Re: IFI Critique
 

So far everyone has focused on the OI, joysticks, and RC. I'd like to make a comment on the KOP frame and drive train. In my opinion the KOP stuff is perfectly usable and helps rookie and low resource teams get something moving. The one qualification is that it is not acceptable in the 2 wheel drive configuration. To be usable it needs to be 4 wheel drive. This is where I believe some changes could be made. Starting with the 28 tooth sprocket. The sprocket is custom made and heavy. The wheel chair sprocket adapters while expensive are well made. For a team to get 4 wheel drive They have to buy 2 28 tooth sprockets, 2 sets of adapters, some chain, master links and 2 trans sprockets. The 2 items that really add to the upgrade cost are the 28 tooth sprockets and adapters. 23.00$ for 2 pieces of plastic? I think somethings wrong here. 19$ for a custom steel sprocket is OK but they are heavy. IFI all ready has Aluminum 28 tooth sprockets for their wheels. For 2007 maybe IFI could drop the steel sprockets and use the aluminum ones with a redesigned wheel chair wheel adapter. Just seams that 2 pieces of plastic should be less than 10$ This would cut some weight out of the system and make IFI's stocking easier and reduce the cost to a team.

Re: IFI Critique
 

My only request is that they change the program port over to USB... and possibly make the whole RC smaller.. other then that I think its pretty good

Re: IFI Critique
 

You know what would really speed things up: a dsPIC or 2 for the OI and RC.
up to 40 MIPS
12 bit A/D(up to 32 Channels)
32 bit timers
Hardware Multipliers
Who knows with one of these, a little inginuity, and new rules, USB joysticks might be a reality.

Re: IFI Critique
 

The cost of the micros is almost immaterial to the cost of the RCs & OIs & radios -- I can tell you that the BOM cost is not what is driving these required items to list for $1100. Even if BOM costs were the driving force behind the price, you would be surprised at the features that would be possible with an extra few tens of dollars added to the BOM.

But, because IFI is the sole FIRST electronics supplier, they are dictators of what micro we get as well as basically everything else with regard to FIRST electronics.

There are micros & peripherals that are much more friendly toward robotic features and functions. This is especially true with respect to rapid code development.

I have been thinking a lot recently about the IFI monopoly. A number of questions keep coming to my mind. This thread seems as good of a place as any to start collecting them.

• Do we have an I2C or CAN or SPI or other serial interface to easily and cheaply implement sensors, etc.?
• Does the RC have any sensors baked in or easily added (e.g. 1D, 2D or 3D gyros and 1D, 2D, or 3D accelerometers)?
• Do we have a robot friendly micro (small example: shaft encoders have to be done with interrupts rather than with peripherals dedicated processing quadrature input in silicon)?
• Do we have a method to expand memory?
• Do we have an IDE with that helps autocode some of the more tedious aspects of setting up peripherals, interrupts, etc.?
• Do we have a processor that is well supported by autocoding tools?
• Do we have state of the art radios?
• Do we have radios that enable wireless operation in the pits?
• Do we have access to non-Windows compiler/loader?
• Do we have an easy way of implementing a gamepad controller on the OI?
• Does the default code encourage and support good programming and robotic practices (e.g. feedback control)?
• Do we have a software library for doing robot type tasks quickly and cleanly?
• Do we have an RC that can interface to standard hobbyist RC sensors and motor drivers?
• Do we have support for brushless motors? stepper motors?
• Do we have a reasonably priced alternative to driving lower current motors?

I have more, but I am tired of asking questions that have NO as their answer. So I will wrap up with just 2 more:

Would all this be expensive? Yeah, I am sure it would be.

Would it be a bargain given that 1300+ teams have to deal with the system? I am sure of that too.

Joe J.

Re: IFI Critique
 

We can get 2 gumsticks with 200Mhz Intel Xscale processors and radios for far less than $1100...

I can buy the 18F8772 for less that $10 a piece. And this is probably the most expensive component in the IO and RC... IFI buys these in bulk so they get them for far less.

They should at least switch to dsPIC's, which are the same price but have a massive performance boost.

Re: IFI Critique
 

Like Joe said, the BOM cost isn't the big cost in the IFI RC and OI. The major cost is the engineering, manufacturing, and support time for it. Think about the low quantity - if they are lucky, they might sell 2000 of them, including the ones which come in the kit of parts. No one likes doing low quantity manufacturing runs, as setup costs are enormous. And the guy answering the phone or the email when you have a problem with the RC likely isn't volunteering his time. Plus, I'm very doubtful that FIRST pays full price for the ones in the kit of parts. The only alternative I could see would be to develop a primary system, which is a very capable robot controller without all the FIRST specifics, and then give a FIRST addon pack, assuming there was a market for the original system to justify the additional work. They are likely still trying to earn back the cost of the engineering time that went into the original PIC solution, so I can't see us getting a total revamp (like Parallax to Microchip) anytime soon.

Look at the vex controller and the mini-rc - they are essentially the same product minus a TTL-RS232 converter and a Darlington pair array to drive solenoids and the differences as far as how they take radio input, yet due to the production quantity differences they are priced drastically different.

I think we should be grateful that IFI puts in as many resources as it does, for a product that they likely make no money on. And unlike some other FIRST contractors, they have shown a clear commitment to a fully functioning product, even if it meant putting in extra hours. If you think its so wrong that they are the sole supplier, develop your own system which is as refined as theirs (must have field control, IO protection (no outputs when disabled)) and demo it to FIRST, maybe they will choose you as the supplier...but I don't see anyone dropping the resources to do this.

Re: IFI Critique
 

for the time being you can either use hot glue to keep them in (works great), or buy one of these to fix your problem. I have no quick fix for metal shavings other than to put a blanket or sheet (or something similar) over your controls before you drill or do anything on your robot that would cause metal shavings to fall, or using one of these in and around the controls normally does the trick....

Re: IFI Critique
 

BING!

Of course, since it seems that so many folks do great things with the IFI system as it is, I have to wonder how long it'd take for someone to develop an IFI-compatible means of doing (or streamlining) some of the things we want out of the RC. I mean, if teams can rig up Gumstix to work or dump data from a Pocket PC into the RC to define an exact path for the robot, how long until we see a business-major-friendly way to make encoders work with the RC without dealing with the wonderful world of interrupts?

In fact, I'm willing to bet that someone could make a nice pile of beans doing it, while also scratching off some of Dr. Joe's wish list at the same time. (Although weren't the accelerometers and gyros this year just a matter of hacking off a PWM cable and soldering three connections?)

The way I see it, I'll be happy working with an IFI RC that gets the job done--controls the robot safely and reliably, lets us move in autonomous, handles some custom programming and trickery. If others want to get fancier with encoders and uber-enhanced PID loops and such, let's make it so that they can do it without too many four-letter words and without having to complicate (or, more importantly for a lot of teams, raise the price of) the RC.

Re: IFI Critique
 

I just took a regular PWM cable and put the pins right into the PCB and soldered it. Didn't even have to cut it! We used the gyro this year and it worked great.

-Chris

Re: IFI Critique
 

1. Stronger DB-9 connections. Too many teams, including ours, have ripped them out of the board. It's a quality rather than convienience request. Replace the rivets with small threaded hardware and the problem would be pretty much fixed. However, I do realize how standard the riveted connections are, and that an actual replacement implementation might be difficult.

2. Would it be possible for IFI to aqcuire the molds, rights, etc. to produce the now discontinued Flightstick? And would it even be profitable, considering also producing a commercially available USB option? Or, does IFI have enough connections that an enterprising team approaching them with blueprints and models could have them produced at reasonable cost? Unfortunately, I doubt that kind of short run production could be done for a reasonable price.

Re: IFI Critique
 

The DB9s are designed as breakaways. Consider what it would do to an RC if they were attached firmly to the PCB, and an errant robot tried to run away with its tether. Right now it's maybe a $100 job to fix it, accounting for time, parts and shipping to and from IFI. If it were screwed in, you'd have to replace the entire PCB, and maybe the RC itself.

Re: IFI Critique
 

I am not ungrateful to IFI for the system we have, in fact, I have defended them publicly on many many occasions. I was happy when FIRST chose their system.

But... ...time marches on. Bob and Tony are not infallible and they are certainly not making "no money" on unless you think numbers with 7 figures between the dollar sign and the decimal place constitute no money.

Even so, they are a for profit company. It is their charter to make as much money from FIRST and FIRST related sales as they can. It is not IFI's job to make sure that FIRST teams have the best products and the lowest prices. That job falls to Manchester.

I will say it flat out and in plain English: I think that FIRST HQ has not done their job well in recent years with respect FIRST electronics generally and managing the relationship with IFI in particular.

Joe J.

Re: IFI Critique
 

Actually, one of your questions have yes as their answer. The OI could in theory control stepper and brushless motors. It's just a matter of creating a new controller. I honestly don't see the point inside of FIRST but it would be useful to someone outside of FIRST.

Re: IFI Critique
 

Stepper motors would be great for some applications -- position control with no feedback required.

As to brushless motors, look at the power ratings of the so called coreless servos from Futaba, Hitec, et al. Those are really brushless DC motors. Brushless DC motors enable significant power density increases (more power in less weight & size - sign me up). And brushless DC motors are getting cheaper all the time.

Just as an exercise for the reader, search TowerHobby.com for "Brushless"

Why can't those motors go on FIRST robot? Well, because there is no way to legally power them on a FIRST robot.

And don't think that you can use any of those brushless speed controllers on TowerHobby.com either. Actually, you can't use ANY of the speed controllers on TowerHobby.com because IFI is the sole source for electronic speed controllers and that is that.

Even if the FIRST changed the rules and made these controllers legal, we STILL couldn't use them because IFI has bastardized their PWM outputs to make them incompatible with standard hobbyist electronics. IFI has no incentive to make them compatible either. In fact, they have strong financial reasons to make them INCOMPATIBLE, otherwise, some day someone may convince one of these speed controller makers to donate a few thousand or so to the FIRST KOP. If that happened, it could jeopardize IFI's Million Dollar Victor Monopoly. So, IFI is better off with incompatible outputs*.

Call 'em as I see 'em.

Joe J.

*Even as it is, IFI is in danger being caught with their monopolistic slip showing. Wouldn't Bob and Tony look like greedy so-and-sos if a speed controller manufacturer went to the bother of designing a special IFI compatible speed controller and offered to donated them to the FIRST Kit of Parts only to have IFI force FIRST to refuse the donation? Making 1000's of FIRST teams pay more so they can keep up their profits? They couldn't be THAT heartless... ...could they?

Re: IFI Critique
 

Heres a hint for you all (real world here): :ahh:

IFI is not in business to supply FIRST with OIs, RCs and Victors. IFI is in business to make money (just like every other company out there).

They make their money selling OIs and such, but that isn't their goal. If it was they would probably be out of business by now.

When a company (in this case FIRST) single sources and goes with a proprietary design (as in IFIs), it usually gets an initial price break. However, the company must be ever vigilant against that single source from then using its propriatery design to increase prices higher than the standard market.

It would be in FIRSTs best interest to reevaluate its position with such propriatery design companies every year to see if it would be better to find a more standarized version...or at least have some bargaining chips with the propriatery vendor.