Challenging question for Michael

[Fridge]

Hehe.. Since you CH Products guys are all over these boards I have a question :-)

I have been considering building my own 'Trim Control' USB device based on the Mini-Phantom USB interface (I can not get to their site right now to post a link). This device has been designed (in my mind) to have three 'sliders' with center dents (center dent pots are more expensive and I do not have the means to create some nice fancy arsed wheels) which would act as three separate axis - potentially for elevator, aileror and rudder trim.

I have been eyeing the 'mini-joystick' part on the CH Products parts site as potentially two more axis (if the USB interface has them available). Throw in a number of simple pushbutton switches as available on the device layout/USB interface potentially in an MFD configuration.

Wouldn't it be sweet to have two of these (three axis, 20 buttons in an MFD square, each)?

Then I noticed on the CH Products OEM site this little gem:

(http://www.chproducts.com/oem/misc_usbkit.html)

My question.. how useable is this module? I'm going to make a leap of faith and even ask if it can (or even will) show up in the CH Products Configuration Manager software?

Ok. That would be a bit much to expect :P - how about more likely question: is there any documentation for the device?

:idea:

Dave 'Fridge' Penney

[JNOV]

That's cool! I never bothered to check out the "other" side of CH's site.

Micheal:

A few questions --

1. When are you guys going to offer a "Hall Effect" version of your Figherstick/Pro Throttle/Pro Peds?

2. More realistically, would you consider offering Hall Effect transducers that could be used to upgrade your existing sticks?

3. Do your Hall Effect sticks have a greater resolution than those that use resistor potentiometers?

- JNOV

[JNOV]

I just noticed this in the "specifications" section of the site:

Hall Effect
Standard HFX
PHYSICAL SPECIFICATIONS:
• Joystick travel: 36° (18° from center)
• Centering: single spring, omnidirectional
• Return to center repeatability: ±1%
• Panel thickness: 0.046 to 0.125in. (1.17 to 3.17mm)
• Housing: high impact glass-filled nylon
• Shaft: 0.25 in (6.35mm) diameter, stainless steel
• Handle: thermoset phenolic (model 1100 ball tip)
• Boot and gasket: thermoplastic rubber
• Operation force: X/Y axes = 2.25N nominal
• Breakout force: X/Y axes = 1.25N nominal
• Weight (2 axis w/ball tip handle) = 0.2 lbs (0.091 kg)
• Weight (3 axis w/one button) = 0.215 lbs (0.098kg)
• Environmental: IP57 above panel (certain models)
• Operating temperature: -40°C to 85°C
• Storage temperature: -55°C to 165°C
• Flammability rating: 94HB
• Mechanical MTBF: 3,000,000 cycles under normal use

ELECTRICAL SPECIFICATIONS:
• Supply voltage: Vdd = 5V
• Center voltage: 2.5V (±25mV)
• Supply current (2 axis): 4.8mA min to 11.0mA max
• Supply current (3 axis): 7.2mA min to 16.5mA max
• Resolution: Infinite
• Response: 40-80uSec
• Output current: ±2mA
• Output voltage tolerance: ±2%
• Electrical MTBF: 1,000,000 operational (power on) hours

I particularly like the part about "infinite" resolution! Also, the supply and center voltages seem to bode well for using these sensors in your existing "gaming" controllers. Oh, and the mean-time-between-failure rating of one million hours is pretty impressive, too! :)

- JNOV

[MichaelCHProd]

Fridge:

You could order an OEM USB Kit but it would not be seen by the Control Manager.

There really isn't any documentation for the USB Kit as it is designed for OEM usage. Meaning we assume you have a controller and all of its electronics and all you need is a way to interface it to a PC.

JNOV:

1. Still looking into it.

2. In all likelihood probably not.

3. Depends on the application and implementation but usually yes.

[Bob Church]

Just a little clarification, some might be interested. There are basically two kinds of pots, wirewound and composition. Every joystick pot I've ever seen is a composition pot of some sort. Wirewounds are typically too large, too expensive, and wear out far to quickly for joystick use. The composition type varies (carbon, cermet, or whatever) but it's basically a continuous coating of some resistive material on a non-conducting substrate. As such, it inherently has infinite resolution. You can always move the contact just the slightest bit and get a change in resistance. Any composition pot that you buy, whether it cost you a 50 cents or fifty dollars, has infinite resolution.

Resolution in a USB stick is really dependent on the A/D converters that are used to read the pots, has nothing at all to do with the pots themselves. A/Ds generally rated by the number of bits they can resolve, e.g. 8-bit A/D (0..255), 10-bit A/D (0..1023), etc. The practical resolution can also depend on how smoothly you can change the pot position, how much play there is in the mechanism, etc. If the mechanics aren't up to snuff, you may not be able to move the pot in a small enough increment to even get the resolution that the stick circuitry is capable of.

To get a feel for just how much movement it is, put the stick into Direct Mode, open the calibration screen, and watch the Raw Value display. First, check that the centered value is around 128 or so. Adjust the trim wheel if it's not, calibration doesn't really matter for this test, we're just looking at the raw numbers. Next, push the stick off-center so your under a little return pressure and out of the deadzone, maybe down to a raw X value of 64 or so. Get the 64 stable, then move the stick just enough to change it by one count (to 63 or 65). That will give you a feel for what the practical resolution is, e.g. the smallest increment of stick movement that the controller can measure is the distance it moves for a one count change.

Probably worth noting that Hall Effect sensors is essentially exactly the same situation with regard to resolution. The only difference is that where the pot has a sliding contact on the resistive material, the Hall Sensor just waves a magnet around in the air near the HE device. The magnet position changes the voltage, which is measured by the same A/D converter you'd use for the pot, and the resolution is again a function of that circuitry more than the device itself.

- Bob

The StickWorks
http://www.stickworks.com

[JNOV]

That's interesting, Bob. I always figured that with high-precision A/D converters so cheap, the limiting factor would be the potentiometer, not the converter. (We were using 24-bit sigma-delta converters ten years ago that weren't prohibitively expensive. I don't recall exactly, but I'm pretty sure that the sampling rate was sufficiently high for use in an application like a joystick. I realize that "cheap" in an industrial application and "cheap" in a $100 consumer product generally mean two different things, but we were designing microcontroller boards to be sold for $150-$200 each, so they aren't that far off. I'm thinking the parts had multiple channels and were less than $5 a piece.)

A/D aside, I'd be surprised (it wouldn't be the first time :) ) if Hall sensors aren't capable of a higher useful precision than typical joystick pots, if for no other reason than their non-contact transducing mechanism. (There's got to be some reason why people are using them!) Whether you can design a consumer-level joystick with a mechanism stiff enough to exploit that resolution in a deterministic way, I don't know. But I would think that an extra couple of bits of resolution would make a big difference around the middle, particularly if you've specified a response curve that has a very low slope around that point.

As always, thanks for the info,

JNOV