Hi Nikodem! Congratulations my friend for your degree! It's a personal great achievement. I'm happy to be among the very first subscribers who rapidly noticed you had energy, skills and the willing to gather the needed skills to build serious devices from A to Z, many YEARS ago. Your projects and your audience as well, will be part of the beneficiary of your complete knowledge. Pretty exciting.
On your previous video I posted references to the patent of using asymmetric pulses in the perimeter wire. This will, if you combine it with a v/2 reference and the envelope detector, allow you to know which side of the wire you are on. Super useful for not crossing the wire, but also great for following the wire as you can see the signal flip polarity as you cross it. (designating v/2 as 0v)
Congrats on graduating, great choice of degree. You should consider making a home-lab AI server to run your own local models. You'll have faster response times and more control.
Congrats on getting your diploma, Mr. ehm, Master Nikodem! I wish you all the best for the future and lots of new (crazy) ideas that come to your mind. 😎
I found a dead husqvarna (motherboard is dead), planning to use it as the base for my own arduino driven auto mower. The hardware is very simple. Here are some things you may want to know about it (happy to make a short video about it if you want to see it in more details for taking ideas from it) 2 wheels in front one in the back. 2 line sensors, one in the front and one closer to the back and they are offset only by a few centimeters so I guess this allows for better tracking of the line when trying to go home. all 3 motors are brushless but with extra wires for sensing "things" (I assume they have feedback on how they are turning or are to make the esc cheaper by providing a feedback signal). to detect "crashing" and even detecting hills the whole body is surrounded by a "shell" that is hold "floating" by some rubber masts, when you move the shell a couple of magnets move with it and two hall sensors inside can then detect the machine hit something or someone is lifting/tilting it (pretty clever). What I found amazing is how well isolated the inside and outside are, this machine was living outdoors in the sun and rain but the inside was pristine.
I am very encouraged by your progress. I want to one day build something similar, but in my case, I need it to use rtk and / or imaging and/or lidar instead of a guide wire. Hopefully, your project will eventually progress to that phase so that I can follow your steps. Perhaps by that time, some of the more advanced components required will also go down in price. One thing you might consider is using a power bank that is wireless for your mower's power source to facilitate your self recharging station. Buying a cheap mower and retrofitting it is not that cheap in many places. It might be cheaper to buy a commercial device fully done. So i like your approach better.
Bardzo interesujący projekt, ciekawa odmiana w stosunku do gotowych rozwiązań. Mając dodatkowo rozwiązanie komercyjne, czyli robota koszącego Navimow, na zasadzie porównań łatwo jest dojść do oczywistych wniosków: - płyta podłogowa - jak już wspomniałeś musi być pełna. Najlepiej ze śliskiego tworzywa, tak by nie przyklejała się trawa - osłony boczne - na przykładzie kamerki GoPro szybko da się zauważyć, że są niezbędne. Przód na zasadzie grzebienia przepuszczającego trawę, a jednocześnie spełniającego funkcję odpychającą. Tak, zgadza się, robisz tego robota na własne potrzeby, ale zawsze może się zdarzyć chwila bez nadzoru (a docelowo bez) i niefajnie by było gdyby kosiarka trafiła na jeża czy inne zwierzątko. Boki mogą być pełne - tarcze tnące - wydaje mi się, że powinny zagarniać do środka, szczególnie jeśli tył będzie bez osłony (bo tu nie musi być, a nawet nie powinno) - regulacja wysokości cięcia - wiadomo, mechanizm musi być dostępny bez otwierania obudowy. Nie musi być wspólny dla obu noży, choć to by był ideał - zasilanie - nie wiem czy nie jest korzystniej użyć gotowego rozwiązania Parkside czy Einhell, zamiast zgrzewać samemu. Przykładowo Einhell 18V 5,2Ah kosztuje na znanym portalu zakupowym niecałe 180zł, w tej cenie ciężko o DIY 5s2p na nowych ogniwach 18650 dedykowanych do elektronarzędzi, na dodatek od razu w obudowie i z BMS-em. Satysfakcja z DIY na pewno większa, ale czasem dobrze jest podejść racjonalnie ;-) - stacja dokująca - obligatoryjna, z automatycznym powrotem robota po określonym czasie lub niskim stanie naładowania - oprogramowanie - oprócz aktualnych funkcji na pewno wskazana jest możliwość zapisania harmonogramu koszenia na poszczególne dni - algorytm koszenia - może coś w stylu odkurzaczy sprzątających lub robotów Bosch - w paski? Większa efektywność ;-) Lista życzeń długa, ale wbrew pozorom do zrobienia ;-) Natomiast sam prototyp bardzo fajny, skuteczny i przede wszystkim lekki, dzięki czemu jest małe prawdopodobieństwo, że się gdzieś ewentualnie zakopie lub utknie. Zapewne jak pozostali widzowie będę z zaciekawieniem czekał na kolejną, ulepszoną wersję :-) Szczególnie, że już na tym etapie w kilku aspektach jest nawet lepszy niż niektóre modele znanych producentów.
Yes, going to the Bambu Lab printers from basic 3D printers is like night and day - my Cocoon Create Touch which was my first 3D printer had a slight bow in the bed so was tricky getting a good first layer.
Hi from the Czech Republic! I'm working on the same thing as my school project, but I want to make the lawnmower wireless. It won't be controlled by GPS, but instead, it will detect where the grass is and where it isn't. You'll just need to charge it and place it on the lawn, and it will do the rest. (My school focuses on mechanics and electrical engineering.)
hi nice work, but you should implement 1 safty feature a degree sensor so that when the thinks flips in any direction over 30degree it goes in safe mode.
concerning the rotation direction of the plates, I believe the direction is determined by the rotation direction of the screws that fix the plates - you do not want to have them losten during operation ...
I used Arduino (r3), my perimeter wire is galvanized steel electric fence wire (cheap, no need for transmitter) and inductive proximity sensor on the mower. I tried 4 blades on my first one, (looked like a hammer head shark), redesigns ended up with just one 400mm disc. I use [2] 25w solar panels to charge a 3s5p li ion battery pack. (power draw for the mower total is 19 watts) Cutter motor is 775 brushed, at 3v-1a. Wheel gear motors are 30 rpm @ 12v. (VT-555-12V- 5000R) Hope this helps. Thanks for your videos!!!
To avoid problems with the traces on the PCB, i would use a permanent marker to cover the copper, use the cnc, then submerge the board in muriatic acid, to dissolve the unwanted traces, then clean the marker with isopropyl alcohol... i love your work!
If you want you could put some smaller, higher rpm motors on this guy for its cutting blades and just increase the weight of the disc. Just make sure you start it on a clean surface. Doing that should make sure you get all the grass in the one pass.
Years ago when I did electronics at school we made PCBs by drawing or printing the design onto the copper then we put the PCB in a acid bath, is that not something that is done anymore?
Yeah, it's far less popular due to safety + complexity reasons. Most hobbyists that want a PCB will just use JLC/PCBWay, they make a 5 boards for literally a couple bucks, and they can do 6 layers (or even more) fairly cheaply. Some people who want boards FAST (without shipping from cn) will use a mill, but the number of people who want to sensitize, print, transfer the print, etch, and then clean is really dwindling.
You can do a thermal transfer of a toner from a laser printer which is a similar method you mentioned but way more accurate. It's a bit of work and you need to play with the acid, sometimes it's tricky to get it right and you need to drill all the holes by hand. Doing it with a CNC makes a lot more sense to me, is easier and takes less time.
@@nikodembartnik for someone like me that doesn't have the CNC I would try and make it with the acid bath. Looking forward to the next episode of this project
Can’t say ChatGPT was wrong. Since it’s a voltage divider, all it needs is a ratio (in this case approx 1:4 ) which can be achieved by either 27k and 10k or 2.7k and 1k if it made sense.
Have you considered using trimmer line instead of blades for the cutters? Would definitely be lighter and probably safer. Easier to maintain too. No sharpening.
Regarding the direction of rotation for the mower blades- should they oppose one another because if they go the same way they could try rotate the mower- it would depend on how heavy they are and how fast they spin. Also the direction they move in could pile the cut grass up or spread it out.
Great project, and the best part is its really useful. I was thinking about create some sort of that for the last year. ps. I need to write something for the yt algorithm
If you remove all of the copper between the traces the copper "hairs" are gone. Takes a bit longer but also makes the PCB (oder MCB?) way cleaner. At least it worked for me.
Not dogging on this design, but I have a question in general for all these robot mowers... why not just use an 18/40v weed trimmer motor/mechanism for the actual 'mower' portion? Everyone has so many issues with razor blades and other odd designs, but a weed-whacker works great, and is basically maintenance-free for months. Self-adjusting mechanisms.
Good point but there are some problems with that. 1.Noise and efficiency: The motor needs to spinn a lot faster to cut the grass and it‘s generally a lot louder if you cut grass with a plastik wire. That makes your robot way too loud and that faster and bigger motor needs a lot more power and that means a way bigger battery for the same runtime. 2. Cut quality: The grass looks a lot better when you cut it with blades instead of a plasik wire. 3. Size: The bigger motor, the bigger battery and the bigger driving circuitry for the motor needs a lot more space, so the robot is going to be bigger. 5. Safety: when you use a weed wacker tiny bits of grass come flying out that can cut you or fly in your eyes. The robot would have the same problem that’s especially bad when children or pets are near the robot. Another reason this would be bad for safety is that robot lawn mower’s are often driving around with no supervision, so if a child reaches under the robot with small blades the damage would be relatively minimal but with a robot with a weed wacker cutting system the damage would be a lot worse.
Good points mentioned by cheesburger2875. Additionally because these mowers cut just a very tiny piece of the grass on each pass it's harder to cut with a line and a sharp blade works best.
@@scrampker not really because the motor needs to spinn a lot faster to cut the grass and in most robot lawn mower’s with tiny blades, the mowing disk is designed that when the blades hit something that is harder than grass they swing under the mowing disks.
@@RoombaMaster6999 Gotcha, so any sort of weed or vine cannot be cut? Anyhow I mention this mostly because 40v Ryobi stuff is dirt cheap. Would be rather trivial to mount the weed head, controller, and battery sled on a 4 wheel platform. Could have the robot run off the 40v, and even have a second sled. Regardless I love to see these projects. I just wonder "what if" someone used a 40v weed eater mechanism. I see ground crews use these all the time to mow entire lots, depending on the size, and the result looks great. I think the main benefit is that a weed eater can easily take down taller grass so this doesn't have to run daily. Keep up the good work!
ChatGPT is not stupid or weak. It's nothing but a repository of text accessed by tokenizing a textual input into a sort of vector into the repository. If you understand that using it is pretty straight forward.
yeah the current versions of gpt struggle with complex maths. its surprising that the mistake for you was a factor of 10 and not something completely different, as it tends to be.
Nie zgrzewaj baterii zgrzewarką bo to bardzo prosty sposób do pożaru. Zrób tak jak to robi tesla. Na każdym ogniwie zrób na plusie i minusie bezpieczniki w postaci cienkiego drucika lub np. trzech bardzo cienkich. W ten sposób w razie zwarcia lub uszkodzenia jednego ogniwa drucik się przepala i reszta jest ok i nie ma pożaru. Zobacz sobie moje filmy tam pokazuję jak prosto to zrobić.
Nie jest to prawda. Punktowe zgrzewanie baterii czyli spot welding to najbezpieczniejszy sposób ich łączenia i tak to się robi w rowerach elektrycznych, hulajnogach czy deskorolkach.
To powiedz mi dla czego w autach elektrycznych są takie bezpieczniki? Np. tesla ,lucid, bmw. Blachy robią Chińczycy bo nie interesuje ich twoje bezpieczeństwo. Jak jedno ogniwo lub tranzystory w balanserze dostaną zwarcia to będziesz mieć piękny pożar. Ale jak to mówią jest ryzyko - jest zabawa. Tylko szkoda będzie potem jak trzeba będzie odbudować garaż lub w najgorszym przypadku dom. Widzę że nie zdajesz sobie sprawy z tego jak niebezpieczne mogą być takie ogniwo i jaki prąd mogą z siebie dać. Przewody od balansera stają się czerwone w sekundę w momencie zwarcia. Pomyśl o tym i zrób to najlepiej jak potrafisz a potem będziesz spał spokojnie. pozdrawiam
What do you recommend for a beginner with 20 yrs C/C++ experience to learn ML? I want to show it 3D computer generated designs, and have it learn to identify a real object. Also, can I get away with not knowing Python but 20+ years in C/C++?
The problem, is people keep thinking that chatgpt is an Ai...its not... It's a large language model that predicts text... It isn't a "computer", it can't do math...
I still think perimeter wire is the wrong way to go. You can only use it if you have permission from the land owner to modify their property, and it can interfere with the pipes, garden, ecosystem, not to mention, most millennials don't even own their land (even land they are responsible to maintain). And that's before HOAs get involved. For many properties, it's just not possible. I think building the navigation in a modular manner to allow GPS, AGPS, or even just temporary radio base stations are a better option imo. Also think about the logistics for perimeter wire, if you are running a business, it's just awful; you can't build a fleet of these and run them on other peoples' properties if all of your customers have to get in touch with the land owners to agree to ruining their yards with a wire install. (Even if they don't garden, and you know where all the pipes are and you think it doesn't ruin the land, then go ahead and try to talk to the 80 year old owner who thinks they can feel Wi-Fi, and tell me how many hours it takes you to convince one, if you ever do; and if you can even get a meeting with one. Most people will decline your business vs even figuring out how to arrange a meeting between you the land owner...) Not to mention contracting for lawn mowing vs installation of a permanent fixture is a completely different level of liability, and will definitely raise your insurance premiums...
I noticed you are setting the wire on top of the grass in your video and my above comment was in thinking it had to be installed and digging up the yard was required. I'm still not sold on it. Like yeah, you can do that.. But I feel some yards will still be more work than just mowing them by hand. Anyway, I still recommend going modular, so that you can support any kind of navigational plugin. And not locking yourself into just one technology.
Chatgpt isn't stupid it's just ignorant. The learning process is still being worked through. It's like calling a baby stupid when it's trying to figure out how to walk. Then falls over, gets up and tries again lol
Please stop this "watch until the end" stuff. I see it more and more and I despise it, if your video is good I'll watch it to the end regardless, this just makes me not continue an other intersting video
Ehhh, I'm not so sure about the insistence on doing it from scratch - I believe the approach of retrofitting a Chinese lawn mower with open / smart "brains" is significantly better. Why not reuse the work already done by Chinese companies? Milling PCBs is also super gimmicky. I also have a CNC capable of it (Carvera) and it's not worth the effort unless you need something done immediately or its just a single breakout board etc. Why waste time fixing your manufacturing process if JLC will make professional PCBs for $20?
Agreed. I already stated in previous videos that he wasted a lot of time trying to reinvent wheels instead of looking at all the mature solutions that have been field tested for 10 years. Obviously I get the somewhat educational aspect of this approach but his choice also feels dictated by the need to feed his YT channel with a time-consuming series… Frankly I’m also surprised that he studied ML as part of his degree but still used ChatGPT like a candid beginner.
Hi Nikodem! Congratulations my friend for your degree! It's a personal great achievement. I'm happy to be among the very first subscribers who rapidly noticed you had energy, skills and the willing to gather the needed skills to build serious devices from A to Z, many YEARS ago. Your projects and your audience as well, will be part of the beneficiary of your complete knowledge. Pretty exciting.
Thanks! Congratulations on the degree!!
Thank you!!
Congratulations on your graduation. You will do many great things, I'm sure. Well done.
Thank you!
On your previous video I posted references to the patent of using asymmetric pulses in the perimeter wire.
This will, if you combine it with a v/2 reference and the envelope detector, allow you to know which side of the wire you are on.
Super useful for not crossing the wire, but also great for following the wire as you can see the signal flip polarity as you cross it. (designating v/2 as 0v)
Yes, a workshop build video would be very interesting.
Congrats on graduating, great choice of degree. You should consider making a home-lab AI server to run your own local models. You'll have faster response times and more control.
Love to see the improvements! I look forward to the spring to see what else you come up with!
Congratulations on your degree! Welcome to our big family of Electronic and Computer Engineers and Scientists! Wish you ll the best!
Congratulations for your degree and thank you for sharing your progress!
Congrats on getting your diploma, Mr. ehm, Master Nikodem! I wish you all the best for the future and lots of new (crazy) ideas that come to your mind. 😎
I found a dead husqvarna (motherboard is dead), planning to use it as the base for my own arduino driven auto mower. The hardware is very simple. Here are some things you may want to know about it (happy to make a short video about it if you want to see it in more details for taking ideas from it)
2 wheels in front one in the back.
2 line sensors, one in the front and one closer to the back and they are offset only by a few centimeters so I guess this allows for better tracking of the line when trying to go home.
all 3 motors are brushless but with extra wires for sensing "things" (I assume they have feedback on how they are turning or are to make the esc cheaper by providing a feedback signal).
to detect "crashing" and even detecting hills the whole body is surrounded by a "shell" that is hold "floating" by some rubber masts, when you move the shell a couple of magnets move with it and two hall sensors inside can then detect the machine hit something or someone is lifting/tilting it (pretty clever).
What I found amazing is how well isolated the inside and outside are, this machine was living outdoors in the sun and rain but the inside was pristine.
Congratulations on your degree. I found your channel recently and I am really impressed.
Congratulations on completing your degree Nikodem!
Congratulations for your degree!! Keep up the good work
I am very encouraged by your progress. I want to one day build something similar, but in my case, I need it to use rtk and / or imaging and/or lidar instead of a guide wire. Hopefully, your project will eventually progress to that phase so that I can follow your steps. Perhaps by that time, some of the more advanced components required will also go down in price. One thing you might consider is using a power bank that is wireless for your mower's power source to facilitate your self recharging station. Buying a cheap mower and retrofitting it is not that cheap in many places. It might be cheaper to buy a commercial device fully done. So i like your approach better.
I'm excited to see what you will build with your knowledge of AI and machine learning.
Bardzo interesujący projekt, ciekawa odmiana w stosunku do gotowych rozwiązań. Mając dodatkowo rozwiązanie komercyjne, czyli robota koszącego Navimow, na zasadzie porównań łatwo jest dojść do oczywistych wniosków:
- płyta podłogowa - jak już wspomniałeś musi być pełna. Najlepiej ze śliskiego tworzywa, tak by nie przyklejała się trawa
- osłony boczne - na przykładzie kamerki GoPro szybko da się zauważyć, że są niezbędne. Przód na zasadzie grzebienia przepuszczającego trawę, a jednocześnie spełniającego funkcję odpychającą. Tak, zgadza się, robisz tego robota na własne potrzeby, ale zawsze może się zdarzyć chwila bez nadzoru (a docelowo bez) i niefajnie by było gdyby kosiarka trafiła na jeża czy inne zwierzątko. Boki mogą być pełne
- tarcze tnące - wydaje mi się, że powinny zagarniać do środka, szczególnie jeśli tył będzie bez osłony (bo tu nie musi być, a nawet nie powinno)
- regulacja wysokości cięcia - wiadomo, mechanizm musi być dostępny bez otwierania obudowy. Nie musi być wspólny dla obu noży, choć to by był ideał
- zasilanie - nie wiem czy nie jest korzystniej użyć gotowego rozwiązania Parkside czy Einhell, zamiast zgrzewać samemu. Przykładowo Einhell 18V 5,2Ah kosztuje na znanym portalu zakupowym niecałe 180zł, w tej cenie ciężko o DIY 5s2p na nowych ogniwach 18650 dedykowanych do elektronarzędzi, na dodatek od razu w obudowie i z BMS-em. Satysfakcja z DIY na pewno większa, ale czasem dobrze jest podejść racjonalnie ;-)
- stacja dokująca - obligatoryjna, z automatycznym powrotem robota po określonym czasie lub niskim stanie naładowania
- oprogramowanie - oprócz aktualnych funkcji na pewno wskazana jest możliwość zapisania harmonogramu koszenia na poszczególne dni
- algorytm koszenia - może coś w stylu odkurzaczy sprzątających lub robotów Bosch - w paski? Większa efektywność ;-)
Lista życzeń długa, ale wbrew pozorom do zrobienia ;-) Natomiast sam prototyp bardzo fajny, skuteczny i przede wszystkim lekki, dzięki czemu jest małe prawdopodobieństwo, że się gdzieś ewentualnie zakopie lub utknie. Zapewne jak pozostali widzowie będę z zaciekawieniem czekał na kolejną, ulepszoną wersję :-) Szczególnie, że już na tym etapie w kilku aspektach jest nawet lepszy niż niektóre modele znanych producentów.
Yes, going to the Bambu Lab printers from basic 3D printers is like night and day - my Cocoon Create Touch which was my first 3D printer had a slight bow in the bed so was tricky getting a good first layer.
I want to make a project like this, this is how I found you channel
Congrats to getting the MoDS. Good luck for the future.
Hi from the Czech Republic! I'm working on the same thing as my school project, but I want to make the lawnmower wireless. It won't be controlled by GPS, but instead, it will detect where the grass is and where it isn't. You'll just need to charge it and place it on the lawn, and it will do the rest. (My school focuses on mechanics and electrical engineering.)
congratulations on the titel, as always its nice to watch you work and figure things out
This is brilliant. Thanks for making this video!
Gratz! And very cool project here!
Have you heard you can make PCB's using etching and resin printer screens? Something to consider.
Winter is coming, convert it into an automatic snow blower :)
hi nice work, but you should implement 1 safty feature a degree sensor so that when the thinks flips in any direction over 30degree it goes in safe mode.
yeah, good job not trusting ChatGPT. I asked it a wiring question, followed the instructions carefully, and burnt an $80 board.
Use copilot instead. It even gives a reference video
Dobra robota! Gratulacje
concerning the rotation direction of the plates, I believe the direction is determined by the rotation direction of the screws that fix the plates - you do not want to have them losten during operation ...
I used Arduino (r3), my perimeter wire is galvanized steel electric fence wire (cheap, no need for transmitter) and inductive proximity sensor on the mower. I tried 4 blades on my first one, (looked like a hammer head shark), redesigns ended up with just one 400mm disc. I use [2] 25w solar panels to charge a 3s5p li ion battery pack. (power draw for the mower total is 19 watts) Cutter motor is 775 brushed, at 3v-1a. Wheel gear motors are 30 rpm @ 12v. (VT-555-12V- 5000R) Hope this helps. Thanks for your videos!!!
To avoid problems with the traces on the PCB, i would use a permanent marker to cover the copper, use the cnc, then submerge the board in muriatic acid, to dissolve the unwanted traces, then clean the marker with isopropyl alcohol... i love your work!
If you want you could put some smaller, higher rpm motors on this guy for its cutting blades and just increase the weight of the disc. Just make sure you start it on a clean surface. Doing that should make sure you get all the grass in the one pass.
Impressive work.
Nice work !
that's awesome! I'd love to see further refinement, focused on open sourced lawn mowing 😎🌞
Thank you!
oh hey i made a servo actuator set for the openroboticplatform it needs some kinks work on but i can wait to post it.
Nice. Please Nikoden, make a tutorial about manufacturing PCBs with the CNC machine.
Will do, hopefully this year!
Years ago when I did electronics at school we made PCBs by drawing or printing the design onto the copper then we put the PCB in a acid bath, is that not something that is done anymore?
Yeah, it's far less popular due to safety + complexity reasons. Most hobbyists that want a PCB will just use JLC/PCBWay, they make a 5 boards for literally a couple bucks, and they can do 6 layers (or even more) fairly cheaply. Some people who want boards FAST (without shipping from cn) will use a mill, but the number of people who want to sensitize, print, transfer the print, etch, and then clean is really dwindling.
You can do a thermal transfer of a toner from a laser printer which is a similar method you mentioned but way more accurate. It's a bit of work and you need to play with the acid, sometimes it's tricky to get it right and you need to drill all the holes by hand. Doing it with a CNC makes a lot more sense to me, is easier and takes less time.
Thankfully it’s not the case anymore. Quite often this acid ended up in a sink
@@nikodembartnik for someone like me that doesn't have the CNC I would try and make it with the acid bath. Looking forward to the next episode of this project
Can’t say ChatGPT was wrong. Since it’s a voltage divider, all it needs is a ratio (in this case approx 1:4 ) which can be achieved by either 27k and 10k or 2.7k and 1k if it made sense.
Ok, I forgot to mention that I supplied the value of the second resistor already so there was only one solution.
Congratulations on your graduation.
Im EXCITED. I love this idea
Have you considered using trimmer line instead of blades for the cutters? Would definitely be lighter and probably safer. Easier to maintain too. No sharpening.
Regarding the direction of rotation for the mower blades- should they oppose one another because if they go the same way they could try rotate the mower- it would depend on how heavy they are and how fast they spin. Also the direction they move in could pile the cut grass up or spread it out.
Great job
Congrats Master.
Pico is also my favourite for now. MicroPython, overclock to 250Mhz, two separate cores.
Great project, and the best part is its really useful. I was thinking about create some sort of that for the last year.
ps. I need to write something for the yt algorithm
Thank you for the nice comment and help with the Algorithm
yes, mower video, stay on the grind
RTK would still be a really nice technology
Interesting project. If you improve the blades and add a bag it would do the job. The soft seem to work well.
Congratulations to your MA.
Throw a PV panel on top of it, and it will keep Wall-E company.
If you remove all of the copper between the traces the copper "hairs" are gone. Takes a bit longer but also makes the PCB (oder MCB?) way cleaner. At least it worked for me.
Split your test area and let 3/4 fallow lest it get exhausted.
PCB was hard to create: Next time, use JLCPCB as a sponsor :)
Not dogging on this design, but I have a question in general for all these robot mowers... why not just use an 18/40v weed trimmer motor/mechanism for the actual 'mower' portion? Everyone has so many issues with razor blades and other odd designs, but a weed-whacker works great, and is basically maintenance-free for months. Self-adjusting mechanisms.
Good point but there are some problems with that.
1.Noise and efficiency: The motor needs to spinn a lot faster to cut the grass and it‘s generally a lot louder if you cut grass with a plastik wire. That makes your robot way too loud and that faster and bigger motor needs a lot more power and that means a way bigger battery for the same runtime.
2. Cut quality: The grass looks a lot better when you cut it with blades instead of a plasik wire.
3. Size: The bigger motor, the bigger battery and the bigger driving circuitry for the motor needs a lot more space, so the robot is going to be bigger.
5. Safety: when you use a weed wacker tiny bits of grass come flying out that can cut you or fly in your eyes. The robot would have the same problem that’s especially bad when children or pets are near the robot.
Another reason this would be bad for safety is that robot lawn mower’s are often driving around with no supervision, so if a child reaches under the robot with small blades the damage would be relatively minimal but with a robot with a weed wacker cutting system the damage would be a lot worse.
Good points mentioned by cheesburger2875. Additionally because these mowers cut just a very tiny piece of the grass on each pass it's harder to cut with a line and a sharp blade works best.
@@RoombaMaster6999 I can agree on many points but not safety. Plastic string vs razor blades. Pretty obvious which would do more damage to flesh.
@@scrampker not really because the motor needs to spinn a lot faster to cut the grass and in most robot lawn mower’s with tiny blades, the mowing disk is designed that when the blades hit something that is harder than grass they swing under the mowing disks.
@@RoombaMaster6999 Gotcha, so any sort of weed or vine cannot be cut?
Anyhow I mention this mostly because 40v Ryobi stuff is dirt cheap. Would be rather trivial to mount the weed head, controller, and battery sled on a 4 wheel platform. Could have the robot run off the 40v, and even have a second sled.
Regardless I love to see these projects. I just wonder "what if" someone used a 40v weed eater mechanism. I see ground crews use these all the time to mow entire lots, depending on the size, and the result looks great. I think the main benefit is that a weed eater can easily take down taller grass so this doesn't have to run daily.
Keep up the good work!
When this will be available on Kickstarter? :D
You're awesome. Subscribed
Cool video
Love it !
Would a Cylinder Lawn Mower work longer? (More power efficient)
was it 5y to complete your master? or both your bachelor and master?
It was both, 3.5 years for the bachelor and 1.5 year for the master
ChatGPT is not stupid or weak. It's nothing but a repository of text accessed by tokenizing a textual input into a sort of vector into the repository.
If you understand that using it is pretty straight forward.
I used to work for a Autonomous mower company (electric sheep robotics)
If you use motor to turn both in the middle of the mower, you will have a line in the middle of pass...
You should put solar panels on it
yeah the current versions of gpt struggle with complex maths. its surprising that the mistake for you was a factor of 10 and not something completely different, as it tends to be.
Nie zgrzewaj baterii zgrzewarką bo to bardzo prosty sposób do pożaru. Zrób tak jak to robi tesla. Na każdym ogniwie zrób na plusie i minusie bezpieczniki w postaci cienkiego drucika lub np. trzech bardzo cienkich. W ten sposób w razie zwarcia lub uszkodzenia jednego ogniwa drucik się przepala i reszta jest ok i nie ma pożaru. Zobacz sobie moje filmy tam pokazuję jak prosto to zrobić.
Nie jest to prawda. Punktowe zgrzewanie baterii czyli spot welding to najbezpieczniejszy sposób ich łączenia i tak to się robi w rowerach elektrycznych, hulajnogach czy deskorolkach.
To powiedz mi dla czego w autach elektrycznych są takie bezpieczniki? Np. tesla ,lucid, bmw. Blachy robią Chińczycy bo nie interesuje ich twoje bezpieczeństwo. Jak jedno ogniwo lub tranzystory w balanserze dostaną zwarcia to będziesz mieć piękny pożar. Ale jak to mówią jest ryzyko - jest zabawa. Tylko szkoda będzie potem jak trzeba będzie odbudować garaż lub w najgorszym przypadku dom. Widzę że nie zdajesz sobie sprawy z tego jak niebezpieczne mogą być takie ogniwo i jaki prąd mogą z siebie dać. Przewody od balansera stają się czerwone w sekundę w momencie zwarcia. Pomyśl o tym i zrób to najlepiej jak potrafisz a potem będziesz spał spokojnie. pozdrawiam
What do you recommend for a beginner with 20 yrs C/C++ experience to learn ML?
I want to show it 3D computer generated designs, and have it learn to identify a real object.
Also, can I get away with not knowing Python but 20+ years in C/C++?
The problem, is people keep thinking that chatgpt is an Ai...its not... It's a large language model that predicts text... It isn't a "computer", it can't do math...
LDR please
did you go to open sauce?
unfortunately not
I still think perimeter wire is the wrong way to go. You can only use it if you have permission from the land owner to modify their property, and it can interfere with the pipes, garden, ecosystem, not to mention, most millennials don't even own their land (even land they are responsible to maintain). And that's before HOAs get involved. For many properties, it's just not possible.
I think building the navigation in a modular manner to allow GPS, AGPS, or even just temporary radio base stations are a better option imo.
Also think about the logistics for perimeter wire, if you are running a business, it's just awful; you can't build a fleet of these and run them on other peoples' properties if all of your customers have to get in touch with the land owners to agree to ruining their yards with a wire install. (Even if they don't garden, and you know where all the pipes are and you think it doesn't ruin the land, then go ahead and try to talk to the 80 year old owner who thinks they can feel Wi-Fi, and tell me how many hours it takes you to convince one, if you ever do; and if you can even get a meeting with one. Most people will decline your business vs even figuring out how to arrange a meeting between you the land owner...)
Not to mention contracting for lawn mowing vs installation of a permanent fixture is a completely different level of liability, and will definitely raise your insurance premiums...
I noticed you are setting the wire on top of the grass in your video and my above comment was in thinking it had to be installed and digging up the yard was required.
I'm still not sold on it. Like yeah, you can do that.. But I feel some yards will still be more work than just mowing them by hand.
Anyway, I still recommend going modular, so that you can support any kind of navigational plugin.
And not locking yourself into just one technology.
🥇
🥈
Dziwi mnie tylko fakt, że pierwszą rzeczą przy DIY autonomicznej kosiarce nie była osłona ostrzy.
no racja
Chatgpt isn't stupid it's just ignorant. The learning process is still being worked through. It's like calling a baby stupid when it's trying to figure out how to walk. Then falls over, gets up and tries again lol
this thing could end up chewing out a lot of animals 😶
Myślałem, że masz z 16 lat
Please stop this "watch until the end" stuff. I see it more and more and I despise it, if your video is good I'll watch it to the end regardless, this just makes me not continue an other intersting video
Ehhh, I'm not so sure about the insistence on doing it from scratch - I believe the approach of retrofitting a Chinese lawn mower with open / smart "brains" is significantly better. Why not reuse the work already done by Chinese companies?
Milling PCBs is also super gimmicky. I also have a CNC capable of it (Carvera) and it's not worth the effort unless you need something done immediately or its just a single breakout board etc. Why waste time fixing your manufacturing process if JLC will make professional PCBs for $20?
what about shipping time
Agreed. I already stated in previous videos that he wasted a lot of time trying to reinvent wheels instead of looking at all the mature solutions that have been field tested for 10 years. Obviously I get the somewhat educational aspect of this approach but his choice also feels dictated by the need to feed his YT channel with a time-consuming series… Frankly I’m also surprised that he studied ML as part of his degree but still used ChatGPT like a candid beginner.
yes as openmower project at the end