Speedo question

[Huzo]

I don' t know how to do the NGC thing so I'll just say Non Guzzi Content.
I'm inviting comment however well informed or otherwise, but I know KevM and a couple of others will have something to offer.
My mate has a 2002 BMW R1100 S that is beautifully maintained and as close to perfect as anything of that age that I've seen. It does however have a speedo anomaly that I'd like some thoughts on.
Starting from rest and accelerating normally, the speedo climbs in unison with the velocity and is very accurate, plus or minus 2 kph up to 100 kph.
However above that speed the indicated speed starts to flicker randomly both faster and slower.
He's into the rear bevelbox and has ascertained that the metal vane that passes over the magnet is a push fit onto it's register and we wondered if the vane is given to slippage and thereby giving an erroneous frequency of pulses and therefore generating a slower reading than actual.
This theory of course goes arse up when you see that occasionally the needle flickers up to an artificially high reading, this would necessitate the vane spinning faster than the pinion shaft on which it is mounted so that theory is obviously a nonsense.
All the connections have been meticulously checked for shorts and poor conductivity and have come up ok.
Given that the vane spins past a permanent magnet on the rear bevelbox sensor, I'm assuming that induces a current that is transmitted up the wire to the speedo, where that current is fed into a coil which in turn creates a magnetic force that deflects the needle..
Am I a mile off..?
Any thoughts as to why the system could behave normally up to 100 k's then begin to fail ?
Here's a shot of the bike in question..

[twowings]

You're too old for one...only for young guys! 

[Huzo]

Nice of you to share your private collection mate..

[tris]

If it helps my B11 did something like that when the speed sensor got water in it

[Murray]

I am assuming we are talking about an electronic speedo? so a magnet and a pickup sensor in the bevel box? Assuming he has gone through all the connections for the sensor and cleaned them its likely the pickup itself is succumbing to vibration and heat its basically a bunch of fine windings that the magnet induces a small current in and creates a pulse for the rotation of the wheel, there is also likely some electronic gubbins in it to regulate the size of the pulse and turn it into a square wave that the instrument micro processor will read. The only other possibility is the shielding for the sensor is rubbed through and its getting some interference from the ignition. if you know someone with a signal generator they might be able to fire a signal up the wire with the ignition on just to confirm.

If its a mechanical speedo then its likely the drag cup in the instrument is slightly warped and its loosing the eddie current as the tension on the spring increases.

[Huzo]

If it helps my B11 did something like that when the speed sensor got water in it

Thanks Tris, it does help a bit, but I think we have started to realise it's the PCB in behind the speedo, but we'll see.

[Huzo]

I am assuming we are talking about an electronic speedo? so a magnet and a pickup sensor in the bevel box? Assuming he has gone through all the connections for the sensor and cleaned them its likely the pickup itself is succumbing to vibration and heat its basically a bunch of fine windings that the magnet induces a small current in and creates a pulse for the rotation of the wheel, there is also likely some electronic gubbins in it to regulate the size of the pulse and turn it into a square wave that the instrument micro processor will read. The only other possibility is the shielding for the sensor is rubbed through and its getting some interference from the ignition. if you know someone with a signal generator they might be able to fire a signal up the wire with the ignition on just to confirm.

If its a mechanical speedo then its likely the drag cup in the instrument is slightly warped and its loosing the eddie current as the tension on the spring increases.

Thank you also Murray.
I will put that in with the other possible root causes and soldier on...

[Murray]

Just had another thought this is assuming its not a CAN bus system which I think BMW used a little latter in the effort of answering a question nobody asked.

[RinkRat II]

   Back in the day, soon after the Brick bikes came out '85 on they had a similar problem but more intermittent. The suggested fix was on the connectors to the board behind the speedo. It worked for a while. The final fix on mine was to take the connector pins and bend them slightly along with tightening up the female portion of the connector. A liberal use of vaseline and Voila. Also on mine the mount system for the speedo was very spongy so I replaced two of the four mounts with stiffer material. Less bounce. So far 31 years later still works well.

           Paul B

[Kiwi_Roy]

Is it one like this?
https://www.ebay.com/itm/99-BMW-R-1100-S-R1100S-SPEED-SENSOR-Z106-/351695119172
Or is it an ABS model?

It looks as though there is no way to adjust the gap.
What happens with many of these sensors like the early CARC bikes, the flexible cable is potted into rigid epoxy. the cable being flexible breaks contact with the epoxy creating a tiny gap that wicks water into the housing where it eventually shorts out the coil between the two soldered joints.
You might try warming it up for a few days then while it's nice and dry sealing it with some silicon sealant. the silicon remains flexible sealing water out.
Of course if it's already corroded through the wires it might stop working altogether when it doesn't have any moisture to conduct the signal.

We used to do this to very expensive pH probes that would be rendered useless in a few minutes, with silicon they would last for years submerged.

From the pictures it looks like the sensor is the usual small magnet surrounded by a coil.

[Huzo]

The thing that bugs the crap out of me, is that it performs beautifully up to 100 k's, then plays up above that every time.
Heat does not change anything either.
I'm thinking it has to be a computation error, if it was some anomaly in the mechanicals, I can't see how it would manifest itself so reliably @ the same speed every time.
The owner has had a toolmaker of undisputed quality, fit a couple of roll pins to the vane to ensure no floatation.

[tris]

When I fixed the sensor on my B11 I refused to part with the cash the MG wanted and went for a home brew solution using an industry proximity sensor thanks to the Aprilia RSV boys here http://apriliarsv.com/guides/16285-inexpensive-speedo-sensor-fix.html

I've found that I don't have it quite close enough to the trigger and the speedo goes a bit wappy at higher speeds +80 MPH ish

Too far away and it can miss a trigger too close and it can get a double hit

Can you adjust it closer/further from to the trigger thing?

[Huzo]

Just had another thought this is assuming its not a CAN bus system which I think BMW used a little latter in the effort of answering a question nobody asked.

I don't know what CAN bus means.

[Huzo]

When I fixed the sensor on my B11 I refused to part with the cash the MG wanted and went for a home brew solution using an industry proximity sensor thanks to the Aprilia RSV boys here http://apriliarsv.com/guides/16285-inexpensive-speedo-sensor-fix.html

I've found that I don't have it quite close enough to the trigger and the speedo goes a bit wappy at higher speeds +80 MPH ish

Too far away and it can miss a trigger too close and it can get a double hit

Can you adjust it closer/further from to the trigger thing?

I will investigate that Tris.
Very worthwhile suggestion

[Paul Brooking]

I don't know what CAN bus means.

It's what Bikers north of the great lakes use to get around in winter

[Huzo]

It's what Bikers north of the great lakes use to get around in winter

[Murray]

Control Area Network two wires run around the machine and they do everything power for the indicators lights fuel injection sensors etc. All good while its working but utter nightmare for the home tinkerer to diagnose if it plays up. Bit like BMW's electrically assisted servo brakes they had for a while stuck with all the complaints no feel overly sensitive loss of power = loss of brakes and then quietly moved away from it when they finally worked out it answered a question nobody asked.

[Huzo]

Control Area Network two wires run around the machine and they do everything power for the indicators lights fuel injection sensors etc. All good while its working but utter nightmare for the home tinkerer to diagnose if it plays up. Bit like BMW's electrically assisted servo brakes they had for a while stuck with all the complaints no feel overly sensitive loss of power = loss of brakes and then quietly moved away from it when they finally worked out it answered a question nobody asked.

Thank you Murray.
I never knew.