• State TI Thinks Resolved
  • Locked Locked
  • Replies 8 replies
  • Answers 1 answer
  • Subscribers 50 subscribers
  • Views 2172 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

XTR111EVM: 4-20mA accuracy issue at high ambient temperature ( > +65C)

Weiran Chen
Prodigy 70 points

Part Number: XTR111EVM

Hello Sir,

Attached SCH shows our 4-20mA conversion circuit  which is part of the circuit in our design. Basically we used design of TI evaluation board for XTR111.

The 4-20mA accuracy of this circuit meets our requirement when ambient temperature is between  -40C to +60C.

However, when temperature rises over +65C, the 4-20mA output current was out of our specification (< 2.5%) if Vin was below 2V.

The accuracy comparison can be found in the other attached file.

We measured the voltage of R226 (shown in attached SCH), this voltage only changed a little (less than 0.5%) compared with the voltage value at +22C.

It looks like that the current passing through B pole and E pole of Q8 significantly increased at +85C so that the 4-20mA output current accuracy was severely degraded. 

Because we used TI's evaluation design, does any one of TI engineers know this issue? Is there any recommendation to get it improved?

Thanks

William Chen

4-20ma accuracy comparison.pdf

4-20mA circuit.pdf

  • 0
    TI__Mastermind 23050 points

    Hi William,

    I just wanted to let you know we are investigating this issue. It is not something I have seen before, but given that the XTR111 is specified up to 85C, I suspect one of the external transistors is responsible. I am going to test our evaluation module and see if I can produce similar results. I will get back to you as soon as I have more information to share.

    Regards,

    Zak Kaye

    Precision Amplifiers Applications

  • 0 in reply to Zak Kaye
    Prodigy 70 points

    Thank you Zak!

    I'm waiting for your good news :)

    Best regards,

    William Chen

  • 0 in reply to Weiran Chen
    TI__Mastermind 23050 points
    Hi William,

    I have acquired an XTR111EVM and aim to begin testing today. Can you please clarify a few things about your setup:

    1) How were you controlling temperature? Was your entire board subjected to increased temperature or were you able to isolate the XTR?

    2) How are you calculating accuracy? The shape of the data suggests your error was not referenced to the full-scale output value.

    3) Have you probed the output of the OPA333? Because of the nature of the output impedance of the device, it may struggle to drive the 100Ohm + 1uF cap load. Are you configuring the amplifier as a buffer?

    Regards,
    Zak Kaye
  • 0 in reply to Zak Kaye
    Prodigy 70 points


    Hi Zak,

    1) My entire board was put inside the temperature chamber and heated up totally. However, I've confirmed that the accuracy issue was only related to the circuit from Vin of XTR111 to 250 Ohm load resister. So I disconnected OPA333 by removing resisters R229 and R245. An adjustable DC voltage supply directly injected voltage on Vin of XTR111. The DC voltage was adjusted from 0V to 3.29V. Therefore, the accuracy issue has nothing to do with PWM to voltage conversion circuit. To reduce the effect of load resistance change, I put the 250 ohm load resister outside of temperature chamber.

    2) Yes, the curve I gave you was not referenced to the full-scale output value. The 4-20mA output was compared with the calculated ideal value generated by the same DC voltage applied on Vin. We call this accuracy as "relative accuracy".
    I used full scale accuracy to evaluate the 4-20mA conversion as well based on 5V full scale. The full scale accuracy was < 1.5% within full temperature range from -40C to +85C. The design specification we were required was < 2.5%. Full scale accuracy was good enough to satisfy the design specification.

    3) Yes the OPA333 was configured as a buffer. However, as I mentioned above, OPA333 had been disconnected to my accuracy test.

    4) Do you mean we have to use full scale accuracy ?

    Thank you,

    William Chen

  • 0 in reply to Weiran Chen
    TI__Mastermind 23050 points

    William,

    For industrial applications you will almost always see things specified in terms of full-scale accuracy. It is typically a more useful measurement than relative accuracy. However, I have conducted my own temperature testing on our XTR111-2EVM and even at 85C I did not observe any relative error in the output current above 0.5% across the entire input range. In your data you show that even at 85C, the current through R226 is still very accurate. This leads me to believe that what you are seeing is an increase in the leakage current through your protection circuitry (D14, TVS1, and TVS2 in your schematic). As a quick test you may try conducting your temperature testing with these components removed and see if you still observe significant error. If this solves the problem then I would recommend using lower leakage components, such as those used in our EVM. 

    Best,

    Zak Kaye

  • 0 in reply to Zak Kaye
    Prodigy 70 points

    Hi Zak,

    Thank you so much for your help!

    I've re-tested our design by removing D14 and D16. You are right! The relative accuracy issue at high temperature was gone!

    The red curve in the attached diagram shows the relative accuracy after protection Schottky diodes were removed.

    Now it is the time to solve this problem. Of cause I use the same type of Schottky diode, SS16-E3/5AT, that used on TI evaluation board.

    However, I have some questions:

    (1) Why do we have to use Schottky diode? Can we use normal Silicon diode? 

    (2) TI's XTR111 evaluation board doesn't have diode connected in the 4-20mA loop. Therefore, a protection diode is needed to prevent 4-20mA circuit is damaged in case high voltage appears on 4-20mA load. However, in our design, we have 1n4148 (D15 and D17) in the 4-20mA output loop, this blocks the high voltage that might appear on 4-20mA load to damage internal MOSFET and 4-20mA IC. Do we still need the protection Schottky diode D14 and D16?

    Have a great one,

    William

    Root cause of 4-20mA accuracy issue at High T.pdf

  • 0 in reply to Weiran Chen
    TI__Mastermind 23050 points
    Hi William,

    1) You do not have to use a Schottky diode, but there are a few reasons you will typically see them used: They have lower forward voltages that remain relatively constant with increases in current, and thus ensure that the output will not significantly exceed the supply. They also have lower capacitance and are able to respond more quickly to transients. However, they do tend to have larger leakage currents, and in a current loop as you have seen this may introduce more error than is acceptable.

    2) This ultimately depends on what you expect to see and what you want to protect against. The Schottky to the supply has the advantage that it will clamp any over-voltages to the supply rail, which stops the XTR from being damaged and limits the stress on the surrounding components. However, the effectiveness of using the Schottky depends on if your source is able to sink the current flowing through the diode during an over-voltage condition. If not, then your supply voltage will rise and if it rises past the absolute maximum of 44V, then this may be more detrimental than beneficial. Something else to keep in mind is that the drop across D15 and D17 will contribute to your compliance voltage, so be sure you have sufficient room to maintain regulation after your load and wire resistances are factored in.

    Best,
    Zak Kaye
  • 0 in reply to Zak Kaye
    Prodigy 70 points
    Thanks a lot, Zak!
    I've got all answers needed to solve this problem.

    Best regards,

    William Chen