CAI Webcontrol Analog-In noise issue

[CAI_Support]

If you replaced C2 and noticed the analog noise reduction, that means your power source has high internal impedance. C2 only uses to filter incoming power line. A good DC power supply should has very low internal impedance to any noise.

 

[roussell]

Guys, do you think we should start a dedicated sub forum for this CAI controller?

Yes, please.

Terry

 

[electron]

Yes, please.

Terry

Done, you can find the CAI Webcontrol subforum here. I moved most of the threads discussing the CAI controller, but if I missed any, let me know via PM.

Back to analog-in noise issues

 

[CAI_Support]

For measureing low level analog input, LM2902 low cost operational amplifier can be used. It is sold two for 99 cents on eBay.
Here is the pin out of the chip and wiring. For using with WebControl, you can have 5V power from the board to power this chip by connecting +5 to pin 4 and ground to pin 11.

For get gain of 10, R2 will be 9 times of R1. Please note R1 will always be 10k. If R2 is 90K, then the total gain is 10, so that full scale of Webcontrol will be 1V. If you need to have even smaller voltage to measure, you can use R2=990K, then you will get gain = 100, WebControl analog full scale will be 0.1V.

LM2902 has four channels.

 

[rossw]

To the original poster: I've helped a number of people (myself included) who are running off-grid and need to measure battery voltages.
Most of us are running 48V, some 24V. A simple divider results in too little resolution of the 0-10V input.
The easiest answer is to use a zener diode to make a suppressed-zero analog input. Eg, a 21V zener in a 24-volt nominal system, will make your analog input read (nominally) 0-10V for 21 to 31V input range. For 48V, I use a 2:1 divider and two 20V zeners in series to read 40-60V.
Adding a capacitor, around 1uF to the analog input helps reduce jitter.
I also read the A/D input and average it many readings.

 

[Machineman]

To the original poster: I've helped a number of people (myself included) who are running off-grid and need to measure battery voltages.
Most of us are running 48V, some 24V. A simple divider results in too little resolution of the 0-10V input.
The easiest answer is to use a zener diode to make a suppressed-zero analog input. Eg, a 21V zener in a 24-volt nominal system, will make your analog input read (nominally) 0-10V for 21 to 31V input range. For 48V, I use a 2:1 divider and two 20V zeners in series to read 40-60V.
Adding a capacitor, around 1uF to the analog input helps reduce jitter.
I also read the A/D input and average it many readings.

I thought the Zener puts out a constant voltage specified by its designed break down voltage. How does the Zener provide a 21V - 31V range? Can I see the circuit. I'm using a Zener in a seperate circuit reduce 24V to 15V to power a 15V presure sensor.

Thanks.

 

[rossw]

I thought the Zener puts out a constant voltage specified by its designed break down voltage. How does the Zener provide a 21V - 31V range?

A zener in parallel with the load will provide a (reasonably) stable voltage output, sure.
But a zener in SERIES with your divider will drop that same bandgap.

 

[CAI_Support]

Thanks Ross! I think this is a very useful information for many people using WebControl monitor their off grid power systems.

 

[fwd03]

I put a 470uF 16V capacitor crossing the temp sensor pin 1 and 3, the analog-in noise dropped to 0-1 only!

 

[ChubRock1]

I put a 470uF 16V capacitor crossing the temp sensor pin 1 and 3, the analog-in noise dropped to 0-1 only!

fwd03 - I am trying to limit the noise as well but it still seems to be jumping around in the 5-9 range. Is all you did was stick one end of the capacitor into the socket for thermo pin 1 along with the one wire sensor for pin1 and stick the other wire from the capacitor into the socket for pin 3 along with the one wire sensor for pin 3? So basically it just looks like a capacitor is plugged in along with the one wire sensor wires? Or am I missing something?

 

[fwd03]

Yep, the terminal 1 of the temp sensor is ground, and terminal 3 of temp sensor is 5V. How big is your capacitor? I use 470uF16V and the analog noises come from 5-9 down to 0-1.

 

[ChubRock1]

I have tried with a new 220uF 35V and a 470uF 35V and even an old 470uF 16V and none of them seemed to have any real impact. I also used two of the 35V ones across 1 and 3 of both the temp and humi sensors and again it didn't seem to do anything?

Also, I have tried the following and still get 0-9 bounce on the inputs:

Run the unit from a 9V battery instead of 9V wall wart - no change to analog bounce
Disconnect just the temp leads and leave both 470uF caps plugged in to 1 and 3 on the temp and humi pins - no change
Disconnect both temp and humi leads, so only the caps are plugged into the ports - no change

Thoughts of what to try next?

 

[CAI_Support]

running on a 9V battery will not help because its high internal resistance. When you disconnect the 16 pin header, does your noise level go lower?
There are two regulators on the board. One regulates the incoming 9V to 5V, the other regulates 5V to 3.3V.
C4 is for filtering 5V power, C6 is for filtering 3.3V power. You can try to put capacitor in parallel of C4, then try C6 to see if any of that position helps.
From what we tested, the A/D noise is from power side. If the noise is lowered when 16 pin header disconnected, it is possible the noise is from the wiring on that side.
If that is the case, adding 1000pF capacitors to the A/D input will help reduce the A/D noise.

 

[ChubRock1]

I actually have nothing plugged into the 16 pin at this point. So with the only connection to the board being the power, I still get 0-9 bounce. So it would seem this has to be coming from the power. What size Cap should I try on C4 and C6? I just held a 220uF and 470uF to the contacts and it didn't do anything to the bounce?

 

[CAI_Support]

You can have the "-" of the 470uFcapacitor attached to the ground pin on the end of TTL output block, then have positive side (longer lead) touch the 5V side or 3.3V side on the small regulator.
Please make sure the 470uF is a big capacitor, not the little disc capacitor.