[I][edited from support email][/I]
Our work group is currently discussing which mouse type and brand we should use for our reaction time experiments. For these experiments we need a precise mouse. What do you recommend?
Printable View
[I][edited from support email][/I]
Our work group is currently discussing which mouse type and brand we should use for our reaction time experiments. For these experiments we need a precise mouse. What do you recommend?
For starters, take a look at the PDF file I've attached to this posting. It's an article from Behavior Research Methods, Instruments and Computers which discusses mouse accuracy.
If the article doesn't seem to help, here are some technical details which may be useful to you. And if you still have questions, please feel free to contact us.
[CENTER][B]* * * Testing Hardware Latency * * * [/B][/CENTER]
It is easy to test this sort of thing if you have access to a dual-trace oscilloscope and are comfortable with electronics. Note that this involves opening the mouse enclosure and possibly voiding its warranty. Also, depending on your skill, it may be partially or completely destructive to the mouse. More importantly, if you are not very careful you could potentially short your computer's power supply which could cause noise, smoke, and a shut down of your system. Of course, we will make clear here that we are not responsible for any hardware or data loss.
All that being said, here's the process:
The overall goal: measure the time elapsed between (a) when the mouse button is pressed and (b) when the data transmission to the computer has finished.
1. Open the enclosure. Locate the switch which is activated when a mouse button is pressed. It's probably mounted to the circuit board with two soldered terminals. This step may be permanently destructive to your mouse, depending on your disassembly skills. But mice are cheap these days.
2. Find 'ground' inside the mouse. Probably the mouse cable contains 4 conductors plus a shielded metal jacket. Ground will be the shielded connector + one of these wires. Connect the ground lead of your scope to this point. The other three wires will be power, data and clock.
3. With the mouse plugged in and your computer running, connect the scope probe to one of the pushbutton terminals. Press the button and watch the screen. You're looking for a change (high to low, or low to high) on the pushbutton terminal. If you don't see anything, try the other terminal instead. This is the part where shorting pins on the mouse circuit board to ground may or may not damage your computer's power supply. So be careful here.
4. Connect your other scope probe to one of the other three wires heading back to the computer. One of these is a 'data' signal and one is a 'clock' signal. Both will be idle unless a mouse click is registered - during transmission they'll oscillate at a few tens of KHz for a few microseconds. The 'power' signal will always be constant, probably at +5 volts.
5. If your 'scope has triggers, set it to catch the falling edge of the signal. Usually clock and data idle high.
Once you've found the proper wires, just use your oscilloscope's time-measuring cursors to compare the signal from the pushbutton with either the clock or data signal. You'll want to move your cursors to the end of the clock or data signal transmission, which is when the entire mouse click has been transmitted.
A fixed delay is a good thing: just add the delay (probably between a few mS and a few tens of mS) to your collected timing data. A variable delay is not a good thing. You may want to run this test a few dozen times to ensure that the delay is more or less constant.
If all of this seems too complicated or frightening, let us know and we can discuss other options with you.
Attached is the article I referenced above.
- JEC