PC trackball tracking poorly in X-axis!!!

[Tiger-Heli]

I recently purchased the trackball shown here http://www.softwareandstuff.com/h_acc_mactrack.html, and am having one minor and one major problem with it.

First off, I am using vanilla (win)MAME (not analog+), I am using the generic Win98SE USB hiddevice drivers (not the floppy disk drivers), and I had both my PS/2 mouse and the USB trackball plugged in during testing.  I experienced problems before and after applying 3-in-1 oil to the trackball shafts.  I also didn't mess with the analog settings in MAME very much and maybe that would help.  I am not sure exactly how to tweak them, though.  Also note, I am not planning to hack the trackball, I plan to use it just as it came for occasional MAME use.

First, a minor odd problem - the trackball buttons register in MAME as Right - J1X-axis - and Left - J1X-axis +.  Not sure about the middle button.  I expected either Mouse B1, or Mouse2 B1, etc, but it's not important, as I wasn't really planning to use these buttons anyway.

Now the main problem:  The trackball works well in the Y-Axis, but poorly in the X-axis, both in MAME and Windows.  For example, lightly spinning the ball vertically zings the cursor to the top or bottom of the screen :-).  But lightly spinning the trackball horizontally, moves the trackball about 3 inches (1/5 of the screen) (17-inch monitor).  Rapidly spinning the trackball horizontally, results in no movement, either due to slipping or backspin.

In use, the trackball worked really well in Golden Tee Golf, and acceptably in Capcom Bowling (except I had to use the arrow keys to avoid delay of game when centering the ball and I had a hook of 3 selected, but this was due to my unfamiliarity with the controls).

Games like Centipede and Missile command played slightly worse than with the mouse due to the poor X-axis tracking.  Cabal was about the same as with the mouse, the better Y-tracking somewhat made up for getting killed more often by the poor X-tracking.

I am considering wrapping electrical tape around the X-axis contact roller to make it thicker.  Is this a good idea, or is there anything else that someone can suggest?

Thanks in advance.

[AlanS17]

Well games like Golden Tee use primarily the Y axis while Centipede uses primarily the X axis so it's no surprise that GT would work better. Have you tried pulling the ball out and manually moving the rollers with your finger to see if it's not a mechanical problem after all (despite the oil). Does the ball spin well on the X axis or does it stick a bit? Maybe it feels like it's rolling well when it's not actually spinning the roller? I'm no mouse expert.

Here is one other thing to consider. What do you expect for $4? Is it possible you got a bum trackball? Maybe faulty interal wiring or circuitry is the culprit. I've had mice before that exhibited "driver problems" before when it was really just a bad mouse. Sometimes you get what you pay for.

[BillyJack]

i don't know if this is even worth trying, but would you consider messing around with the little sensor LED thingy by the wheel of the x-axis?  maybe it's vision is poor?  

BillyJack

[Tiger-Heli]

Well games like Golden Tee use primarily the Y axis while Centipede uses primarily the X axis so it's no surprise that GT would work better. Have you tried pulling the ball out and manually moving the rollers with your finger to see if it's not a mechanical problem after all (despite the oil). Does the ball spin well on the X axis or does it stick a bit? Maybe it feels like it's rolling well when it's not actually spinning the roller? I'm no mouse expert.

Here is one other thing to consider. What do you expect for $4? Is it possible you got a bum trackball? Maybe faulty interal wiring or circuitry is the culprit. I've had mice before that exhibited "driver problems" before when it was really just a bad mouse. Sometimes you get what you pay for.

The X-axis rollers spin fine (physically).  The ball feels like it's rolling well, but I suspect it is not spinning the roller.

One thing I need to try (which you implied) is removing the ball and spinning the rollers and comparing the X- and Y- response in Windows.  This would tell me if it's largely backspin (encoder wheel) problems, or slippage (mechanical problems).

I don't think it's faulty internal wiring as the X-axis does respond, just not well, and the Y-axis does work well.  It's probably poor contact between the x-axis rollers and the ball, but I need to try without the ball to confirm.

[Tiger-Heli]

i don't know if this is even worth trying, but would you consider messing around with the little sensor LED thingy by the wheel of the x-axis?  maybe it's vision is poor?  

BillyJack

As I said, the next step is to spin the rollers in Windows without the ball attached.  If I get good results, it's a mechanical problem.  If not, I'll look at giving it some vision correction.

Thanks for the tip.

[Tiger-Heli]

UPDATE:  Ok, I took the ball out and was still having the same problem.  So it's not slipping of the ball.  And it's not so much stuttering or skipping, so it's not something with the encoder.  I think I found the problem:

I took the T-ball apart:  If it ain't broke, take it apart anyway and lose some of the pieces, then it will be . . .  Anyway, the plastic thingy that holds the vertical roller is firmly attached to the PCB, but the plastic thingy that holds the horizontal roller is able to shift about 1/16 of an inch top and bottom, I suspect that this shifting causes the roller to spin less for a given input.

I was able to get Centipede playing decently by increasing the X-axis Speed setting to 75 while leaving the Y-axis speed setting at 10.

The slight problem is I was hoping to buy two of this and use them for CABAL and then use either one for Centipede, etc.  But if this is a QC issue, and it appears to be, then the next one I get might have a horizontal axis that spins freely and a vertical axis that stutters.  So I will probably have to label them as 1 and 2.

Question:  Does anyone have a good explanation for the MAME Analog Controls Sensitivity and Speed Setting?  I checked EasyEmu, but didn't see anything.

[u_rebelscum]

UPDATE:  Ok, I took the ball out and was still having the same problem.  So it's not slipping of the ball.  And it's not so much stuttering or skipping, so it's not something with the encoder.  I think I found the problem:

 . . .  Anyway, the plastic thingy that holds the vertical roller is firmly attached to the PCB, but the plastic thingy that holds the horizontal roller is able to shift about 1/16 of an inch top and bottom, I suspect that this shifting causes the roller to spin less for a given input.

Is there any way to reinforce the "plastic thingy"?  

Question:  Does anyone have a good explanation for the MAME Analog Controls Sensitivity and Speed Setting?  I checked EasyEmu, but didn't see anything.

Analog key/joy speed: "Mouse speed equivent for a digital input (ie keyboard or arcade stick), with the number as mouse delta per frame."  That's my definition, but pretty close, since mice can be use for AD stick inputs, too.  Anyway, "mouse delta" is the number mame would get from a mouse as the amount of movement.  Mame reads the input once per frame one basically all games; a speed setting on a 45 hz game would be slower than the same number on a 60 hz game.

However, sometimes the joy/key speed seems to effect mouse and analog joystick inputs, even though I dont' think it's designed to.  I haven't pinned down how it effects these, though.

Analog sensitivity: "A modifier of mouse or analog joystick input values, with a 100 setting the same as not modifying the input."  Again, my def.  The two formulas sensitivity seems to do the most are (simpified, input > 0 shown):

[size=+1]((current  / 128.0) [/size]^{(100.0 / sensitivity)}[size=+1] ) * (max) * (100 / sensitivity)[/size]

and

[size=+1](current * sensitivity / 100) + 1/2[/size]  <== cleaned up formula

I simplified assuming default (center for joystick) is 0.  The first equation is only for stick inputs, not spinner or trackball inputs.  If input < 0, substitute min for max in first equation, and - for + in second.  <== added this paragraph

I hope these help.

(edit - cleaned up second equation, and added more info on them)