My name is Tyler Warren and I'm working on an invention which calls for a device to be built that will interface with the computer. I need an engineer to develop this for me. It's not at all complicated.
I need to connect one button to a computer via USB. The hardware needs to take a signal from a simple switch and tell a computer "a button has been pressed" And that is basically it.
I know there are plenty of products on the market which can do the same thing for 20-100 buttons, but I just need one. It has to be cheap to reproduce, I will eventually need to make hundreds of them.
If you are capable and available please contact me at snipped-for-privacy@hotmail.com or snipped-for-privacy@mcrdpi.usmc.mil for information. I truly hope someone can make this happen.
The simple solution is to use an off-the-shelve USB-serial or USB-parallel chip, connect the switch to one of the pins and voila. A next step would be to do the same using an FT232BM or FT245BM chip. Beyond that: use for instance a Cypress or Microchip microcontroller with build-in USB. A very cheap solution: if your PCB can do without a mouse you might just get a mouse and wire your switch parallel to one of the mouse buttons.
Tyler, I believe that you underestimate the sophistication of the USB inteface, which requires some smarts on the peripheral device attached to it.
For a simple switch closure, I'm sure that it will be far more easy to implement by using one of your computer's serial ports, than its USB interface. While this may be very cost efftive for large quantity production runs, on a prototype scale involving less than say 1,000 units, it's going to become very costly.
For USB, you're pobably talking about a simple software design costing maybe less than $25K, but for the hardware, you could be asked from $50K to $75K for the first working, production prototype, at least at the going rates. For production, figure on about $100 a pop for runse of less than 1,000.
By contrast, detecting a switch closure on a serial interface is child's work and cheaply implemented.
I think you're way, way off base here. First of all, PCs are losing their serial ports. All of us hardware nuts might hate that, but USB is going to be the only thing going in the near future. Macintoshes already have only USB and Firewire. And, to be honest, there's a lot of nice things about USB. You can daisy-chain them, once you have a USB setup working you can make it do a lot of stuff with relatively little extra work, and you can connect everything from a low-speed mouse to a 120Gig hard drive to the same port. Power's nice too.
Second, I think you're way overestimating the cost and complexity of implementing the desired function. Motorola makes USB-capable microcontrollers for less than three dollars each. They are ready to go, simply wire the USB D+ and D- lines to the chip, apply power and a crystal and they are up. The USB startup and handshaking is straightforward to code even in assembly. And on the PC software side, you don't even have to write drivers for this application. You can program the microcontroller to emulate a keyboard, mouse, or joystick and convert the button press to a keypress, mouse click, or joystick fire button event. The HID drivers are standard fare on Windows, Mac OS, and Linux operating systems; even most BIOSes these days support a standard USB keyboard. The programming is only a couple day's worth if you're being careful, the board design is whatever you need to support a 28-pin SOIC, a 6Mhz crystal, a few capacitors, some resistors, and places to wire in. If someone charged me $50,000 to design that, I'd call my lawyer. You need a USB Vendor ID, of course, but you don't need it while prototyping and they are available for $1000-$2000 from USB.org. The per-unit cost is not going to be prohibitive.
This is also a better solution than the FTDI one above, as no special drivers need to be loaded and the IP belongs entirely to you.
Sorry, but you have NO IDEA what you are talking about. Look up FTDI's RS232-USB chips, this sort of thing can be done for a LOT less then the figures you've given.
The problem is that USB interfaceing is not a trivial project. It would be far easier and simpler to look at adding a switch the serial port, parallel port, or even a joystick port.
Take an FT232BM, connect it to your USB port (OK, that part is not trivial but there are small PCBs that do this for you - I sell one). Install the drivers provided by FTDI, and your PC has an extra serial port. In your software you just use that serial port, so it is
Not a hacker Rev, an experienced engineer like myself and who, unlike you, realizes that increased interface sophistication adds severe design cost and success risk to what should be a simple interface design.
It's the hacker that foolishly believe that he is going to take an off-the-shelf, complex interface chip, stick it in a box, and make it work RELIABLY without expending weeks or months to accomplish the simple goal sought.
I believe in the 'KISS' principle, and from experience know that the use of complex, active devices by amateurs can lead to serious and sometimes catastrophic problems. In this particular case, there is no method more simple and reliable than feeding a contact closure into a RS232 port on a computer, unless of course a discrete digital input is available which, on most PCs is a bit unusual.
Your cheap shots are often clever and humorous, but at the cost of making you appear to be a total technical/engineering fool!
The Lord alerted my mind to the presence of this EVIL article by Harry Conover, and I thusly replied:
I didn't know they called programmers "engineers" these days. Harry seems to know about coding, and so I was just wondering whether he is a hacker too (in the true sense of the word!)
It's more expensive to produce than directly using the USB controller. If you add even a dollar onto the cost, that eats into your profits a lot. Also, you have would to write a program on the computer-side of things, a different one for each operating system, to convert the input into a keypress. A custom USB solution can emulate a generic keyboard directly and will also work on all major operating systems without additional drivers. You can't do this with a keyboard wedge, either, because Macintoshes don't use PS/2.
And if an individual, for example, was already familiar with using a specific USB microcontroller and perhaps had done most of the hard work already, then such a project begins to look a lot simpler. USB microcontrollers, especially when dealing with Low-Speed HID devices, are not complex. The RS-232 port is heading the way of the dodo, it's time to get with current technology or get out if you can't take it.
Of course, but an FT232BM is not that mich more expensive than an USB controller.
The OP did not ask for that, just a way for a program to detect the keypress/switch closure. And that can be done in an OS-independent way, just open the serial port and check the status of a handshake pin. Been there, done that.
If USB offers a cheaper or more feasible sulution I agree that it is the good choice, but otherwise your last scentence just shows a lack of real arguments.
It's unneccessary complexity to perform a simple task. Why not simply attach the switch to the computer's RS232 or parallel port, if either exists, rather than going though the effort of needlessly serializing and then then decoding a simple contact closure?
Why unnecessarily convert a contact closure into a logic level on a chip in a computer, filter through several hundred traces and through several layers of controllers, eventually reaching a massively complex processor chip where millions of calculations per second are used to read the changed value, and change thousands of additional bits to register in video memory that a button has been pushed, where the information filters all the way back out the video controller, and then into modulation of an electron beam flickering across the internal surface of a vacuum tube? Why not just use a battery, a switch, and a light bulb? Or, a battery, switch, and solenoid positioned over the keyboard?
The complexity of it is not the issue. To a typical user of any USB, serial, parallel, or ethernet device, the complexity level is the same. You plug it in, and the thingamabob is connected to the computer. USB and RS232 are not much different in complexity when compared to the rest of the computer system. Yes...it is more complex than it absolutely has to be, that is why industrial computers have digital i/o pins. Most people don't have an industrial computer, though, so to them the USB port is just a hole they plug a wire into. They won't feel any guilt using up a port capable of
480MBps to read the value of a single button, if that is the application they want.
PolyTech Forum website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.