Op Amp with a/c output

[HomerDodd]

A few weeks ago you guys helped my understand op amp circuits, to help me with a car sensor. A new question and one I should have asked before: can anyone give me a simple circuit design or schematic that shows how to take 12 VDC from the car wiring and convert it (maybe through an op amp or similar) to get 12 VAC output? I really need to send 5, 10, 12 VAC to the signal harness to "fool" the cars computer (ECM) into thinking the sensor is working correctly, which will turn off the "Check Engine" light long enough to get my car to pass smog. Again, much thanks to all of you.

 

[davenn]

can anyone give me a simple circuit design or schematic that shows how to take 12 VDC from the car wiring and convert it (maybe through an op amp or similar) to get 12 VAC output?

an op-amp will only give an AC out if it has an AC input

what you really need is an oscillator

and what frequency does the AC signal need to be ?

I really need to send 5, 10, 12 VAC to the signal harness to "fool" the cars computer

do you really need each of those voltage levels or just say, the 5VAC level ?

 

[HomerDodd]

The sensor must send at least 5 VAC to the computer, but it can be higher. I think it is somewhere around 600 Hz, let me double check. Is it possible to do this?

 

[AnalogKid]

If the car computer only really needs to see a frequency, but not necessarily a sine wave, then any square wave oscillator circuit will work. Opamp, 555, 1/6 of a CMOS Schmitt trigger hex inverter, 1/4 of a CMOS quad NAND gate, two transistors - lotsa options.

What is your experience level with any of these technologies, and building a small circuit on perf board?

ak

 

[dorke]

You should define the "shape" of the AC signal, as well as it's frequency and amplitude.
If you have a photo of how it should look like that would be great.

 

[kellys_eye]

I'd refer posters to the original thread....

The signal required comes from a speed sensor and is (probably) used to adjust timing, fuel injection etc so needs to follow the engine rpm else the emissions produced would 'fail' the vehicle anyway.

It's not an 'AC' signal either. It's DC (with an AC component) i.e. it's a squarewave.

From the information already given there is (I surmised) a supply voltage fed to the sensor and a signal coming back from it. What I don't know is the supply VOLTAGE (could be 5V could be 12V?) but the signal seems to come back from a hall effect device - make your own mind up as to what that consists of.

Any comparator with a 'set' input lower than the signal source maximum (but higher than it's minimum) will produce an output that is as big as the comparator supply voltage (give or take) but without knowing what the MAXIMUM is allowed to be we can't simply guess and use (say) a 12V output for the potential of ruining the follwing signal processing circuitry.

Of course, an op-amp will do the same but the circuitry is (ever so slightly) more complicated - more so if it requires a dual-rail supply or offset biasing - hence my suggestion to use a comparator.



Here, the potential divider R2/R3 set the switching level and the voltage at the slider of R1 is the signal being compared.

As far as the OP is concerned R2/R3 are chosen such that the voltage at pin 7 of the IC is just lower than the lowest voltage produced by the shaft sensor and R1 is adjusted until the output at pin 1 switches between the supply and zero.

In a PRACTICAL circuit R1 would be removed and the original sensor signal would go to pin 6 and pin 1 would then go to where ever the sensor signal was meant to go.

R2/R3 could be replaced by a potentiometer (100K) to allow for adjustment of 'sensitivity'.

 

[HomerDodd]

If the car computer only really needs to see a frequency, but not necessarily a sine wave, then any square wave oscillator circuit will work. Opamp, 555, 1/6 of a CMOS Schmitt trigger hex inverter, 1/4 of a CMOS quad NAND gate, two transistors - lotsa options.

What is your experience level with any of these technologies, and building a small circuit on perf board?

ak

Thank you. Yes, I have constructed small circuit boards a few times, have installed resistors, etc. I have the dexterity to build them. If it is no trouble, can you get me a schematic or drawing showing how to lay out the circuit? I have two choices for supply: can use the cars readily available 12 VDC coverted to 5 VAC or more, or the sensors supply wire of 1.5 VAC to be stepped up to 5 VAC or more. The returning signal to the cars computer (PCM) must be at least 5 VAC to keep the "Check Engine" light off. Bottom line, whatever small circuit I can build to send at least 5 VAC to the PCM is what I need. Thanks !!

 

[HomerDodd]

I'd refer posters to the original thread....

The signal required comes from a speed sensor and is (probably) used to adjust timing, fuel injection etc so needs to follow the engine rpm else the emissions produced would 'fail' the vehicle anyway.

It's not an 'AC' signal either. It's DC (with an AC component) i.e. it's a squarewave.

From the information already given there is (I surmised) a supply voltage fed to the sensor and a signal coming back from it. What I don't know is the supply VOLTAGE (could be 5V could be 12V?) but the signal seems to come back from a hall effect device - make your own mind up as to what that consists of.

Any comparator with a 'set' input lower than the signal source maximum (but higher than it's minimum) will produce an output that is as big as the comparator supply voltage (give or take) but without knowing what the MAXIMUM is allowed to be we can't simply guess and use (say) a 12V output for the potential of ruining the follwing signal processing circuitry.

Of course, an op-amp will do the same but the circuitry is (ever so slightly) more complicated - more so if it requires a dual-rail supply or offset biasing - hence my suggestion to use a comparator.

View attachment 36208

Here, the potential divider R2/R3 set the switching level and the voltage at the slider of R1 is the signal being compared.

As far as the OP is concerned R2/R3 are chosen such that the voltage at pin 7 of the IC is just lower than the lowest voltage produced by the shaft sensor and R1 is adjusted until the output at pin 1 switches between the supply and zero.

In a PRACTICAL circuit R1 would be removed and the original sensor signal would go to pin 6 and pin 1 would then go to where ever the sensor signal was meant to go.

R2/R3 could be replaced by a potentiometer (100K) to allow for adjustment of 'sensitivity'.

Thank you for the info. Yes, the cars computer (PCM) sends 1.5 VAC (it is AC per the manual) to the input speed sensor. The sensor, if working correctly, will return 5VAC or more (some guys have measured 9,10, 12 VAC) to the PCM for smooth shifting calibration only. So, any circuit that take that 1.5 VAC and step it up, or take the cars general 12 VDC power and convert it to AC is all I need. This is a very common problem on the older Cadillac's. Many guys have "fooled" the computer with an op amp, so they said. However, I cannot reach any of them, they may not be on the Cadillac forum anymore. Here is a link from that forum from over 10 years ago describing the problem with the P056 code (Input Speed Sensor). Thanks for everything.

http://www.cadillacforums.com/forums/cadillac-forum/t-25774.html

 

[Alec_t]

Yes, the cars computer (PCM) sends 1.5 VAC (it is AC per the manual) to the input speed sensor

Can you post a pic of or a link to that section of the manual? It would help to understand the calibration procedure and might give a clue as to how best to amplify the sensor's 1.5VAC and inject a higher voltage.

 

[HomerDodd]

Can you post a pic of or a link to that section of the manual? It would help to understand the calibration procedure and might give a clue as to how best to amplify the sensor's 1.5VAC and inject a higher voltage.

I don't have the actual manual. I was quoting what guys on the Cadillac forum quoted from their manual and posts. All of them who have correctly functioning sensors have measured the supply voltage signal (from the PCM) as 1.5 VAC approx and over 5 VAC output back to the PCM. Whatever I can do to get 5 or more VAC back to the PCM is what I need to figure out. Anything would help. Thanks !

 

[Alec_t]

All of them who have correctly functioning sensors have measured the supply voltage signal (from the PCM) as 1.5 VAC approx and over 5 VAC output back to the PCM.

I'm puzzled as to how the output can be greater than the input. How many wires connect to the sensor? Presumably one wire is a supply input, another wire is the output and the sensor case is a ground connection?

 

[kellys_eye]

I'm puzzled as to....

I asked for clarification of the voltage and signal in the original thread on this subject.

The schematic provided (only part of, not a complete schematic) shows two terminals, one marked 'V' and one marked 'S'.

My inference was that 'V' was the supply to the (presumably) hall effect sensor (against splined shaft or keyed wheel) and 'S' was the signal return to the ECU. I requested that 'V' be measured to ascertain the validity of my assumption - ideally I'd like to have confirmation of 'S' but that would require an oscilloscope.

It seems that we are unable to convince the OP that the signal is not AC...... (in the conventional sense) but a simple squarewave referenced to...... dunno??? 5V??? 12V???? We need to know!

Quite obviously the hall effect device is 'lacking' and that its output isn't switching at the levels the ECU requires to detect it reliably hence the need for amplification.

I'd love to resolve this - but without the feedback requested this is going to go nowhere.....

 

[HomerDodd]

There are two wires going to the sensor. I scraped away a slight bit of insulation on the wiring near the trans, about a foot from the actual sensor. I then scraped away some insulation on the two wires in and out of the PCM which is located inside the car near the glove box. Definitely AC voltage. One measures .5 VAC at idle, varying slightly, where the other measure 1.5 at idle, raising to 2.5 VAC at low speed as I drove down the block with the multi-meter connected to wires closest to the PCM. So there is some response going on in there. I am wondering if my sensor is actually OK, and there may be high resistance in the wiring somewhere. I think tomorrow I will run bypass wires directly to and from the sensor, and then test. If the voltage goes to within spec, I know where the problem lay. If not, the sensor will most likely be the culprit, and back to square one, needing the amp or comparator circuit. Thanks !

 

[kellys_eye]

If it was true AC then both wires would read the same voltage.

A squarewave with a DC component will also read as AC on a meter.

Measure between one wire and chassis - what do you read?
Measure between the other wire and chassis - what do you read?

Use your meter on the DC ranges.

 

[HomerDodd]

If it was true AC then both wires would read the same voltage.

A squarewave with a DC component will also read as AC on a meter.

Measure between one wire and chassis - what do you read?
Measure between the other wire and chassis - what do you read?

Use your meter on the DC ranges.

I grounded the negative probe of the DMM to the frame and clipped the positive lead to one wire. At first on the AC setting, it bounced all over, flashing on and off, almost like trying to read a frequency or wave, and funny enough 7-8 volts flashing off and on. I changed the range and it read steady, no flashing, 1 -2 VAC. I will try DC today. I wonder if the sensor itself generates a voltage (like an alternator) proportional to it's rotational speed?

 

[Alec_t]

Have a read of this. If your car is old-ish I suspect your sensor is a variable reluctance type, rather than a Hall type, in which case it generates an AC signal but is passive (doesn't have a supply voltage provided by the ECU). What year/model is the car? Does your car service manual identify the sensor type?
BTW, an inexpensive DMM won't give accurate readings on AC ranges if the waveform is significantly different from a pure sinewave. It won't be a pure sinewave if the teeth on the pick-up wheel are widely spaced.

 

[HomerDodd]

Have a read of this. If your car is old-ish I suspect your sensor is a variable reluctance type, rather than a Hall type, in which case it generates an AC signal but is passive (doesn't have a supply voltage provided by the ECU). What year/model is the car? Does your car service manual identify the sensor type?
BTW, an inexpensive DMM won't give accurate readings on AC ranges if the waveform is significantly different from a pure sinewave. It won't be a pure sinewave if the teeth on the pick-up wheel are widely spaced.

It is a '95 Seville. From what I've read, it is a Hall sensor; the trans shaft has a toothed wheel which picks up the signal only when in gear and moving. The gear does not move when in park or neutral. Interesting, that I can clear the code, sit at idle and never have the code comeback. Once I put it in gear, still not code returns. About 1 block after I start moving normally, the code comes back. I am just wondering if I can "substitute" a fixed DC voltage to satisfy the computers desire. Would clipping one of the wires and inserting a 9V battery solve the problem temporarily. Look at the schematic from the service manual below and let me know what you think.

 

[kellys_eye]

Do you have the schematic of where the other ends of the wires go i.e. the wiring to the ECU (or any device/circuitry that is in the 'line' between sensor and ECU)?

 

[HomerDodd]

Do you have the schematic of where the other ends of the wires go i.e. the wiring to the ECU (or any device/circuitry that is in the 'line' between sensor and ECU)?

No schematic, other than the harness picture already posted. The two wires simply terminate into the ECU/PCM. Nothing else branches from them. Today I measured the voltage on the harness going into the ECU. One had nothing, the other had battery voltage, 13 VDC. Could that wire be shorted to voltage or ground? BTW, is it true that a Hall sensor (wheel and magnetic sensor) create's it's own AC voltage? If that is the case, then I doubt the sensor is dependent on a source or supply voltage. Would that make sense?

 

[kellys_eye]

If the sensor is a 'coil' pickup of a rotating magnetic slotted disk then you'll have an AC signal output. If the sensor is a hall effect device then it MUST have a supply voltage to one side of it ergo one of the wires will have a fixed DC voltage on it.

The 13V (battery) you measured fits with my previous assumption - the other wire will have the return (speed) signal that is only present when the gearbox input shaft is rotating. That signal being a squarewave with a frequency proportional to the rotation speed of the shaft.

The comparator circuit I showed above would therefore accept the 'signal' (the NON 13V wire) on pin 6 and R2/R3 could be replaced with a preset potentiometer (wiper to pin 7).

You then take the output signal from the comparator at pin 1 and feed it back into the ECU, adjusting the preset so the circuit switches at the level of the input signal.