Not making an argument that automation in cultivation isn’t a phenomenal tool. Also, I like TIA because safety integration is much easier than with Rockwell and given the value of the assets, it’s worth the extra cost to have hardware that won’t fail. TIA is only like $1200 a seat, and I’m willing to bet that wouldn’t be a major contributor to the cost of your system. Happy to hear if competition is bringing the price of options down, but it doesn’t seem you’re very well versed in the alternatives in the space.
1 Like
Are you building anything covering outdoor cultivation?
These proprietary hardware devices and complicated implementations are exactly what we are trying to replace, so we tried to verse ourselves with accounting spreadsheets and cost benefits analysis to make sure we are adding value with a method that is 100% safe and like 95% reliable without any extended period of downtime in our edge IOT controller collected metrics.
Also, Matlab Simulink or AWS microservices are what i’m more versed in when it comes to automating hardware signals, but I wouldn’t say i’m not well versed ;-). $1,200 a seat is more along the lines of the Matlab seat, but AWS wIll do this for you for pennies on the dollar, check out the diagram in their blog link i posted above
Yes! Most sensors are hard to protect outdoors but cheap to replace. Our goal was to find that magic formula of maximum reliability but with a reasonable price point which we have tried to implement in our choice of sensors
That 5% is what makes every business owner wake up in terror in the middle of the night. Sil-3 safety devices usually have B10 ratings of 10^9 cycles. I understand we all do hacky stuff and cut corners to get started, but as soon as cost as allows, you’d have to be on crack to not jump ship from Arduino consumer electronics for a more industrially proven option
1 Like
Be aware, you are not going to have great connectivity in a lot of places. While cloud services are wonderful for being able to do post-sales support and monitoring, actual hardware control and logic should be at the edge, with the cloud serving as a data-storage device and possibly configuration pushing.
Imagine that 4G signal going down and your lights not turning on, or having an unregulated pH swing. That can destroy a crop worth tens / hundreds of thousands of dollars. Saving $150 in hardware to push to the cloud is not going to cut it for industrial control systems.
Augment your systems with the cloud for sure, but never rely on it for operations. I even hesitate to use it for configuration management.
This 100000 times over.
That said… It looks like you have some cool tech! Would love to connect and chat with a fellow hardware hacker =) I wonder if a F4200 virtual happy-hour would be possible.
3 Likes
For sure, that’s completely understandable and relatable from a business standpoint. Our goal is to get to that 100% by printing so many cheap backups that if one fails, you have redundancy and plenty of time to replace it the next day, similar to how our cattle vs. pets software philosophy works in product development. Also, notice how i included the ethernet interface next to the arduino board part. Wifi coverage is only available once we accurately survey the environment’s wireless, radio, and low power signal connectivity coverage similar to a GIS surveying rover. Also, Arduinos are typically used in prototyping but that’s as far as we have gotten. Our goal would be to appropriately shield and protect battery-backed up controllers that we print on a mass scale to share the price of PCB and IC design of building a reliable controller. My guess is, we could probably spread it more evenly than the cost of that controller linked put together with all the sensors.
I think your target market has it right, I’m also wondering if there’s a market for what I’ve described above.
Well i’ve hit my maximum replies for the day so for the next 21 hours i might have to reply here:
@VapeJet
You definitely have some good points here. I would never put my crop in the hands of my cloud. My router fails to find my smart switch on a regular basis.
We’ve tried to design shielding to protect the device in many scenarios, even outdoors. You’ll notice in the diagram I posted above, you can get cheap ph sensors in each soil plant and broadcast them to wifi, that should be fine. For ph controller of the nutrient reservoir, however, that should definitely not only be wired but appropriately shielded and you’d need a reliable Atlas Scientific PH probe wired with a BNC interface to an appropriate smart controller with the software defined logic. For reliable configuration management our hope was to leave the latest controller plan in the hands of the user’s local redundant network design in case of fail over, with a private wifi network for control in case of total network switching failure.
Edit #2, hopefully i’m not abusing the system here:
@SidViscous
That’s definitely interesting you mentioned PLC redundancy because I was just reading this the other day:
Which seems to differ with you on their first point, or maybe I misunderstood the market segment. But even for a commercial grower, a certain amount of redundancy may be guaranteed in cheap systems that could reuse open designs to cheaply recreate the hardware’s stability. We could integrate new control systems and base them on existing control systems as well. See Siemens TIA Portal Support from Simulink PLC Coder - Hardware Support - MATLAB & Simulink. To summarize that page the value add would be:
Simulink and TIA Portal integration lets you:
- Model and simulate your control system using Simulink models and Stateflow® charts
- Detect design errors early and reduce physical prototyping time and cost
- Generate IEC 61131 structured text from the same control algorithm model (using Simulink PLC Coder™ targeted for TIA Portal)
- Download the structured text into TIA Portal (using Simulink PLC Coder)
You can then work within TIA Portal to deploy the control application on your PLC.
I appreciate your feedback and I don’t think it’s rambling because it does affect our market analysis. I look forward to connecting soon.
@Dirteagle
yeah you’re right, actually, this would be a better data collection goal, less lofty and simpler than pulling sensors and reprogramming controllers. Thanks for helping us with what would be a good entry point to outdoor commercial cultivation. You’re first on our list with this. I’ll try to curate some hermy pics, but we are definitely looking for some commercial growers with the data to help build out culling.
1 Like
So, I think your focus might be better aimed towards people trying to get off the ground that want an automation system that is subordinate to manual monitoring. Which is actually a very large market segment but a fairly saturated one.
I LOVE the “fuck it just replace the parts more often” mentality, I really do. But I think when you start to look at architecture, I think that building to avoid common causes of failure make fore very complicated systems. Before we had safety PLCs with redundant processors, we ran full PLCs in parallel, with contactors in parallel and so forth. To get the overall system reliability you can get with modern control systems out of what frankly are consumer grade electronics, you’ll be stepping back to that.
The alternative for the capable and budget minded is buy the hardware used/new-old stock and build it yourself. For those who aren’t, maybe you have an option. But ultimately, if you own a grow or an automated production facility, a catostrophic failure of revegging your plants might as well be a nuclear meltdown, so why not use the hardware and software architecture used to make sure that doesn’t happen.
I also realize I’m probably rambling well in a direction that isn’t related to what you’re asking about so I apologise for derailing slightly
I think you’re missing the forest through the trees on this one. Instead of trying to be everything to everybody, find one pain point that a majority of the industry has and come out with a turn key solution.
No one wants to hear ‘well we need to survey your site and check for wifi connectivity’. While Arduino controllers are inexpensive, the cost difference is trivial compared to an industrial PLC for growers needing a no-fail solution covering many acres.
I hear a lot of outdoor growers that are row cropping have a huge issue identifying and culling males when they use regular seed or transplants. It’s also a huge time suck. If you want to put a rover, data processing, IOT, and Machine Learning to work, build a culling machine. Daily plant inspections will give you a lot of looks at the plants and can help to find the dreaded missed hermes.
I look forward to receiving one when you need beta testing.
2 Likes