Only related, but the Apple Lisa had a capacitance keyboard and we had one of the original demo units in our computer store ... but we soon discovered that with hands on the keyboard, random characters would be 'typed' sometimes. Trying to solve this mystery we checked connectors, outlets, building earth grounds. Finally we were picking up the table it was on and sliding it around the room. We found a position in the middle that seemed to minimize it but for certain people at certain times, the ghost typist would jump in and add things to what they were typing.
Next day I was experimenting on an empty document with my hands resting on the keyboard watching ghost characters appear. A radio was tuned to 1000kHz AM and someone was speaking. The rhythm of the words was a perfect match. Jingles and music generated a continuous stream of gobblegook. Not surprisingly, the station's transmit tower was ~1000ft from our building.
I ran a ground wire to the table and while holding it the ghost typing stopped. We replaced the chair with an all-metal folding chair with ground wire attached, so as we demo'd the machine customers would not experience the problem while they were sitting on it.
It turned out to be un-sellable, not even one unit sold. It was smooth and pretty but everyone was already used to the choppy but crisp and fast CRT block fonts and static character arrays of the era for text apps. AutoCAD people were already used to the characteristic 'redraw and blink' stuttering behavior of other computer brands' graphic cards and high resolution displays, which typically displayed on their own additional monitor.
With its extremely high price point and crude software available Lisa was unable to win anyone over.
I've actually seen this before where a board would run all night with no problems but as soon as someone was sitting next to it in a pneumatic office chair, it crashed.
Drove everyone mad until someone figured out it was the chair itself.
How does sitting on a gas lift chair generate an EMI spike (per the article)? The metal piston sliding an inch or two inside the cylinder collar is enough? Or is the hydraulic fluid itself metallic?
The article cites/links-to a whitepaper from 1993 on the topic of EMI producing furniture. I haven't actually read the whitepaper myself, but there might be more information in there....
The paper is mostly about measuring the magnitude of the EMI. They observe that dry air and metal legs are required, but not much more than that - I don’t think they mention gas lifts at all. The last few paragraphs have some guesses but they basically said that finding the cause is for future work.
I actually see this on my home Dell 4K connected to a MBP13 via Thunderbolt to DP. It happens when I'm not using a chair (adjustable desk) but now I'm going to be very observant when sitting down. I'd lean towards my cable being the issue however.
I saw this as an intern working on internal dev boards for some wireless gear. A coworker had set up some debugging tool to trigger a pin on some condition and it was being set off whenever he sat down in his chair.
Only related, but the Apple Lisa had a capacitance keyboard and we had one of the original demo units in our computer store ... but we soon discovered that with hands on the keyboard, random characters would be 'typed' sometimes. Trying to solve this mystery we checked connectors, outlets, building earth grounds. Finally we were picking up the table it was on and sliding it around the room. We found a position in the middle that seemed to minimize it but for certain people at certain times, the ghost typist would jump in and add things to what they were typing.
Next day I was experimenting on an empty document with my hands resting on the keyboard watching ghost characters appear. A radio was tuned to 1000kHz AM and someone was speaking. The rhythm of the words was a perfect match. Jingles and music generated a continuous stream of gobblegook. Not surprisingly, the station's transmit tower was ~1000ft from our building.
I ran a ground wire to the table and while holding it the ghost typing stopped. We replaced the chair with an all-metal folding chair with ground wire attached, so as we demo'd the machine customers would not experience the problem while they were sitting on it.
There's one lesson here about debugging problems with unintuitive solutions.
There's a second lesson here about not trusting sales people.
It turned out to be un-sellable, not even one unit sold. It was smooth and pretty but everyone was already used to the choppy but crisp and fast CRT block fonts and static character arrays of the era for text apps. AutoCAD people were already used to the characteristic 'redraw and blink' stuttering behavior of other computer brands' graphic cards and high resolution displays, which typically displayed on their own additional monitor.
With its extremely high price point and crude software available Lisa was unable to win anyone over.
I've actually seen this before where a board would run all night with no problems but as soon as someone was sitting next to it in a pneumatic office chair, it crashed.
Drove everyone mad until someone figured out it was the chair itself.
tweet w/ video: https://twitter.com/royvanrijn/status/1214162400666103808?s=...
How does sitting on a gas lift chair generate an EMI spike (per the article)? The metal piston sliding an inch or two inside the cylinder collar is enough? Or is the hydraulic fluid itself metallic?
The article cites/links-to a whitepaper from 1993 on the topic of EMI producing furniture. I haven't actually read the whitepaper myself, but there might be more information in there....
The paper is mostly about measuring the magnitude of the EMI. They observe that dry air and metal legs are required, but not much more than that - I don’t think they mention gas lifts at all. The last few paragraphs have some guesses but they basically said that finding the cause is for future work.
1 reply →
ah thanks, had missed that.
Huh. I've heard of PEBCAK errors, but this is the first PEIC I've seen.
I actually see this on my home Dell 4K connected to a MBP13 via Thunderbolt to DP. It happens when I'm not using a chair (adjustable desk) but now I'm going to be very observant when sitting down. I'd lean towards my cable being the issue however.
I saw this as an intern working on internal dev boards for some wireless gear. A coworker had set up some debugging tool to trigger a pin on some condition and it was being set off whenever he sat down in his chair.
I don’t think we ever found a work around.