ADS1115 excessive noise in differential mode
Moderators: adafruit_support_bill, adafruit
Please be positive and constructive with your questions and comments.
- johnclulow
- Posts: 10
- Joined: Mon Aug 21, 2017 9:40 pm
ADS1115 excessive noise in differential mode
I ordered an ADS1116 and UNO recently and am having difficulty with my application. The 1X gain range for ADS1115 says +/- 4.096VDC. I have a +/- 15VDC signal and used a voltage divider to attenuate to +/- 3.9VDC. Putting the tranducer common on A1, I'm able to read from -3900 to +3900 mVDC from the LVDT output signal, but the readings are extremely noisy peak- to-peak, 200mV even with ample low pass filter. If I tie the UNO ground to the transducer common, the readings are very stable, but I lose most of the negative range and I am, in effect, only getting 15 bit resolution. Do I need to further attenuate my signal to +/- 2VDC and then tie the common to UNO ground so that the entire range fits within the 4.096 Volts? I read elsewhere that "all voltage inputs must be between ground and Vcc", and maybe that means that the +/- 4.096 range for 1X must occur with the maximum difference between A0 and A1 being 4.096? If so, perhaps that is the source of the noise I'm seeing. I'm not clear on what the limitations are.
- adafruit_support_mike
- Posts: 67485
- Joined: Thu Feb 11, 2010 2:51 pm
Re: ADS1115 excessive noise in differential mode
Post a photo showing your hardware and connections and we'll take a look. 800x600 images usually work best.
- johnclulow
- Posts: 10
- Joined: Mon Aug 21, 2017 9:40 pm
Re: ADS1115 excessive noise in differential mode
I could do that, but photos would not be of much use in this case. Here is a word picture: I have a position sensor that is powered with +/- 15VDC from a linear supply. Within the stroke measurement range, it provides +/- 13VDC with linear change based upon position. These voltages are relative to a common. Four wires -- +15, -15, common, and signal. I have that signal going to a voltage divider to (1) a 5000R and (2) a 2200R to common. At the junction 1 and 2, a 10KR connects from there to pin A0 on the ADS1115 with the total 50uF ceramic caps from this signal line to common. My application changes position slowly, and this low pass should take out virtually everything over 1 Hz.. The common (to A1) from the transducer is connected to case / earth ground at a star point to avoid ground loops. I have a connector on the board that accepts the ADS1115 pins, and when I use my multimeter to measure voltage on the connectors that correspond to A0 versus A1, I get +/- 4000 mV as expected versus +/- 13VDC from the transducer.
With the Arduino power supply ground separated / floating relative to earth ground, when then I plug in the ADS1115, I get a +/- 4000 mV range reading also, but there is high frequency oscillation with a range of about 200 mV that makes the data unusable.
If I tie the Arduino ground to the earth ground, I then get a rock solid result from the ADS1115, but it only covers a total range of about 4 VDC -- from +4 to about -0.2 VDC. It is no longer reading the negative range from the transducer.
So my question and working hypothesis at this point is that I am trying to measure an actual range of 8 VDC (+/- 4VDC as I interpreted the Adafruit ADS1115 library I'm using) but the module appears only able to provide a measurement _range_ of 4VDC at 1X, Early this morning I am going to replace the 2200R in the divider with 909R which should further attenuate the signal to +/- 2VDC. We'll see if I can then measure it with the Arduino gnd tied earth and get a stable result. If so, it would mean that there is a constraint that the total range of voltages at 1X must be equal to or less than 4V, and if not, it means the problem lies elsewhere.
Perhaps you can provide that information, but if my question remains unclear, I will just report the results of my experiment here for information purposes. If the reduced attenuation works, I'll be able to measure the full transducer output with 16 bit resolution and negligible noise, and that will work OK in my application.
With the Arduino power supply ground separated / floating relative to earth ground, when then I plug in the ADS1115, I get a +/- 4000 mV range reading also, but there is high frequency oscillation with a range of about 200 mV that makes the data unusable.
If I tie the Arduino ground to the earth ground, I then get a rock solid result from the ADS1115, but it only covers a total range of about 4 VDC -- from +4 to about -0.2 VDC. It is no longer reading the negative range from the transducer.
So my question and working hypothesis at this point is that I am trying to measure an actual range of 8 VDC (+/- 4VDC as I interpreted the Adafruit ADS1115 library I'm using) but the module appears only able to provide a measurement _range_ of 4VDC at 1X, Early this morning I am going to replace the 2200R in the divider with 909R which should further attenuate the signal to +/- 2VDC. We'll see if I can then measure it with the Arduino gnd tied earth and get a stable result. If so, it would mean that there is a constraint that the total range of voltages at 1X must be equal to or less than 4V, and if not, it means the problem lies elsewhere.
Perhaps you can provide that information, but if my question remains unclear, I will just report the results of my experiment here for information purposes. If the reduced attenuation works, I'll be able to measure the full transducer output with 16 bit resolution and negligible noise, and that will work OK in my application.
- johnclulow
- Posts: 10
- Joined: Mon Aug 21, 2017 9:40 pm
Re: ADS1115 excessive noise in differential mode
With the ADS1115 is connected to the voltage divider center (V0) and to the UNO GND, and V1 to the divider ground, the transducer readings from the voltage divider are -2.012 to +1.888 VDC. on the multimeter, the same as the multimeter reads without the UNO and ADS1115 connected.
If either the UNO Ground or the A0 wires (or both) are disconnected from the voltage divider, the result is the same. But when both wires are connected to the divider, the transducer readings range from -0.710 to +1.890 VDC on the multimeter. (The A1 connection does not affect this). So the ADS1115 appears to be pulling the voltage on the divider down from -2 to -0.7 when V0 and GND are connected to the divider.
At the greater attenuation, I am seeing the same thing I did at the +/- 4VDC level: With the UNO Gnd disconnected from the divider ground, I read -4 to +4 but with over 100 mV oscillation noise in the measurement. With the UNO Gnd connected to the divider ground, rock solid values but with the - side truncated.
Any ideas would be appreciated.
`
- johnclulow
- Posts: 10
- Joined: Mon Aug 21, 2017 9:40 pm
Re: ADS1115 excessive noise in differential mode
I have kept experimenting and may have a workaround / solution. With the passive filtering on the signal, I thought maybe the ripple was somehow associated with the grounds, so I decided to try sticking a large cap in series with the two grounds as an alternative to connecting them directly together. I used a 4700 uF electrolytic. So the connection in the above circuit has a wire going from the UNO ground to the cap + with the cap - wire going to the common / earth ground. Since you get an open if you measure with an ohm meter between the two grounds, I figured maybe this would not drop the negative voltage while passing (negating the effect of) most of the ripple (high pass configuration).
So I just checked and see that with the cap in series with the UNO and LVDT common/Earth grounds, I get a range of -4 to +4 on the VOM and ADS1115 (returning to my original divider values of 2200R and 1000R which draw the maximum 4 mA current from the signal that the LVDT allows. With a plain wire substituted for the series cap, I see the voltage drop from -4 to -0.7 on that side as before. The ADS1115 readings now match my VOM at both ends of the +/- 4VDC range.
The ripple in the data has dropped from 100 - 200 mV peak-to-peak to about 0.5 mV peak-to-peak, which I can easily live with.
The only problem is the 4700uF is reacting too slowly, so now I'll experiment with lower capacitance values on the breadboard and then make the optimized changes in my build and keep my fingers crossed.
I obviously don't know much about electronics, and it would be nice to understand why all this is happening, but I'm glad to at least have a path forward that looks viable. Once I get the cap value optimized, I'll post that for anyone interested. Any insight into the underlying electrical theory would be most appreciated.
So I just checked and see that with the cap in series with the UNO and LVDT common/Earth grounds, I get a range of -4 to +4 on the VOM and ADS1115 (returning to my original divider values of 2200R and 1000R which draw the maximum 4 mA current from the signal that the LVDT allows. With a plain wire substituted for the series cap, I see the voltage drop from -4 to -0.7 on that side as before. The ADS1115 readings now match my VOM at both ends of the +/- 4VDC range.
The ripple in the data has dropped from 100 - 200 mV peak-to-peak to about 0.5 mV peak-to-peak, which I can easily live with.
The only problem is the 4700uF is reacting too slowly, so now I'll experiment with lower capacitance values on the breadboard and then make the optimized changes in my build and keep my fingers crossed.
I obviously don't know much about electronics, and it would be nice to understand why all this is happening, but I'm glad to at least have a path forward that looks viable. Once I get the cap value optimized, I'll post that for anyone interested. Any insight into the underlying electrical theory would be most appreciated.
- johnclulow
- Posts: 10
- Joined: Mon Aug 21, 2017 9:40 pm
Re: ADS1115 excessive noise in differential mode
Having tried a wide range of capacitance, it seems there isn't much difference. I saved a sample of 100 means of 25 samples each at 1X in differential mode. This is 20 averages per second for five seconds. Then a calculated the standard deviation as a numeric index for comparison.
With no capacitor -- a direct connection between UNO ground and the common / earth ground of the transducer / enclosure, I consistently got values around 90 mV. I don't think the values form a normal distribution, but at 2 Sigma, you'd be around 200 mV which is the approximate range I was seeing.
Standard Deviations for several caps were:
0 F - direct connect - 93.5 mV
4700 uF electrolytic - 0.17 mV
10 uF ceramic (106) - 0.24 mV
1 uF ceramic (105) - 0.26 mV
So I'm planning on using a 1 uF as a high pass couple for the ground planes and move on even though I don't understand what's going on with the direct connect phenomena.
With no capacitor -- a direct connection between UNO ground and the common / earth ground of the transducer / enclosure, I consistently got values around 90 mV. I don't think the values form a normal distribution, but at 2 Sigma, you'd be around 200 mV which is the approximate range I was seeing.
Standard Deviations for several caps were:
0 F - direct connect - 93.5 mV
4700 uF electrolytic - 0.17 mV
10 uF ceramic (106) - 0.24 mV
1 uF ceramic (105) - 0.26 mV
So I'm planning on using a 1 uF as a high pass couple for the ground planes and move on even though I don't understand what's going on with the direct connect phenomena.
- adafruit_support_mike
- Posts: 67485
- Joined: Thu Feb 11, 2010 2:51 pm
Re: ADS1115 excessive noise in differential mode
The basic problem you're dealing with is that of tying a floating signal to a reference voltage in some other circuit. To make things more fun, the floating signal is differential and the circuit where you want the reference is single-ended.
The standard circuit for that looks like this:
The 1k-1k voltage divider all the way to the left creates a reference voltage at 2.5v between the Arduino's supply rails. The capacitor network keeps the voltage across the two 1uF caps more or less equal, with the voltage of the 1k floating resistor split between them. That forces the voltages at A0 and A1 to stay centered around 2.5v, going up and down symmetrically.
There's no DC connection between the 2.5v reference and the floating signal, but there is a low-pass filtered AC connection.
Without that kind of reference, the floating signal can swing all over the place between the Arduino 5v and GND rails, mostly under the control of parasitic connections between the Arduino and the floating supply. If those connections pull one side of the floating signal too close to a rail, the other side of the signal will get clipped by the input protection diodes on the ADC's input pins.
You can boost the sizes of the 1uF caps and the resistors in the floating voltage divider to lower the frequency response. Any parasitic connection that wants to pull the average voltage of A0 and A1 away from 2.5v basically has to push charge out of one 1uF cap and into the other. The only path where that can happen is through the floating 1k resistor, which limits the current.
Making the resistor larger reduces the amount of current that can flow from one cap to the other, but you have to remember the input impedance of the ADC.. the ADS1115's differential impedance at 4.086v is about 15M, so a the load on a 1k resistor will be about 67ppm, or one LSB at 14-bit resolution.
Making the caps larger reduces the rate at which the caps can change voltage at a given amount of current.
The standard circuit for that looks like this:
The 1k-1k voltage divider all the way to the left creates a reference voltage at 2.5v between the Arduino's supply rails. The capacitor network keeps the voltage across the two 1uF caps more or less equal, with the voltage of the 1k floating resistor split between them. That forces the voltages at A0 and A1 to stay centered around 2.5v, going up and down symmetrically.
There's no DC connection between the 2.5v reference and the floating signal, but there is a low-pass filtered AC connection.
Without that kind of reference, the floating signal can swing all over the place between the Arduino 5v and GND rails, mostly under the control of parasitic connections between the Arduino and the floating supply. If those connections pull one side of the floating signal too close to a rail, the other side of the signal will get clipped by the input protection diodes on the ADC's input pins.
You can boost the sizes of the 1uF caps and the resistors in the floating voltage divider to lower the frequency response. Any parasitic connection that wants to pull the average voltage of A0 and A1 away from 2.5v basically has to push charge out of one 1uF cap and into the other. The only path where that can happen is through the floating 1k resistor, which limits the current.
Making the resistor larger reduces the amount of current that can flow from one cap to the other, but you have to remember the input impedance of the ADC.. the ADS1115's differential impedance at 4.086v is about 15M, so a the load on a 1k resistor will be about 67ppm, or one LSB at 14-bit resolution.
Making the caps larger reduces the rate at which the caps can change voltage at a given amount of current.
- johnclulow
- Posts: 10
- Joined: Mon Aug 21, 2017 9:40 pm
Re: ADS1115 excessive noise in differential mode
Thanks, Mike. I'll take a look at the schematic you provided. What I'm finding is that I cannot use the differential capability of the ADS1115 if I attach the A/gnd reference to the UNO ground. When I do that, the voltage is essentially truncated on the negative side of the bipolar differential mode. Here is an update from a trial in the application yesterday: The system worked fine using my notebook computer to power the UNO, with a separate power supply to the transducer linking earth ground to signal ground on the LVDT -- I was getting well-centered, full range +/- 4.016 V from the transducer output. (As I mentioned, connecting the UNO ground plane with the LVDT ground plane/earth ground using a capacitor in series fully resolved the noise problem -- I'm using 1 uF, but just about anything in the range 100nF to 100 uF seems to work equally well. ).
When we switched to the customer's desktop, however, the (-)Voltage side was again almost fully truncated. The reason for this turns out to be that the earth ground for the notebook stops at the power brick -- the notebook itself isn't connected to earth ground so that the USB ground from the notebook is not connected to earth ground. In the desktop, the USB ground is connected to earth/case ground by the manufacturer so that connecting the USB cable from the desktop to the UNO also connects the UNO to earth ground, and with both the UNO and ADS1115 A1 reference grounds connected to earth, the - side of the bipolar differential was truncated.
If only the USB shield were connected to earth ground from the desktop, I could disconnect that from the USB B plug that goes into the UNO. But since the USB ground pin is what is connected to earth, there is no way to disconnect it. So what I am now going to try is to reverse the grounding so that the UNO ground is connected to earth ground via the USB cable but the LVDT signal and common lines will be floating and connected to A0 and A1 of the ADS1115, respectively. The LVDT cable shield from the metal UNO/ADS enclosure will, of course, remain connected to earth ground at the enclosure end but the signal common will be disconnected from it. I will again use the series capacitor between the LVDT common and the UNO ground/ earth ground as before and hope this arrangement continues to give me rock solid, centered full bipolar 16 bit range results with the desktop as I was getting with the notebook. I'm going to try this tomorrow and will post my results in this thread for informational purposes. I'm hopeful it will work. If not, I'll give the circuit you suggested a try.
Thanks again for taking the time to comment, Mike. I really appreciate it and all of the support Adafruit Industries provides in so many ways.
When we switched to the customer's desktop, however, the (-)Voltage side was again almost fully truncated. The reason for this turns out to be that the earth ground for the notebook stops at the power brick -- the notebook itself isn't connected to earth ground so that the USB ground from the notebook is not connected to earth ground. In the desktop, the USB ground is connected to earth/case ground by the manufacturer so that connecting the USB cable from the desktop to the UNO also connects the UNO to earth ground, and with both the UNO and ADS1115 A1 reference grounds connected to earth, the - side of the bipolar differential was truncated.
If only the USB shield were connected to earth ground from the desktop, I could disconnect that from the USB B plug that goes into the UNO. But since the USB ground pin is what is connected to earth, there is no way to disconnect it. So what I am now going to try is to reverse the grounding so that the UNO ground is connected to earth ground via the USB cable but the LVDT signal and common lines will be floating and connected to A0 and A1 of the ADS1115, respectively. The LVDT cable shield from the metal UNO/ADS enclosure will, of course, remain connected to earth ground at the enclosure end but the signal common will be disconnected from it. I will again use the series capacitor between the LVDT common and the UNO ground/ earth ground as before and hope this arrangement continues to give me rock solid, centered full bipolar 16 bit range results with the desktop as I was getting with the notebook. I'm going to try this tomorrow and will post my results in this thread for informational purposes. I'm hopeful it will work. If not, I'll give the circuit you suggested a try.
Thanks again for taking the time to comment, Mike. I really appreciate it and all of the support Adafruit Industries provides in so many ways.
- johnclulow
- Posts: 10
- Joined: Mon Aug 21, 2017 9:40 pm
Re: ADS1115 excessive noise in differential mode
Problem Solved, I'm happy to report. This is a summary recap that may hopefully save others some headaches:
The Problem: I found I was unable to use the ADS1115 in differential, bipolar mode if the analog reference (A1 on the ADS) was hardwired to the UNO ground. When the grounds were connected, only half the bipolar range (+) was obtained. Separating the UNO ground plane from the ADS1115 ground plane, however, while it allowed full differential bipolar access, resulted in excessive signal noise with a standard deviation of 90, making the system unusable.
Noise Solution: Placing a capacitor (value seemed relatively unimportant, I ended up using a 1 uF ceramic) between the two ground planes in series fully resolved the noise problem. In series, the cap is open to DC current but will conduct alternating voltage current.
Solution A - laptop only: I first tried to connect earth ground (case ground for my metal application enclosure) to the analog signal common/return and make sure the UNO ground and all associated connections were not directly connected. (the ceramic cap was in series and open insofar as DC is concerned). This worked quite well with my laptop.
Solution B - When trying this with a desktop computer, however, I witnessed the same truncation of one side of the ADS bipolar range. I surmised that now the UNO ground was also connected to earth ground and subsequent testing showed that this was the case on two different brands of desktop computer: not only the USB shield but also the USB ground pin are connected to earth ground / case ground for the desktop. (This is not the case for the laptop since the ground pin on the line receptacle plug only goes to the power brick, not to the laptop itself. )
So I decided to try swapping things around and switch the earth ground to the UNO and disconnect earth ground from the sensor common, leaving earth ground connected to the sensor cable shield, however. This arrangement worked perfectly well on both the laptop and on the desktops. Hopefully this account will be of value to someone else who is as perplexed as I was at this, not having a background in electrical engineering..
The Problem: I found I was unable to use the ADS1115 in differential, bipolar mode if the analog reference (A1 on the ADS) was hardwired to the UNO ground. When the grounds were connected, only half the bipolar range (+) was obtained. Separating the UNO ground plane from the ADS1115 ground plane, however, while it allowed full differential bipolar access, resulted in excessive signal noise with a standard deviation of 90, making the system unusable.
Noise Solution: Placing a capacitor (value seemed relatively unimportant, I ended up using a 1 uF ceramic) between the two ground planes in series fully resolved the noise problem. In series, the cap is open to DC current but will conduct alternating voltage current.
Solution A - laptop only: I first tried to connect earth ground (case ground for my metal application enclosure) to the analog signal common/return and make sure the UNO ground and all associated connections were not directly connected. (the ceramic cap was in series and open insofar as DC is concerned). This worked quite well with my laptop.
Solution B - When trying this with a desktop computer, however, I witnessed the same truncation of one side of the ADS bipolar range. I surmised that now the UNO ground was also connected to earth ground and subsequent testing showed that this was the case on two different brands of desktop computer: not only the USB shield but also the USB ground pin are connected to earth ground / case ground for the desktop. (This is not the case for the laptop since the ground pin on the line receptacle plug only goes to the power brick, not to the laptop itself. )
So I decided to try swapping things around and switch the earth ground to the UNO and disconnect earth ground from the sensor common, leaving earth ground connected to the sensor cable shield, however. This arrangement worked perfectly well on both the laptop and on the desktops. Hopefully this account will be of value to someone else who is as perplexed as I was at this, not having a background in electrical engineering..
- adafruit_support_mike
- Posts: 67485
- Joined: Thu Feb 11, 2010 2:51 pm
Re: ADS1115 excessive noise in differential mode
Glad to hear you got it working to your satisfaction. A lot of the time, it's just a process of finding a solution you can live with.
It sounds like getting the signals floating properly relative to each other did the trick.
It sounds like getting the signals floating properly relative to each other did the trick.
- johnclulow
- Posts: 10
- Joined: Mon Aug 21, 2017 9:40 pm
Re: ADS1115 excessive noise in differential mode
Thanks, Mike. It's great to know that Adafruit Industries supplies great tech support by folks like you in addition to product support such as the numerous libraries and product tutorials your company has shared with the community. I will certainly continue to support Adafruit.
- adafruit_support_mike
- Posts: 67485
- Joined: Thu Feb 11, 2010 2:51 pm
Re: ADS1115 excessive noise in differential mode
Glad to hear it..
I happen to be the analog circuits geek, so problems like the one you posted are pretty much catnip for me. ;-)
I happen to be the analog circuits geek, so problems like the one you posted are pretty much catnip for me. ;-)
- johnclulow
- Posts: 10
- Joined: Mon Aug 21, 2017 9:40 pm
Re: ADS1115 excessive noise in differential mode
Good to know going forward. Nice to meet you, Mike.
Like many who've become involved in various electronics products, for me the journey is the destination, or at least a big part of it, and that would likely not be possible were it not for the rich resources available via the internet provided by others doing similar things. This particular learning curve for me was primarily about the ADS1115, a module that is truly mind-blowing in its capabilities, form factor, and price: I'm amazed that we can obtain this resolution at this speed for this price. I'm running it at 16-bit depth at 500 samples per second at 1X. I'm sure I'll find many additional applications for this module. Thanks again for your help, Mike. Great to know folks like you are here to help us.
Like many who've become involved in various electronics products, for me the journey is the destination, or at least a big part of it, and that would likely not be possible were it not for the rich resources available via the internet provided by others doing similar things. This particular learning curve for me was primarily about the ADS1115, a module that is truly mind-blowing in its capabilities, form factor, and price: I'm amazed that we can obtain this resolution at this speed for this price. I'm running it at 16-bit depth at 500 samples per second at 1X. I'm sure I'll find many additional applications for this module. Thanks again for your help, Mike. Great to know folks like you are here to help us.
- adafruit_support_mike
- Posts: 67485
- Joined: Thu Feb 11, 2010 2:51 pm
Re: ADS1115 excessive noise in differential mode
It's a pleasure to meet you too John, and thanks for the kind thoughts.
This really is what Adafruit is about.. Ladyada went to MIT and was part of the knowledge-sharing culture there. It's a culture that used to be everywhere in the US when you could find Popular Electronics and a dozen competitors in any magazine rack, but which kind of faded away through the 1990s and early 2000s. The current resurgence is called the Maker movement, but it's the same old "test my ideas by putting them into hardware and seeing if the universe agrees with me" mindset that's driven people to tinker and share knowledge for centuries.
The support and expansion of that culture is our main product. We pay the bills by selling devices.. mostly ICs that only come in packages that are too small for any sane human to solder by hand (I'm good down to TSSOP, but think I'm gonna need new tools before I can do DQFN reliably.. conventional soldering irons just don't work with 1mm-square packages).. but the tutorials and the teaching are where we really roll up our sleeves and have fun.
This really is what Adafruit is about.. Ladyada went to MIT and was part of the knowledge-sharing culture there. It's a culture that used to be everywhere in the US when you could find Popular Electronics and a dozen competitors in any magazine rack, but which kind of faded away through the 1990s and early 2000s. The current resurgence is called the Maker movement, but it's the same old "test my ideas by putting them into hardware and seeing if the universe agrees with me" mindset that's driven people to tinker and share knowledge for centuries.
The support and expansion of that culture is our main product. We pay the bills by selling devices.. mostly ICs that only come in packages that are too small for any sane human to solder by hand (I'm good down to TSSOP, but think I'm gonna need new tools before I can do DQFN reliably.. conventional soldering irons just don't work with 1mm-square packages).. but the tutorials and the teaching are where we really roll up our sleeves and have fun.
- johnclulow
- Posts: 10
- Joined: Mon Aug 21, 2017 9:40 pm
Re: ADS1115 excessive noise in differential mode
Thanks for that background on your company. Growing up in the '50's, the exponential rates of change in so many areas are pretty stunning to me, and particularly seeing they all interact to further accelerate the acceleration; Github is a good example in one area. It's good to know now how Adafruit Industries fits into this picture. From my perspective, in spite of momentary local setbacks, there are many reasons for optimism.
And speaking of acceleration, one other thing I neglected to mention on my project was that I changed the conversion delay for the ADS1115 in the library header H file to 1 ms:
#define ADS1115_CONVERSIONDELAY (1)
and the register for the default to 0X00E0 :
#define ADS1015_REG_CONFIG_DR_1600SPS (0x00E0) // 1600 samples per second (default)
These changes to the H File seem to be giving me stable results in sending 20 averages of 25 samples each per second back to Excel via PLX-DAQ (Jonathan Arndt's 2.11.01 version) at 19200 baud. I think that Jonathan's contribution is definitely worth looking into for many who are interested in ADS1115 applications as Excel offers so many processing alternatives.
(PS - just ordered a second UNO and ADS1115 from Adafruit Industries to build a backup unit :)
And speaking of acceleration, one other thing I neglected to mention on my project was that I changed the conversion delay for the ADS1115 in the library header H file to 1 ms:
#define ADS1115_CONVERSIONDELAY (1)
and the register for the default to 0X00E0 :
#define ADS1015_REG_CONFIG_DR_1600SPS (0x00E0) // 1600 samples per second (default)
These changes to the H File seem to be giving me stable results in sending 20 averages of 25 samples each per second back to Excel via PLX-DAQ (Jonathan Arndt's 2.11.01 version) at 19200 baud. I think that Jonathan's contribution is definitely worth looking into for many who are interested in ADS1115 applications as Excel offers so many processing alternatives.
(PS - just ordered a second UNO and ADS1115 from Adafruit Industries to build a backup unit :)
Please be positive and constructive with your questions and comments.