Current to Voltage Op-amp problem

[Jdo300]

Hello,

Lately, I have been venturing into the wonderful world of analog circuits and am currently working on debugging a circuit I made which is supposed to convert the current flowing through a very small sense coil I wound, into a voltage signal that I can read with my scope. The circuit consists of a simple Current to Voltage converter op-amp circuit, where, in my case, the Op-amp outputs 1V for every 1mA of current flowing through the input side (See attached circuit).

The circuit consists of a LT1206 current feedback op-amp (I used this because it is the only fast op-amp I happen to have in my collection). To test the circuit out, I used my function generator with a 1k resistor in series so that I could limit the current into the circuit. I then clamped my DC-50MHz current probe on the input so that I could measure the actual input current to the circuit. The voltage output from the op-amp was then shown on the scope also for comparison and I noted that with a sine, triangle and square wave that the circuit worked flawlessly within the bandwidth of 0Hz to 15Mhz before significant attenuation of the output signal occurred.

However, once I attached the small sense coil to the inputs (as shown in the diagram), the op-amp started drawing over 100mA through both the input rails and oscillating randomly at about 47Hz more or less. Thinking that the op-amp didn't like the low resistance of the coil, I added the 10k resistor I used with the function generator and it still did the same thing though the frequency changed. I think took the coil out completely and just put the resistor across the leads. Even in that case, it still was drawing a bunch of current. I'm sure there is probably something silly that I am missing here so I was hoping someone here could give me some insight into what I'm missing here .

Thanks,
Jason O

 

[arithmetix]

Hi
I have experience with op-amps though not specifically current sensing ones... I wonder whether the power supply is causing this phenomenon? I'd look at the rails.

 

[uart]

Also you don't need a 50 ohm load. Most op-amps won't properly drive a low impedance load like that, so change it to 10k use a regular scope probe (if you're not already doing so) and ditch the load capacitor (unless that's just the cable parasitic).

 

[vk6kro]

Maybe you just haven't shown them, but there should be bypass capacitors on the power supply leads.

These are wide band devices and they can oscillate readily with so much gain. So multiple bypass capacitors would be better than just one on each lead. Say 10 uF, 0.01 uF and 470 pF on each lead. All cut with very short leads (or surface mount) and connected to a copper ground plane by the shortest route.

 

[Bob S]

Hi-
To minimize the input offset current, you should put a 1k resistor between the sense coil and pin 2, and a 510 ohm resistor between the ground and pin 3. The extra 1k on the negative side has very little effect on input current for a current-to-voltage op-amp with high gain.
But, any inductive output of a sense coil is a Faraday Law induction voltage, not a current. The current comes about only because of the external load resistance. If the coil has an internal resistance, the induction voltage across the load is decreased. Do you have any guidance concerning the optimum resistance load for the sense coil?
Bob S

 

[berkeman]

Try putting a buffer amp between the opamp output and the coax. Opamps don't generally have enough phase margin to drive capacitive loads.

 

[Jdo300]

Hi Everyone,

Thank you all for the responces and for the great tips. I'm going to be playing with the circuit some more tonight so I will definitely try out the things you suggested. To answer some of your questions I did originally have the output connected just to my scope probe rather than the 50Ohm terminated coax cable. I just did this to keep the noise to a minimum though the LT1206 is supposed to be able to source a minimum of 250mA for driving cables and even has a compensation pin to help with driving capacitive loads.

At the moment, I didn't try to compare the behavior with or without the coax line so I will try that first. Also, I have the circuit built on a solderless breadboard with semi-long leads going to my power supply. I did add bypass capacitors locally on the board near the ICs but I"m sure that cold be much, much better. I'm going to actually solder the circuit together on a perfboard as soon as I get this issue resolved to make it work much better but I didn't want to do anything semi-permanent until I at least got the basic setup functioning reliably. I'm using the little 2" x 2" square Rat Shack boards which don't like to be re-soldered more that about once or twice before the pads peel off so I want to (Ideally) get it right the first time.

I will also try the idea with adding the resistors on the input to see if that improves performance. I'll report back as soon as I get some results.

Thanks!
Jason O

 

[Jdo300]

Hi Everyone,

I figured out what the problem was. The original op-amp I was using was fried from some earlier tests that I did. Once I replaced it and also tried some of the tips posted here, it now works like a champ! Thanks you all!

@Bob S,

How did you determine that a 510Ohm resistor is good to fix for the offset current problem? Is this just a standard value that is used or is there additional rationale for that?

Thanks!
Jason O

 

[Bob S]

@Bob S,

How did you determine that a 510 ohm resistor is good to fix for the offset current problem? Is this just a standard value that is used or is there additional rationale for that?

I first chose a 1 k resistor between the senser and the neg input; that shouldn't hurt the signal because the neg input is a virtual ground. To balance the input bias curents from the two inputs, I had to choose a resistor on the pos input to ground that equaled the shunt resistance of the two 1k resistors on the neg input. The input offset current will flow through the sense coil and creat an input offset voltage, so a little extra resistance in series with the coil will limit that.
Bob S