Jellybean Wireless

[I][edited from support email][/I]

I am considering using Jellybean wireless buttons to get children’s responses on a DirectRT program. Is there any issue with response times and wireless devices?

While there shouldn't be any reason not to use a wireless product, I would be curious about the wireless system latency.

Here's how most of these systems work:

[Pushbutton -> Microcontroller - > Wireless Transmitter]

~~~ air ~~~

[Receiver -> Microcontroller -> Computer Interface]

The unknown is how long the microcontroller (on both ends) waits to transmit a signal after a button has been pressed. There also may be a slight delay on the receiving side.

The manufacturer can probably quote you a 'system latency' for their product. If you can find that out, we'd be very interested in knowing!

John

Well, the manufacturer wasn't able to go beyond "instantaneous" in terms of quotes on latency. Therefore, a colleague and I took a dual channel oscilloscope to the device, as recommended. The shortest latency we found was 100ms, and the longest was 350ms. Therefore, it was longer than expected AND relatively unreliable.

It looks to be a great device if one is only concerned about accuracy, but our lab is going to decline as a tool for reaction time.

Outstanding!

Thanks for the feedback.

-John

[I]This is the test procedure used to measure the system latency:[/I]


It's actually easy to test this sort of thing if you have access to a dual-trace oscilloscope.

It involves opening the enclosures (and possibly voiding a warranty), but it's not too bad.

Open the enclosure surrounding the first pushbutton. Connect one lead of your oscilloscope to the pushbutton and connect the other lead to the output on the receiving side. Connect the ground clips for both probes to an appropriate ground point. This may be the negative edge of a battery, the 'shield' of a shielded cable or something similar.

You should be able to press the button and see a change in the oscilloscope trace. It might drop from 5 volts to 0 volts, or it may step up from 0 volts to 5 volts. The important thing is that there's a measurable change in voltage.

If you don't see a change on your screen, try probing the other terminal of the switch. There are probably two terminals for each switch.

You should see a similar change on the output of the receiver assembly.

If you're using a digital storage scope, you can use the 'cursor' function to measure the time between the button pressing and output. This measurement will give you the *total* time elapsed between the pressing of a button and the reception of the signal on the other end.

If this time is always constant (or at least, constant +/- some reasonable degree), you can just add the transmission time to the reaction time recorded by your software.

However, if the transmission time varies significantly, you'll need to decide if the variability will significantly affect your data collection.

If you don't have a 'scope, someone from across campus in the engineering department should be able to help you. This whole process will only take a few minutes.

If you're really clever, you can use this same procedure to test the timing delay and latency of regular computer mice. There are four wires inside a standard mouse cable: power, ground, clock and data. Just compare the signal from the pushbutton with either the clock or data signal. You'll want to move your cursors to the [I]end[/I] of the clock or data signal transmission, which is when the entire mouse click has been transmitted.