***** Thanks John - Hope you're well.... Cheers!!

Thanks M4ST3RC4TS - I really must get on and make some more progress on this project - It's been sitting on my desk for many months now with very little progress - I got to the point where I needed to build some decent chassis/framework to hold all the bits together, and then I stalled. I want to build a better system for holding all the pumps, and I have an idea of exactly how I want to do that... I've got a CNC router in my shed which I built about a decade ago which would be absolutely perfect for cutting patterns out in acryllic... so I just need to refamiliarise myself with the CAD package I used back then, spend a day or two getting the design right and a day or two in the shed actually building the chassis.... For one reason or another, I never seem to find the time to get properly stuck into that job. I did know when I started this one that it would be a long-term project.... If you start one, I might be able to copy your ideas, so I'm going to fully encourage you to do that :-) Cheers!!

Thanks Patrick - not sure it will ever make it to 'product' status, but it sure is a fun project. Cheers!!

Jim Conner How about a Kickstarter Jim ?

I'd certainly consider doing a kickstarter.. if I can get the prototype built and working right... I wonder how many we'd need to get a decent production run ready... and I'd have to put in a good chunk of effort on designing something which can be mass produced (rather than the current jim+soldering iron solution). Cheers John!!

Jim Conner Well I'm sure you know where to get your front panels cut :) I'm impressed with your plastic bending BTW ...Maybe a vid on that as it can be a bit of a black art to get right!?!

Hi Bryn - Sorry for taking so long to get back to you... Sometimes life is busy. A chap called Devin recently got in touch with me via the comments to let me know about their own project where they'd used stepper motors and syringes already - they've got a project up on github with all their code... github.com/devinrayolsen/DRO-Matic ... It might be worth you getting in touch with Devin (you should be able to find him down in the comments section somewhere) as it sounds like there's a lot of similarity between what you're planning to buid and what he's already built. My own mixologist project has been on-hold for a couple of years now as I got busy with other bits and bobs and haven't put in the time needed to get the project finished (and also I switched to using dry powdered nutrients instead of pre-mixed, which totally changes the way that dispensing would need to be done.... I'm currently thinking that I should try a simpler project to begin with... Perhaps just a measured water inlet so that I can tell the system that I want exactly 20 litres of water dispensed into a container instead of just relying on my guesswork to work out exactly how full any container actually is. Cheers!!

I liked your idea for load cells Doug. Then I thought "Level Sensors, surely" My internal jury is out on that. I like what is being said here about designing "to avoid the possibility of things going wrong", but part of designing for what I would call "robustness" (and "reduced cost") is designing out unnecessary components, and of course adding failsafes and backup systems. So I am thinking why not solenoid control and single delivery line to the mixing tank? An array of solenoids at the end of a single, any length, small diameter, delivery line for all nutrients suggests to me you might even use one high quality peristalic for all nutrient mixes. These could then be stored a long way from your mixing tank(s), and a single mixing tank might service multiple, independent, root zones on a time-share basis. For consideration: (1) Could you abandon the "one peristaltic per feed" design? (2) How would a single mixing system work with multiple root zones? (3) What are the wider benefits and cost implications of the sophisticated simplicity being advocated here? (1) Could you abandon the "one peristaltic per feed" design? I am pretty sure you could use the two wires of an I2C bus or even RF to control any number of solenoids, perhaps with compressed air to move stuff faster and flush lines. Nine (goodness me!) nutrient mixes would not then be a limit on the design, and more mixes need not add to the cost and compexity as the number of "ferts in the mix" increases. To deal with the "don't mix concentrates" issue and eliminate any hysterisis that a long delivery line to the reservoir implies, you would designate one solenoid to "flush nutrient delivery line with water" and/or have a local mixing tank. Of course this fewer peristaltics introduces an element of additional time elapsed, but I don't think time is limiting - 1 ml per second from one peristaltic will deliver 3.6 litres per hour and we are only talking about the dosing, not the mixing or flushing of the delivery line since that can be handled by a different pump. Either way there is certainly a case for bringing the dosing under software control because growers might want, for example, to buy commericial solid N ferts and disolve them, use low (monitored) volumes of specialist organics, or play tunes on urea vs nitrate vs amonia during the course of a single day (possibly not a 24 hour day) AND do so for multiple root zones - for which read everything from a single flood and drain pot to a greenhouse bay 1000m2 in area. (2) How would a single mixing system work with multiple root zones? Such a system would sample each zone in turn by filling the mixing tank with that zone's growing solution and, on the basis of both analysis and anticipated demand (I feel an algorithm coming on that takes account of weather, maturity of crop and other factors), prepare a top-up mix for that zone by dosing a variable quantity of each fertiliser in turn, mixing and then dumping the diluted mix to the zone, flushing and repeating for a different zone - all with one array of sensors. The pH adjustment stage can be combined with flushing, so the low pH solution can be expected to sanitize the mixing tank, reducing any possibility of pathogen cross-transfer between root zones. (3) What are the wider benefits and cost implications of the sophisticated simplicity being advocated here? If systems are cheap enough then duplication of the entire system as a failsafe has a cost benefit justification. You can monitor and calculate and log the whole process, using your PC and have real time up-to-date information on cost of inputs for each root zone. If such inputs are measured and recorded accurately and reliably then they can inform a "blueprint" type growing regime that is replicable anywhere.

Would the load cells drift too much over time though?

Thanks MrDutchgrass. I really like that people are enjoying my videos. I really enjoy this little youtube community that we've got going on here. I think what Ryan's doing over in Thailand is awesome (wish it was me doing it). Awesome that I gave you some motiviation. Would love to see some details of your project - are you going to put a vid up when it's done? Cheers!!

Thanks MiFraDeLa88 - That's a very good thing to have on your to-do list, and there's some great tutorials and exercises out there now to help you along the way. Obviously I've no idea of what your skill level is from just one youtube comment, but if you're looking for a place to start at the very beginning with programming, then 'Hour of Code' (google it) is a great place to start. I've used their online tutorials with both of my kids and they went down very well. Cheers!!

Jim Conner Wow thanks a lot! I have a technical background (Chemistry), but I'm a complete beginner when it comes to computer programming. I'm definitely going to be trying this website out. It seems a lot more fun than reading the old C++ text book I bought used on Amazon.

MiFraDeLa88 C and C++ are great languages when the only alternative is writing assembler code (i.e. when you need maximum performance, or you're running on some tiny microcontroller with hardly any memory), but with the big fast computers, tablets and phones we have these days, it's much easier to use a higher level language (python/ruby/java/lua/golang etc) which lets you express things in easier to understand ways and not have to worry about how memory is getting allocated and and how the bytes get moved around. Lots of people seem to get on well with either Python or Ruby, but it's a bit of a marmite choice... I'm a Python fan myself (some ruby statements seem back-to-front to me), but I know a few Ruby fans who find Python hard to understand. Anyway.. the hour of code is a great place to make the first steps... Hope you enjoy it. Cheers!!

Thanks! I'll definitely keep that in mind when I sit down to invest some time in learning. As it is now the most useful language to learn would be something called Labview, by National Instruments, but I can't really find any free resources for learning it that are geared towards beginners.

MiFraDeLa88 Labview is something I've been wanting to learn myself as it can plug into my power supply, signal generator and scope, and then I can write programs to control them all. That one's on my to-do-list, but like yourself, I haven't really found any good resources for labview that take you through from the start. Cheers!!

Hi +yagoa - I'm planning on putting all of the nutrient bottles onto a scale and measuring the weight difference as I dispense the nutrients - that gives me a very level of confidence that the nutrients have been dispensed, and let's me detect when bottles are empty etc - The load-cells for digital scales are crazily accurate - I've got myself four of these load cells and now I need to build a weighing platform/enclosure to hold the bottles in place. Thanks for taking the time to comment. Cheers!!

Hi Murat Acikgoz - Sorry it's taken so long for me to notice your comments - KZfaq decided that they were all spam as they contained links, and didn't bother to alert me to them - I only got the notification when I came to this video to reply to someone else's comment. You've got an impressive looking setup there - great work - I particularly like that it's showing you the exact water levels. My own Mixologist project has been sadly neglected for the last few months (got a new job... it's keeping me far too busy)... but someday soon I'll find some more time for this project - It's very much needed for me as I accidentally killed all of my strawberry plants a week ago. My pumps are on timers, but I switched off the timers when I was changing the nutrients and forgot to switch them back on again. This wouldn't have been an issue if I'd fully automated my nutrient changes. Cheers!!

Thanks asuspower . Just a quick update on what's been going on with this project... I've now switched over the microprocessor board as planned to the one with more memory and I now have all 9 of those little OLED screens working. I'll try and get screen no. 10 (the full colour LCD) up and working this week. Cheers!!

Dominique Dupont Gotta have a beer.. .it's the rules :D

Hi +JHinkleJr - I really need to get one and make some more progress on this project - it's been gathering dust on my desk for too long now... but fortunately I've got some holiday time coming up, so I hope to be making some more progress (and another video) next week. Although I was using Canna nutrients when I made this video, I have since got hold of some Masterblend (same stuff MHPGardener uses) which was shipped over from the states.... The dry nutrients sure do work out a lot more cost effective than the pre-mixed stuff.. I'm hoping to do a Canna vs Masterblend experiment next year to see if there's any noticable difference between the two. (The guys at my local hydro shop reckon there's no difference between any of the brands they stock (which doesn't include masterblend). My masterblend recipe is 0.6g/litre of masterblend, 0.6g/litre of Calcium Nitrate and 0.3g/litre of Magnesium sulphate.... Of course... modifying my planned dosing system to use dry nutrients instead presents a whole different set of chanllenges, so I might press on with desgining for liquids for now and worry about modifying it later. The problem with relying on EC level for nutrient measurement is that it doesn't tell you the relative proportions of the different nutes... for example, if the plants are using lots of nitrogen, but no P or K and you keep on topping up, then the strength of P&K will gradually increase over time. I don't know how real-world of a problem this nutrient imbalance is. I understand that massive-scale commercial growers get water samples analysed with a mass spec so that they can work out which particular nutrient needs topping up on any given day (and the biggest growers even have their own mass spec machines and do it in-house). I'd already been thinking about using air pressure to measure the water level - I've saw it done that way in a dishwasher I once dismantled and was always rather taken with the idea. A DIY inductive EC probe is a project I'd be very interested in... I read up on that when I was working out how EC probes worked, and it seemed like a fantastic idea... Please keep me posted if you make any progress with that. Thanks for taking the time to write such a long and well thought out comment - Happy Christmas!! Cheers!!

+Jim Conner Use your solid Masterblend components. Figure how much is required for 10 gallons as an example and then mix it with only 1 gallon of water. Now you have a 10 times solution to use in you system. You just need to account for less solution using the concentrate than with your premixed purchased liquid solutions. The notion of nutrient unbalance is propagated thru the use of fertilizer (like Masterblend) that is NOT designed for hydroponic use (when used in soil - these impurities don't have an adverse affect on the plants -- they just add to the gross weight of the fertilizer). Fertilizer like Masterblend have a LOT of components that effect EC and PPM readings that are NOT consumed by the plants. Over time these unusable components build up. The typical explanation as to WHY an EC issue exists is to call it a nutrient unbalance instead of stating the fact that the fertilizer contains garbage and needs to be purged often to keep the system running with some degree of proper EC or PPM values. In your investigation on how EC meters work, you should have noticed a range of different material used for the probes. Probe contamination is one of the biggest problems with leaving an EC probe in solution for extended period of time -- as in an automated system. THAT is why I'm exploring an inductive probe. Inductive commercial probes sell for $600+ which is well outside my pain limit for an automated lettuce grow system. The pressure sensing water level work well if the proper sensor is used. I used it to track Atlantic tides in the Florida Keys. The problem I found in salt water --- was salt water. The sensing tube would contain a hi humidity environment of salt which tended to corrode things quickly. When used to sense fresh water levels -- the corrosion issues is greatly reduced. Cheers back at you!

Pre-mix the masterblend in concentrated form.... That's so obvious... D'oh... Facepalm... bumps hand on forehead... Yes of course... What a superb suggestion... How on earth did I fail to think of that one. Thank you, thank you, thank you :-) With the water level sensing thing, I'd actually been thinking of using some lay-flat tubing filled and sealed at atmospheric pressure, inside a rigid outer PCV tube. As the water level rises, it will attempt to compress the lay-flat tubing inside, increating the air pressure inside the sealed tube - I think that would avoid any corrosion issues. Cheers!!

+Jim Conner I'm just quick thinking off the top of my head --- but if the pressure sensor was placed in a plastic baggie -- sealed where wire exit the baggie and the baggie is placed in the tube -- as the pressure in the tube changes -- the same pressure change would be seen in the baggie -- there may be some lag in response but levels don't need millisecond response times.. I'll have to try this as all humidity issue would go away. Cheers!

Yep... that's pretty much what I was thinking.... Can you tell me what pressure range you I'd need to measure from/to for a reasonable depth of water (let's say up to 50cm of depth)... I've got some of those dirt-cheap sphygmomanometer pressure sensors of ebay kicking around which can measure 0 to 5.8psi ... If I'm really lucky, that could be perfect for the job. Cheers!!

Hi +applemarkwantsvids ... The project is coming along so slowly that 'standstill' is probably the best word to describe it... I did a few more weeks work on it after making this video... I added a 320x240 pixel Colour TFT display to run a menu system, and I worked out that the most accurate and reliable way of dosing out the nutrients was to use a load cell to weigh the contents of the bottles during dispensing... That way I'll know 100% for sure that the nutrients have been pumped out of the bottles...I got a load cell hooked up and working giving me weight readings, and I added a few buttons so I could control a menu system but after that, I ground to a halt... I realised that because the load cells are also sensitive to twisting, I'd need a proper weighing platform with a load cell in each corner, and to make that I'd need to spend a day or two in the shed cutting bits of plastic and drilling holes for screws and sensors... This was around the same time that I started a new job, and since then (mid February) I haven't managed to find a couple of days spare time to tinker in the shed. I'm planning on getting myself a 3D printer for christmas this year as that'll let me print out parts for this project. I've not given up on this project... it's just slow going for now. Cheers!!

Hi Daniel. Ultrasonic sensors could work as long as I'm always mixing nutrients in the same container. It would be OK for the bulk liquid (i.e. 20 litres of water), but I doubt that measuring the surface height would be able to tell me if I'd dispensed 5ml or only 2ml of one of the nutes. I'll look into the ultrasonic options a little more. Thanks. I found some excellent looking flow sensors today (no idea of the cost yet), They work by having a heating element placed directly between two temperature sensors. When there's no liquid flowing past, both temperature sensors will read the same level, but as the liquid flows past, it will cool the first one in the flow more than it will cool the second. The sensors say they're ridiculously accurate (up to flow rates of 120ml/minute). I bet they cost a ridiculous amount. With the DS1820, the data line is normally held high except for when a device is tranmitting or receiving. You can run the DS1820 without the power line at all and it will parasitically draw power off the data line, with each device storing just enough power to perform a conversion and send the results back. (nobody ever seems to use them without the power line though as it's just more hassle than having an extra wire). So yes... seeing 3.3v on the GPIO line would be perfectly normal. Cheers!!

Hi Special Agent - unfortunately I never finished building this project (though I still plan to at some point), so I'm afraid there's nothing for you to try out. Cheers!!

Good question, and I'm really not sure of the answer as my Mixologist project never made it to completion. One thought though.. Silicone tubing is actually slightly porous and will allow air into the tube over time. If you're using that then maybe switching to vinyl tubing will help. Cheers!!

@@ThingsWhichArentWork Thank you Jim, I'll try with other tubes to see if it's the same ;-)

Thanks ***** - Would love to see what you come up with... I need to make some progress with this project myself... I haven't touched it in the last six weeks as I got busy with work and having a social life. Cheers!!

Hi Eduard Mazo - I'm using the Adafruit lib which someone ported to mbed... developer.mbed.org/users/nkhorman/code/Adafruit_GFX/ ... It worked OK for me, except for sometimes getting the top-of-screen set at somewhere other than the top of the screen... I have a fix for that, but I have to dash out for work now. Hopefully the Adafruit ported lib will get you started - I'll try and dig out the fix later today. Cheers!!

Hi stlaunder. Evaporation from the tanks is pretty minor compared to the amount of water that the plants will transpire for themselves. Also, you want to cover over the tanks to prevent the sunlight getting into the nutrient mix and allowing algae to grow. With a lid on top of the tanks then evaporation ceases to be a problem. Top tip... If you suddenly notice that your plants are consuming more water than normal, have a check for pests on the plants... When the plants are under attack by sap-sucking insects they'll get through much more water than normal. I've never worried too much about totally draining the tanks every time. I figure that as long as 95% of the solution gets changed over, any nutrient imbalance that was left int he remaining 5% of the liquid is diluted down by the fresh nutes anyway. Cheers!!

@@ThingsWhichArentWork Great, thanks for the quick reply. I should be able to get away with one tank for water storage and one for mixing and storing in that case. I'm using node red if you haven't heard of it already. It's very cool!

Hi W Francisco - Although that's a good idea for checking the dispensed amounts, the two-part nutrients that I use are supposed to be added seperately to the water, as if they come into contact with each other when they're undisolved then it can form some other salts which are undesireable (I can't remember which ones), so I'd have to have multiple dump mechanisms for at least four of the measuring containers. I'm currently working (very slowly) towards the idea of using a load cell (i.e. digital scales) for measuring the weight of the nutrient bottles, and only ever dispensing from one bottle at a time - I'll can measure the starting weight before dispensing anything, and then watch the weight drop as the nutrients are pumped out of their containers. That way I'll know for sure that the nutrients have been dispensed, and because load cells are ridiculously accurate, I'll have a high degree of confidence in the measured amounts. I've actually already hooked up a load cell since I made this video - it's a 5KG device, with a 24-bit Analog to Digital converter attached to it. That'll gives me sixteen million different levels of weight measurement between 0 and 5KG, which if I've done my sums right is a resolution of 0.00029 grams. Sadly though, this project has been gathering dust on my desk for the last few months as I never seem to have any free time. I'll get back to it sooner or later though (most likely when the days grow short and the nights grow long). Cheers!!

Hey Jim, I was thinking in so many aplication this device can have, auto coffe maker, dump chlorine in a well, mixing tints... there are plenty of things to do, keepup the good work.

Hi Bill - The pumps were just from ebay - If you search for 'peristaltic pump' and go for the cheapest prices, you should find them. You might also want to check out Devin Olsen's DRO-Matic project ( kzfaq.info/get/bejne/fdKKe7d-0Jm0cps.html ) - He's got far further along with his design than I did - my mixologist project has had no progress for the last 18 months, where as Devin has some working code up and running. Cheers!!

Hi Alexis - The code is up on developer.mbed.org rather than github. developer.mbed.org/users/jimconner/ ... I'm afraid there is no parts list or diagrams as this project hasn't been finished yet (it's been on hold for the last 2.5 years so far... one day I'll carry on, but don't hold your breath). Cheers!!

I'm afraid this was an unfinished project - I got the OLED screens and some input buttons up and working, but after that I got myself a new job and have never finished The Mixologist. The electronics part is still on a shelf beside me, so i might one-day revisit it.. but don't hold your breath. Cheers!!

Hi Don - I'm planning on doing exactly that, but with the small optimisation that I can weigh all the bottles at the same time, but only dispense from one bottle at a time... so the reduction in total weight will be exactly the same as the dispensed amount. I've been very slow at progressing this project (don't know if I will ever finish it). I got as far as building the weighing platform with four load cells, but haven't got around to hooking them up to a microcontroller yet... (I built the weighing platform over a year ago now). Cheers!!

Jim Conner what about stepper motors with maybe a syringe. You could count the steps to pull up a ML.

Sadly it is an unfinished project. I switched over to using dry powdered nutrients and never finished the automation for liquid ones. Maybe one day.. it's still on a shelf next to me. Cheers!!

@@ThingsWhichArentWork: I am in the process of building a garden and what you described is far better than anything out there in the market. I would have purchased one from you (That is how good it is). Than you for answering

Hi BringItHard - I wish I had better skills with 3d-design packages. I'm still bumbling along with Autodesk's 123d design at the moment, which is really rather limiting... I really must invest some time in learning a proper 3d package (tempted by Autodesk Fusion 360 because it's free for hobbyists). When it comes to learning electronics, my advice would be to only worry about learning enough to solve the exact problem you're trying to solve at the time. You can go a long long way just knowing Ohms law. If you get youself a copy of 'The Art of Electronics' by Horrowitz and Hill it has every possible thing you could ever need to know, but it's a huge book and not really readable from cover-to-cover. It's just a great reference when you decide 'I need some simple oscillator circuit' and you can look up a few different examples. You should also check out Dave Jones' EEVBlog channel on KZfaq. Dave's done some fantastic tutorial videos on there (mixed in with product reviews etc) so his channel's well worth a look. You might also want to check out Julian Illet on KZfaq as he makes videos at more of a beginner level. Hope that helps. Cheers!!

Great! The book as well as the robotic builder bonanza, I also sub those 2 channels as well. In the manufactioring world I'm in we use Solidworks for modeling and print making. You can find those files online for free as well if you would like to know more about that PM me and Ill point you in the right direction. Thank you for getting back to me!

Thanks +Amit Patkar ... What would be even more awesome is if I actually put in some more time on this project and got it finished... It's been gathering dust on my desk for months as I got sidetracked with other stuff... I'm planning on doing some more on it soon though. Cheers!!

Thanks Ryan...Strikes me that it would be an even better asset to have around if I added a 10th pump on there for dispensing beer :) Cheers!!

Jim Conner While I agree completely, 1ml/sec is not going to cut it!

Ponic Wars Respect... 3.6 litres per hour would leave me paralytic :)

Ponic Wars LOL ; )

Keyplayr61 Greenhouse Hydroponics And Gardens Jim Conner I was only looking at it from the perspective of it taking 568 seconds to get a pint! lol

Hi +Chris Johnson - Nice guess, but I'm afraid I'm not a chemical engineer... I spend my days battling with computers instead. I do often call upon the help of a friend of mine who's a biochemist though. I'd love to see your system when you get it up and working... Cheers!!

Graham Wilson Hi Graham... I feel it's a bit too soon to start getting critical of the design just yet. I'm still very much at the stage of buying up bits that I think I'll be needing and figuring out how to get them to work. I'm very aware of the opportunities for this system to screw up massively, which is why I'm more worried about metering the liquids being dispensed than actually dispensing them. I recently saw a teardown of a blood sample analyser (over on mikeselectricstuff ), For the most accurate dosing requirements in that system, they were using syringes driven by stepper motors on a threaded bar... For the slightly less accurate requirements they were using 5ml vacuum driving diaphram pumps which dispense 5ml with each cycle of the pump. They didn't actually monitor a single fluid for flow rate. I was rather surprised by that. Even the coffee machine at work uses one of those rotary hall sensor flow meters, but for analysing blood samples they didn't feel the need. Amazing.I like your idea of having the sight glass stage, but we couldn't release all the liquids at the same time as they need to be mixed in the right order, so it would take just as long as doing it by hand. I do however want the system to be able to run without oversight as I need it to be operable by one of my neighbours when I'm off on my holidays, so if I can managed a one-button-does-it-all approach then that's what I'm going for. The big problems I have with my current manual approach is that a.) I don't actually know how much liquid is in my reservoirs.... I make an educated guess... (45-litre tub... half full... take a big off for the pump and heater.. Let's call it 17 litres in that one)... and then inaccurately add my nutrients based on that guesstimate and some lazy maths... Most of the time I get it right... but I got it wrong often enough to want to do this project.I'm looking at some interesting options for flow monitoring... I've seen some awesome 'thermal mass flow' sensors which would be perfect for the job if it wasn't for the fact that they cost $1500 each... I've currently got some bits on order so I can have a play aroud with capacitive flow sensing in a few weeks time when they arrive. Thanks for watching and taking the time to comment. Cheers!!

Hi TinkerTom - You are totally correct that EC reading only gives you an overall reading and doesn't allow you to identify the levels of individual nutrients. You are also correct that different plants will use nutrients at different rates and if you're just topping up with nutes the solution will eventually become unbalanced. As far as I'm aware, the only way to identify the levels of individual nutrients is through mass-spectromotry which generally involves a laboratory (or tens of thousands worth of equipment). If the nutrient mixture becomes too inbalanced then the plants can suffer from nutrient-lock, where the pores in the roots become clogged with a particular nutrient, preventing uptake of the ones it needs. Without the lab tests, the only way to be absolutely sure what is in your nutrient mix is to make a fresh batch every couple of weeks. An alternative approach (if you want to use the same bacth of nutrietns for longer) is to try and identify deficiences in the nutrient mixture from the plants leaves. Take a look at the Canna deficiency guide at www.cannagardening.com/deficiency-guide - If your leaves look like the pictures then you'll know which mono-nutrients your plant is lacking. Thanks for taking the time to leave a comment - Hopefully my answer is helpful. Cheers!!

Thank for the prompt reply Jim greatly appreciated 👍 It gets even more complicated if you add different plants to the mix as every plant is going to require different nutrient profiles and to further complicate things when they flower and fruit they require a different nutrient profile again. I mean I have no idea how much it would cost here in the UK to have it tested every couple of weeks but doesn't sound cheap especially if you plan on doing it for a long time. It would be okay if money grew on trees though haha 😂 I've checked out the website very informative but apart from the few deficiencies listed is there a way to tell if a plant is deficient in other elements? I can see why they don't add the trace elements but it's a real shame though just because they are used in small quantities doesn't make them less important. I mean I'm no expert but I'm sure our bodies need these trace elements I watched Cody's lab and he done a mix for algae he planned on consuming and he put a large variety of elements in his. Yes it was very helpful thanks 🙏

I haven't even finished making one of these for myself yet. You might want to take a look at Devin Olsen's DRO-Matic controller which he's been wokring on. He's a lot further along with his project than I've managed. kzfaq.info/get/bejne/fdKKe7d-0Jm0cps.html ... Cheers!!

Yes i see that my apologies i may have commented aloud! Thank you for your response

If you're worried about EMI, you can use an optocoupler to isolate the higher voltages and feedback from the arduino. They're really cheap. Just drive a N mosfet chip to run the motor.

I relay not going to not let pass the EMI if any. Are you sure that arudino was freezing up because this.

Hi Dominique Dupont. I've been fascinated by electronics and computing since an early age and studied them at college/university. The hydroponics stuff I learned entirely by watching videos on youtube and a few articles I found with google. Cheers!!

Jim Conner Cool. I'm also learning hydroponic from youtube and I realize the technology is still old. For example, I wanted to make a cheaper CO2 controller. Why does it cost 500 $ at the moment ? It has not been designed and mass produced properly yet.

Hi Dominique Dupont - The guys at my local hydro shop tell me that they reckon 95% of their customers are all growing one single highly lucrative (and illegal in the UK) crop... I think a lot of these addons are so ridiculously expensive because they know that they can sell to growers who are making $500 per plant every few months. They're really not targetted at people growing veg. I can see no other way that they could justify charging £150 for an injection moulded plastic tray when I can pick up a similar injection moulded plastic children's sledge of roughly the same size and shape for £10. I reckon you can DIY a CO2 controller much more cost effectively... Ebay tells me that a CO2-Sensor costs £28, and Arduino pro mini is £1.48 and Pneumatic Solenoid Valve is £4.65 .... I believe that works out to somewhere around $50 for that lot... After that it's just wires, hoses, soldering and code. Cheers!!

Jim Conner I don't completely agree with you, that is a too broad assumption. But parts of it is true. I think the main reason why I can't buy a CO2 controller under 100 $, is because no one like me took the time to make it ! But yes, I believe it's technologically and economically possible. Science is sparsely applied, even today !

FesterWerks Thanks Marc. Oooh... nice idea using the Imp's photocell... I'll definitely be stealing that idea from you (particularly as I haven't got the BH1750 light sensor up and working yet)... though I see from your graph that you're plotting it as a 1 or a 0 - does it read off an analog value too? I started off thinking solely about doing the pH balance part (since that's the bit I screw up so often), but decided that if I could get that to work, it wouldn't be much more complicated to dispense all the other solutions too,,, it's just more of the same over and over again. Cheers!!

Jim Conner Yes, it does do analog (I think only analog) but I only wanted to know if the light was on or off so I did some real quick math to set the value to a 0 or 1. And unless the sensor is misbehaving, my light turns on briefly around 5am ... odd. My biggest concern in the whole thing is what you're working on... reliability of the pumps to turn on/off and be accurate. I'd hate to be at work and see my plot.ly chart off the scale on the pH level and not be able to get home to fix it.

FesterWerks I've now added the light level (divided by 10,000) to my own graph at plot.ly/~Fecn/340/hydroponic-greenhouse-temperatures.embed - Thanks for letting me know about that one. It was an easy win.Nothing can screw things up quite as well as a computer... I may sometimes overdose my nutes by 5-10ml, but I'd never empty a whole 500ml bottle into there by mistake... which of course a computer will happily do. That's why I'm more worried about dealing with the failure cases I can envisage than I am about actually making the pumps dispense some nutrients (I know I can do that... raise an output high, the relay switches on and the pump starts running). I contacted a sensor company yesterday who make some tiny 120ml/min rated 'mass thermal flow' sensors,,, They've designed a microchip with a tiny heater element in the middle of two temperature sensors... when there's liquid flowing past, one sensor will read hotter than the other one (and the difference between the two sensors can be used to calculate flow rate)... If I can get hold of them at the right price (and I'm betting they cost a fortune and it's not going to happen), then they'd be totally ideal for ensuring the right amount of nutes got added. Cheers!!

Jim Conner Been curious what happened to your Growbed Nutrient Tank (green)... It's up at 85 pretty steady. Did you throw it in the oven? I do kind of like the full scale of the light as opposed to on and off. When I get this working outside, I may do that. What if you made it a bit longer of a process, adding just a little at a time, waiting, checking, adding again for each process. Your relay would only be closed for a couple seconds, then Perhaps run a loop that checks for any I/O whose state is 1 for more than x seconds, if it is, shut the whole process down or some such. If that thermal flow sensor you spoke of is a bit pricy, perhaps all liquids go through that one sensor, with a fresh water flush in between each. Just throwing stuff out there.

That 85C reading is a failed sensor... The 1820 gives a reading of 85C if it's failed to perform it's temperature conversion for some reason. The 85C is the start value on the 1820 before it deducts the number of counts from one of it's oscillators. I've had one of these sensors fail the same way before (back in 2001)... never did work out why it failed. I've got some more on order anyway so when they arrive I'll pull the faulty one off the bus and test it on it's own... it's possible there's some other reason for it failing to measure the temperature (i.e. water somehow got into my wiring loom etc) The thermal mass flow sensors were apparently around $1600 each.... so there won't be any of them going in there :).... I've ordered a little cheap capacitive touch sense board from ebay whcih I'll have a play around with... I'm hoping that it'll be able to detect the difference between liquid in the tubes and an empty tube.... Guess I'll find out in a few weeks when it gets here. Cheers!!

Hi mnelson... I still remember the days when being a nerd wasn't cool :).... The guys at my local hydro shop say that 90% of their customers are all growing the same crop.... I still don't think they really believe that I'm not growing weed - they try to put everything I buy in black plastic to hide what it is in case I get stopped.... but I have nothing to hide. I actually chose the Canna stuff because it seemed to be the least canabis-specific stuff they had in stock... It came with growing guides for peppers and tomatoes etc... Cheers!!

I like the good 'ole General Hydroponics 3-part, but I don't know if it's available in the UK. Here in the States, it's kind of the de-facto standard for research type hydroponics. I think it's also pH buffered... I never had to fiddle with pH when I used it.

mnelson10000 Having watched all of Bobbly/MHPGardener's videos, my number one choice was going to be Masterblend, but the only way to get that in the UK was to order it of Ebay US and by the time the postage was added in the price was ridiculous. My local choices seem to be 'Dutch Pro', 'Canna' or 'BioBuzz', and Canna had the helpful grow guides online, so that's what I bought... I'll give something else a try when the current 5-litre bottles I bought run out (and now that I have a bit more of a clue what all the different things are for)... Cheers!!

The plants do not use up all of the nutrients at the same rate, so if you're just topping up you tend to get high concentrations of one particular nutrient (i.e. potassium) but not enough of the others. The EC will still measure correctly because the conductivity is kept correct by the high levels of potassium, but the plant won't be able to get enough of the nutes it's actually seeking. I did look into how large commercial growers deal with having thousands of litres of solution, but they're doing full laboratory water quality analysis with spectrometers so they can measure how much of each individual nutrient is in solution and top up accordingly. Cheers!!

@@ThingsWhichArentWork Thanks

@@ThingsWhichArentWork If you watch some videos about trace elements you'll learn that our body tends to balance minerals by taking only what it needs. Maybe plants do the same think.

No progress for the last three years I'm afraid. I started this project when I had a five-week break between contracts and have never found the time to continue with it. It is still sitting on my desk gathering dust, but it's been there for so long now that I've started to scavenge components from it for other projects. Cheers!!

Hi Devin - Yes... I've left this project gathering dust for far too long. I started the project when I had a few weeks available between swtiching jobs, but then I got into my new role and left this on the desk gathering dust. I've actually switched jobs twice since I started this project (I do short-term contract roles). One day I plan to get back into it again, but the demands of work, children and life in general have meant that I've been doing other smaller projects instead. (plus, I swtiched from using pre-mixed nutrients to getting bulk powders, so I have to re-think how to do dosing from powder instead). I'll try to take a look at your DRO-Matic this evening... Well done for being far more productive than I have been. I will most likely abandon my project in favour of yours if I can get it all up and working. Cheers!!

Excellent sir! Sorry for not fully understanding where/why the project was left off. I had a vision in my mind of you possibly still working hard on it with not much success after all this time and wanted to see if I could help! Give the OS a run, and again not sure about it working with your pumps as it was designed for a syringe doser, but I really want to tweak it to work with both types.. just don't want to spend more money on parts ;)

Having said that I'd look 3 evenings ago, I'm only just having a little look at your DROMatic OS now... (past midnight on a Saturday night)... I like your syringe doser setup. I've subbed to your channel so that I won't forget about your project. I really must get on and make some more progress on my own one... not likely to happen this weekend as I wasted today reviewing a 3Doodler 3D printing pen (it was crap). Cheers!!

No worries, I've still got to get it out of a beta state myself. My daughters birthday is around the corner so I know how tight time can get (Thanks for the heads up on the pen heh). I'm hoping to get few more commits out this weekend myself and possible instructional video of the OS and wire configuration. I would definitely stick with the pumps you have, my setup was over kill and I could of saved myself a ton of money had I purchased the pumps you have. Any-who, I'm very active on the tube, so just keep in touch with what your results and I'll be more than happy to change code to accommodate. Thanks Jim!

Hi +Demetrius Holmes - I'll be putting all of the source code up on Github when there's something useful to put up there. This project has been gathering dust on my desk for far too long now (most of this year), but I've just ordered myself a 3D printer so that I'll be able to design the custom bits of plastic I'm going to need to hold all the parts together... I'll put up all the designs on github too (when I get around to doing them). I'll take a look at Robo-tank's videos this evening. Thanks for the pointer. Cheers!!

+Jim Conner what is your github name? can you provide a link? thanks, I'm interested in the code you are using to display the data to the multiple screens

Hi Kaniick - The code isn't on github, but on mbed.org instead (as that's the development environment I was using). You can find the goodies at developer.mbed.org/users/jimconner/ - Hope that helps. Cheers!!

101blog That's one hell of a slick interface he's got going on his phone there. I never thought about using AR to control the greenhouse.... Now the important question... Who's going to write it for me :)

Jim Conner I feel volunteered all of a sudden ...Maybe we can fob it off to Ben seeing as he's the expert :)

101blog I'm always up for more people helping out. Sounds like a great idea to me. Cheers!!