PDA

View Full Version : OSCAR



Adrenalynn
06-02-2008, 10:19 AM
I've been talking for awhile about OSCAR off and on. I'm going to start fleshing him out as I move forward here.

OSCAR is an autonomous garbage can for events - he was spawned during the Maker Faire when we had to go searching for a refuse bin. "Gee, you'd think here the garbage can would come find you!"

He will drive around, ask for garbage, when he full, he'll deposit it in a full size dumpster and return to his patrol route.

OSCAR will be driven by two large (160A stall @ 36v) cart motors. The lifter design is still up for grabs, but my friend and I are starting to nail it down - that's really his piece, where base, sensors, brains, programming are mine. He will likely turn with that big bad Vantec SSP105 servo.

Motor controllers are 2x Robot Solutions 160A dual channel controllers, and battery today is 6x lawn tractor batteries (spillable), two sets of three in series on each "side" - but it's planned that we'll go to gel-cells once the base is proven out.

There's a couple additions we've come up with, we'll see how many survive. One's an animatronic "driver" on the back, randomly turning a steering wheel and banging on the horn. The second is an animatronic Oscar The Grouch in a trashcan on a pedastal, singing the "I love trash" song, asking people for garbage, and yelling at them to "SCRAM!"

I had originally welded-up something like a battlebot base for it, but then happened upon a gorgeously designed (and on sale!) garden cart at Lowes. 1200lb rating, bright yellow, sides that can be rolled down and/or removed entirely. 12" wheels. It's pretty awesome!

So I'll start posting some photos as things progress. The biggest concern today is figuring out WHERE to construct him. Without the lifter and trashcan he's already approaching several hundred pounds (lots of battery weight)! Even now he doesn't want to fit in my lab... ;)

metaform3d
06-02-2008, 11:17 AM
That's a very cool idea! Looking forward to watching your progress.

Do you have plans yet for route planning, navigation and obstacle avoidance? An autonomous vehicle that large will need to be especially safe, especially one with a Sesame Street character to act as child magnet.

Adrenalynn
06-02-2008, 11:59 AM
Totally agreed!

The plans at the moment ("no battle plan ever survives contact with the enemy**") are for home-brewed 3D laser ranging and obstacle plotting at about 7" height, on all four sides, as well as a full complement of IR and ultrasound.

Route planning via visual SLAM backed by subscription high-res GPS. The intelligence will be weighted towards "brake first, decide later". ;)

Even with all that, I expect it will be escorted. It has triple-redundant runaway kill switching as well. 315Mhz long-range (4000+ feet), 433Mhz medium-range (1500'), and a super-long-range over amateur radio (2m)

We're talking about several hundred pounds capable of moving at least a thousand pounds, and able to accelerate to nearly 30mph. This isn't a toy and safety is being designed in from the ground up. We're even looking at weight-shift in case some kid jumps on the bot - we'll be able to have it automagically shut down instantly and issue an alarm.

Processing planned right now includes 2x PicoITX, 2x duocore MiniITX, 3x Arduino, and some custom-built stuff I'll talk about later (super-dense) analog and digital input channels.

Great questions, Meta! Please keep 'em coming! As you mention, this is a potentially hazardous project - more input is ALWAYS appreciated!


** Field Marshall Helmuth Carl Bernard von Moltke

[edit: Forgot to attribute the quote...]

ooops
06-02-2008, 12:35 PM
Very cool project!
When it comes to “bots” big is beautiful! :veryhappy::veryhappy::veryhappy:
I gathered from previous posts that you were using golf cart motors. Do you plan on using the whole rear axle assembly? I assume not since you are “dual motoring”, or was that in order to handle the “skid-steering” that you decided against?
As you can tell I am excited for you on this project that said, I promise I will try and keep the questions in check.:wink:
Please keep the pictures and details coming!

Adrenalynn
06-02-2008, 12:46 PM
Not at all! Thank you for your interest!

The plan is to direct-drive each wheel via chain-drive. As it stands right now I'll bolt-up a cog on each wheel and run the chaindrive to the motors. Even though we've scrapped the slip-skid, I figure a "pseudo-slip-skid" will assist the turning servo which I'm still not confident of. Really I need to load up the cart and put a force-meter on the turning scissor and see how much torque I'm really going to require. Also experimenting with the hybrid skid I proposed above vs just straight drive.

It's pretty young yet, so there's still a lot to flesh-out. I haven't attempted anything quite this large before, nor quite this specialized. The garden cart is beautiful, but it did create more questions than it answered, me thinks ;)

4mem8
06-02-2008, 01:17 PM
Hey Adrenalynn, This project is absolutely gorgeous, i can't wait to see how this pans out. You are going to a lot of trouble here and cost to get this project going. I admire your goals here as this is a huge project. Are you doing ALL your own hardware? I know your doing software, Like me I can only do hardware and a small bit of software, If you can do both from the ground up that's great, But it is also nice to have input as I find you can get side tracked on projects with one train of thought until another person says, well how about this for an idea. Then it snowballs again. Great work.

ooops
06-02-2008, 01:34 PM
Really I need to load up the cart and put a force-meter on the turning scissor and see how much torque I'm really going to require.

Hmmm… A couple of quick thoughts
Have you determined (best guessed) what the all up weight will be loaded before dumping the rubbish? That will go along way toward determining how much force will be needed.
It is easier to turn moving wheels than wheels that are static, and the footprint of the tire will greatly impact that force as well. You can affect the footprint by over inflating the tires (unless they are foam filled). That said, I would recommend doing your turn tests at 10% under inflated, for real world conditions.

I gather you have already acquired your motors and controllers?

This project parallels one of mine pretty closely (your dumping and I am mowing). Both are 36v, both are heavy, and both share dangers inherent with turning something big loose on the world. It will be with great interest that I follow your progress.

Adrenalynn
06-02-2008, 01:53 PM
Hi 4mem,

Thanks! I do have another person helping me, a friend of mine from way back. But yes - I do my own hardware and software. He's my hydro/pneumo expert though.

A huge reason for posting these early to the forum is to get input from people such as yourself and Ooops! More heads are always better!

-------
Ooops,

I really value your input too! You obviously have a lot of "hardware" experience - the term used as not commonly used here. ;)

I had a work-up but now I need to rework it with the new cart. With trash I'll be shooting for approx. 500lbs (overestimate/hyper conservative). I've simulated the forces on the new tires, but I don't have time today to simulate the turning mechanism (which is always open for redesign). Sometimes it's just easier to jam a force-meter on something and measure it than it is to simulate it, you know? ;)

The motors are in the vices right now and the motor controllers are connected to 'em, powered by lawn tractor batteries. They have high-current ammeters inline at the moment so I can build power/current graphs for them when I get time...

With your mower, saying "Ooops" won't be a good thing. :) How comes your project?

ooops
06-02-2008, 03:34 PM
With your mower, saying "Ooops" won't be a good thing. How comes your project?


Ooops is never a good thing … except here (hopefully)
Progress is being made. The big mower has dual 36v motor and axles mounted – but looks funny due to the size of the axles. I will be working on that.
The little ATV mower has a main frame all welded up and is ready to have the motors/wheels mounted.
I need to figure out how to post the pictures and then I will start a thread.

Sienna
06-02-2008, 04:54 PM
Eesh on the skid steering. We had a 200+ lb bot in college that was a 4 wheel drive skid steer. Anytime you commanded a turn on pavement it would jump around like you couldn't believe!

Maybe we were just incompetent (its possible, we were in college after all :P) but I would personally be wary of skidding something that heavy.

Adrenalynn
06-02-2008, 05:24 PM
It depends a lot on the geometry. In fact, look at a Bobcat or other Slipskid loader operate on pavement. I've operated them and they are totally a thing of beauty. I rent 'em every chance I get because they are a *BLAST*. If a project can even marginally be defined as requring a loader, you'll see me at the rental place slapping my plastic down to pick up one of those babies!

The trick is to get close to a square if they're not turning fast enough for you, or elongate the square front-to-back if they're turning too fast, vice-versa gives you faster turns. Sorry - does that make sense?

The garden cart is too long to be an effective slipskid, most likely.

ooops
06-04-2008, 07:20 AM
I rent 'em every chance I get because they are a *BLAST*. If a project can even marginally be defined as requring a loader, you'll see me at the rental place slapping my plastic down to pick up one of those babies!

Nothing puts a smile on my face quicker than playing with heavy equipment:veryhappy:

One alternative to the skid steer approach would be the zero turn - front wheels function like casters. One warning: they don’t work well at all on slopes or grades.

Adrenalynn
06-04-2008, 10:07 AM
Casters also get cranky out in the rough. Even patchy grass doesn't make them happy, let alone sand or a bed of pebbles.

Of course, with slip-skids, those conditions are when they start to get "fun" ... :)

LinuxGuy
06-04-2008, 10:53 AM
Casters also get cranky out in the rough. Even patchy grass doesn't make them happy, let alone sand or a bed of pebbles.
It's really tough for a two wheeled robot outside, because it needs two casters to help keep it from falling forward or backward.

W.A.L.T.E.R. really isn't an outdoor robot, but he does have 5" off road tires and I have allowed for up to 6" tires in the new bottom deck design. I just can't really see any way to do casters that would work well outside and not impair movement in some way. I've purchased a pair of 2" DuBro airplane wheels for this purpose, but I think I would need at least 3" or 4" wheels to be workable outside. Hmmmmm..

8-Dale

Adrenalynn
06-04-2008, 11:12 AM
Balloon casters tend to be a better choice outside than none at all, but in the rough, it's hard to beat slip-skid.

LinuxGuy
06-04-2008, 11:22 AM
Balloon casters tend to be a better choice outside than none at all, but in the rough, it's hard to beat slip-skid.
I'm not sure exactly what "slip skid" is, or if it would work on a two wheeled robot.

8-Dale

Adrenalynn
06-04-2008, 11:24 AM
A slipskid is like a Bobcat (wheeled) or, in a treaded design, a tank. One side wheel-set skids whilst the opposite drives creating the turn.

LinuxGuy
06-04-2008, 11:59 AM
A slipskid is like a Bobcat (wheeled) or, in a treaded design, a tank. One side wheel-set skids whilst the opposite drives creating the turn.
So one wheel would be stopped while the other would roll to make a turn?

8-Dale

Adrenalynn
06-04-2008, 12:14 PM
Yup. Or two wheels, or three wheels or four wheels or... depends upon how many you have on a side.

For really fast turns you reverse one side's wheels while driving the other side's forward. That decreases the traction as well and allows for what's called "super spin". Works best with 4WD or above, or tracked vehicles.

Lynxmotion's 4WD rover platform is an example of slip-skid.

darkback2
06-04-2008, 12:20 PM
I've never understood how someone can program factoring for Ackerman stearing...it just seams like a total mess...but I guess when you actually look at it it may be easier than I think. Any thoughts on using a combination of skid and ackerman? For example, when you want to turn, the left and right wheel in front go in different directions, while the back wheels pivot around a verticle axel?

Also I had some students design a much more basic version of a garbage bot at my old school. It had IR and bumper sensors...but only weight 20 or so pounds. It was based on the ER-1

Adrenalynn
06-04-2008, 12:42 PM
Well, the whole idea behind Akerman Steering is to prevent the inside wheel from scrubbing and decreasing effective traction. [as I understand the concept, but my understanding does come from auto racing] You angle the inside wheel more to compensate for the decreased track or radius.

Given that, Akerman would be, by my definition, incompatible with slip-skid - where the whole point to the concept is to scrub the insides.

The calculation is actually fairly easy. It's just one more triangle built either from the apex or from the center-point of the scribed turning circle.

True Akerman is most effective on preset ovals. At the speeds our bots tend to go with the terribly limited traction they tend to have - I just don't see an advantage.

Now, four wheel steering might be worth considering on larger bots like mine. For the smaller/lighter bots, slip-skid is just terribly effective and trival to implement.

ooops
06-04-2008, 06:16 PM
Casters also get cranky out in the rough. Even patchy grass doesn't make them happy, let alone sand or a bed of pebbles.


I was suggesting something like the zero turn lawn mower … casters on steroids:veryhappy:
http://forums.trossenrobotics.com/gallery/files/1/7/7/8/zero_turn.gif

But I realize it wouldn’t work well for what you are doing. Likely “step” on a few toes in tight turns.
Seems like robbing the steering off of something with a power steering set up would be a possible solution. That said, I haven’t come up with any examples.

Sienna
06-05-2008, 06:32 AM
Well, you could always go with those designs that allow four wheel independent steering!

Then the bot could strafe :P

ooops
06-05-2008, 07:29 AM
Ok, possibly the best solution would be articulating!
Thinking about all the truly heavy moving done in the world, most is done by articulating equipment (with exceptions like forklifts, trains, & tanks). Heck even the tractor trailers on the highway utilize a type of articulation.
I am thinking a setup similar to a frontend loader. If you strategically locate the liner actuator (or two – one on each side of the pivot) you can have a short stroke with “big turn action”. The short stroke should equate to quick response, and the “big turn action” is always good. Now I haven’t given any time to considering the force required, and or the cost of the actuators. But it gives you another option to explore.

LinuxGuy
06-05-2008, 08:12 AM
Well, you could always go with those designs that allow four wheel independent steering!

Then the bot could strafe :P
In my hybrid robot design, I have four wheel independent, and fully retractable, wheels. :veryhappy: This is going to be a very expensive robot to build though.

8-Dale

Adrenalynn
06-05-2008, 11:49 AM
Articulation is cool, but it does present some concerns when it comes time to program it. You open up a whole lot of angles which makes interpreting your sensor data a nightmare.

Besides, if you park it on a couple small kids you lower your static coefficient of friction vs that of concrete, pavement, or even grass.

ooops
06-05-2008, 12:23 PM
Strap a couple kids to it and that solves the steering issue, but it will probably head straight to the arcade everytime!

Adrenalynn
06-05-2008, 12:25 PM
If you staple them to the front, you have an organic European siren:

"Momeee Momeee Momeee Waaaaaah WAAAAAAAAAAAAH!"

LinuxGuy
06-06-2008, 10:56 AM
Articulation is cool, but it does present some concerns when it comes time to program it. You open up a whole lot of angles which makes interpreting your sensor data a nightmare.
Hmmmm, this sounds like a challenge.. Muahahahaha..


Besides, if you park it on a couple small kids you lower your static coefficient of friction vs that of concrete, pavement, or even grass.
Wouldn't the small kids have a problem with this?? ;)

8-Dale

Adrenalynn
06-06-2008, 11:58 AM
>> Wouldn't the small kids have a problem with this??

Not for very long... :robotindifferent:

ooops
06-06-2008, 01:10 PM
Besides, if you park it on a couple small kids you lower your static coefficient of friction vs that of concrete, pavement, or even grass.

Seems like the squish angle would be hard to calculate.

Adrenalynn
06-06-2008, 01:27 PM
Depends upon the weight of the 'bot, now doesn't it? At some given weight it's really just drivetrain lubricant.

DresnerRobotics
07-09-2008, 11:37 PM
I demand updated pics!

Adrenalynn
07-10-2008, 02:18 AM
Thanks for digging this back up...

OSCAR has an acronym now - Oversized Scrap Can Autonomously Roving.

Pics and video next week!

Electricity
07-10-2008, 12:13 PM
No, pictures and video NOW!




:p

archcvd
07-15-2008, 06:12 PM
I am most excited to see how you will implement your localization with OSCAR. Eagerly awaiting news!

Adrenalynn
07-15-2008, 06:14 PM
I'm actually starting with a much smaller platform for that. OSCAR is just too darned large to let him safely roam around.

A-Bot
07-16-2008, 11:08 AM
OSCAR has an acronym now - Oversized Scrap Can Autonomously Roving.


Reverse-engineered acronyms are cool. :veryhappy:

Adrenalynn
07-16-2008, 11:11 AM
I alternate between "Oversized" and "Obnoxious" since the animatronics will be grouchy...

And originally it was "Sh**-Can" - but that's just not as family friendly, ya know?

I spent hours reverse-engineering that acronym, thank you. :p I could have reversed a nuclear reactor faster. ;)

DresnerRobotics
07-16-2008, 11:14 AM
"Pics or it didn't happen"

Matt
07-16-2008, 12:02 PM
http://cyborg.namedecoder.com/

O.S.C.A.R.: Operational Synthetic Construct Assembled for Repair

Adrenalynn
07-16-2008, 12:15 PM
or "Rubbish" - I'm going to go visit that site though - could be a real timesaver in the future! :D

Adrenalynn
07-16-2008, 12:19 PM
BWAHAHAHAHAHA!

"A.D.R.E.N.A.L.Y.N.N"

"Artificial Digital Replicant Engineered for Nocturnal Assassination, Logical Yelling, and Necessary Nullification"

AWESOME!!!

JonHylands
07-16-2008, 12:28 PM
Got you beat:

"J.O.N.H.Y.L.A.N.D.S."

Journeying Operational Neohuman Hardwired for Yearly Learning, Accurate Nullification and Dangerous Sabotage

DresnerRobotics
07-16-2008, 12:36 PM
T.Y.B.E.R.I.U.S.: Transforming Ytterbium Being Engineered for Repair, Infiltration and Ultimate Sabotage

Mine isn't nearly as good as the previous two :(

ScuD
07-16-2008, 02:28 PM
S.C.U.D. : Synthetic Construct Used for Destruction.

That's actually pretty accurate... :D

archcvd
07-16-2008, 04:22 PM
A.R.C.H.C.V.D.: Artificial Robotic Construct Hardwired for Ceaseless Violence and Destruction

Hahaha

Edit: Sorry for the hijack, back on topic!

Who are you working with on this, Adrenalynn? (you mentioned something about there being some help)

JonHylands
07-16-2008, 04:27 PM
A.R.C.H.C.V.D.: Artificial Robotic Construct Hardwired for Ceaseless Violence and Destruction

Hahaha

You win. Everyone can stop playing now...

ooops
07-16-2008, 05:13 PM
But I didn't get a turn.
O.O.O.P.S.: Organism Optimized for Observation and Potential Sabotage

Adrenalynn
07-16-2008, 05:31 PM
Oh man, I had to read "Organism" like five times before the word morphed into what it actually was. :o

I blame my new monitor. Too contrasty...

LinuxGuy
07-17-2008, 04:19 AM
W.A.L.T.E.R.: Wireless Artificial Lifeform Trained for Efficient Repair

L.I.N.U.X.G.U.Y.: Lifeform Intended for Nocturnal Utility, Xpert Gratification and Ultimate Yelling

Ha ha ha Muhahahahahahahaahah!

8-Dale

Alex
07-17-2008, 10:13 AM
T.R.B.O.T.H.E.A.D. (my name on pretty much every other place on the net): Transforming Robotic Being Optimized for Troubleshooting, Hazardous Exploration and Accurate Destruction

edit:


You win. Everyone can stop playing now...just saw this now. Good point, enough hijacking (Matt! :mad:).

:p

ooops
07-18-2008, 07:43 AM
Oh man, I had to read "Organism" like five times before the word morphed into what it actually was. :o

I blame my new monitor. Too contrasty...


You should probably get that monitor out of the gutter before it rains!!!

Adrenalynn
07-18-2008, 10:34 AM
Good idea, thanks! It's a pretty nice monitor. Makes everyone seem much brighter.

ooops
07-18-2008, 11:24 AM
Makes everyone seem much brighter.

I have some goggles made by Adolph Coors that do that for me.

Adrenalynn
07-18-2008, 11:38 AM
I thought they made people more attractive?

If they work for brightness, I may try them on my television. It doesn't matter how much I crank up the brightness, those shows just don't get any smarter.

ooops
07-18-2008, 11:58 AM
Now that you mention it, maybe they aren't brighter ... but when the goggles are "properly adjusted" I just care less. I am pretty sure the same mechanics apply to the "more attractive as well".

Electricity
07-22-2008, 12:34 PM
E.L.E.C.T.R.I.: Electronic Lifeform Engineered for Ceaseless Troubleshooting and Rational Infiltration
Wouldn't take the full name.
So Jodi, any pictures yet!?

A-Bot
07-22-2008, 09:24 PM
Oh man, I had to read "Organism" like five times before the word morphed into what it actually was. :o

I blame my new monitor. Too contrasty...

I can't believe no one went to the Sexy Name Finder. :rolleyes:

A-BOT: Adonis Bestowing Orgasms and Touches

:veryhappy:

Hephaistos
08-19-2008, 07:58 AM
This looks like a very cool project. Any updates? How are you planning on dividing up the workload for the various on-board PC's and microcontrollers?

Adrenalynn
08-19-2008, 11:07 AM
Hi Hephaistos, welcome!

I've had to switch to a smaller platform for testing. Several hundred pounds isn't the safest thing to play with SLAM invention. ;)

Hephaistos
08-19-2008, 12:18 PM
Well, I suppose that depends on the type of SLAMming you're referring to. : )

So on your new bot, are you using the same computational platform or have you gone with something different? I'm curious as to how you went about dividing up the processing tasks across multiple systems, if that's what you're doing. My bot has a similar layout to the one you proposed (dual pico's, several Atmega168's, and a dual core mini-itx).

Hephaistos
08-19-2008, 12:18 PM
Hi Hephaistos, welcome!

And thank you very much for the welcome : )

Adrenalynn
08-19-2008, 12:58 PM
Wow - big bot processing there! Any photos and description you can share?

My smaller bot has a mini-ITX for video processing, ranging and detection, a mini-ITX for SLAM, and acquisition and preprocessing are done with an ATMEGA640. Shortly I'll be adding a Serializer to it, not sure why yet. ;)

4mem8
08-19-2008, 01:02 PM
Hephaistos: Welcome to this forum, Man that's serious processing power for your bot, As Adrenalynn has stated, pics would be awesome please. Hope you enjoy it here, great place to be.

Hephaistos
08-19-2008, 01:28 PM
Wow - big bot processing there! Any photos and description you can share?
I'll look at setting up a project page to post some pics and descriptions to. I don't want to go and hijack your thread.


My smaller bot has a mini-ITX for video processing, ranging and detection, a mini-ITX for SLAM, and acquisition and preprocessing are done with an ATMEGA640. Shortly I'll be adding a Serializer to it, not sure why yet. ;)

The Atmega640 is an impressive chip. Are you using the Axon (http://www.societyofrobots.com/axon/) perhaps? I thought about working with the 640, but I want to build some of my own PCB's and I figured I'd try a TQFP-32 rather than a 100 in the electric skillet (http://www.sparkfun.com/commerce/tutorial_info.php?tutorials_id=59) first... : )

Hephaistos
08-19-2008, 01:32 PM
Hephaistos: Welcome to this forum, Man that's serious processing power for your bot, As Adrenalynn has stated, pics would be awesome please. Hope you enjoy it here, great place to be.

Thanks for the welcome. I've been lurking for a few weeks now and have been very impressed with everyone and their projects. I'll post some pics in the coming days. A lot of the processing power is being used for image processing with OpenCV, but a good chunk goes to SQL Server (storage of "common sense" data, facial images, etc) and some of the higher level "cognitive" processes as well.

4mem8
08-19-2008, 01:40 PM
I look forward to seeing more on this project, Thanks for sharing it with us.

Adrenalynn
08-19-2008, 01:53 PM
Guilty as charged! Yup - I'm employing the Axon. I think it's a great board for the features-to-money. I have a couple Arduinos and a couple tubes full of loose '168's. When I have little stuff, I proto on the arduino then whip-up a light-weight 168 board from the arduino proto. But for getting real work done, I think it's hard to beat the Axon - and I'm way too lazy to lay out a board for the 640 for what the Axon costs. I couldn't save more than a few dollars and it's just not worth it with such a great resource there. I'll end-up picking up a few more, I'm sure. ;)

4mem8
08-19-2008, 02:00 PM
Adrenalynn: Who is the main supplier of the Axon board? Link if you have one please.

Adrenalynn
08-19-2008, 02:07 PM
http://www.societyofrobots.com/axon/

TR is supposed to start carrying them "soon" - in the meanwhile, John's a good guy to work with.

Hephaistos
08-19-2008, 03:10 PM
Guilty as charged! Yup - I'm employing the Axon. I think it's a great board for the features-to-money. I have a couple Arduinos and a couple tubes full of loose '168's. When I have little stuff, I proto on the arduino then whip-up a light-weight 168 board from the arduino proto. But for getting real work done, I think it's hard to beat the Axon - and I'm way too lazy to lay out a board for the 640 for what the Axon costs. I couldn't save more than a few dollars and it's just not worth it with such a great resource there. I'll end-up picking up a few more, I'm sure. ;)

Yes, the Axon is a great board for the money. I need more than 16 ADC inputs though, so I hooked up 4 Atmega168's via I2C and then hooked the master up to the PC via USB. The 168's are so cheap, it's hard to beat them for quick and dirty port expansions.

Adrenalynn
08-19-2008, 03:31 PM
Don't you chew up two of the five with SCL/SDA anyway? And can you live with that little memory?

Hephaistos
08-19-2008, 03:41 PM
Don't you chew up two of the five with SCL/SDA anyway? And can you live with that little memory?
Yes, I do lose two of the ADC lines with the I2C, but the TQFP version has 8 ADC as opposed to the 6 on the DIP version. That said, my plan is to move to SPI and regain the two lost ADC's with I2C, but I haven't done that yet.

Regarding memory, it hasn't been an issue thus far because I mainly use the 168's as a cheap capture device for the PC and let the PC do all the heavy lifting.