Gallons Per Hour Error?

[consaka]

Originally Posted by dcb

do you have a picture of the sensor and/or part number?

yep.. both are in the link I gave you in the previous message.
csla2cd is the part number for the one I have.

[dcb]

Ah, missed that.

So the lead going to the injector, it looks like you will have to "break" it to get it through the hole. I would say wrap that injector lead around the sensor as many times as you can stand it, the more it looks like this, the stronger the magnetic signal to the hall sensor should be when current is flowing through the injector:


Or just wrap some appropriately thick magnet wire around the sensor and put it in series with the injector lead.

[consaka]

Originally Posted by dcb

Ah, missed that.

So the lead going to the injector, it looks like you will have to "break" it to get it through the hole. I would say wrap that injector lead around the sensor as many times as you can stand it, the more it looks like this, the stronger the magnetic signal to the hall sensor should be when current is flowing through the injector:

Or just wrap some appropriately thick magnet wire around the sensor and put it in series with the injector lead.

Aye, hooking the sensor lead up isnt my issue.. thats easy.. My problem is the supply. I dont want to burn the sensor up and it says 20mA max current to the sensor. the sensor has to have power to it before it will read the current in the injector wire. It is a three wire sensor similar to TPS sensors.

I know my battery is rated at over 600cold cranking amps. Im rusty on some of my electronic terms but it would seem to me that 600 amps exceeds the 20mA rating,,,,, just a little. To my way of thinking the supply should come on with the accessory link when you turn the key. I would really rather test this on the bench first though. Either way if my thinking is correct I need to limit the amps to the csla2cd first. not sure how. resister? fancy diode?

[dcb]

Well, I know that they hook up 5 volts straight to them, no biggie (I will be doing this on my e-bike project shortly):



And the out is going to the high impedence mpguino connection, so not much worries of high current there (<0.00024 amps at 12 volts).

If I were worried about it, I might try hooking up the sensor with 5 volts at the source with a millampmeter first and if it is less that 1/3 of max current then assume I could hook it up to 12 volts?

[consaka]

Originally Posted by dcb

Ah, missed that.

I would say wrap that injector lead around the sensor as many times as you can stand it, the more it looks like this, the stronger the magnetic signal to the hall sensor should be when current is flowing through the injector:

Ok this has been bugging me for a while, If I am going to have to amplify this signal with electronics anyway i.e. transisters, why does it need coiled through the sensor as long as the sensor picks up the signal?

[dcb]

I don't know that you will have to amplify it with electronics anyway. I don't have any real experience with halls, but my understanding is that you just need to get enough magentism to the hall that it can switch on and off and make a signal for the guino, weather it is in peak or hold current levels.

If you need more magnet signal to detect hold current, than add more coils. If it doesn't even pick up peak current then add even more Don't know if you can over-magnetize a hall or not though.

But it looks like *possibly* the only components needed are the hall effect and maybe a 10k pullup (to 12v) resistor on the hall signal line. Then the hall *should* switch low when current flows through the inductor and create a low output on the signal line, which connects to the guino inj input. I'll draw a pic later, but the questions are:

1. How much injector current at peak
2. how much injector current at hold
3. how many turns needed to activate hall at hold current
4. will that many turns be too much gaussess (sp?) for the hall at peak current? If it is even possible to over-magnetize a hall (did see a max rating).

[consaka]

Originally Posted by dcb

I don't know that you will have to amplify it with electronics anyway. I don't have any real experience with halls, but my understanding is that you just need to get enough magentism to the hall that it can switch on and off and make a signal for the guino, weather it is in peak or hold current levels.

Ok obviously I have been unable to make it clear how this type of sensor works. It is not a wheel speed sensor or a cam sensor signal. This doesn't get switched so to speak it just puts out a voltage relative to the current running through it. It is used to power meters that measure current. Thanks for trying to help. Ill figure it out..

[dcb]

Here more info on your sensor, from the source:
Datasheet -- CSLA2CD

So just do the math. You might have to measure the actual values on your rig but lets assume for examples purposes that you have a 6 ohm injector, 2 volts hold voltage, 12 volts peak voltage:

You need a 2.5 volt rising signal and a 2.2 volt falling signal at the cpu to trigger on and off respectively (regardless of how the signal gets there, the input is a binary switch to the cpu). There is a 50k resistor in series with the pin.

sensitivity of sensor is 32.7 mV N* ± 3.0 @ 8 Vdc , 72 amps max (N=number of turns). Offset is vcc/2.

So all I can figure is that at 0 amps it reads 4 volts (8/2). and at 72 amps (one turn) it reads 8 volts. @32.7mv sensitivity that is 122 steps for half travel (hmm, sounds like an 8 bit conversion going on somewhere maybe).

Ok, so in this example for peak current we have 12 volts/6 ohms = 2 amps. So if we want maximum output signal we divide 72/2, and come up with 36, so 36 turns will have this sensor full on at peak. output should be about 12 volts, no extra amplification needed there it seems.

So how "on" is it at hold? Well 2v/6ohm=.333 amps. So the answer is "not very" and now it has an offset.

This particular device, while handy for you, may not be an appropriate choice, if there are hall effect switches that won't require extra circuitry to undo the circuitry built into the device.

In order to keep this simple we need to locate one that can be on from, say, .2 amp on up, preferrably clamp on. That way anyone can use it instead of tapping injector wires.

[consaka]

How come we cant count and discredit the first pulse and only count the second pulse at the mpguino via software so to speak.?
I need a better scope, sometimes it looks like there is only one pulse untill you put your foot into it a bit.

[dcb]

It's a little complicated.

Main issue is that the injector is a one or a zero to the software. a "solid" one with the current design is going from 12 to zero volts. The hold "one" is going from 12 to 11 or so, not 12 to 0.

And the CPU measures in reference to ground, so 11 volts might as well be 12 (or 14 if the alternator is turning). So we want something that can turn a threshold amount of current (or voltage drop across the injector) into a solid one or zero for the CPU, and it has to be simple as possible/inexpensive/durable.

There are ways to measure voltage in analog in software but it is significantly slower than the digital inputs and would be a large bit of coding and a little bit of rewiring.

Basically we are looking for something simple (optoisolator and resistors maybe) that will turn any floating voltage drop across the injector above say 1 volt, into 3-5 volts and anything below into 2-0r zero volts (or vice versa, the software can invert it now).