If you have noise issues, it may be the RPI or the connection to the RPI.


 

John Bostwick

TVWBB Wizard
I never really ran into this before so I thought I would make a post about my recent build.

I was making a HM and I got it up and running and the Thermocouple was just giving me all kinds of noise. I first thought it was the new amps I recieved from a vendor I never used before, so replaced that and no difference.

Again, I thought it was Thermocouple circuit and I replaced all the smd parts and even purchased new ones, as some of my parts I have had for many years, no change in the noise. At this point, I was pulling out what was left of my hair. I ordered some solder flux to make sure my joints were good and soldered up another board and the same issue, temps fluctuating and noise.

I looked at every solder joint with my digital microscope and I also had to replace my soldering/air station during all this because the solder iron finally stopped working after close to 10 years :(.

Everything looked good. My Hm boards are custom made board so the traces different then the regular boards and I have added a 3 pad switch attachment for use with the Adapter boards, I was starting to think they were the cause, even though I had been using the same design for many years now and zi always get them made at OSHPARK. Maybe they changed something, as the last time I had boards made, I also had some noise issues.

Oh, I even purchased a K-thermocouple digital probe unit that came with 3 or 4 probes for $30, on Amazon, just to see if my Thermoworks probe was causing any issues and the noise was also the same on them, also.

Recently, I made an Adapter board for a customer and I tested it on the same board that had the noise and it worked perfectly fine, when I switch over to it, and when I switched back to the main Thermocouple connection right back to noise again.

When it comes to noise, I am basically at a loss without have a scope to test and to help locate the noise, but usually I have had successfully fix noise issues and usually it's just a bad solder joint.

Since, I had multiple Rpi's I figured I would try out different one, and will doing that I realized when I was touch the Rpi connection the noise went away 😳. Hmmm, I started looking at the connection and when I held the connection the noise would go away, and when I let go, it would come back.

Did some resoldering and it seems to have fixed the issue. I always assumed that the noise was coming from the HM board, power supply, or the Thermocouple circuit. Never even thought about the Rpi connection. I mean the Pi worked fine, so I figured it was good to go.

So if you have noise issues and you can't figure it out, check your PI connection, and it can save what's left of your hair.
 
Can you post a picture where you were touching? The thermocouple unit I have has always had noise. Brian spent a lot of time trying different things as well as myself. I would mess with thermocouple circuit if I had a great SMD soldering station. There are some combinations of resistors I would like to try to see if they would help. Interesting you touched a connection and problem went away. That could be telling us that the capacitance to ground might not be correct or if no capacitor on that connection, maybe one to ground would solve this problem that's been an issue for years.
 
Can you post a picture where you were touching? The thermocouple unit I have has always had noise. Brian spent a lot of time trying different things as well as myself. I would mess with thermocouple circuit if I had a great SMD soldering station. There are some combinations of resistors I would like to try to see if they would help. Interesting you touched a connection and problem went away. That could be telling us that the capacitance to ground might not be correct or if no capacitor on that connection, maybe one to ground would solve this problem that's been an issue for years.
Sorry I took so long to reply,

I have already tried multiple values of caps and on the thermocouple circuit. I bought a smd 805 cap kit(book) that has just about every value you would ever need. I didn't see any improvement.

Basically, I was holding the right side connections for the Rpi header. But, after I did some rsoldering of the connection, the noise went completely away.

What was throwing me off, was my Adapter board worked perfectly fine(no noise), it only effected the HM thermocouple circuit.

Speaking of soldering stations, my soldering rework station I bought almost 11 years ago, finally took a dump, a few weeks ago. I found a fairly cheap 3 in 1 unit that works pretty good. It has the hot air and a power supply, something I have always wanted.

I have been working on an old Peter F. Smd board for someone. This board had tons of noise, and I could not figure it out and I figured it was the caps, since they were soldered on around 201. I tried new caps, noise would not go away. Today, I realized I have the same board saved on OSHPARK.COM, so I am going to order 3 of them to just redo the whole board reusing the components that should be good, hopefully that will fix the noise of that board.

God I hate noise, even when I'm sleeping I have to wear earplugs, lol.
 
No problem. I have spent way too much time looking for noise source. I did read through the data packet for the 328 chip and on page 254, there is a power suggestion that makes sense to me on how to get rid of noise. When I was working as a tech, one of the golden rules was never feed the ADC power rail directly from the same rail that you feed your processor chip. Always use a separate source. On page 254 they show the avcc fed through a 10uh choke and a 100nf cap to ground. This definitely would block clock pulses from the avcc line. I suspect Brian might have turned on noise cancelling in the program, but I don`t know that for sure. If I had another board for a backup, I would cut the trace and add that recommended lc network and see if that's the solution. With that said, have you ever cut the avcc rail and added this lc circuit to feed power to pin 20? If I have another assembled board for backup, I would cut the avcc power feed trace and add the recommended lc network and feed avcc through it. I suspect that would eliminate or greatly reduce the internal clock pulses from getting into the ADC.
 
No problem. I have spent way too much time looking for noise source. I did read through the data packet for the 328 chip and on page 254, there is a power suggestion that makes sense to me on how to get rid of noise. When I was working as a tech, one of the golden rules was never feed the ADC power rail directly from the same rail that you feed your processor chip. Always use a separate source. On page 254 they show the avcc fed through a 10uh choke and a 100nf cap to ground. This definitely would block clock pulses from the avcc line. I suspect Brian might have turned on noise cancelling in the program, but I don`t know that for sure. If I had another board for a backup, I would cut the trace and add that recommended lc network and see if that's the solution. With that said, have you ever cut the avcc rail and added this lc circuit to feed power to pin 20? If I have another assembled board for backup, I would cut the avcc power feed trace and add the recommended lc network and feed avcc through it. I suspect that would eliminate or greatly reduce the internal clock pulses from getting into the ADC.
Im not as proficient at electronics to figure stuff out like this, lol. If you do decide to give that a try, what the worse that could happen, if it does not work, then just remove the circuit and solder it back together again
 
Going to give it a go. Only need to add the inductor since the cap is already installed from pin 20 to ground. Will post if it works in a week or so. John, do you still print cases? Mine is showing what living in dry, hot climate does to cases. Getting a little wavey. I got one of Brians last versions.
 
Going to give it a go. Only need to add the inductor since the cap is already installed from pin 20 to ground. Will post if it works in a week or so. John, do you still print cases? Mine is showing what living in dry, hot climate does to cases. Getting a little wavey. I got one of Brians last versions.
Yeah, i can print in any color and material
 

 

Back
Top