PGA308EVM: One Wire Calibration Connections PGA308EVM

Ryan,

I am working on a comprehensive answer to your question.  I will get back with you tomorrow.

Art

Art,

I made it through page 12 steps and get an error message "Zero DAC is out of range: -2.5<Fine_Gain<2.5" and sometimes "Fine Gain Out of Range: Fine Gain Range: 0.33333<Fine_Gain<1.0000"

I am connected identical to pages 5 and 6 with the exception of the 9V supply and 12V floating supply using different outputs of the same power supply.  The floating supply and INA power are isolated to separate power supplies.  I noticed your bottom meter on page 6 is reading 14V.  Is the target +-12V here?

I double checked all of the jumper settings. 

I'm running approx. 15mA on the ammeter during cal and approx 25mA after cal. 

Any suggestions?  Thanks for the help!

-Ryan 

Ryan,

• I think the problem is the "9V supply and 12V floating supply using different outputs of the same power supply.".  This may or may not work depending on the supply.  In some power supplies the ground of the different channels are tied together and for other supplies the channels really are floating relative to each other.  If you have an extra power supply I think it would be good to use a seporate supply for these two items.  Note that the +/-12V supply and the 9V supply don't have to float from each other, so maybe you can just rearrange your connections.  
• The actual voltage that each input requires is a range of voltages.
  • T2 - The connections labeled +/-15V on the USB-DAQ-Platform could range from 12V to 16V (I used 14V)
  • T1 - The connection labeled +VL an -VL are the loop supply.  This needs supply needs to float.  This voltage could range from 12V to 24V.  I set it to 12V in my experiment.
  • T3 - The 9V input can range from 6V to 10V.  This was originally intended to be used with a "wall-wart" (wall power supply ).  That supply was removed from the EVM kit at some point, so now you need to find an additional supply or purchase a wall-wart from digikey or other.  I used a 9V wall-wart.
  • no other external connections are required.
• The Zero DAC out of range and/or the Fin_Gain out of range happens when the calculated sensor output span and/or offset is too large to correct for.  The calibration algorithm calculates the sensor voltages by taking the PGA output and dividing by the gain.  So, if the PGA is not operating properly (e.g. it its output is saturated), the software calculates incorrect values for the sensor output and the algorithm says that it cannot correct these large errors.  This error could also happen if the sensor output are truly too large.  I set the simulated sensor outputs on page 11 to 0.5m and 5m.  This is the normalized sensor outputs so the sensor output voltage is 0.5mV/V and 5mV/V multiplied by the excitation voltage.  You should double check this to make sure you entered these numbers correctly.  Note that the units "m" need to be lower case for milivolt.  This software is case sensitive so it needs to be lower case "m".
• You should double check the jumper settings.  It only takes one incorrect jumper setting to make the calibration fail.
• I can't think of any other suggestions.  If these suggestions don't help, we may want to do a webex / phone call.

best regards,

Art

Art,

Thank you for looking into this, we certainly appreciate it.  I'll isolate the 9V and 12V supply as well as do all of the recommended settings and jumper checks.  Hopefully we find something obvious!  I won't be back in the lab until tomorrow but will follow up ASAP.  It's encouraging to know we can do a call if we need to.

-Ryan 

Ryan,  Happy to help and looking forward to hearing your results.

Art

Art,

Great news, we ran a successful 4-20 cal on the internal chip and are ready to write to our external chip.  I isolated the 12V and 9V with no change.  It turned out to be that my +14 and -14 source ground termination was incorrect.

Our external circuit is identical to figure 6-3 above.  Can you assist with similar guidance for writing calibration gains to that chip?

Best,

Ryan 

Art,

Thank you, this certainly helps!  We're just about done in the lab for the week and will give Option 2 a go when we pick up next week.

This looks pretty straight forward for the most part.  I assume there are software settings I'll need to adjust.  Specifically the one wire control within the load preset tab? (image below)

Art,

We really appreciate the detailed and thorough help!  

I'll give this a go later this week when I'm back in the lab and follow up.

Best,

Ryan 

Ryan,

Thanks!  I look forward to hearing the results.

Art

Art,

Okay so not quite there but made more progress.  Getting the error message "Fine Gain Out of Range: Fine Gain Range: 0.33333<Fine_Gain<1.0000" after applying upper limit to sensor (full bridge load cell).  Can you help verify my variable (4) below and look for any other suggestions?  Thank you! 

Sensor module connections: connected my sensor module (containing XTR116, PGA308, and other components identical to figure 6-3 above), using page 11 connections to J3

Sensor connections: double checked and wired via module output diagram

DAQ power connections: double checked and wired via page 10 connections  

DAQ jumper settings: double checked and wired via page 5 settings (same as internal settings from before)

Variables I'm looking into: 

(1) My sensor bridge resistance/span/gain are within PGA308 tolerable ranges.  Currently using a small puck style 250lbf full bridge load cell with a nominal gain of 2.0+-1% mV/V, an input impedance of 750ohm, and an output impedance of 700ohm.

(2) Ensuring exact application of load to puck.  Was using weights in the lab which were only approximately 2/3 of sensor capacity but consistent.  I assume this is okay for a proof of concept, giving an inaccurate mA/lbf output, but okay for just initial go-no go system checks.  I did run a cal with the sensor disconnected and received the "Zero DAC is out of range: -2.5<Fine_Gain<2.5" so I know the sensor is being seen by the 308.  I did make it through cal at one point to the second low point step and then errored out on that step.  The DAQ measures the upper limit @ 18.76mA and lower limit @ 18.77mA which doesn't make sense.  

(3) Verify sensor module board traces and pinouts/footprints match figure 6-3.  I've double checked these but will check again.  The sensor excitation is correct and 4-20 loop is running in the right ballpark.

(4) Verify the XTR Scaling tab within the load presets tab does or does not need to match sensor module circuit.  I changed these component values from default to match the sensor module circuit with no change either way.  Images below.

Default Settings:

Modified Settings To Match 6-3 sensor module:

Error message and outputs:

Ryan,

Thanks for the excellent notes.  I hope we can resolve your issue quickly.

1. Do you know the sensor offset?  The sensor span is 2mV/V.  With a 4.096V reference this is approximately 8mV full scale.  Some sensors have very large offsets (e.g. the outptut would move from 10mV to 18mV.  For some of these unusual cases, the PGA308 may not have enough correctoin range.  I doubt this is the case, but if you can measure the raw sensor zero-scale output and full-scale output that would be helpful to confirm.
2. Once you can tell me the measured zero and full scale sensor output I can emulate that condition to make sure the PGA308 can calibrate that sensor output range.
3. Based on the last image, I think there is a hardware issue.  The measured offset and measured full-scale read the same.  This tells me that one of the following is probably true:
   1. The sensor is not connected to the input or the sensor excitation is not connected.  Can you measure the raw Vexc and sensor outputs on the PGA308 input pins.
   2. The PGA308 output is not connected to the XTR, or the XTR has incorrect resistors for the sailing options.  When I look at my PGA308 schematic, the resistor values are different from my software.  So, I need to go into the lab and measure what is actually populated and make sure that the software matches the actual circuit.  It looks like you already did this verification.  I think the algorithm would work ok even if the resistor values were not correct as long as they were in the proper range.  This is true because the calibration algorithm corrects for gain error.  There is a limit to how much correction it can to so Ideally these resistors would be the same in software and hardware.  I don't think the error in your case could be from a small difference between the software and hardware values for these resistors.  The reason it needs to be a very significant issue is that the result is the same for zero-scale and full-scale.
   3. During calibration, you should include a debug ammeter in series with the current loop.  As the load is adjusted and the calibration procedure proceeds, you should see variation in the 4mA to 20mA.
   4. The PGA308 is communicating, so we know the XTR powers the device and it is basically functional.  So the question is why does the output stay at only one level.  If possible you should monitor Vin1, Vin2, Vs, Vref, and Vout on the PGA308 during calibration.   Does Vout stay at one specific value?
4. Another possibility is that the software is not configured correctly.  This seems unlikely to me as the settings should be almost identical for the EVM with sensor emulator test and the sensor module test.  If the debug above doesn't yield any good results, you could wire your sensor to terminals T5 and T3 on the EVM.  Thus, you would be using your actual sensor, but the PGA308EVM Test Board for the PGA308+XTR.  If this test works, than the issue is the difference between the EVM and your PCB.

I hope this helps!  Best regards, Art 

Art,

Thank you!  This is becoming quite the long support thread for you and we really appreciate the help.  I won't have time back in the lab until Monday but will start following up on these now for a placeholder.  I ran through all of the PCB traces, datasheet pinouts, and component placements with no problems found. 

1: Sensor offset is listed as <1% of rated output (~8mV like you mentioned) so calculation yields a very low 0.08mV.  Will follow up and measure this in the lab.

2: Will follow up and measure this in the lab.

3a: Vexc was running right at ~4.096V (measured), and didn't get a voltage across Sense+ and Sense- which I guess would mean offset is low. Will follow up and measure this in the lab and verify Vexc at the actual 308 (pin 10 Vref) as well as grab 308 output voltage from 308 (pin 9 Vout).

3b: I measured/configured these both ways with same error but will certainly not lose sight of this as calibration develops.

3c: Original results on the ammeter yielded strange values; sometimes would drop to ~7mA for both low and high cal points and then ~18mA for both high and low on other runs...couldn't make sense of it.  Will follow up and measure this with actual values.

3d:  Will follow up and measure these in the lab.  I would be mostly curios about Vout as well...

4:  Great idea with using the EVM board as troubleshooting logic!  I'll give this a go if all else fails above.

Thanks again and I'll plan on following up Monday.

Best,

Ryan 

Art,

Thank you, this logic will help me run through diagnostic measurements as I'm debugging.

I'll be back in the lab tomorrow and will hopefully be responding back with something definitive!

Best,

Ryan 

Ryan,  Excellent.  I look forward to hearing the progress.

Art

Art,

Okay so I made some progress but still fail with same "Fine Gain Out of Range: Fine Gain Range: 0.33333<Fine_Gain<1.0000".

Responses to measurements/questions:

1: Sensor offset is 8mV.  measured zero scale = 0mv full scale = 8mV

3a: raw Vexc @ pin 10 on the 308 is 4.086 which matches sensor Vexc at connector.

3b: image below, adjusted to external module circuit.

3c: amperage (I didn't get good results here and may need to reconfigure my meter, levels fluctuated around 1-1.5mA which doesn't make sense)

3d: (value = zero scale-full scale): (Vin1 = 2.044-2.048), (Vin2 = 2.044-2.040), (Vs = 4.853-4.853), (Vref = 4.086-4.086), (Vout = 0-3.990)

4: Sensor "passes" calibration using EVM board but automatically switches the output mode to Vout.  (image below).  Is it possible that I'm missing a software setting somewhere for scaling to my sensor module instead of EVM?  Images below 

Art,

Thank you, I really appreciate the detailed help.  This is encouraging!  I was really hoping it could be software related as I've triple checked the PCB, components, and wiring.  

One question:  Do I configure the block diagram settings to the same settings shown on page 7?

I'll be back in the lab in the morning and will run through these scenarios.  I loop back with my results.

Best,

Ryan 

Art,

Thank you, I ran again this morning with mixed results.

Excluding the load post cal feature I was able to calibrate and measure post cal results successfully and accurately with the EVM board connected to my load cell, and not revert to Vout settings.  So we're good to go there as far as software settings go there.

 Unfortunately when connecting the external 4-20 board configured identical to figure 6-3 on page 78 of the PGA308 manual we still fail with the same fine gain error.  Something is going on with my current loop.  Measurements of the EVM vs external 4-20 are as follow.  Voltages look nearly identical, while the current is not.  I think at this point I'm leaning towards my XTR being damaged/faulty or capacitors being misplaced.  I'm going to layup another board with new components to rule this out.

I'll loop back with results once I layup a new board and test. 

Thanks again for the help!  Hopefully this will be the missing link. 

Art,

No luck with swapping components.  Swapped the XTR116, capacitors, and BJT with new and correct value components and have the same results as listed in the table above.  A few questions listed below.

1: Is it possible the PGA chip is damaged and still has the correct voltage outputs above?   I'm going to start laying up an entire new board to verify that.

2: Are there specific jumper settings on the DAQ that need to be changed when switching from the EVM board to the external 4-20 module?

3: Are there specific software settings that need to change when switching from the EVM board to the external 4-20 module? 

The calibration passes when sensor is connected through the EVM board but fails with fine gain error when the EVM is replaced with the external 4-20 module.

Best,

Ryan 

Ryan,

If you like, send me an E2E friend request so that we can switch to email, and set up a meeting.  I think a call might be the quickest way to debug this issue.

Best regards, Art 

Art,

Okay great, I appreciate it!  A call might be best.  I sent a friend request and I can put together the responses to your last thread realistically on Monday when I'm back in the lab.

-Ryan