Magnetic reed switches can have maximum operating frequencies in hundreds of times per second, which on a bike wheel would be hundreds of miles per hour. So I think that having the magnet at the rim will make no difference whatsoever with respect to the magnet going by too fast for the sensor to read. If anything, I'd expect better performance at the rim because the faster attenuation of the magnetic field would mean less debouncing issues.

Reed switches have several advantages over hall effect switches, less power draw, cheaper, simpler, unaffected by electrostatic shock. The hall effect might last forever, but are more sensitive to current in wires, ESD and so on. Since a reed switch can last for billions of cycles, that's not really a problem. Convert a billion turns of the wheel into miles, and you'll see what I mean.

I think that the orientation of the magnet is a bigger concern than its position.

Wheel magnets are plenty accurate, no need to worry about where they are placed. I always had accurate mileage when I used them. I now use rotation sensors and they are just as accurate, especially when combined with a GPS head unit. It can calculate wheel size based on GPS miles/revolutions, and just take it from there. I switch my sensor and head unit from a 700c road bike to a 26" MTB and I get the same miles on either, and the same MPH, both based on the rotations, not the GPS, without doing anything. And these miles equal the miles my old magnet system would report. I had them running together for a while just to be sure and it never varied by more than .1 over a 50 mile ride.

Don't know of any current bike speedometers that use a Hall effect sensor, but Avocet made some that did back in the '80s. I still have one of theirs which uses a little ring mounted on the spokes at the hub instead of the single magnet used by the reed-switch models. One advantage of the Avocet approach was that it got multiple readings per wheel revolution and therefore a bit faster response at slow speeds.

The little Avocet pickups that worked with the ring magnets were just coils of wire. They were nice, simple, and somewhat robust (the connectors that slid onto the pickup's pins were not all that reliable). I'm still using these Avocets on three vintage bikes, and have a new Avocet 20 in the box. Very minimal, and a bit of a conversation piece.

These old Avocets do go through batteries quickly... mine last about a year. Maybe that's why there was an assumption that it used hall cells?


Steve in Peoria

I never thought about the sensor tech in an avocet, but a hall would just require a single magnet just like a reed switch. I used my avocet for years, the ring eventually got to be too much of a pain and I stopped using it. I hope I threw it out, but it might be in a junk box somewhere.

GPS speed is good enough for me, I know it's unreliable, but close enough. Don't the new ones use an accelerometer?

I use a Sigma bike computer on my inlines, wheel dia 74 mm.
I chicken out shy of 40 kmh/25mph, at which speed the computer is still doing fine.

Kinda, sorta, maybe!

It's a hypothesis, and a very reasonable one. Another might be that the wheel under such loads it experiences at that speed deflects and deforms more at the outer diameters. So maybe not so much the time of engagement between the magnetic flux and the reed switches, but maybe more that the magnet for that revolution was further away from the sensor.

But who knows. I'm not going to go out and test it. But likely there are other possible reasons too.

Jobst Brandt designed that Avocet computer. He claimed it was more accurate than the kind with the single magnet. He said the ring created a sine wave, and the computer counted the number of times voltage went to zero. I don't know enough to say he's right or wrong, but it sounds reasonable to me.

that's interesting that he designed it. The difference in accuracy is questionable in my mind. And the disadvantages of the design from a mechanical viewpoint outweigh them.

Maybe those are the reasons the design didn't stick around. Brandt reported that at high speeds, the one-magnet design would give spurious spikes, which is why you see people claiming speeds of over 50 mph. I don't know if he's right, but even if he is, it happens rarely enough that it's not a big problem.

only time I've been over 50 on a bike was with his computer. I'm pretty sure it was correct, long steep straight stretch on a mountain.

I don't see how more data makes things better on a cheap computer. But I'm pretty sure it went away because it was a lot more expensive than a magnet. At least in manufacturing terms.

50 mph on a bike!? You have more manhood than I do or will ever have. ;)

Regarding the subject of why Avocet didn't survive in the market, I've got some thoughts.....

I like the old Avocet computers, partly for nostalgia and partly because I like the way that the pickup and magnet are unobtrusive. Call me shallow, but I find the pickups on conventional "magnet & reed switch" computers to be less pleasant visually. Here are some examples of my bikes with Avocets, and one with a Cateye.....

Olmo
https://c1.staticflickr.com/5/4514/3...398ceb_z_d.jpg


Raleigh Team
https://c1.staticflickr.com/1/591/21...e24bc1_z_d.jpg


Raleigh International
https://c1.staticflickr.com/9/8120/2...6d1776_z_d.jpg


Hetchins with Cateye Mity, using a reed switch pickup
https://c1.staticflickr.com/9/8243/2...73636d_z_d.jpg


But.... (and there's always a "but")... I would say that the Avocet computers had some tech problems that contributed to their failure in the market.....

1. poor electrical connections at the pickup.
It looks like they used some cheap Molex stamped, tinned, connector sockets to connect to tinned wire at the sensor. As a guy who both spent years fixing aircraft electronics, and then later a career as an electrical engineer, I would say that connectors are a major source of problems on vehicles. I've had my share of troubles with the connectors on these Avocets, and just last year fixed a problem on the International by just soldering the wires to the sensor pins and got rid of the connectors. Problem solved! :)

2. poor electrical connections at the computer bracket.
It looks like they used the same family of cheap Molex contacts, but used the pins at the mounting bracket. In fact, it's fairly clever. Unfortunately, those pins weren't built for all of the mate/de-mate cycles that are typical for computers. The plating wore off quickly, resulting in the typical corrosion that degrades the signal strength.

3. poor housing design... on the model 30, at least
The top and bottom of the housing snapped together, creating a compression fit of the circuit board, LCD (liquid crystal display), and the "zebra strip" that connected the board to the display. After some period of time, the plastic housing relaxed a bit, reducing the compression and causing the display to loose contrast or causing digits to fade away entirely. A rubber band around the top of the housing takes care of this, however. Later designs used ultrasonic welding of the two halves of the housing, which appears to have fixed the problem.

4. poor battery life.
The battery lasts about a year. Compared to other computers, such as the Cateye Mity's that I've got on other bikes (about a 7 year battery life), this is pretty poor. It would be interesting to know the reason for the poor battery life. Was it using bipolar circuitry instead of CMOS? Is it related to the sensor technology?

5. The gap between the magnet ring and sensor has to be kept fairly small.

6. Fitting the magnet ring to the hub wasn't as trivial as sticking a magnet to a spoke.
I recall that there was large ring with 4 tabs that was 52mm in diameter that was used to clip onto high flange hubs (this is what I have on my International). For my bikes with small flange hubs, there is a smaller ring (48mm) with 3 tabs that is used. It works well with my 36 hole hubs. Not sure if it works with 32 or 28 hole hubs, since you can't divide 32 or 28 by 3. I think that was when you used the eyes on the ring, and you zip-tie'd the magnet ring to the spokes. Yeah... not a very easy way to attach the ring to the hub.

These are just the problems that I've experienced with the Avocets. Like a lot of quirky technologies, I've found fixes to most of the problems, and have just gotten used to the rest. Clearly, I don't represent the average bike computer purchaser from the 80's and 90's. :)

Steve in Peoria

if you go fast enough you can achieve a time warp

I had an earlier version of the avocet, I want to say I bought it in '85? The ring was much bigger, probably more than 3" in diameter. Pretty sure I got rid of it even though it still worked.


actually was faster than that. Then I realized that if something happened, the results would not be good and sat up.