Noise graph firmware?


 
B

Brent Cooley

TVWBB Member
I wired up a small break out box to run my probes over cat5. It works but it introduced noise into the probes... Ui showing noise icon and some jagged temps. Probably not a big deal but I'd like to fix it if possible.

I remember a while back a special firmware which included a noise graph that may help with debugging. Does anyone know whether that works with the latest HM software release?
 
Brent Cooley said:
I wired up a small break out box to run my probes over cat5. It works but it introduced noise into the probes... Ui showing noise icon and some jagged temps. Probably not a big deal but I'd like to fix it if possible.

I remember a while back a special firmware which included a noise graph that may help with debugging. Does anyone know whether that works with the latest HM software release?
Click to expand...

IDK if the noise graph firmware is available and working with the latest HM software release, but I do have a couple questions for you since I run my probes over a 50+ foot CAT5 cable that has been out on the weather for over a year and I do not get the noise indicator popping up (unless something is wrong probe wise), so maybe I can help you out...

First of all, what version HM board are you running?

Are you running just the probes over the CAT5 or the probes, blower and/or servo?

If you are running more than just the probes through the CAT5 cable are you sharing a ground between power components and the probes? How long is your CAT5 cable?

On the probe end of the cable do you have a CAT5 connector (so you can use a standard two ended CAT5 cable) or do you have the wires soldered directly to the probe jacks?

Let me know a bit more about your setup and I think I can help you eliminate the noise....
 
Thanks Ralph.

I have a hm 4.1with the white wire and led 1 mods. I'm not sure the specific version number but it was an early 4.1 board.

I'm running 3 probes, servo and blower over a short 6' cable for now. The probes themselves share a ground and wire in the cat 5. The servo and fan share a different cat 5 ground wire.

I'm using the cat 5 jack on the hm and a surface mount cat 5 box from home depot.

I'll upload some pics in a little while.

I've tried thermoworks probes and braid grounded et732 probes and results are similar. They all work noise free when plugged into the hm directly.
 
Brent Cooley said:
I remember a while back a special firmware which included a noise graph that may help with debugging. Does anyone know whether that works with the latest HM software release?
Click to expand...
You can always try it and if it doesn't work just switch back. You need v11 linkmeter for the graph to show up though, and to manually download the HEX file and use the "from local machine" option to flash it.
http://capnbry.net/linkmeter/snapshots/trunk/hm-noise-20140214.hex

EDIT: I'm pretty sure this won't work though, now that I think more about it because the noise firmware may use the old debug logging mechanism?
 
Last edited:
Pictures:

HM board, CAT 5 section. Wiring is as follows:

1 - Pit Probe +
2 - Probe 2 +
3 - 5v for servo
4 - Ground (for servo & fan)
5 - Blower +
6 - Servo pulse
7 - Probe ground (all "front" pins of jacks connected here)
8 - Probe 3 +

IMG_20140809_170539.jpg




CAT 5 Surface mount "breakout box", with lid removed for now

Jack holes are drilled through the underside.

IMG_20140809_170602.jpg



If I were to hazard a guess, the close proximity of everything inside the breakout box is the cause. The first step of debugging I plan (unless you have a suggestion) is to wire another probe jack, by itself, "outside" of that little CAT 5 box, away from everything else. I'd disconnect all the other probe jacks too.
 
OK, seems you are pretty much on the right track. The HMv4.1 does not have the RC filter on the probe lines, so if I were you I would remove the jumper wires you have on the HM board for the probes and install a 100K resistor there instead and then add a 0.1uf cap from each probe lead to ground, that will add the RC filter on your probes that come in through your CAT5 wire like the new HMv4.2 boards have and should filter out the noise. Reference the HMv4.2.4 schematic to get a look at the new RC filters in place on the probe leads...

Edit: also pick the wires so the probe leads and ground are together in the same twisted pairs, don't mix the servo or blower wires in the same twisted pair as any of the probe leads or probe ground.
 
Last edited:
Thanks Ralph. The twists, if my cable is standard, are not mixing servo/fan with anything probe-related. I will try another CAT 5 cable just to make sure it's not that.

On the RC filter idea... that seems like a good idea and I had read about that in another thread. When I read the schematic of the 4.1 board and the mini-schematic Bryan posted in another thread re: RC filter, it seems to do it, one trace per probe needs to be cut and replaced by the 100k resistor. I'd love it to be as simple as your description so I hope I'm wrong (I'm far from a schematic-reading expert).

Here's a picture of what I think needs to be done. Perhaps you and/or Bryan could comment.

tracechange.jpg
 
You can cut the trace(s) and that would add the filter on both the probes that come in via the CAT5 cable AND the jacks that are built onto the HM board. If you install the filters in the jumper wires between the CAT5 jack and the probe jack(s) it will filter out the noise that is picked up in the CAT5 cable, but will not filter out any interference that is generated on the HM board ...
 
Last edited:
I *think* the noise firmware Bryan listed is working.

I get this for my pit probe (ET-732 probe with grounded braid):

pitnoise.png



This is food 1 (Thermoworks probe):

foodnoise.png


In both cases, this is with fan plugged in but not running.

This is with fan @ 100% (which is 45% max in config):

fanrunning.png



P.S. This is without any RC filter like Ralph suggested above. I'm going to try that later.
 
Brent Cooley said:
On the RC filter idea... that seems like a good idea and I had read about that in another thread. When I read the schematic of the 4.1 board and the mini-schematic Bryan posted in another thread re: RC filter, it seems to do it, one trace per probe needs to be cut and replaced by the 100k resistor. I'd love it to be as simple as your description so I hope I'm wrong (I'm far from a schematic-reading expert).
Click to expand...
So close! You've got the trace right but the capacitor needs to come after the resistor and if you put it across the probe jack terminals the cap will be before. Run the 100k resistor to replace the trace then attach one leg of the 0.1uF cap to the ATmega end, and the other leg to ground.
 
So if I understand your post right, is the following correct (for pit probe only)?

1. Cut trace shown by dark blue arrow
2. Replace trace with 100k resistor
3. Place cap shown by light blue curve (marked with CAP) - I assume I could use a ground point closer...

tracechange.jpg
 
I would use a closer ground but yup that's correct. You can use the analog ground pin on the ATmega which is something like 5 or 6 pins to the left.
 
Well, I cut the 3 probe traces and soldered in the 100k resistors and capacitors. It actually was pretty easy once I knew what needed to be done.

The most important part is: It worked! The noise graph, only "bounces" between 2 maybe 3 ADC values now with the fan running... before it was 10-15. In the 5 minutes I tried it, the HM GUI never showed the little noise icons at all, where before it was either yellow or red all the time for every probe.

Bryan and Ralph - Thanks for all the help getting this issue resolved!



Here is the end result. It's not pretty but it works.

final.jpg
 
Before I forget Bryan, I noticed one thing this morning when I was testing this final modification out... I kept bouncing between the current/stable HM firmware .hex file to see the noise icons (or lack-thereof) and the noise-graph .hex file to see the actual noise range. I noticed that the probe temps differed by ~1 degree. I didn't change my config/probe types at all, nor touched any of the probes.

ex:
1. current/latest HM firmware flashed
2. Pit probe reads 68.0
3. flash noise graph firmware
4. Refresh HM web page and pit probe reads 69.2
5. reflash current/latest HM firmware
6. Refresh HM web page and pit probe reads 68.0

This isn't a big deal since it's more or less within the probe's accuracy range anyway, I was just curious if there is some underlying software difference that may have caused this.
 
Looks great to me! It looks like every HeaterMeter board I have with wires and components all over the place and even the partially melted button cap.

Very astute observation about the temperature difference too. This is because of the ADCmax change. I've always been sort of on the fence about if the maximum ADC value is 1023 or 1024. If you check the datasheet it says ADC = Vin*1024/Vref or rearranged Vin = ADC * Vref/1024. This means that 3.3V in would be calculated out to be 3.29677734375V and I feel like it should be 3.3V especially considering ET-732 thermistors at room temperature are 3.245V.

So the change I linked lowered the maximum by 1 point (times 2 to the power of 6), which has the side effect of making the input voltage appear higher and give a slightly different result. The difference is a lot less at BBQ temperatures, on the order of 0.1-0.3F I think.
 
You must log in or register to reply here.

 

Back
Top