Re: IFI Critique
 

Ok. So you don't have to load drivers.

You still have to make a USB host device and support all the associated specifications.

Also, how do you map the input and output on an HID device to IFI's axis/buttons? Looking at http://www.usb.org/developers/devclass_docs/Hut1_12.pdf, there's a huge number of possibilities.

If IFI did implement this, but just for "joystick", "gamepad", and "multi-axis controller". Teams would then complain of the inability to use mice, keyboards/keypads, and tablet inputs. (Imagine, a touch pad which has a diagram of the field on it. Touch it, and the robot goes to the spot you hit.)

The options are:

1. Support limited devices. See teams' disappointment when they find that ____ doesn't work with the OI. Recieve more complaints
2. Support all devices. Increase complexity of the system exponentially. Increase time to develop. Increase cost of production. Increase chances of something going wrong.

I think that allowing adaptors for non-gameport controllers (PS, PS2, Sega Dreamcast, N64, etc.) is much more probable. The ability to implement this is dependent on primarily one thing: the power source for such a device. IFI may not need to change anything; FIRST just needs to change its rules.

BTW, that document is not the HID specification. It describes some of the specifics of the specification.

Re: IFI Critique
 

Well, it was primarily based on the assumption that no specification existed. In which case, the time it would take for such a spec to become widespread would likely exceed the 4 years it takes most people to get through high school.

In light of the other post, if IFI were to begin today on USB HID support, I would say that it would take them about 2 years to develop basic HID support, assuming that the current processor(s) and libraries could handle the extra load.

If they used another device and/or chip for actual USB "hosting", they may be able to develop basic support before January.

If it turns out that they have to radically change the processor, the bandwidth, or anything else in the control system to support it, I would say that the 2 years is a minimum. 3 or 4 is more likely, IMHO.

That is all based on what came in last year's kit and my knowledge of it. It is possible that the PIC18F8722 (or even the whole PIC18 line) is not capable of handling USB, or only very poorly. (This is all on the USB host end, not the device end.)

If they were to fully support the USB HID spec in its entirety, I would not be surprised to see a change of the control system comparable to the 2004 switch to PICs.

Disclaimer: I am only an 18-year-old software guy. I have no experience in hardware like this, so the numbers above are, at best, guesses. Use this information at your own risk.

Re: IFI Critique
 

I think by all means that we should stick to analog inputs and not support usb. They provide us with usable joysticks every year, so no team will be left out, but if your needs are more then that, then start figuring out how the device you want to use works, and how to interface it to your OI. Is if really going to help kids learn anything if all they have to do is plug in a new joystick to expand their capabilities?
If FIRST relaxed the rules about what you have supplying the analog signal into the OI somewhat, I know for sure I could do much more with the analog inputs then with a USB port.

Re: IFI Critique
 

I know many who'd disagree... Anyway, the thing I like about analog sticks is they're so simple, a team can quite realistically and easily build their own interface devices. A connector, some wire, some stuff from Radioshack, a soldering iron, that's all you need. I don't know much about making USB devices, but it seems like it'd be hard for them to be simpler.

Re: IFI Critique
 

Like with joysticks, USB is harder to implement. Also keep in mind that the program port serves as a second serial port (in addition to the TTL serial port) besides being used for program transfer. The programs don't get that big, and it wouldn't be worth the better speed for the problems it could cause. A better solution to having no serial port on a laptop would be a converter.

Hm, could next year's game object be a USB cable?

Re: IFI Critique
 

Implementing a USB programming port is one thing, implementing a USB port with a universal joystick software driver is a big deal. A very basic driver stripped down may fit on a pic, but you may not like the feel and response of a modern PC joystick with out the bells and whistle driver. CH makes a line of industrial joysticks. They are available with 100k pots. I have no idea of the cost. I've used a portable indoor crane that used them for the control and was very impressed with the feel and response.

Re: IFI Critique
 

USB Host isn't exactly a walk in the park, but its certainly not outside the reach of FIRST within the next 3 years. I bet that there are several teams, who have mentors with the knowledge and students with the drive, that if they were to take a season off FIRST and put in as much effort and resources into developing this system as they do into building a robot, would be able to finish it in 6 weeks. It is a tough task - but a: I don't see any reason why they should be limited to an 8-bit PIC, there are hundreds of other embedded processors out there (the only thing going between the OI and the RC is serial data, it doesn't matter what source it comes from), and b: theres enough knowledge out there to get it done. The RC itself is relativity simple as far as embedded systems go - its just two micro-controllers with their IO brought out to headers, with a couple serial busses brought out to the right ports and some buffers to enable and disable the relay outputs. Thats overly simplistic view - but the point is that it is no where near the extent of what you can get done with an embedded system.

All they have to do is give us 5 joysticks they support that won't go OBS/EOL too soon - which would be an improvement over the generally commercially available game-port joysticks right now.

Edit: Oh yah - as far as the original topic - like others said, USB Host on the OI would be nice, as would some high speed serial BUS (not single port) like I2C, SPI, etc - which should be possible as the PIC on last year's RC has a 2nd I2C port, but I think it goes to one of the relay ports, which AFAIK goes through a buffer which is enabled and disabled by the master controller. I2C could always be implemented in software, but letting the hardware handle the buffering and bus arbitration would make it easier.

Re: IFI Critique
 

Just to throw this out there...

Maybe you guys should take your ideas and apply them to college? You see, you are a sucess in my eye. You are interested in technology and you want to learn more... you should sponsor a team next year. As a part of people in your school, you should g0 far.

As a "teacher/engineer" I need elimentary topics to teach my students.

Re: IFI Critique
 

John,

Are you referring to the patents that the various gamepad makers have on their connectors?

I am not sure it is true, but I have heard that basically that some sort of intellectual property concerns prevent un-approved manufacturers from making a connector that plugs into a standard Nintento gamepad.

If this is the case, I am almost certain that some sort of home brew solution could be made where teams cut off proprietary connector A and plumb in commercially available connector B.

Alternatively, there are Linix boxes that sell for much less than an OI that have USB ports capable of reading a USB gamepad. I say not that we should have a Linix based OI but that there are open source drivers out there in the wild. I think that a determined effort could definitely implement a method to read a gameport from an OI.

There are other methods and hacks possible to get a gamepad controller to the FIRST masses as well (look at these pictures) but, as long as IFI is the only electronics supplier allowed under FIRST rules, I wouldn't hold my breath.

Joe J.

Re: IFI Critique
 

So we can expect folks to have to deal with joining wires from 4 different 15 pin D-sub connectors in order to gain access to all the pins for various switches & pots & LEDs as they build their operator interaface but we can't expect them to solder 4 wires from a gamepad?

I don't want to pick a fight, but I am not convinced.

People have been asking for a way to drive their robots via a gamepad type controller for many many years.

It seems to me that we don't have an easy way to do a FIRST legal gamepad type controller because it has not been a priority for either IFI or for Manchester.

Joe J.

Re: IFI Critique
 

Following up the above:

Digging around on the web it took me only a few minutes to find this chip:
CY7C67300 a.k.a. EZ-host. It cost under $7 from Digikey and it can be configured to read USB gamepad controls like the Logitech Dual Action gamepad and report that data to one of the PICs that IFI loves (via a number of protocols easily implemented, even on a PIC).

Read this paper if you have any doubt about the feasibilty of this chip to be able to implement a 4 gamepad interface for an embedded controller like that used on the OI. It can be done. It has not been done.

Again, I repeat my contention that the only reason we don't have a gamepad interface on the OI is that there has not the commitment to implementing it.

Joe J.

Re: IFI Critique
 

Joe and John both have valid points. The IFI system is showing it's age. There are better ways to do the whole system. However there is a tremendous amount of work to change and legal issues that come about when a product is sold as opposed to a hobbyist kludge. I have a Pelican wireless PS2 controller running a Basic programmed pic 18 with a servo chip for my VEX bot and it's better than the VEX system. But, it's not an off the self plug and play system. Yes I would like a fancy robot controller with all the bells and whistles. Then again I look back to the 2006 year and I'm not as enthusiastic. The only thing that allowed our team to survive this year was the KOP trans and frame and easy C. If things were any more complicated, we would have failed. I mean flat out not have been able to ship. The current stuff has matured into a base system that allows low resource teams to get some thing that moves to the competition no matter how resource strapped they are. What ever changes do happen in the future please remember that these are high school students and not all team members are AP fast track kids. There has to be a base level system that everyone can do, but also allows advanced use. The one thing I really liked from IFI was the 2005 breaker board. I was disappointed that it wasn't in the 2006 kit.

Re: IFI Critique
 

Would you be able to give some more information about your ps2 controller controlled vex robot?

Re: IFI Critique
 

USB robot controllers/operator interfaces would go together nicely with the new lego mindstorm kits using USB :rolleyes:

Re: IFI Critique
 

A USB host isn't entirely out of sight. Atmel has usb host chips for their AVR sub family.

Also if you want to use a gamepade to control something, don't PS2 controllers just use a USART interface. I know it is a 9 pin connection, but I have never looked at the specifcs of the system.

If all you people want a USB interface to a joystick so badly instead of having an easy solution from IFI why not develop something yourselves, then make a white paper. This way all teams can benifit from it with having to have a huge chane in IFI equipment.

I don't want to bash IFI but their system could use somechanges. On the other hand, look at what they have given you a sworking system. IFI provides motor controllers, radio, and a RC and OI. Plus many other fun things. Why do you want a easy solution any ways? Isn't FIRST about learning and being innovative?

As far as changes go I think the OI could provide more support for outputs to displays or whatever. they only have 4 outputs and I htkn the most you can use those for is LEDs.