TC4 + HTC roast controller for Hottop available - Page 10

Discuss roast levels and profiles for espresso, equipment for roasting coffee.
_OlTimer
Posts: 54
Joined: 11 years ago

#91: Post by _OlTimer »

I installed the TC4+HTC controller recently ( thanks to Jim and Randy for all your help) in my new HotTop B. I used the Omega XCIB style 1 T/Cs as I read on another board that this was "the one". Turns out the author apparently meant style 3, but my order was in route when I discovered my error. Rather than return, I kept these as they are grounded and tipped which I assume makes them more robust

I attached by filling a couple of 5/16 nuts with hi-temp silicone, let set and then JB welded to the back of the roast chamber. I drilled out the silicone enough to allow a firm fit of the leads. I then put heat shrink wrap on the leads where they went through the nut and wall, about an inch into the roast chamber. The leads were then in silicone tubing from the back of the wall to the panel.

I roasted two batches of 225g each yesterday and all worked well. Today, I ran another set back-to-back of the same size, allowing the roaster to cool between each. Using Roast Logger for control, again the first roast was fine, but the T1 (bean temp) sensor started showing brief fluctuations of a few degrees, lasting only a few seconds. At about first crack, T1 shot to a reading of 2,000+ for a second, then down to 100, and fluctuated wildly like this for several minutes. I dumped at 2nd crack and the readings seemed normal as the roaster cooled

I'm not at all electronically inclined, so I welcome suggestions as how to proceed. Could this be a software glitch, a bad TC, improper insulation at the roaster wall?

User avatar
GreenBean
Posts: 53
Joined: 16 years ago

#92: Post by GreenBean »

Congratulations on completing the installation and you first roasts. It must be very frustrating to be experiencing the problems you describe. The problems suggest that the thermocouple is intermittently being grounded to the roaster frame. I am afraid that the Omega XCIB style 1 T/C's are not a good choice for this service for the following reasons:

1. They are grounded so need carefully insulating from the roaster frame.

2. They are flexible so they are difficult to insulate satisfactorily and they can move around in the roast chamber and possibly come into contact with the drum vane, the drum, or even the rear wall.

3. They will conduct a significant amount of heat to the rear wall which will result in low readings unless they protrude a significant distance into the roast chamber which would exacerbate 2. above.

I assume that the shrink wrap you mention is heat shrink tubing. This is not suitable for the temperatures in the roast chamber and will break down quickly. This will have to be removed which will require disassembly of the roaster. I would strongly suggest that you take this opportunity to change the thermocouples for Omega XCIB style 3. The style 3 thermocouple comes insulated from it's sheath and metal collar and will cure the problem with the least heartache and can easily be fixed in place. If you decide to go this way please see the posts dated 23 September 2012 on TMC for details.

Thermocouples work on very small voltages and currents. Small externally induced currents can have a significant effect on the readings as you have found. The easiest way to avoid problems is by using insulated thermocouples or carefully insulating grounded thermocouples from the roaster frame using silicone tube or other high temperature insulation as described in a number of posts in this thread.

There is an alternative to ensuring good insulation between the thermocouples and the roaster frame. This is to ground the thermocouples and the TC4C to the roaster frame. This can be more difficult to achieve which is why I suggest the above approach. If you wish to try this alternative approach I suggest you contact Jim for further advice.

JimG
Posts: 659
Joined: 18 years ago

#93: Post by JimG »

Very little to add to Tom's good advice, here. I wish I knew why, but thermocouple sensors with metal (yours is Iconel, many others are SS) overbraid/shielding just seem to have issues :x

The more normal behavior of ground loop problems has been a relatively constant offset between actual temperature and measured temperature, maybe combined with smallish noise spikes. The very large spikes you are reporting may point to a different cause, particularly if the effect is only observed on one sensor.

If you elect to go the route of running a ground wire, the TCGND pins on the HTC board are provided for this purpose. On your board there should already be a 10K resistor soldered in place near the TCGND header pins. This provides a little isolation for the benefit of the USB connection to your computer.

You may find that 10K is too much isolation to effectively kill the voltage difference between the roaster frame and the TC4C ground plane. In this case, you can solder a smaller resistor (4K7 or 1K at min) in parallel with the 10K that is already there.

Feel free to PM me for assistance.

Jim

_OlTimer
Posts: 54
Joined: 11 years ago

#94: Post by _OlTimer »

Thanks Tom and Jim. I ordered the new style 3 sensors and will swap out. My suspicion that the current probe was contacting the frame seems likely.

ronsil
Posts: 44
Joined: 16 years ago

#95: Post by ronsil »

I have just completed my second conversion on 2 Hottop Ps. I used 8 mm BLACK ACRYLIC COATED GLASS SLEEVING to enclose the Omega XCIB-K-3-5-3. In the first conversion I used a piece of sleeving on each TC lead.

However on the second conversion I found it possible to squeeze the 2 leads through a single 8mm sleeve making it easier to handle through the Hottop with a much neater finish.

Both ways work equally well for me.

Ron

User avatar
JohnB.
Supporter ♡
Posts: 6579
Joined: 16 years ago

#96: Post by JohnB. »

GreenBean wrote: I am afraid that the Omega XCIB style 1 T/C's are not a good choice for this service for the following reasons:

1. They are grounded so need carefully insulating from the roaster frame.

2. They are flexible so they are difficult to insulate satisfactorily and they can move around in the roast chamber and possibly come into contact with the drum vane, the drum, or even the rear wall.

3. They will conduct a significant amount of heat to the rear wall which will result in low readings unless they protrude a significant distance into the roast chamber which would exacerbate 2. above.
They may or may not work well with your software but I've had the type 1 XCIB T/Cs in my Hottop since 5/09 & have had zero problems with them. I've also installed them in several other HTs & the owners have been very happy with their performance and have reported no problems.
LMWDP 267

User avatar
GreenBean
Posts: 53
Joined: 16 years ago

#97: Post by GreenBean »

Hi John, I did not mean that there is anything inherently wrong with these thermocouples, only that they do not lend themselves to being easily insulated from the roasters frame. I have suggested changing the thermocouples to insulated thermocouples as the easiest robust solution.

There are a number of threads on this and other forums describing the considerable problems people have had with various thermocouples in various circumstances. In my experience problems arise where two or more thermocouples are installed and read by one device especially if the device is installed in the roaster. Generally the problems are solved by either carefully grounding all the thermocouples, reading devices and the roaster together or by carefully insulating them from each other. Grounding them all together reliably can be difficult to achieve and has some risk if the roaster and reading device are powered separately. If any thermocouple is not satisfactorily grounded you run the risk of spikes in readings. Insulating them from each other can be achieved relatively easily but must also be done well as any intermittent grounding of one thermocouple can cause spikes in readings.

The spikes are read as high or low temperatures at the device reading the thermocouples. The software used, RoastLogger or otherwise, is simply reading the data sent by the device reading the thermocouples and does not care what thermocouple is used.

_OlTimer
Posts: 54
Joined: 11 years ago

#98: Post by _OlTimer »

ronsil wrote:I have just completed my second conversion on 2 Hottop Ps. I used 8 mm BLACK ACRYLIC COATED GLASS SLEEVING to enclose the Omega XCIB-K-3-5-3. In the first conversion I used a piece of sleeving on each TC lead.

However on the second conversion I found it possible to squeeze the 2 leads through a single 8mm sleeve making it easier to handle through the Hottop with a much neater finish.

Both ways work equally well for me.

Ron
Ron,

I have silicone tubing for each lead separately. It does make for a little tighter fit. Where did you find the sleeving you used?

Dave

ronsil
Posts: 44
Joined: 16 years ago

#99: Post by ronsil »

I found it on ebay UK.

Ron

User avatar
mikepetro
Posts: 16
Joined: 11 years ago

#100: Post by mikepetro »

GreenBean wrote:particularly interested in the drum vane/thermocouple clearance and the clearance between the thermocouple and drive motor in the rear compartment.
Finally got all my parts sourced, one part turned out impossible to obtain and I had to fashion it my myself. Total project, including the 2 thermocouples and mounting the HTC/TC4 board, took about 10 hours from start to finish once I had all parts in hand. Could have done it in less but I took a "lot" of photos of every step along the way.

Anyway, you asked for pics showing the probes clearance, here you go......

After trying out my mock probe on every placement I could think of I landed on this: I drew a straight line through the center of the 2 HT sensor screw holes, then drilled my 5/16th hole 10mm above and below those sensor screws. I needed that 10mm clearance for nut on the backside to clear the sensor screw.


You can see here that the nuts are right up against those two screws.


Plenty of clearance behind the probe, a good half inch gap between the probes backside and the motor housing.


And a solid half inch between the probes front and the bottom drum vane. I dont think I will have any problems with beans jamming.


The probes stick out a 1/2" from the compression fitting. Thats 8 diameters of the 1/16th probe. This should minimize any possibility of stem effect.

At 1/16th diameter the probe responds very fast, allowing me to get away with using an ungrounded probe. No ground loops, yippee!

This particular probe was 2.5 inches long, if I ever have to replace them I will buy them in a 1.5 inch length, thats as short as they come. That would give even more clearance in the motor area.

No JBWeld or glues used, assembled from off the shelf parts with the exception of the backnut for the fitting which I had to drill and tap myself, I tapped a 5/16th nut. Try to find a 1/16th NPT backnut, I couldnt anyway, the smallest I could find was 1/8th.

Now I have to learn to roast all over again, I had just gotten used to the funky binary control of that P controller. What a waste of money, wished I would have went for the B instead. Anyway, should be an easier learning curve this time around..........
Mike Petro
Martinsville, VA
PLC/PID'd Lady Silvia, Stepless Rocky, HotTop+TC4C-HTC