Re: IFI Critique
 

Actually, they make most of it selling server racks, and you don't see them putting server racking in the kit to boost profit margins. (Mind you, server racks might be very usable construction materials) Anyway, let's try and get back on topic here, I think sparksandtabs wanted a thread about how to improve the current IFI system, not the company and its relationship with FIRST. A more accurate thread title might have been IFI Control System Critique.

Re: IFI Critique
 

Mind you... nothing has improved GM more than Toyota!

Re: IFI Critique
 

This is true, but I'd argue that in the market of FRC robot parts, you're all but required to provide the most bang for the buck in order to keep getting business. According to Brandon's statistics, there were just over 7,000 active users on CD. (For comparison, FIRST's 2005 annual report indicates that about 25,000 students participated in FRC in 2005. The two numbers are borderline apples-and-oranges, but you get the point.)

Now, let's suppose that Billfred Industries starts selling two-speed gearboxes, trying to go head-to-head with AndyMark*. A lot of those customers are either going to be on CD, be part of some local FIRST scene, or know someone who is. If these gearboxes are awesome, business will pick up next build season. But if the things just don't work, that same network will spread that news just as quickly. (How many AM Shifters did you see on robots in 2005? How many did you see in 2006?)

From what I've seen, there's been two ways to get a product out there in FIRST: team demand leading to it going in the KOP (think kitbot, CMUcam, the lighted target for the CMUcam), or word of mouth among teams about it (think roughtop, wedgetop, AM Shifters). A brand new control system can't really be done through the latter route, and convincing other teams that this new system you've developed is worth FIRST making the switch from IFI a bit of an uphill struggle. It'd take a heck of a salesman to get enough teams calling for the switch for it to really show up on the radar, but I believe it can be done if the product is quality. We are talking about the same folks who can scout a whole regional and collect enough data to make my head explode, after all.

*Now really, who would want to try and go against Andy Baker? That's like trying to single-handedly outscore 25, 254, and your choice of triplet with a shopping cart.

Re: IFI Critique
 

I jumped into this thread with a discussion about why we don't have an easy way to do a gamepad type controller on the OI. The answer was not customer focused but IFI support focused. Well, actually, first it was implied that it was illegal to do so and when a perfectly legal work around was discussed, it was claimed that it would make IFI answer more phone calls than they'd like. They have not responded to using a USB host chip such as EZ-host.

I think the two are inseperable. The relationship between IFI and FIRST one of the reasons things are the way they are.

Joe J.

Re: IFI Critique
 

Hmmm.....I wonder when the AM control system will be shipped out ;)

But seriously, the quatily and cost of the control system to me is outrageous. I have seen in many occasions where somebody has used a basic stamp to control something far more difficult that what we do in first. Go to the parallax website and look at the "Nuts and Volts" articles they have. They are hooking their BASIC stamps up to EVERYTHING you can imagine. Granite the programming languange is really easy to understand and it gives rookies an advangtages, it still held back more expirienced teams from acceling. I have seen manythings done in BASIC as we do in C. ALso that control system if bought direct from parallax is far cheaper. 2 BOEs and 2 transcievers with a few ADC chips and external circuittry and voila you spent only 300$ if that.

Not to bash anybody's control system but I have a development board sitting 3 feet from me that cost 80$ that can do EVERYTHING that the IFI system can do. Sure it may take a little more coding but it is cheaper and more raw (more room for development and expansion by individual teams) that the IFI and parallax solution.

<bashing>How hard could it possibly be to design a new control system?</bashing>

of course all IMHO

Re: IFI Critique
 

I issue the same challenge to you that I issued in my other post - if you think it can be done better and cheaper, then do it and sell it to FIRST. The BOM price for a IFI RC is likely less than $75, including the PCB, but engineering (including the base software dev - both master and user in thier configuration) and support time are not free. Basic Stamps are overpriced and under powered - and would be a move backwards - as they certainly don't provide the level of refinement necessary for FIRST. What is it that the stamp can do that the PIC can't? I don't see anything in your argument of your dev board being more capable - there are plenty of systems that cost less which are more capable, but they aren't as refined and don't provide the output control. You seem to be arguing two different things - going simpler in one sense (BASIC) and more complicated in the other sense (here's a dev board with nothing done for you). The IFI platform provides a mix of these - and with easyC, the simple side is covered pretty well.

Remember, in the end, the system doesn't have to be used by just you - it must be used by 1000+ other teams, and must provide a safety factor that others who aren't on a team know they won't be hurt by another team's misprogramming. None of those dev boards can provide that without some level of custom IO protection. Plus who wants to slap 8 pcb's on their robot - your 6 pieces of parallax gear does nothing about packaging.

I'm not saying it can't be done better and cheaper - I'm just saying that I'm not doing it with my time, and I don't see you doing it. If you think it is such an outrage, then be a do-er and not a talk-er. Even if you don't have the resources to do it yourself, there is nothing stopping you from convincing FIRST to accept a new system and getting together the resources to create it (even if by a sponsor, etc).

Re: IFI Critique
 

I am not neccessarily saying that the the Stamps are better, but I am saying that from what I have seen they can do basically everything we need to do. Maybe not so much in the upcoming years but right now from my expirience I would say they would run fine.

My dev. board isn't a stamp it is an atmel AVR. Also again I am not saying to give the teams a dev board and expect them to do anything. But I mean how hard would it be to take a socket for a 40 PDIP and create the neccessary ports for it on a PCB. I am not saying it needs no engineering but I also dont think it needs a whole heck of alot to get the job done. (then again what do I know I am only in highschool, feel free to bash me for this comment.)

As for your challenge. Sure. what the heck it is summer and I have nothing better to do. Just to set some rules for this new system I will try to develop what do you want. How about everybody who want to helps or has an idea as to what I could include in this system to make it something that they think a majority of FIRST teams could use or want in a system to post here.

I will work on a simple one that does everything that the IFI system does right now and make periodic updates on everything.

<edit> Due to me being poor right now. and not having the money to prototype something like this it will have to wait for a little bit. I do not currently have the mony to put into the right kind of socket to mount the chip I was looking at to a PCB let alone my dev board.

Though I will still do conceptual planning and post it on here</edit>

Re: IFI Critique
 

STAMP2 is not a step in the right direction to be sure. The problems we had with STAMP2s as a control mechanism were many and various -- complex & unweldy programming slots, unsigned 16 bit math, no interrupts, poorly implemented UART, and so on.

It was basically easy to do easy stuff using PBASIC, but anything complex took a guru to implement.

I think there are better micros/languages to use.

Joe J.

Re: IFI Critique
 

Once again, I am not saying that using the BS2 is a step in the right direction. I never said anywhere that I wanted to use it. I simply said that I have seen it used in a setting that needs more complex coding than what we do. So it is possible to use it. But I don't necessarily think that using a microchip processor is the best thing either.

Re: IFI Critique
 

I see from your profile that your rookie season was 2004. You realize that prior to 2004 the robot controllers used a BS2, right? Many, many people complained about the lack of power and other features (hence the move to the PIC-based RCs by IFI). And yes, there were many teams that completely maxed out the performance of the BS2. Teams were most definitely being held back by the BS2-based RCs.

Re: IFI Critique
 

I assume you're speaking of the STK500 board, in which case, you are flat out wrong. That board cannot do anything the IFI system does save for program a Flash based microcontroller. It is a very fancy, albeit useful multi-AVR programmer with some builtin debug features, and a builtin array of LEDs and buttons for testing. The obvious features such as PWM generation, I/O and ADC hardware require a chip with far more pins than a 40 pin PDIP AVR could ever supply, not to mention it doesn't support the ease of connection ability or any number of little things, such as pulldown resistors, multiple ADC inputs (16 last I checked), relay outputs, multiple serial ports (the DB-9s on the board must be jumpered by ribbons to the proper pins on the AVR, and even then only one of them is directly accessible, the other talks to the STK500s processor array.) You can't fully emulate the IFI control system with any combination of chips on that board, not in full functionality with all the extras you probably don't even realize you have and take for granted. There's also the cost of the development IFI put in to the pre-existing controller libraries and default code, which gives you much of that functionality you arbitrarily claimed the STK500 could do.

In short, there's just no way that $80 development board could emulate the IFI system properly, and have all the pre-existing software support and safety features, plus it definitely does not have a built-on 900MHz radio system.

Before you go bashing IFI and making wild claims, make sure they have some foundation.

Re: IFI Critique
 

While I do agree that the PIC processor has to go, I know that is not going to happen soon. But my biggest gripe with the RC is the use of PWM cables. They are way to fragile, and the connectors suck.
What IFI should use is standard CAT 3 cable with RJ-11 connectors.

Re: IFI Critique
 

I am referring to the stk500. But I also have the stk503 that can run 100 pin chips. Also doesn't the IFI system have 2 processors? So if I were to actually try to make a new system I could use 2 procesors, right? So I could run a ATMEGA128 or ATMEGA2560 and with one chip I could emulate most of the features on the IFI system. Also I would outsource somethings, such as PWM because I think it would make the code run better, though using a USART you would only need a set of 3 charaters to set the pwm. It would require less timers and less general code to run them. Also I would want to use the IFI radios just because I think it would make it easier then finding a whole new wireless communication system.

I don't think I could do this myself, that is why I asked a couple of electrical and software engineers to help me work things out with this. I don't think it will be completely easy but I do think it is possible.

Re: IFI Critique
 

A single atmega 128 won't be able to re-create what they do with 2 chips - mainly due to the lack of UART's (Atmel doesn't seem to understand the synchronous thing) - we use 3 of the 4 as far as i can tell on the microchips. (Radio on master, program and TTL on user). Especially if you use one of the UART's to "outsource" PWM generation, you would have to find a way to share the program and radio ports, unless you forced everyone to use ISP. I think you would be better off using I2C or SPI for the secondary PWM generation.

I did have the pleasure of using Brian Dean's MAVRIC (www.bdmicro.com) on the real time scoring system I developed for Battlecry, and it was a nice change from the Microchip's - but I don't think the Atmega128 provides significantly more than the high-end 8-bit PIC in the RC. (And yes it was overkill for the RTS system, but we had a board laying around).

Re: IFI Critique
 

That is the only thing I haven't been able to work out yet. I have only found 1 real user friendly PWM board.I am sure there are more but I haven't found them yet. I know for FACT that one chip wouldn't beable to do everything. It only has 8 ADCs, 2 USARTs, and a few other things such as IO and such. Now on the other hand the atmega2560 has 4 USARTs