Need to find a compatible magnetic speed sensor for speedometer

Alec_t

Joined Sep 17, 2013
10,715
Tried to run without any resistor at all and it wouldn't work.
That implies the speedo input doesn't have any pull-down resistive path of its own. So keep the 47k (or 10k) in place, since it provides a path for the capacitor's charge/discharge current.
The link to the Hall sensor data is helpful. It shows the sensor has an internal pull-up resistor Rpu (as we assumed), and gives the max sink current as 20mA. The 3k3 resistor limits the current to ~4mA worst case, so you're well within the limit.
 
Last edited:

Thread Starter

MB107

Joined Jul 24, 2016
103
OK I think I now have everything figured out. The circuit below shows what appears to work the best. But I have a few questions while I wait to get this car back on the lift.

1) There are many capacitors to choose from. I choose this one TDK FA26C0G1H104JNU00 because it said used in the Automotive industry and somewhere in dealing with motors I heard the capacitor voltage rating should be at minimum 3 times the motor voltage. So this is a 50 V 0.1uF ceramic capacitor. May there have been better choices?

2) I assume that in the event that the capacitor shorts out I am going to get full voltage to the speedometer. How likely is that and how likely is it to damage the speedometer if it should happen.

3) I have attached a couple of pictures of how I plane to get these components into a box. I have been considering taping the Cap and 3K3 resistor and just putting them in series inside a covered blade type fuse holder like the one shown in the last picture. And then putting the 47K resistor into a second micro blade type fuse holder. Dose this sound like a viable solution, do I need to be concerned with heat build up inside the enclosed boxes.


Picture.jpg Hall Effect Circuit 1.jpg Hall Effect Circuit 2.jpg Hall Effect Circuit 3.jpg Fuse Holder.jpg
 

Alec_t

Joined Sep 17, 2013
10,715
1) That cap looks fine.
2) Even if any/all the components shorted I now think there's no chance of harm to the speedo, given that it's designed to cope with a VRS at the input. VRS voltage, as you know, is AC which can get to 100V+.
3) Good idea to use the fuse holders. I presume you've checked their internal dimensions are sufficient. Now that the circuit has been reduced to just three passive components you could dispense with the circuit perf board. The heat generated by the circuit is negligible - only a few microWatts, so no problem with totally enclosing it or even potting it in epoxy.
 
Last edited:

Thread Starter

MB107

Joined Jul 24, 2016
103
The big fuse holder shown has plenty of room for a cap and resistor. If I can pot them then I should be able to get both into a small standard size fuse holder. The real problem might be replacing the 10ga wire with 20 ga wire.
 

Alec_t

Joined Sep 17, 2013
10,715
If it were me I'd probably just solder the 3 components as required, as compactly as possible, then pot the lot in a moulded 'box' of epoxy with emerging 20ga fly leads having suitable crimp-on connectors.
 

Thread Starter

MB107

Joined Jul 24, 2016
103
Very successful. Cars speedometer is smooth just like Mercedes made it. The fuse holder came in and it was way too big so I went back to my old circuit board plan, which worked very well.

Many Thanks to Alec_t for his contributions to this task.

The next task will be calibrating the speedometer I will start a new thread for that so stay tuned.

Compared to large fuse holder.JPG Making The Circuite.JPG Soldered Joints.JPG With Cap Installed.JPG Wiring into the PN 009 545 53 28 Plug Housing.JPG Plugged in.JPG
 

Thread Starter

MB107

Joined Jul 24, 2016
103
Just be sure to mark what it is, so the next guy doesn't replace it with something that isn't made for doing this.
Not really possible since it's soldered directly to a non stock sensor. That is exactly why I opted not to use a relay box with a plug in socket.

It would have been perfect but????
 

Thread Starter

MB107

Joined Jul 24, 2016
103
Not out of the woods yet. I finally got the car on the highway. Above 80 MPH the speedometer just shuts down. I'm going to have to find a way to spin this wheel on the bench at 5000 RPM to represent the full 170MPH range of the speedometer. I believe my drill press will spin that fast but its shaky and has no way of feeding sensor position. Also I will need to make a means of chucking it up.
 

Alec_t

Joined Sep 17, 2013
10,715
Above 80 MPH the speedometer just shuts down.
Oh. What is the ~ pulse rate at 80mph (or at some more convenient known speed, e.g. 40mph)?
I'm going to have to find a way to spin this wheel on the bench at 5000 RPM
Not necessarily, since you have a pulse generator. It should be possible to knock up a test circuit using that, to judge what the output of the combination of the Hall sensor plus your built circuit looks like at 180mph. Clutching at straws now, but the internal pull-up resistor of the sensor may be limiting the capacitor charge current a bit too much.
Any chance you could sketch the interconnections of those components visible in the speedo pics in your other thread? It would be nice to have some idea of what the speedo input stage circuit probably looks like.
 
Last edited:

Thread Starter

MB107

Joined Jul 24, 2016
103
Oh. What is the ~ pulse rate at 80mph (or at some more convenient known speed, e.g. 40mph)?
Pulse rate at 80 is ~ 175 Hz. I'm Not sure but I think the problem might be that the tone wheel is really designed for a VRS sensor. The VRS system gives off short pulses of +/- Voltage with long duration of low voltage in between. The hall effect is giving me wide + voltage with short - in between. The current tone wheel is 90% open and 10% closed. I may need to make one 50/50. Those are pretty hard to make so if you have an electrical solution I'm all ears.
IMG_20190217_240343382.jpg
Oh. What is the ~ pulse rate at 80mph (or at some more convenient known speed, e.g. 40mph)?

Not necessarily, since you have a pulse generator. It should be possible to knock up a test circuit using that, to judge what the output of the combination of the Hall sensor plus your built circuit looks like at 180mph. Clutching at straws now, but the internal pull-up resistor of the sensor may be limiting the capacitor charge current a bit too much.
Any chance you could sketch the interconnections of those components visible in the speedo pics in your other thread? It would be nice to have some idea of what the speedo input stage circuit probably looks like.
I was actually thinking about that on the way into work. It's a really cheep Chinese unit with directions that are practically in Chinese. Not sure if it has the ability to make an uneven square wave like the hall effect but it has an option to be able to get wave forms from an excel file. Not bad for $65.00 on Amazon. Hopefully I will be able to figure it out.

I was thinking that a different capacitor might allow voltage to bleed off faster and appear as a more symmetrical wave.

Fortunately I have a spare hall effect sensor. I will set it up and get data at various speeds on the lathe. Then I can take apart the speedometer and try to figure out how its made.
 

Alec_t

Joined Sep 17, 2013
10,715
If you have a spare cap of the same value as the one in circuit, then connecting it in series with the present one will effectively halve the total cap value and shorten the pulses. Putting a 10k (or less) resistor between the sensor output and the 12V rail will also shorten the capacitor charge time. I'd try both mods.
The 80mph pulse rate figure should suffice to play with simulation.
You shouldn't need to take the speedo apart. The tracks on its pcb are hopefully visible as is, so you could sketch how the Rs/Cs/D are interconnected.
 

Thread Starter

MB107

Joined Jul 24, 2016
103
If you have a spare cap of the same value as the one in circuit, then connecting it in series with the present one will effectively halve the total cap value and shorten the pulses. Putting a 10k (or less) resistor between the sensor output and the 12V rail will also shorten the capacitor charge time. I'd try both mods.
The 80mph pulse rate figure should suffice to play with simulation.
You shouldn't need to take the speedo apart. The tracks on its pcb are hopefully visible as is, so you could sketch how the Rs/Cs/D are interconnected.
Unfortunately trying anything on the car is extremely difficult installation so the simulation route seems to be the best approach. Speedometer goes up to 170MPH but its limited to 155 so I'm half way there. I do have spares of every component I purchased so lots to play with.
 

Thread Starter

MB107

Joined Jul 24, 2016
103
OK the results are in. It turns out my wave form generator has the capability to change the waveform just as I need it using a feature called duty cycle. As I suspected it does appear that this system is looking for waves with shorter on times and longer off times. Doubling the caps had no effect infact it was slightly worse.

Using 3.3K resistors to ground did help the situation. They increased the usable duty cycle range from 64% to 76%. I need to get to 89%, So I was wondering what would happen if I used say a 500 Ohm resistor.?

Capacitor Comparison.jpgDuty Cycle Comparison.jpg
 

Alec_t

Joined Sep 17, 2013
10,715
So it looks as though you need to stretch the negative pulse duration.
This re-jig of the adapter circuit might do it (if the speedo input is something like my assumption). The simulation is for an input of ~400Hz.
SpeedoSensorInvert7.PNG
C2 can discharge quickly via D2 and C1 when the Hall sensor output goes low, but hopefully charges slowly via the input of the speedo (which seems to have some pull-up arrangement). If you can cannibalise some electronic scrap you might find a 10nF cap (or thereabouts) for C2.
 

Thread Starter

MB107

Joined Jul 24, 2016
103
That does appear to be the problem but how much appears to be function of that resistance to ground.

The attached file shows 6 solutions that are at the extreme limits of duty cycle before they fail. The left side simulated the system with the smallest window opening that will work successfully at 400 Hz. The right side simulated the system with the largest window that will work at 400Hz. For reference my tone wheel has windows open 80° over 90° of arc, equivalent to 89%

The top two traces were using a 47K resistor to ground the, middle two 10K and bottom two 3.3K. I thought I might find some in common constant between them all like the system needs a minimum of some time at negative values, but I'm not seeing it. The only thing I am seeing is that 40% duty cycle is about optimum.

The other thing of interest I noticed is that if you look at the traces on the left with low duty cycles there is almost no zero crossing it almost looks exactly like the raw sensor output with a decreased amplitude. Previous assumptions were that a zero crossing was required and when the circuit was adjusted to do so it did work, but the left side data is showing that extremity little if an zero crossing is required at low duty cycles.

So the conclusion I'm coming to that a 40% or 50% tone wheel might be the optimum solution and it may in fact eliminate the need for any additional circuitry at all with the exception of the capacitor to dissipate the voltage when stopped.

I may try the circuit you have proposed for educational purposes but at this point I will commit to cutting a new tone wheel. I will need to get a 10n cap as well as the 220K resistors. While it may work, its becoming a little over complex to fix the real problem.

This will probably take some time, but I will get back when I close this solution.

In any case, I certainly appreciate all the help you have given me.

400Hz Extream Duty Cycle.jpg
 

Alec_t

Joined Sep 17, 2013
10,715
Depending on the speedo's input characteristic, the 'adapter' in post #95 (it's only 4 components :)) could convert a 90% duty-cycle 400Hz signal from the sensor into a ~30%-40% duty-cycle output for the speedo. Would that avoid the need to re-do the tone wheel?
BTW, you don't need the 220k resistor. That's just part of the model of the guessed input stage of the speedo.
We might get a better handle on your (spare) speedo input if you could bench-rig temporary connections to do the following (without the adapter and sensor):
1) connect a 47k resistor between the speedo input and +12V,
2) supply the speedo with its usual 12V,
3) measure the speedo input voltage (relative to ground),
4) connect the 47k between the speedo input and ground instead,
5) measure the speedo input voltage again.
 

Thread Starter

MB107

Joined Jul 24, 2016
103
Depending on the speedo's input characteristic, the 'adapter' in post #95 (it's only 4 components :)) could convert a 90% duty-cycle 400Hz signal from the sensor into a ~30%-40% duty-cycle output for the speedo. Would that avoid the need to re-do the tone wheel?
BTW, you don't need the 220k resistor. That's just part of the model of the guessed input stage of the speedo.
We might get a better handle on your (spare) speedo input if you could bench-rig temporary connections to do the following (without the adapter and sensor):
1) connect a 47k resistor between the speedo input and +12V,
2) supply the speedo with its usual 12V,
3) measure the speedo input voltage (relative to ground),
4) connect the 47k between the speedo input and ground instead,
5) measure the speedo input voltage again.
I will try what you said. I think I understand. I will report back.

The decision to cut a new wheel is not based on one car. This is not a 1 off sensor. I may only make 10 of them but its getting to be time for me to get the other 9 so I might as well redesign the slots and order 10 with the new configuration.
 

Alec_t

Joined Sep 17, 2013
10,715
Can you sketch the new slot configuration? From your comments in post #96 I think you're hoping to get a Hall sensor output with ~ 50% duty cycle? That won't necessarily result in a ~50% duty cycle being achievable from the adapter output. The problem is that the capacitor will still be needed to get negative pulses from a Hall sensor, and the pulse shape/duty-cycle is governed by cap and resistance values and the speedo input characteristics. Those characteristics are presently unknown, but may be clearer after the post #97 test.
In the end you may need a more sophisticated adapter than the one we've considered up to now, as the present simple one is never going to be able to give a ~constant duty-cycle output over a 10:1 input frequency range, if that's what the speedo is looking for.
Before making new tone wheels I suggest you try the post#95 four-component adapter.
 

Thread Starter

MB107

Joined Jul 24, 2016
103
The old wheel had 80° window. 80/90 = 89% it produced a signal that appears as ~80% duty cycle as shown in the second picture.
The wheel proposed below has a window opening of 45/90 = 50%. I'm expecting that one to produce a equivalent duty cycle of ~45% which should be optimal.

Capture.JPG


Time Scale.jpg
 
Top