Sensing of the Rload

[alok1982]

Hi All,
Attached is the schematic of a comparator circuit.
Where R9 = Resistive load. Change in the resistance changes the output.
A inverting circuit using NPN has been used for getting other state.

Its working fine.

Now I want to add a circuit or a way to identify that my "Rload" is present in circuit.

Can you all please share your ideas?
Let me know if more details required.

Thanks!

 

[Diver300]

If R9 is missing, it will have the same effect as it having a high value, so the voltage on it will go up. If you want to sense that as being a different condition, you need a separate comparator. The separate comparator would have a different threshold set by another voltage divider.

I've got a couple of comments about the circuit. If you want this circuit to work in real life, your resistors values are too high for the op-amp input. 100 MΩ is going to get swamped by leakage currents. I would suggest staying below 100 kΩ unless you look carefully at leakage.

You've not got enough current driving the base of the transistor. The 120 kΩ resistor gives you 100 μA. The transistor gain is only guaranteed to be 100 at 10 mA, and less at the 25 mA that you are running, so you need at least 250 μA for reliable operation. Also, even 100 μA through the LED could make it glow, so you should put maybe 2.2 kΩ in parallel.

I can't see why you are using the transistor at all. You could just have the the green LED and its resistor connected between the comparator output and ground.

Can you explain what the circuit does? If R9 is just a resistor, the circuit will never switch, so what is R9 in your application?

 

[alec_t]

I can't see why you are using the transistor at all

Because a LT1017 comparator output can't pull up to drive a LED satisfactorily.

 

[Diver300]

Because a LT1017 comparator output can't pull up to drive a LED satisfactorily.

OK, I hadn't spotted that.

 

[alok1982]

Thanks for replies!
Actually this circuit I want to use for ferrite testing. A ferrite having resistance of Mega ohm.
Idea is to segregate the low ferrite resistance value and high ferrite resistance value.
The band between both the values will be pre-decide by comparator resistors.
So, RED led will indicates low resistance and GREEN indicates high resistance.
Here in schematic for green indication, after a band I am inverting the values.

Now just want to include the "Ferrite sensing" which will indicate the presence of ferrite in the circuit, with different color led.

I know the resistance is very high, but any idea can you please suggest which can work with this?

Thanks!

 

[alec_t]

Idea is to segregate the low ferrite resistance value and high ferrite resistance value.

Do you have some numbers for 'high' and 'low'? From my limited search, ferrites typically have a resistivity some 12 orders of magnitude greater than metals, but I didn't see any major distinction between the resistivities of different ferrites.
If you're trying to measure resistance of 100MΩ or more, contamination and atmospheric moisture are going to mess with your measurements.

 

[MikeMl]

What is the lowest sample resistance that is OK? The posted circuit suggests that it is 300megΩ?

If so, have you considered the input bias current of the LT1017? It is much too large compared to the leakage current through the ferrite.

 

[alok1982]

Low cores are of 10 to power 7 and High core are of power 11.
Input bias current for LT1017 is 5nA.

 

[MikeMl]

...Input bias current for LT1017 is 5nA.

And therein lies the problem!

The theoretical trip point of your circuit should be when R9 is exactly 300megΩ; its actual trip point is when R9 is ~318megΩ, a significant error. This is due to the input bias current of U1 being about 10% of the current through R9.

Look at this:

The current through R4 is the yellow trace, the current through R9 is the green trace and the difference between those two is the input bias current shown as the blue trace. In a "proper" design, the input bias current to U1 should be less than 1/1000 of the current through R9. The x-axis of the this plot is the resistance of R9.

There are opamps available that have input bias currents on the order of 1pA....

 

[MikeMl]

Here is a very low input bias current opamp configured as an Ohmmeter. V(c) is ~1v when R9=90megΩ, and is about 11V when R9=1.1GΩ.

Hang a window comparitor on V(c). Now you can detect "too low"= [V(c)<3V], "ok"= [3V< V(c)<10V], and "open"= [V(c)>11V].