Is there a consensus solution on erratic readings in artisan

[jyalpert]

I wonder if that's it.  One surefire way to get my graph to spike, was to touch my USB cable to my computer chassis.  And I don't even have a touchscreen.

I think there might be some super weirdness going on with your particular laptop . . .

[yukoncornelius]

Nope, tried a friend's computer - still terrible during the roast.  The last roast I did where I actually tried to keep the thermocouple wires noticably away from any power cords - it did better.  Still something wrong, its not making roasts impossible.. its just still very annoying.  Maybe I didn't actually ground the roaster properly...

Sidenote - why when I was preheating was BT so much higher than ET for this roast?  Are you guys doing your charge according to BT or ET (i always used BT on my hottop but I've heard both on here). 

[.]

.

[jyalpert]

My BT is usually 20-30F higher than ET at drop.  It's because the BT probe is closer to the flame on the perf drum.

That still looks bad.  I'm sorry this is happening to you.  I think you still have grounding issues, frankly.

In fact . . . I'd actually suggest you move your entire rig to someone else' house and try roasting there, with your galv grounding wire installed.

FYI, a GFCI does not mean you have a proper ground.  In fact, in my house (built in 1936), most of my outlets are ungrounded, and my electrician installed GFCIs so that we could have 3-prong outlets that are at least protected.  But those 3-prong outlets are still not properly grounded.

Also, the entire thermocouple wire run from sensor > phidget > usb > computer should not touch anything.

One way to debug is to get a cheap multimeter thermometer with a thermocouple input (you can get these for like $15 at hardware stores).  See if the erratic readings are happening when you use that multimeter.  That would at least allow you to isolate the issue to the phidget/laptop vs. the thermocouples.

I also recommend changing your sampling interval in Artisan to 3s, if you're not there already.

[jyalpert]

My BT is usually 20-30F higher than ET at drop.  It's because the BT probe is closer to the flame on the perf drum.

That still looks bad.  I'm sorry this is happening to you.  I think you still have grounding issues, frankly.

In fact . . . I'd actually suggest you move your entire rig to someone else' house and try roasting there, with your galv grounding wire installed.

FYI, a GFCI does not mean you have a proper ground.  In fact, in my house (built in 1936), most of my outlets are ungrounded, and my electrician installed GFCIs so that we could have 3-prong outlets that are at least protected.  But those 3-prong outlets are still not properly grounded.  In other words, we have GFCI outlets BECAUSE they are not grounded.

Also, the entire thermocouple wire run from sensor > phidget > usb > computer should not touch anything.

One way to debug is to get a cheap multimeter thermometer with a thermocouple input (you can get these for like $15 at hardware stores).  See if the erratic readings are happening when you use that multimeter.  That would at least allow you to isolate the issue to the phidget/laptop vs. the thermocouples.

I also recommend changing your sampling interval in Artisan to 3s, if you're not there already.

[yukoncornelius]

FYI, a GFCI does not mean you have a proper ground.  In fact, in my house (built in 1936), most of my outlets are ungrounded, and my electrician installed GFCIs so that we could have 3-prong outlets that are at least protected.  But those 3-prong outlets are still not properly grounded.

Also, the entire thermocouple wire run from sensor > phidget > usb > computer should not touch anything.

Robert - I've already done all of the artisan setup... the graphing in real time is quite bad, its not right whatsoever.  My smoothing of curves and deltas is already set at 5/5.  My sampling interval is 3s. 

So your wires aren't even touching the huky?  That seems very difficult (at least the ET one).  I do have the thermocouple wires wrapped in one place with one of the nylon zip ties Mr Li supplied - that couldn't do it could it?!? 

As far as the multimeter thermometer - do you mean a standard logger with K thermocouple input? I actually do have one of those from when I started with my freshroast sr500 and had a K prong plugged into it to measure temp. 

Thanks for the tip on BT vs ET - I did something during preheat to make BT to around 100F higher than ET.  500 vs 400....  I need to ping kenneroast and ask him why he still pays attention to ET for a charge temp in his full pound roast guide (that costa rica roast was a full 454g charge weight).

What a pain this is though.. I still don't undersatnd why the huky is so sensitive to these issues and I had no issues ever with my hottop setup.    By the way, usb isolator made no difference.  Tried that a roast or 2 ago...


Thanks for feeling my pain and continuing to offer up suggestions guys - At the very least I have a helluva use case for how to elminate this for anyone in the future.

[yukoncornelius]

With all respect ... why do you have such a sour attitude? I've taken every step of advice offered to me in this thread. If you didn't read...despite these logging issues ..the Ethiopian aricha roast was quite spectacular and one of the better roasts I've ever done so I'm trying to stay positive.

Usb isolator - bought it and didnt work
Different computer - didnt work
Checking terminal connections to phidget - triple checked them and theyre solid.
Attempting to ground roaster per jeff's very helpful guidance - tried it and I think something isn't right so I'm going to revisit it.
Making sure thermocouple wires are not even close to power cords - tried that, think there may be something there, would like to see other people's setups.


Am I missing something??? I VERY MUCH appreciate  every single person trying to help me and I've tried every suggestion presented to me thus far. Was your rude remark in regard to my response to Robert who's a newer huky owner? I was merely telling him that I've tried all artisan settings. There's a hardware issue going on... that seems pretty clear cut. Researching noise and interference in general with other applications...it's a very hard issue to solve in many cases. If it was easy...this thread wouldn't be on the 4th page...there would also be clear cut solutions.

As far as plug and play. Was dismantling my hottop to the bare bones. Drilling holes in the rear wall, finding proper thermocouples, fiberglass sheathing, brass fittings, and connecting them to anot arduino board that over took the hottop board plug and play? I'm not new to roasting...I'm not new to data logging... and I'm not new to modifying something. 


If you don't have something to contribute to the thread, please don't leave a comment. If you have an issue with me, I welcome your private message.

Huky may very well not be the right roaster for me because I think its quite a bit more kludgey than I expected ...and now I very much understand tomc 's harsh description of it on home barista. I do think it exhibits excellent roasting dynamics though so I want to figure it out and not rule it out.

[.]

.

[edtbjon]

Yeah, it seems like you're in the "it drives me crazy territory". You have been very communicative about it, which in my book is good. As we're not neighbours... please take it back to square one. From what you're written you ought to be very capable of sorting out this problem. Uninstall and reinstall etc. One step at a time. Calibrate your TC's, etc... I.e now that you've got all the pieces, pick them apart and reassemble them!
Again, you seem to have had a bit of a tough ride with getting started with the Huky. I had my rough spots too, some were sorted in this forum, others by surfing this and other forums. I for one doesn't think that you should give up. Just hang in there and things will eventually get sorted out. Good to excellent roasts come out of my Huky nowadays, but it took me months to get on par with my Gene Cafe roaster and litterary years to get where I'm at.

[yukoncornelius]

Actually - no you didn't give me links on how to ground the roaster.  You started your response to this thread by being snarky and telling me to utilize the search as if to  imply that I had not tried to search or do not know how to search.  That single link you then gave after editing said snarky response gave no real information on how to ground the roaster.  Other search results simply give grounding as an option.  There were no real pictures... until Jeff helped me out. 

How am I wasting YOUR time?  I'm genuinely looking for answers .  You still haven't told me how I haven't followed the direction of every step in this thread.  The frustrating thing about this issue is like RobertL and others said, it almost seems to be purely coincidental and subjective....

 and ha thats funny, of course I don't get to decide who can reply...  but why even continue posting, just get out of the thread already, you just said you don't care.  Your recent responses are uncalled for, not helpful or needed.


BACK ON TOPIC - I bought an outlet tester from lowes - every outlet is properly grounded.  I also bought new wire that I"ll use to try to ground the huky again, as well as ground the thermocouple wires in a thread I found on home-barista. 

[jyalpert]

Did you try the multimeter k-type thermometer yet?  I'd be interested to see if that one gets jumpy as well.

[.]

.

[.]

.

[yukoncornelius]

Did you try the multimeter k-type thermometer yet?  I'd be interested to see if that one gets jumpy as well.

Alright friends... some more tests...

First, RobertL - I"m using a variac, I'm roasting in my garage, wind is not an issue.  I have done tests without beans in, I'm not sure if I've tried it without the heat on but I look for spikes when fan is turned on, and when drum is turned on.  From what i remember, ROR spikes would occur during preheat in both those cases, and then level out.  The real action or jaggedness in the BT curve would happen when the beans were charged into the roaster.


Ok so tonight, I tried a different wire, ground from the rear huky screw to my power strip, tried a 230g charge.  I thought things were going to be good, but the spiky ROR curve and jagged BT curve started when the beans charged into the roaster.

So then I figured heck, let me try Jeff's suggestion - kick it old school with the k-type logger I used when I had my freshroast (that was a nice flash back).  I took the thermocouple wires out of the phidget and just stuck them in the slots of the logger.  This time I paid attention to any big number jumps that occurred during preheat when I would put fan on  and drum on.... I observed nothing and I thought hot dang thats it!  I then charged the beans in...  this is where my excitement droped but I don't know what is natural as far as temp observation.  Basically, the temps would fluctuate within a +/- 5 degree variance, I'm guessing this cooooould be normal due to the natural behavior of the probes now getting hit with beans and monitoring the resulting temperature???  I seem to remember that same temp fluctuation happening when I logged manually with my freshroast.

So this brings me to a question about how artisan graphs the ROR curve.  At what frequency is it taking the measurement?  If its calcuting RoR (per minute) by taking 2-5 second intervals - you could have BIG spikes.  If its calculating RoR by taking 15-30 sec intervals, you're going to have a muuuuch smoother curve.  Right? 

If you look at my artisan graphs, the BT curve isn't that bad... its the RoR that looks horrific.  Any thoughts from these latest findings?

EDIT - did some searching and found an article on artisan blog.  RoR sampling is measured by DeltaBT interval.  Mine is set at 6, why not set it to max rate of 15?  What do you guys have that rate set at?  I'm thinking I should also change my sampling interval from 3s to 6s.    Referencing this article - https://artisan-roasterscope.blogspot.com/2014/01/sampling-interval-smoothing-and-rate-of.html   When I change DeltaBT interval to 15, the ROR does look better but it still has way too many valleys.  Does changing sampling interval after the roast, not change the data displayed? 


THanks again
-Joe

[hankua]

Not sure if it was mentioned; the Artisan tuning guide in the download section is the best on the web. My sampling interval is at the recommended interval; 3 seconds. If there are any erratic readings; it's going to show up in the ROR graph. Artisan has several tweeks which are covered in our tuning guide.

Some of the other data logging programs seem to have less issues with this; I assume they have more smoothing built in. Roastmaster comes to mind; not sure if it works with a Phidget; Typica does support the Phidget. Its necessary to load a program first (PostgreSQL) to coordinate with the data logging; my little laptop was unable to run PostgreSQL.

http://www.randomfield.com/programs/typica/