Recent
Users
Profile
Discord

Extraction

DIY

Tips and Tricks

Equip

Lounge

Gardening

DIY HarvestRight (A...
 
Share:
Notifications
Clear all

DIY HarvestRight (Analog - 2015) Upgrades

7 Posts
3 Users
1 Reactions
1,610 Views
(@lordevolde)
Posts: 9
Active Member
Topic starter
 

Will I ever complete one project before starting another?

Will I ever buy a piece of equipment and NOT tear it apart because I dont like the EXACT way it functions?

The answer to those questions and more below! (hint NO)

Anyways, this is my next project, fucking with this freeze dryer. But why you say? Well I got this thing for a steal, paid $2k for the dryer, 2 vac pumps, and the giant table you see it on (not to mention a bunch of mylar bags, a thermal sealer, and some other shit) - the downsides (or so I thought, see later) is that its an early consumer model from 2015 (or 2013? I cant remember) - either way it uses a multi-line LCD display and two rotary encoders on the side. Literally the only two options on this thing are to set the freeze cycle time and finish dry time.

So let's start with a little background, how this thing works, first it freezes whatever you put in it at -25F or so (cant remember exactly, but its a good 20-30F below the standard -5F freezer), then once ran for the set time, engages the vacuum pump. Once the pump gets below 500mtorr the plates are turned on - this is the key difference with the new models, from what I understand the newer units use a thermocouple to measure tray temp and moderate the heater voltage (or cycle them entirely) to keep the trays from reaching too high of a temp - this older unit does not - it simply runs the heater for a preset interval, then it waits for vacuum to return under 500mtorr and repeats the process, when the heated tray no longer causes vacuum to rise (largest amount of water sublimated) the final dry process engages. During this it seems the trays are still cycled but at much less frequency. Ive essentially just "decoded" this by observing the freeze dryer and using a datalogging DMM.

So onto the mods...

The first thing I wanted to add is a webUI so that I can see how much time is remaining and the vacuum level. After checking out the display, it turns out it's one from Newhaven Technologies that is well documented. It turns out Harvestright used the display in RS232/TTY, so I was able to connect the serial rx and sig ground to an adapter on the pi and "grab" the data being transmitted to the display. Since its just ASCII and they refresh the display every 1s, they essentially made this task extremely easy for me. Ive got a little bit of python work to do here, but this shouldnt be difficult to build a simple UI to display this information. Ive considered also replacing the rotary encoders (or passing them through) the pi to control it remotely as well, but I really dont NEED this functionality so I wont mess with it for now.Β 

Oh for anyone wondering the USB RS232 adapter is to protect the pi's native circuitry as the freeze dryer uses 5v logic.

The second part obviously is adding tray temp control. This was the worst part as the passthrough NTP fitting was inaccessible. The back of the unit is basically "glued" to the drum, so removing wasnt really an option - drill an access hole it is! Then the best choice to loosen the fitting was a large impact gun - as it can avoid putting unnesecary torque on the secondary fittings weld to the barrel, this was successful and I was able to see how they passed the wiring through, a simple fitting with epoxy (similar to what I had planned!)

So now onto Amazon, I purchased two T Type thermocouples and I2c based interface boards. As well as Loctite Hysol 1C - this is the same formula as Torr Seal a specifically designed vacuum epoxy with very low outgassing. Heres the BOM below for doc purposes (if anyone finds this and decides to pursue as well)Β 

Amazon.com: Loctite 1373425 Hysol 1C, White, 4oz: Industrial & Scientific

Amazon.com: PerfectPrime TL0024TT, Thermocouple T-Type Low Temperature Sensor/Probe, -200~260Β°C / -328~500Β°F, 2 Meter Long: Home Improvement

Amazon.com: Adafruit Universal Thermocouple Amplifier MAX31856 Breakout: Industrial & Scientific

And the following to replace the NPT 3/4->1/2 bushing that was epoxied previously -

Taisher 2PCS Forging of 304 Stainless Steel Reducer Hex Bushing, 3/4" Male NPT to 1/2" Female NPT, Reducing Forging Pipe Adapter Fitting: Amazon.com: Industrial & Scientific

Ill continue to update the thread as this moves forward. The basic idea is to attach the thermocouples, read them with the pi, then use a simple averaging of the temp to control a relay with a multi-step delay to insure the plates arent cycled too frequently. This really shouldnt be too difficult - on the order of projects Im working on this is probably one of the most simple lol. I contemplated trying to map out the logic and just replace their controller entirely, but that would just be too much work right now with everything else I have going on.Β 

Few pics for now



 
Posted : 13/05/2021 1:15 pm
(@lordevolde)
Posts: 9
Active Member
Topic starter
 

Additional discoveries - so the controller inside the unit has two pots on it, these set the "window" for the heating cycle once vacuum is achieved, it appears to currently be around 500/625, I see no need to change this as the unit works great, I just want to limit the max tray temp. Just noting things here as I go.Β 

 
Posted : 13/05/2021 2:44 pm
(@lordevolde)
Posts: 9
Active Member
Topic starter
 

Update number 2, completed quite a few tasks today after taking yesterday to mow and clean up some more brush on the land that will be my future business. Anyways the updates...

Received the T-Type thermocouples and the new MAX31856 boards (SPI based chip that works automatically with basically every type of thermocouple). The headers had to be soldered but the new station made that easy enough. I also created a mounting platform for all of the components in FreeCAD and printed it (too bad I didnt have screws small enough for the 31856's, had to use mounting tape, thats why they arent perfectly flat.

Once assembled I tested some basic python using circuitpython lib to test the thermocouples and all is well there, resolution is good and accuracy seems spot on compared to other reliable sensors. I also tested running the lead wires along the power cable for a t-shirt press with 600w heater to test ac induced emi on the thermocouple signal - this worked better than expected even wrapping the leads around the power cable caused no issue. My fear was that the heaters operating could cause interference in reading the thermocouple as it uses an extremely low voltage signal, however it doesnt appear to be problematic - worst case we can always interrupt the heater cycle to check temp.

Second major milestone is I recreated the "electric interconnect" that HarvestRight used at this time - a 1/2 MNPT plug drilled for the 2 heater cables inserted in a 1/2->3/4 Stainless NPT bushing. I replaced the components and drilled the new plug for the 2 heater cables as well as the 2 thermocouples pulled through one hole (made the tightest fit due to the cables being flat). Finally I mixed the EA1C epoxy on a scale - something Ive never done before hah, but best to insure accuracy with this. The idea is this "special" epoxy has minimal outgassing and is perfect for vacuum applications. Anyways I started with trying to use a syringe to fill the joint, but it was too viscous. Ended up using a small stainless rod to "dip" the epoxy into the fitting, after a bit it settled down into the bottom and worked pretty well, just had to do some expedient cleanup.

The epoxy recommends a 72hr full cure at room temp, so well be waiting until Wednesday night to do any vacuum testing. For now theres a few things to accomplish, but I should be able to get this done tomorrow.

1) Run freeze dryer through cycles (other than truly running vacuum pump to trigger dry) so that we can capture the serial output and make a loop that detects what cycle its in when it changes, then updates the times remaining and vacuum level when applicable - well have to finish this once I can run full test cycles Wednesday, but can get most of it done now.

2) Develop a small flask app (I have some experience with flask, not a alot, its a good opportunity to learn more) that uses a template to render the current values captured from serial and the thermocouples, as well as a simple form to POST the desired tray temp back to the main python loop as a var.

3) Wednesday test it all and debug the app/webui.Β 

Some pics of todays work...




 
Posted : 17/05/2021 1:50 am
(@lordevolde)
Posts: 9
Active Member
Topic starter
 

Got some free time this afternoon/evening and spent it (all) writing what is now called HarvestRightEnhancer lol. Basically a simple python app that captures the serial data to the display (Time remaining for each cycle, vacuum when displayed) as well as the thermocouple temps, displays it via flask, and logs it in influxdb (so we have a datalog we can look at in grafana). Finally it also controls the heater relay which the target value can be set in the webui. It really probably needs some tweaking, but so far it works pretty smooth.Β 

Tomorrow Ill get the fittings back in place and wiring cleaned up - then itll just be waiting until Wed evening with fingers crossed before we do vacuum testing!

Once vac testing is underway Ill be able to capture the final mode (Drying) and configure it to display properly - after that the code will be available on my github with instructions here to follow.

Im really not too worried about the seal, it seems to have filled in the fitting nicely, however I am sitll unsure if the control board uses any type of current sensing on the heater output that could trigger a fault when my relay opens the circuit. This seems extremely doubtful, and Ive certainly powered it up with the heaters unplugged, but I wont be able to fully test (send it into drying mode under 500mtorr) until the epoxy fully cures!

Worst case Im feeling pretty confident I understand the logic here and could just do away with their controller entirely πŸ˜€ (That may still be a project down the road)

 
Posted : 18/05/2021 1:35 am
(@huntermfg)
Posts: 1
New Member
 

I created an account just to follow this post. Love that you figured it out... inspiration for me to tear more shit apart! How’s everything going with it now? Did you figure out the last piece to the puzzle? Would also love the GitHub repo when you post it, I’m a dev so it’d be interesting to look at. Thanks for sharing your project!

 
Posted : 04/06/2021 5:50 pm
dR reacted
(@lordevolde)
Posts: 9
Active Member
Topic starter
 

Unfortunately I ran into some other issues with a bad door seal and different flare fittings sending me on a wild goose chase, so I havent updated this in a bit. Everything from the mod side is working smoothly and has been tested for several cycles now, I added some pulse functionality like what the scientific models have in dry, I meant to add those settings into the flask ui and just put settings on a separate tab, was going to try to get that done before I posted it. So maybe this week! And thanks for the interest!

 
Posted : 06/06/2021 10:30 pm
KLIKWIN188 Daftar Situs Games Anti Rugi Link Alternatif Terbaik
(@klikwin188-daftar-situs-games-anti-rugi-link)
Posts: 3
New Member
 

Is this thread still updated?

 
Posted : 03/04/2024 9:37 pm
Share: