Op-Amp buffer problem

[Scruit]

I am trying to make an op-amp based circuit to read the tach signal from the ECU in my car aqnd supply the signal to a z8-based tachometer that I am building. (hooking the input pin direct to the tach signal wire made the tach in the dah stop working, so I'm using the op-amp as a buffer)

I have a working circuit as below;

**broken link removed**

When I use a 741 single op-amp it works great, giving the following outputs;

VS 3.3v = VO 1.08v
VS 0v = VO 2.57v

R1 and R2 are both 100k

I have a working prototype of this circuit hooked up to my car - works great, dash tach readout is good and my z8-based tach read correctly etc.

The problem comes when I try to is an LM747CN (dual opamp, because I'm also going to read the vehicle speed), the same opamp inverting buffer circuit gives me these outputs;

VS 3.3v = VO 1.93v
VS 0v = VO 2.64v


The problem is that 1.93v is too high to be seen as a logical low by my Z8-based circuit (input pin confiogured for falling-edge interrupt).

I have tried 2 different LM747's and get the same result. How do I get the correct output range?

Any ideas?

 

[ljcox]

Your original circuit is not performing as I would expect. If R1 = R2 = 100k, then the amp has a gain of -1.

When you apply 0V to the input, you see approx 0V at the output.

And if you apply 3.3 Volt to the input, you should read about - 3.3 Volt at the output.

But given that this is not so, then I suspect that the input bias currents are affecting the output.

In order to eliminate this effect, you can either put a 50k resistor between the + input and 0V or use a FET input Op Amp such as the TL071. (there are also the dual and quad versions TL072 and TL074)

Len

 

[Scruit]

I have tried 4 different 741s and get the same. It's puzzling - my voltage source is the VDD ping on the Z8 prototyping board. Even if I switch to using a 9v battery instead I still get 1.5-7.5 swing.

This is getting annoying - all I want to do is be able to read a tach signal from an ECU wire without drawing enough current to kill the factory tach in the dash. I spent 3 hours writing the code to read the pulses and everything else, and I'm stuck on the simplest part - just hooking my circuit up to the ECU.

What about using an optoisolator? Would that work out better?

 

[Roff]

Tell us what your input voltage levels are, what output levels you want, the available supply voltage(s), and whether the circuit needs to be inverting or non-inverting, and we can design a simple solution. From what you have said so far, I doubt that a 741 will be involved in the solution. It sounds to me like you need a comparator or some other form of level translator.

 

[Scruit]

The tach signal wire produces pulses. My multimeter registers about 6v when the engine is running, and the frequency shows about 200hz at 4k rpm.

I have a z8 Encore microcontroller programmed with an input pin set up to count the pulses and display the output on an LCD readout.

The Z8 is expecting to see about 3.3v for a logical 1, and 0v as a logical 0.

The biggest issue with the z8 chip was that it seemed to draw too much current from the tach signal wire, as the dash tachy stopped working.

I posted this before and somone suggested using an op-amp as a buffer/amplifier. I've spent about 10 hours trying to get that to work and when it finally did apparently it was only by luck. I have 7 op-amps here that I have tried, all give exactly the same result, to within a few mv of each other.

So back to sqare one - I need to read the signal from the tach signal wire from the ECU without drawing enough current to kill the dash tach.

 

[John Sorensen]

I'm going to try to attach some circuits. One is an Op Amp as a non-inverting buffer. Since you didn't post your power supplies, input signals and desired output, I'm just guessing. I looked at some datasheets for the 741 and I don't see that it is recommended for single ended supplies. If you don't have split supplies, an inverting buffer is not going to work. The other circuit is using the Op Amp as a comparator. Comparators, like the LM339 or LMV331, generally work better as compartators than do Op Amps.

j.

 

[Scruit]

I just got back from testing... Thinking back to the theory that current drain from the tach signal wire being the problem... I used just a 10k POT and that worked fine...

I adjusted the pot to about 3.5k between tach signal and z8 input, and 7.5k between z8 input and ground (any naturally, 10k between tach signal and ground).

This obviously did the trick. The z8 is seeing a high enough voltage to trigger it's input pin interrupt, and the resistance is enough to stop the z8 from being enough of a current draw to zero out the dash tach.

Do you guys think this would be a good enough solution - or am I doing something stupid here?


(BTW, There is nothing more depressing than going into RadioShack and when you're asked "Can we help you find something?" you say; "I just need a couple of POTS" just to have the kid behind the counter look at you blankly and then suggest the *Gardening* section at Walmart...)

 

[Roff]

Obviously, the pot works. It is not a very reliable solution. If signal polarity doesn't matter, use a transistor as a level translator, as below. It will invert the pulse, which shouldn't be a problem.
Forget about using a 741 op amp as a level translator. The output will not get nearly close enough to zero volts with a single supply. There are comparators that are designed for this purpose, but they are overkill for this application.

 

[Russlk]

The problem with the original circuit is that the 741 will not operate from a single supply unless the inputs are biased above ground.

 

[Roff]

The problem with the original circuit is that the 741 will not operate from a single supply unless the inputs are biased above ground.

That's not the only problem. The output will not pull down to GND, according to the specs and a simulation of two different models. I don't have one to experiment with. The schematic in **broken link removed** implies that a resistor from the output to ground might give you a "soft" zero volts. I certainly wouldn't count on this to work, especially from all mfrs. If you used the 741 as a comparator instead of an op amp in this application, a voltage divider on the "reference" input would solve the common mode voltage problem. As I said, I think any sort of a comparator is overkill in this case.

 

[Scruit]

Ron,

Thanks fro the diagram - I'm gonna put that togerther on my breadboard and try that. I figured the POT solution was too simple...