Blocking Voltage

[morizabal]

Hi,

My application consist of two parts. The first part of the circuit puts a 0-20mA signal through a 500 ohm resistor for an equivalent voltage of 0 to 10 VDC. The second part of the circuit consists of a voltage divider that accepts a ±10 VDC signal. My circuit is part of a larger data acquisition system that can accept both a 0-20 mA signal and a ±10 VDC signal. In between the two circuits I placed a diode to allow the voltage from the current circuit to pass, but not allow the voltage from the voltage input to pass back to the current circuit. The problem is that when I place a negative voltage on the voltage input is passes through the diode. My only immediate solution is to put a mechanical switch in place of the diode. The switch will either be manual or a relay. I tried using a solid state Analog switch which worked, except for the when the signal was negative. Is there a device I can use that could act as a switch to allow the current signal to pass and block both the positive and negative signal. The device must be able to be controlled by a logic signal, I tried using an SCR which worked, but the only problem is that it only triggered when the current input was around 8 mA, I need to cover the whole range of 0-20mA. Any help would be greatly appreciated.

Thank You

 

[morizabal]

Pebe,

The pot is used for calibrating the circuit. I will inputting a known 10 VDC value and adjusting the pot until I get the desired reading. I will be adjusting the pot until I get a reading of 3 mV/V. (The A/D boards are calibrated to mV/V) This is an equivalent of 15mV of voltage on pin 5 of IC23B, this is an equivalent resistance of 15.075 ohms for the pot. I know that due to resistor tolerances that there will be an error present between the two circuits which I will have to correct using software. Eventually my plan will be to replace the pot with a fixed resistor and using software for calibration.

Mario

 

[electricuwe]

With both inputs on a connector (and not on a screw-type terminal strip) using a jumper as suggested on Jan 1, 2003 isn't a good solution.

I also try to avoid jumpers inside of a device because of the reason you already mentionend (and because this would require additional documenation whereas jumpers on terminal strips are documented in the overall schematic of the equipment).

But keep in mind that the solution you have choosen now is even more prone to introducing errors by offset and tolerance than the first approach.

Scaling down a signal and amplifying it again is never a good idea. So I would recommend to analyse the errors resulting from you design and compare it to the specification.

You have choosen a precision opamp with very low offset and offset drift, but you should also consider that radiated electrical noise can be rectified at opamp input stages and look like a dc input voltage to the opamp.

 

[morizabal]

electricuwe,

I mentioned previously that voltage and current inputs are on a screw terminal connector on separate terminals. I will be using low TCR 0.1% resistors (except for the feedback resistors) to minimize errors. You mentioned that it is not a good idea to scale down a signal and ampliflying it again. Since my A/D boards will not accept a high level signal, what would be a better approach?


Morizabal

 

[a31ford]

Hi Guys, I think he was on the right track using a "Analog switch" BUT, the type he used was incorrect....

analog switches can be biased 2 ways, the dallas I'm not famil. with BUT, 4016 and 4116. 4016 is single ended only, BUT the 4116.... can be placed in a +/- supply config, and will NOT lockup when used with - voltage...

hope this helps. Gawd, hope I have the numbers correct

 

[tdbeng]

Can you just remove the voltage input from the circuit altogether? Make the current input your only input connection. When the input signal is 0-20mA, the circuit is fine as it is. When the input signal is +/-10V, remove R63 (either by removing a jumper, turning off a relay, digital pot, etc.).

 

[morizabal]

a31Ford,

I will check out the 4*** devices you mentioned, for now I will be using the resistors that appear to be doing the job. Like I mentioned on a previous post I will be using software error correction which I have tested and it works fine. I tested the circuit for stability, and once its warmed up it will remain very stable over time with no drift. If anyone has a better idea for converting a ±10 volt signal to a ±30mV differential signal any suggestions would be appreciated.

Thanks To All
morizabal