Interfacing XTR116 with DAC0808

I want to drive the DAC with a SN74HC590 counter which will have a count clock input from a reed switch and a reset clock input from a 555 timer output.

Hi Tom,

Thanks for your questions. To start, I understand that you are looking to power your DAC with the Vreg from the XTR116. This is certainly possible; however, I was pretty busy today, so please allow me until tomorrow to put together a couple block diagrams that will illustrate how this can be done appropriately.

Thanks and regards,

Hi Ryan,

Thanks for getting back to me. We have decided to go a slightly different route- still with the xtr116 though.

Instead of all the counter chips We have decided to use a PIC or similar micro-controller to count the pulses from the reed switch, do some math, and send the data via serial (spi) to a DAC through a digital isolator, and then take the DAC analog signal into the XTR116. (See similar schematics here: http://www.mikroe.com/click/4-20ma-t/)

Perhaps the xtr116 could power the PIC and we could just do an analog voltage output to the XTR input all from within the current loop?

We are leaning towards a PIC12F, 16F or 18F controller due to very low cost.

Is there a better option? We have sensors that use multiple xtr116 IC's (we generally have between 4 and 32 current loops in one system for data acquisition) so the transmitter circuit cost is a major concern.

Thanks for your help. Happy Thanksgiving!

Powering the microcontroller from the XTR seems like the obvious choice now that I review the idea. Then use a PWM output to charge a cap with resistor load into XTR current input; 0-5V through a load resistor of 25k into sensing input (pin 2) should do the trick?

We will also have applications where multiple sensor values will be read by a single microcontroller, information will be processed, and PWM signals will be sent out in proportion to the numeric sensor values. For each PWM output channel, a digital isolator will transfer the values to a low-pass filter on the current sense input of the XTR116. Is there a more economical or better way to convey a PWM signal to the XTR116 or to drive a 4-20mA loop from a PWM signal?

Are there any frequencies to avoid with the XTR? We will most likely be operating in the 1MHz range for our PWM frequency.

Thanks!

Now that I think about it; feeding the PWM signal to a low-pass filter for conversion to a current value into the XTR would be highly temperature dependent with the resistor; what is a better way to do this?

What about feeding a variable frequency from the microcontroller (instead of PWM) through a digital isolator/optoisolator into a F/V or F/I converter such as the LM2907 or LM2917?

So- currently looking at two configurations:

1) Loop powered microcontroller; uC sends variable frequency to a freq-to-voltage or current converter, to XTR input.

2) Three-wire configuration with uC feeding multiple variable frequency output signals through unique isolation to different XTR116's with same F/I conversion process as configuration 1.

Last post for a while... I think!

Hi Tom,

I was able to review your posts and have been evaluating the different options that you presented:

1.)    First, let me comment that your latest suggestion, which includes the use of the LM2907/17 frequency-voltage converters, is not a viable option for your application. They require too much quiescent current to run, between 3.8-6mA at a minimum. This would leave no available current to power the XTR116, the microcontroller, or any required isolators, and would thus limit your output current capabilities accordingly.

2.)    The PWM method would work quite well actually, as it is a commonly adopted method in this industry. You can reduce the margin of error over temperature simply with a low TC resistor component. If you chose to proceed with this method, remember to keep the resistor in the low-pass filter significantly smaller than Rin.

3.)    Yesterday, my colleague and I came up with a couple other solutions using SPI or I2C interface between the microcontroller and a DAC. You seemed to be arriving at similar conclusions on your own in your second post (nice work!). They involve more parts, but will get the job done more reliably than the PWM method. I’m recommending that you review the following part combos below. These designs isolate on the digital side of the design between an external microcontroller and a loop-powered isolator and DAC. 

1. SPI – check out the DAC7311 (Iq = 110uA) and ISO7641FC/FM
2. I2C – check out the DAC7571 (Iq = 135uA) and ISO1541

A similar option would be to loop-power the microcontroller and use an integrated DAC or directly communicate with an external, loop-powered DAC. This would eliminate the need for digital isolation.

4.)    Regarding your latest post, we have some other comments about the configurations you were looking at for your application. The “Loop-powered microcontroller” configuration makes sense; you can certainly power the microcontroller from the XTR116 and share the same reference (IRET) with each device, as well as with your reed inputs.

In the “three-wire configuration,” it sounds like you are attempting to send multiple microcontroller outputs to several independent pairs of isolators and XTR116 devices. If so, how will the respective reed inputs for each XTR116 interface with the microcontroller? Are the reed switches local to the XTR116 circuitry, or not?

I think we are close to deciding on a proper configuration for your application. Please review the options above, then let us know how you would like to proceed and we’ll take it from there :)

Thanks and regards,

Hi Ryan,

Thanks for your detailed and speedy reply!

We are using the XTRs to transfer sensor values to PLC analog inputs. There are two basic sensor configurations we are working with:

Configuration A: Single loop-powered analog sensor such as temperature, light level, speed, etc. (For a single sensor PCB/product.)

Configuration B: Multiple of these analog sensors on one pc board with unique current outputs for each sensor.

In configuration B we can't place the microcontroller inside of the current loop because there are multiple current loops! We could use a separate uC inside of each loop- while this would work for some sensors, we have other sensors which draw more current than the loop can supply so either way we need a three-wire option. Also it would be a lot of programming time and maintenance during production; we would rather write (and maintain) firmware to one IC than multiple IC's for configuration "B."

I believe it makes the most sense to go with your option #3 with the microcontroller to DAC via SPI or I2C. In configuration "A" we can have the DAC driven directly by the microcontroller and in configuration "B" we can simply place a digital isolator in between (such as those you selected) to remove the microcontroller and sensor current from the loop.

Hi Tom,

It seems like you nailed it in that last paragraph. Option #3 is the best option for either configuration “A” or “B.”

My only remaining question is how much distance do you expect there to be between your microcontroller and the XTR116 circuitry? I ask because SPI and I2C both have their limitations over distance. Large distances may introduce unwanted capacitance and delay into your system, which can cause communication issues. While SPI can run over larger distances than I2C, neither will work very well over distances more than a few meters and require shielded cables to accomplish this.

If your application requires long distance, I would suggest returning to your PWM idea. Attached is a simulation model for TINA Spice that mimics the behavior of the XTR15/6/7 family. You can use this to experiment with difference discrete component values and PWM inputs. If you do not already have TINA Spice, it is a free download from http://www.ti.com/tool/tina-ti.

8156.XTR116_PWM model.TSC

Otherwise, it sounds like you’ve got it well under control!

Best Regards,

Hi Ryan,

The microcontroller and XTR will be on the same pcb in close proximity. Distance shouldn't be an issue in either of our cases; thanks for pointing out the serial distance limitations. We'll experiment and further evaluate both the PWM and serial options for cost and and performance. Thanks for the TINA model and all of your fabulous support!