Eddy current position sensor circuit

5 replies · 3K views · last reply
Rx7man
Messages
427
Reaction score
194
I started a project, and where I thought I has a potentiometer position feedback, best I can tell I have an eddy current position sensor, which I'm absolutely not familiar with

The application is a diesel engine injector pump (retrofitting to a different engine, it was from a Komatsu V12, no, I can't look at the original ECU for ideas)

There are two big solenoids (fed by brown wires) (one at the top for changing the timing which does not have position feedback, one below it for varying fuel quantity, which does have position feedback, the plunger of the solenoid moves outward (toward camera))
on the right side there is a dual coil sensor system, the lower one is picks up a rotating nub on the shaft (Max9924 should interface that just fine), but the upper part has a copper ring that slides up and down it, and to the best of my knowledge, that's an eddy current style sensor
Interior of a diesel injector pump housing showing a dual-coil eddy current position sensor with a copper ring on a lever, a shaft-mounted sensor, and wiring with brown and green leads.


different angle on the sensor, right part is the shaft position nub sensor, left part has the moving copper ring
Photo of a diesel injector pump's electrical sensor assembly, showing a black coil housing with a copper ring sensor and brown-insulated leads, beside a white plastic rotating-nub position sensor on a metal plate.


Here's the copper ring
Photo of a diesel injector pump with a solenoid plunger, a copper sliding ring on a dual-coil sensor housing, and a rotating shaft nub sensor on the right side.


Shaft nub
Photo of a diesel injection pump housing with a round metal sensor cover, a copper ring visible on its left side, and a shaft nub with a small hole on the right.
The only circuit diagram I've been able to find, in miserable resolution and with poor details is figure 8 in this article, but it tells me nothing of the output frequency or how it changes, but from my reading it would be between 500khz-1mhz, which is far faster than I'd want to interface to a microcontroller interupt... maybe I could put a 14 bit binary counter inline with the output to get ~50-100hz output?
https://www.fierceelectronics.com/components/designing-and-building-eddy-current-position-sensor
Schematic of an eddy current position sensor oscillator: a target, coil L(x) with Rl, capacitor Cp, Rs = 10K, 10 Mega Ohms feedback resistor, two 74LVC04 inverters, and a square wave output.


I have a function generator and a 2 channel scope and was playing around feeding a sine wave to the coil and watching the waveform on the scope, I never did get much for phase shift, but at 75khz seemed to be the resonant frequency in my setup, and at that frequency, moving a copper loop over the sensor did have a large effect on amplitude... my peak-to-peak amplitude went from 1.64V down to 160mV, (when feeding 5V) which is a fair dynamic range, I could increase the feed voltage a bit, rectify it, and put into an ADC... does this sound like a reasonable way to do it?
 
Rx7man said:
The application is a diesel engine injector pump (retrofitting to a different engine, it was from a Komatsu V12, no, I can't look at the original ECU for ideas)
Too much complexity, alternatives and information in the picture, with no arrow to the item you are actually referring to.
Is it an in-line pump? Which shaft, there are many possibilities?

If you insist on using the Komatsu V12 pump, then you will need to get the non-obvious information from a Komatsu technician or engineer. Alternatives would be to find a running Komatsu V12, then measure the signal frequency and amplitude, or consider replacing the sensors completely with alternatives.

Who manufactured the injection pump?
What is the part number of the pump?
What engines was it used on?
Why not use a common rail with electronic control?
 
Baluncore said:
Too much complexity, alternatives and information in the picture, with no arrow to the item you are actually referring to.
Is it an in-line pump? Which shaft, there are many possibilities?

If you insist on using the Komatsu V12 pump, then you will need to get the non-obvious information from a Komatsu technician or engineer. Alternatives would be to find a running Komatsu V12, then measure the signal frequency and amplitude, or consider replacing the sensors completely with alternatives.

Who manufactured the injection pump?
What is the part number of the pump?
What engines was it used on?
Why not use a common rail with electronic control?
Sorry for not making things more clear.. so for now, let's only talk about the eddy current position sensor.

This is the moving part
Close-up photo of a fuel pump's internal mechanism: a rusty brown metal lever with a square opening is circled in red, next to a cylindrical metal component with a shaft.


and it slides over this part of the coil/inducer
Photo of a Bosch inline pump mechanism with a red box around a dark plastic sensor housing and a coil, surrounded by orange wiring and steel mounting screws.


It's a Bosch inline 6 pump.. I was able to find specs on fuel delivery, etc, but nothing on how to control it.. This pump design was used in the later Komatsu SA12V140 and Cummins QST30 engines, and I've heard of similar pumps used in Scania trucks... unfortunately non of any of them are to be found around here.

I have toyed with the idea of a potentiometer type sensor, but from my reading, the eddy current system is extremely accurate over large temperatures, and being non-contact, it has nothing to wear out which is kinda nice.. the interface is my only stumbling block
Close-up photo of a black metal nameplate on a grey cast pump housing, reading BOSCH, 4892400003H, 0 402 796 207, PES6H120/320RS1008-1, 309 4347 0203, Made in Germany.
 
This video found by Google,
https://www.google.com/search?&q=bosch+pes6h120
Also check other links on the youtube page.

From the descriptions there and in the video below, I gather that:
  1. The Red lead of the coil is the center tap, and I suspect the Common
  2. One half would be driven with AC (or chopped DC) voltage
  3. As the Copper 'Reluctance Loop' moves around the core, it acts as a shorted turn, reducing the voltage induced in the other half of the winding

The circuit you supplied in the first post seems to be variation on the above.
It appears to be an oscillator with the frequency dependent on the coil reactance; in other words a position-to-frequency converter.

  1. The first stage (pins 1&2) is an AC coupled inverting analog amplifier and the second stage (pins 3&4) is acting as an inverting comparator to give logic-level output.
  2. With two inversions, there is positive feedback thru the 0.01uF and Rs 10K
  3. As the Copper loop moves and changes the effective coil inductance (Lx), the tuned circuit composed of L(x) and Cp changes resonant frequency

Anyhow, that's my take on it. Looks like you have some computer work to do if you want to use that pump!

Let us know how it goes... and Have Fun!

Cheers,
Tom
 
Tom.G said:
This video found by Google,
https://www.google.com/search?&q=bosch+pes6h120
Also check other links on the youtube page.

From the descriptions there and in the video below, I gather that:
  1. The Red lead of the coil is the center tap, and I suspect the Common
  2. One half would be driven with AC (or chopped DC) voltage
  3. As the Copper 'Reluctance Loop' moves around the core, it acts as a shorted turn, reducing the voltage induced in the other half of the winding

The circuit you supplied in the first post seems to be variation on the above.
It appears to be an oscillator with the frequency dependent on the coil reactance; in other words a position-to-frequency converter.

  1. The first stage (pins 1&2) is an AC coupled inverting analog amplifier and the second stage (pins 3&4) is acting as an inverting comparator to give logic-level output.
  2. With two inversions, there is positive feedback thru the 0.01uF and Rs 10K
  3. As the Copper loop moves and changes the effective coil inductance (Lx), the tuned circuit composed of L(x) and Cp changes resonant frequency

Anyhow, that's my take on it. Looks like you have some computer work to do if you want to use that pump!

Let us know how it goes... and Have Fun!

Cheers,
Tom

I've studied that layout and those vids a fair bit, this only uses a single coil for the sensor, but it shares a ground with the pump position sensor, though they're magnetically isolated.. the blue side circled here is a completely different sensor
Close-up photo of an automotive transmission internals with a black plastic sensor housing; red and blue hand-drawn outlines mark two separate sensor components, and copper-colored wires run across the metal casing.


Here's the circuit I drew up in Spice, r1-3, c1-3 smooth the square wave to about a sine wave, U1 and resistors provide fixed impedence, L2 is the sensor, and the rest of the lower right side takes the peaks of the square wave, smooths them, and will feed the ADC

After about 2 days of fighting with the microcontroller, poring over the datasheets, etc I finally was able to get it to output an adjustable frequency square wave, though I may yet go with a LTC1799 oscillator chip... At this point I can at least breadboard something together to test it...Also working on a current sensor circuit that drives the control solenoids
Simulation schematic of an eddy current position sensor circuit: a 5V supply, an RC low-pass filter of R1-R3 and C1-C3, op-amp stages U1 and U2 (OPA353NA), a 25mH sensing inductor L2, a 1N914 diode D1 and a function generator XFPG1.