You seem like a knowledgeable bloke (to put it mildly). Did you go by another name at the Cougar World / Frugal's World forums? Trying to keep everyone straight in this discussion...
I will ask Thrustmaster about the resolution. Your technical information seems rather indisputable. Can you explain the Nyquist Sampling Principle to a scientist who has never heard the term before? I could Google it, but I think an in-context explanation from you would be more useful.
I go by the same name on Frugal's.
Nyquist's Sampling Principle is also known as Nyquist Criterion or nyquist Freqency, Horowitz, P., and H. Winfield. 1997.The Art of Electronics, 2nd ed. Cambridge Press, pp. 775., it states: "If you make periodic brief samples of the amplitude of a continous analog waveform of finite bandwidth (where the maximum frequency is fmax), you preserve the input information as long as you sample "
Says you are a scientist, so I follow the format of proper bibliography reference. I Now, if you view the x-y coordinate system of the joystick movement like a screen and apply a steady velocity scanning beam, like a CRT, you basically turn a spatial domain problem into a time domain one. Thus, Nyquist Criterion applies. That is you need at least twice the "frequency/resolution" to sample and faithfully reproduce the analog signal/movement of the stick.
So, since the ADC is 15 bit, only 14 bit can be faithfully relied upon.
14bit is plenty, don't get me wrong. The 4th axis of my CNC machine I am building using a 131:1 Harmonic Drive will only get 1.8/131 =0.01374 degree resolution. To put it in perspective of this resolution, say I mount an 8" diameter of round stock on the 4th axis. One step of 0.01374 degree moves the surface of the cylinder by 0.01374/180*pi*4 =0.000959", the arc length = r* theta formula. That's about 1 mil resolution! Let's assume WH's stick is roughly 4" long, the same as the radius of this cylinder. WH has about 3X rresolution of this. If you tell me your hand can feel resolution difference below this precision, I have to invoke Carl Sagan - "Extraordinary claims require extraordinary proofs.", even though I am not from Missouri.
More resolution than this is like audiophiles -- I used to be one, until I got old. Now, I buy $250 Sony AV receivers that come with 5 speakers.
Nyquist's Sampling Principle is also known as Nyquist Criterion or nyquist Freqency, Horowitz, P., and H. Winfield. 1997.The Art of Electronics, 2nd ed. Cambridge Press, pp. 775., it states: "If you make periodic brief samples of the amplitude of a continous analog waveform of finite bandwidth (where the maximum frequency is fmax), you preserve the input information as long as you sample "
Let me take it from here, if you will.
The Nyquist Criterion simply states that when trying to get a discrete time signal from a steady one, through periodically sampling (checking) the value of the constant signal, one has to at least use twice the frequency for sampling compared to the steady signal. So, say if you want to digitalize a signal with frequency range 0-22kHz, you have to sample with at least 44kHz (44100Hz is used on the CD). If this criterion is not met (because for example, you did not filter the input signal to not contain any frequencies above 22kHz) there will be aliasing, that is, the frequencies above the Nyquist frequency will be mirrored into the 'base band' (0-22kHz) around the Nyquist frequency. Example: Take the above and assume there is a 24kHz signal component, this will then be mirrored into the base band at 20kHz (44kHz minus 24kHz=20kHz).
Thanks, guys, makes sense to me. I also agree, having use the Warthog hands on, that 14-bit is good enough.
I know that the throttles didn't get the same "every screw" disassembly treatment as the stick and stick base did. Any point in taking a close look at them?
We now know the stick's physical limit of movement is circular, further bounded by a plastic bit with a rounded square hole in it.
In discussing the 15 degree offset idea and the associated problem with the 'box' the stick shaft hits at it's limits being also offset, Bwaze suggested one could just change the plastic stick limiter, offset it 15 degrees to the right. Much nicer (read easier and cheaper) solution than my stick shaft adapter idea. Could you post some pics of what the "plastic travel limiter thingy" looks like when uninstalled?
If it looks like what I'm hoping (circular), the solution may be as easy as drilling some more mounting holes, 15 degrees off the originals.
Thanks, guys, makes sense to me. I also agree, having use the Warthog hands on, that 14-bit is good enough.
I know that the throttles didn't get the same "every screw" disassembly treatment as the stick and stick base did. Any point in taking a close look at them?
Would be nice to confirm what chips they are using there in the throttle axel, but not really necessary, as I am almost sure it is MLX90316, the sister chip, exactly the same as in DiH. Why bother with other Hall chips? For me, it would cost $11 apiece at quantity of 10 from Digikey or Mouser. With TM's volume, it goes down to about $3 apiece. And the circuit needed is almost identical to the one for 90333, the chip programmer is the same, the code to drive it is exactly the same, and the documents of the two are like 80% copied from the other (understand one and you can pretty much skim thru the other). Not only big saving on design / test resources, simplifies logistics too. TM would be dumb to use any other Hall chip.
Well, at least I came to the same conclusion 3+ years ago. Actually I had my eyes on MLX90333 as soon as it came on the market years ago when I was struggling with Cougar's pots and searching for Hall chips. But it was too new, and Melexis as a Belgian foundry-less design house wasn't exactly established in the market that well, like other big players in this market, like Honeywell. They could go down or EOL the line overnight. BTW, the Honeywell Hall chips suck! Very difficult to use, you need more than a Ph.D. to use it. So, I waited for years until I hear the news that BMW was using 90333 in their swivel headlamps (thanks to all you BMWer drivers). That was when I knew this chip would survive the market and bite the $1700 bullet purchasing the programmer. Once I made the 90333 circuit work, prototype PCB for 90316 was like done in a day once the order of the mangets arrived, including design the circuit, layout, mill PCB, solder, program, test, and hook up to Cougar PCB to test (w/o the DiH aluminum casing). It was that easy!
The only complaint about MLX, the Atmel ATMega based PTC-04 programmer is waaaaaaayyyyy too expensive!!!
We now know the stick's physical limit of movement is circular, further bounded by a plastic bit with a rounded square hole in it.
In discussing the 15 degree offset idea and the associated problem with the 'box' the stick shaft hits at it's limits being also offset, Bwaze suggested one could just change the plastic stick limiter, offset it 15 degrees to the right. Much nicer (read easier and cheaper) solution than my stick shaft adapter idea. Could you post some pics of what the "plastic travel limiter thingy" looks like when uninstalled?
If it looks like what I'm hoping (circular), the solution may be as easy as drilling some more mounting holes, 15 degrees off the originals.
Jocko,
I am going to put this discussion topic on hold. I have been exploring the script for changing the angle, and as you've correctly pointed out, the square plastic piece sort of foils this. I have been messing around with axis zoom and then identified a problem with it. Thrustmaster sent me some new files to try out, but time is short right now. I will get to it within the next few days, so stand by.
We now know the stick's physical limit of movement is circular, further bounded by a plastic bit with a rounded square hole in it.
In discussing the 15 degree offset idea and the associated problem with the 'box' the stick shaft hits at it's limits being also offset, Bwaze suggested one could just change the plastic stick limiter, offset it 15 degrees to the right. Much nicer (read easier and cheaper) solution than my stick shaft adapter idea. Could you post some pics of what the "plastic travel limiter thingy" looks like when uninstalled?
If it looks like what I'm hoping (circular), the solution may be as easy as drilling some more mounting holes, 15 degrees off the originals.
If you can rotate the square window and do software rotation instead of rotating the sensor without measurement blank corners, it means the sensor is configured to cover more area than is necessary for the default "square" of measurement, i.e. There are resolution wasted.
Think about it this way. The ideal situation is the sensor is configured to only respond within the square the stick "draws." now, rotate the stick Counter-clock means the sensor chip also rotate w/ it, thus the sensing "square" as well. Say, we rotate 45degree CC-wise. Then, we rotate the square limiter window 45 degree C-wise, so that two sides of it are horizontal. Now, the two squares are 45 degree to each other and the limiter square has 4 corners that are outside of the sensor square, i.e. no-sense areas. In these areas, inside MLX90333, you have to configure them to fixed values, most sensible values will be either 0%(left,bottom), or 100%(right,top). This would not work for you, as moving in these areas produces no change in value. Dead zones!!! Rotate whatever in Target software, and you get the same thing, garbage in garbage out.
For this scheme to work, the MLX90333 has to be configured to respond to an area bigger than the actual square limiter (directly related to magnet travel, so I am using these two interchangeably). How much over sized? Sqrt(2) in both x and y directions than the square limiter. Perhaps WH is factory programmed this way, perhaps not. I don't know yet.
Maybe TM thought of it and thought since we have plenty of res. available, it is ok to use only 50% of the sensing area, just in case somebody might want to mount the stick rotated and rotate the limiter square. Possible, why not? Beats having to have them send sticks back for reprogrammimg, or replace the chips.
I will report back on the programming of the chip. But, if you do modify it as you described and hit dead corners, you now know why you might need to rotate the sensor too.
User manual for Warthog describes how to engage the afterburner detent (by opening the cover above it). It also describes how to detach the base plate (for the cockpit mounting) - which exposes insides of the stick base. And on the next page:
Quote:
This warranty shall not apply: (1) if the product has been modified, opened, altered,...
Has anybody seen the dimensions of the joystick canister (diameter and height), Joe did mention the total height of the stick + the dimensions of the base plate in his review but no data for the canister was given. Need to make a center mount for the gaming chair.
by default, there's 4 screws used to fix the base to the plate. You can use the plate as a stencil, drill the holes and fix the canister on the surface you desire.
Wooohooooo. Good boy Cougar helo Warthog.
