UniChris

So I had this little coin cell wired cycle computer, that's always been a bit flaky. Last summer I broke the cable and so the weekend before my first Century I took apart the base and re-wired it, and it seemed to work.

Unfortunately, the next morning when I twisted it into the holder in the 5am dark, the screen cut out. Tried it a few more times, then undid the ties and electric tape, stuffed it in my bag and rode un-instrumented.

Just found all the pieces again tonight, and tried to fix it. Same behavior - works fine, twist it in the base and the screen cuts out.

This was really puzzling. I tried putting tape over the electrical contacts, so it shouldn't even "know" if it's in the base - but same issue.

Squeezing it in various ways couldn't cause the problem.

Re-bent the battery contacts, taped some cereal box cardboard on the foam battery pusher in the coin-screw cover, nope, still has the problem.

Noticed the watch crystal inside might be touching the battery contact, so carefully pushed it out of the way, but no dice.

Finally gave the little spring-contact pins on the back that connect to the base a good solid push individually, and... gotcha! the screen cut out.

Looking through the battery opening, the little springs that push out the nubs sit on the circuit board, and one looked to be just sometimes touching the flat ring that is the base of the battery contact.

Couldn't get anything under the spring through the battery door, so got a knife and cracked the plastic at the mold line all the way around. Put a couple layers of tape between the battery contact and the spring. Re-positioned the spring. Taped the watch crystal, too.

Then I had to spend a good 10-15 minutes trying to find the superglue, that I had just seen last week.

Victory is Mine!

(in all seriousness, I'm sure it will find another way to break tomorrow)

Aubergine 

Great story!

wphamilton

You win! But it would have taken a minute to buy a new one on Amazon for $12.50 and your time is worth more than that. Clearly, given your troubleshooting skills.

UniChris

I actually looked but wasn't finding any wired ones and I've been cautioned against wireless as my wheel to bar distance is larger than usual.

But it was more a draw than a win. I had a speedometer for the first segment of the ride to tell me I've been being lazy, but then it quit and I realized the base had come apart again. Kind of think I should make a new semi permanent mount with the 3d printer which can also make it less of a squeeze between that, the bell, and the brake lever.

I keep meaning to make one from scratch with the software features I want, but then neglect to start the project.

wphamilton

You got me there, because I kept that same project on the back burner for a year (or more ) before finally setting one up. Arduino with a gyro in the wheel, BT to my phone displaying with an Android application. It works, but I've never finished it off because the gyro cannot detect rotational acceleration beyond that corresponding to 26 mph.

UniChris

Wow, that's actually pretty close to some of what I was thinking. I was probably going to use an nRF51822 BLE chip for the brains as well as radio. And may well stick with the magnetic reed sensor. But a six axis IMU sensor does give the interesting possibility (for me) of tracking pedal-induced wheel wobble, too. I've sort of toyed with the idea of logging every last wheel revolution in a 16 MB SPI flash, and then post processing to split into riding vs walking vs. stopping.

Topping out at 26 MPH is not personally an issue - as I freak out at about 13. But I wonder, can you switch over to using radial acceleration instead of gyro rotation?

The irony is that this is exactly the kind of stuff I get payed to do, but then I don't get started on the effort to do it for fun, and instead take the slowtime/downtime and go for a ride. Was going to do it over the winter, somehow that didn't happen.

wphamilton

You can track some other interesting things also.

You can but it's trickier in that you have to subtract other accelerations - in fact I think, going the other way around, I can get instantaneous slope by subtracting the radial acceleration from the accelerometer readings. When you're not turning anyway. But you lose the precision and sampling frequency when you use the radial acceleration (gravity vector flips) for speed which was what I wanted it for, so it's not very exciting to do that.

But if all I want is more-or-less speed, https://www.amazon.com/Bell-Dashboar.../dp/B01BNFZ0H8 I don't know why most of them are wireless on Amazon, which sucks IMO.

UniChris

Oh, thanks! That's probably going in the "I still won't get my act together before next weekend" order.

I have put some thought into how to do wireless right, specifically counting revolutions and timing recent average intervals at the transmitter and sending that, so it doesn't matter if a packet or two are missed - attaching the wireless remote display to my helmet gets tempting.

Have you considered just adding a reed switch to yours to augment the in-wheel sensing? You could probably even do it backwards, with the magnet on the fork and sensor on the wheel...

wphamilton

Somehow it's hard to accept adding a reed switch when you're trying to be clever with gyros and accelerometers, but that's a really good idea to get over that last hurdle, and pretty easy compared to the ideas I've been avoiding. I even already have that implemented using a different arduino, that I was just recording to a memory card with and not displaying anything. I could just copy-paste my code, the software de-bounce and smoothing already all worked out, and plug the reed switch in. Thank you for prodding me with that.

UniChris

The Bell 150 just arrived - think it is actually the same firmware as the Sunding except without the somewhat convenient backlight. But better physical form.

You've got me thinking Arduino now too, generally I work on things that need to run for half a year, but if I can keep the current under say 20 milliamps I'm willing give it a 400 mAh battery for the present... The display I want to use will probably double that, but I can make it active for 10 seconds at the push of a button or something.

Somewhere around I have a sketch where I just experimented with mocking up what data I wanted to put where on the screen, I should dig it out and see if I can make it actually work...

79pmooney

On topic "man vs. computer" but not bike computers - read in the NY Times today that software is being developed that can look for the trademark markers for lung cancer in the data acquired from CAT scans. Slightly better accuracy than trained human eyes but with the capacity to go through the massive data acquired by the scan that no human could process.

I had my first and second CAT scans in 1977 on a very early machine that allowed the neurosurgeon a view of my brain that not long had never been seen. At the time my dad's company was making full body bone scanners to give much cruder pictures showing cancer in I presume much more developed stages. I owe my CAT scan to my dad's knowledge of them through his work.

Best news I read today.

Ben

wphamilton

I'm just going to buy a $10 android phone and use that for the display. I'd have to charge it up pretty frequently though.

fietsbob

The Original Avocet computers were a perfect shape to skip several times ,
across the surface of a pond,

before sinking ..

UniChris

Bell computer was fun for the one ride it lasted on Saturday. Come Monday, no registration of rotation. And then I noticed the base that the zip ties goes through had cracked. I'm sure it has nothing to do with my bar ends occasionally impacting the trail, or with the fact that I remove the handlebars for transport, undoing and re-doing the electrical tape holding the cable and sensor to the frame ;-)

So now I have two computers that need re-wiring and re-visioned bases.

Entire span of handlebar to the outsides of the bar ends is under 5" so things are tight (I prefer to go around the bars rather than over them) but mounting the little ding bell upside down cleared some space, now what I think I need to do is fashion something that goes between the bar ends and has the computer sheltered between them in their plane and the bell underneath with lever tab facing forward where I can get it with either fingertip. Also will probably give in and drill a tiny hole in the bar to permit internal routing of the cable.

ChesterWallace

Nice Story.