posted by Andrew Rudolph

I would like to see some gearboxes made for arms and things like that using the FP motors. Gear them down to like 15 RPM, maybe even have them with no backdrive. About the Same cost as andymark gearboxes.

Check out AndyMark for these. If they have not been added to the list I know that they have been developed and should be available soon. This is another blatant plug for AndyMark and their quality items.

[Ryan Foley]

1) shifting dogs, like AndyMarks of the square type that Andy Brockaway uses in the 716 gearboxes

2) shafts made for the shifting dogs

3) a variety of gear types/ sizes that have cut outs so the above mentioned dogs can engage them

Having a supplier of these would allow teams who dont have advanced machining capabilities or dont have a local machine shop to design and build their own shifting transmissions. The AM Shifters are amazing, but in some cases, perhaps the team needs different high and low gear ratios. These parts would allow them to design/ build their own custom gearbox that fits their needs. Or they could be used in 4-speed gearboxes.

Two crazy ideas (could take away the fun/ innovation that these drives already have, but thats a debate for another thread):
1) mecanum wheels like 190 used? Those could be fun.

2) turntables like Team 79's for use in a swerve drive? Just a simple turntable with a rectangular hole in the middle for the wheel, and some bolt holes so the gear system could be mounted on top. I dont know if 79 made the ones they use, or bought the, but I havent found any company that sells a similar product yet. Maybe they could even have large sprockets included with the turntable so teams dont have to search around for sprockets big enough to fit onto the turntables.

[Paul H]

Originally Posted by Ryan Foley

2) turntables like Team 79's for use in a swerve drive? Just a simple turntable with a rectangular hole in the middle for the wheel, and some bolt holes so the gear system could be mounted on top. I dont know if 79 made the ones they use, or bought the, but I havent found any company that sells a similar product yet. Maybe they could even have large sprockets included with the turntable so teams dont have to search around for sprockets big enough to fit onto the turntables.

They're not the exact same, but lazy susan bearings work well as turntables for swerve drive, and they're pretty cheap and available at Home Depot.

[Karthik]

Originally Posted by Andrew Rudolph

Maybe sell a frame with a prebuilt 4 or 6 wheel drive, or where you could buy a frame, wheels and gearboxes from you to quickly make a 4 or 6 wheel drive. Like the kit bot but better. Of course you take a big chance in case the kit bot is improved and your product isnt any better than the kit bot. The frame should cost around what a simmilar frame of 80/20 would cost.

I'm curious as to what sort of improvements you would like to the kit bot? The kit bot provides all the functionality you describe. Considering that it's part of the KOP, I can't imagine that there'd be much of a market for this type of product unless you have some sort of idea on how to establish differentiation.

[Conor Ryan]

An adaptor for that you could use a servo to switch the AndyMark tranny instead of the Pneumatic shift. Also a servo shifter for the Dewalt Transmission too. I'd price them about $30-$45 range depending on the price of the design and the servo.

[Paul H]

Originally Posted by Joe Johnson

3) Although IFI is probably about to redo the IO, I think it would be great to have some easy breakout board that would make customizing OI easier. I am thinking something that would plug into all the joystick ports but it would allow you to plug in 4 joysticks but you could use jumpers to map any of the joystick switches and pots to any of the joystick ports.

Also, I'd put a few LEDs on the board and some switches, along with a few slider pots and a few knobs and let those be patched in easily too.

I really like that idea.

What about dongles? $10 or so for a box with 2 switches (auto on/off, robot enable/disable) to plug into the competition port. I bet a lot of teams would buy that, especially rookies (1504 still hasn't gotten around to making ours).

[BillCloyes]

I understand that many teams already use some form of optical rotary encoders so the market might not be as large...

but a gear tooth sensor setup with custom mounts for each type of COTS gearbox/transmission (IFI or AM).

Having good sealed sensors and having them integrate seamlessly into the housings for either gearbox would be worth many times more than that of the kit gear tooth sensor ($5)

say... even $25-40 each (with sensor).... granted this is something that could be done with plastic injection molding.... but then one faces many of the same issues in higher manufacturing costs discussed by Paul Copioli about the casting

-Bill

[Cory]

Originally Posted by Paul H

I really like that idea.

What about dongles? $10 or so for a box with 2 switches (auto on/off, robot enable/disable) to plug into the competition port. I bet a lot of teams would buy that, especially rookies (1504 still hasn't gotten around to making ours).

Pretty sure that's an upcoming AndyMark product.

I forgot to grab a brochure at nats and bring it home, but I seem to remember that on there, along with some other cool new stuff.

[BillCloyes]

Oh yeah... one more thing

I would love to see the hassle and mess that comes with the PWMs be traded for a system that has a positive-locking connector

Like them or hate them, I would like to think that we can all agree that something that clicks or snaps in would be better...

Off the top of my head, something similar to phone line or ethernet connectors (RJ- something something) would be awesome... (only this type of connector, not necessarily actual phone or data cables)

Without hesitation I would take an RC, relays, Victors (and any other IFI stuff I am forgetting) that incorporated this kind of feature regardless of almost any increase in component size or shape Even a reasonable increase in cost would acceptable.

The apparently accepted norm of hot gluing the PWMs into place for last 5+ years seems a bit sad compared to all the other advances that have come about.

-Bill

[Andy A.]

Originally Posted by BillCloyes

Oh yeah... one more thing

I would love to see the hassle and mess that comes with the PWMs be traded for a system that has a positive-locking connector

Like them or hate them, I would like to think that we can all agree that something that clicks or snaps in would be better...

Off the top of my head, something similar to phone line or ethernet connectors (RJ- something something) would be awesome... (only this type of connector, not necessarily actual phone or data cables)

Without hesitation I would take an RC, relays, Victors (and any other IFI stuff I am forgetting) that incorporated this kind of feature regardless of almost any increase in component size or shape Even a reasonable increase in cost would acceptable.

The apparently accepted norm of hot gluing the PWMs into place for last 5+ years seems a bit sad compared to all the other advances that have come about.

-Bill

Bill-

Check out These available from robotmarketplace.com. I've never used them, but they are a clever little bit of metal. At 4 dollars each, they seem a little expensive.

I still prefer hot glue, though these are about a +2 to robot blingbling.

-Andy A.

[Joe Johnson]

As long as we are wishing, I suppose we should ask for IFI to go to a serial link (I2C ?) with perhaps 2 RJ-47 (phone jack) connectors that would let us daisy chain Victors, fuse blocks, Spikes, etc.

I would much rather have a single line going from Victor to Victor than what we have now.

While we are redesigning the Victor, how about a PID feedback loop built right in. There is a very good paper from Circuit Cellar Inc here that tells how we could all have PID control without encoders, using Back EMF of the motors to determine speed and position of the motors.

I would love to have this built into the Victor control.


Joe J.