Please help - difficult analog circuit.. Signal conditioning

Hi to all...
I'm having troubles while triyng to find a way to convert a signal to another...
Explaination:
I have 2 devices with 2 different acquisition systems (both automotive ecu). I need to let both share a NTC temperature sensor.
The first acquisition system use a strange (but nice) way to condition the signal...
It uses 2 bias resistors, and switch them at a certain temperature for gain accuracy, giving us a signal like this:

This is a warmup sample.

I need to not touch or modify this system...
I've computed that the first bias resistor (low temperature) is 3,9k, and the second is 350ohm.

The second system want a linear response from the NTC, and use a normal partition resistor conditioning scheme. A bias resistor to +5, the other lead of the bias to acquisition and with a terminal of the NTC, and the other lead of the NTC is at ground.
On the second system i can disable the bias resistor, and let it to acquire the pure signal conditioned by the first system, but the signal is not suitable...

This evening i try to find a point on the original system where i have a 0-5v signal, for know when there is the bias resistor swap...

I'm sure that it's possible to done it using only analogic devices, without using uController, ADC and DAC...

Anyone has an advice on how to slove this problem?

Thanks for any help...

Re: Please help - difficult analog circuit.. Signal conditio

Here are the possibe solutions I can think of:

1. Disable the bias resistors on system 1 when measuring with system 2.
2. Use system 1 bias resistors always, but perform a software conversion (i.e., mid-level driver?) to make it compatible with system 2.
2. Use system 1 bias resistors always, but modify the resulting signal to make it compatible with system 2.
3. Use another sensor that will be for dedicated use of system 2.

1) is possible, but how to made the both systems to not detect the switch?
2) software conversion probably is what i'm thinking... ADC, uC and DAC...
3) I think that you mean in analog... this is what i'm trying to do...
4) Probably the easyer, but i not like the idea of drilling in my engine head a new hole

Re: Please help - difficult analog circuit.. Signal conditio

Can you please post a small sketch of sensor connections for system 1 and system 2? Thanks.

Re: Please help - difficult analog circuit.. Signal conditio

Of the first circuit, for the moment i not know much...
I need to follow some traces But in the 4 layer layout is an hell!
But i prefer to not touch the first system... Apart to route a wire for detecting the switch point.

Ecu1 is the one that leave tha strange signal...
Ecu2 is the one that needs the linear signal...

Attached there is what i think to do (in analog or -worse- digital...)

Thanks for the attention

Re: Please help - difficult analog circuit.. Signal conditio

So as I understand it:

1. ecu1 switches bias resistors whenever it wants to and you have no control over that.
2. You cannot interfere with the sensor circuit when ecu1 is measuring or you run the risk of causing a bad measurement on ecu1.

If these are both true then your only option is to adapt ecu2 so it will work with the sensor signals that are controlled by ecu1, correct?

If this is correct then you can do this:

1. In your above schematic, the Conditioning block could be an opamp (voltage follower) or, perhaps even better, an instrumentation amplifier. Then it is up to the ecu2 digitizer and adc to utilize the sensor signal generated by ecu1.

Re: Please help - difficult analog circuit.. Signal conditio

But with an opamp, how i can transform that signal in a linear signal? Something like this (only for example):



This is what the ECU2 want...

An another approach is to simulate a signal compatible for ecu 1...
Something like this:

Re: Please help - difficult analog circuit.. Signal conditio

I'm sorry, but I still don't completely understand.

The schematics and descriptions indicate that fixed resistors are placed in series with the sensor to form a voltage divider, but the waveforms do not show dc levels. What causes the voltage curve? Shouldn't it be a dc level? What exactly do the acquisition inputs expect to see?

The resistor named R i think that it's only a protection, in fact, assuming that the adc has an input buffer, it has an input resistance really high, so virtually there is not voltage drop across R.
C is to smooth the curve and filter (temperature doesn't change too fast).

Forgot to say, the signal is changing because the ntc is warming up, so it lower it's resistance (lowering the voltage.

This what the ecu sees during a warmup of an engine...

Re: Please help - difficult analog circuit.. Signal conditio

How about something like this?

Doh!
Good mspaint job!!!
This seem what i need!
A question: The S block, what is? a switch?
But probably i need a little help to understand it...
The first stage is a noninverting buffer, the second, a non inverting amplifier...
So, when the switches closes, the output of the second stage decrease... Nice!
If i need 2 different gain (for first and after), i can put together 2 'final'stages, with different gain...

This idea is GREAT!

I must try this!
Something to work with in the week-end!

As first as i have a result, a communitate it to you!

Really thanks for your time!!!

Re: Please help - difficult analog circuit.. Signal conditio

S is an analog switch, such as ¼ of a CD4066 or similar device.

Glad I could help!

Please help - difficult analog circuit.. Signal conditio

Hai..
I'm new to ECU ..
I want to do project in ECU
can you tell some gud web links or material to read.
thanks a lot