SuperHouseTV #1: Home Automation System Architecture 2012
Vložit
- čas přidán 25. 08. 2024
- * Episodes: www.superhouse...
* Discord: www.superhouse...
* Gear I use: www.superhouse...
* Patreon / GitHub Sponsors: www.superhouse...
A basic introduction to the approach I'm taking with linking high-voltage devices in my house to the automation system, and how logical inputs are associated with outputs using MQTT. See www.superhouse.tv for more information.
Apology: The "threat" I attribute to Andy in the video is not something he's ever said, or something I'd expect him to say. He's a peaceful soul, and it was purely hyperbole on my part when I was recording the video!
SuperHouse:
- www.superhouse.tv
- superhousetv
- superhousetv/
Jonathan Oxer:
- jonoxer/
- jonoxer
I just finished your entire SuperHouse series up until 16. Great videos, very educational and very well explained. I look forward to the ones to come! Cheers
My home automation setup uses software called Open Source Automation (OSA). Not only is the software open source, but it is designed to be a bridge between nearly all of the home automation technologies out there. When I started, my setup was primarily X10 based. I have since added an RS485 thermostat, some 1-Wire temp sensors, and some magnetic reed switches that I have run through a custom built parallel port interface. And all of these technologies are controlled through OSA.
Great intro video! Nicely done! Published in 2012, well before it's time!
Also, great upgrade path from POST->MQTT, get your 'stakeholder' concerns, if I took all the lights offline for a week I'd get destroyed by the wife.
This channel is pure gold
What a great series. I watched them in reverse order so it was an interesting journey back in time. Really well explained and lots of good ideas. Definitely more motivated to move my home monitoring system forward!
Thanks! And please share your projects if you get a chance. I love to see what other people have built.
Good videos, well explained and very educational for those who knows nothing about home automation, even if they are old now.
Not "YOU-bun-too", it's not pronounced as the letter U. Should be "oo-boon-too" (but with the middle syllable shortened)
The centralised-control approach has another cost consideration that's more important than the cable: labour. With a retrofit the cost of dragging cable can be very high, more than the cable itself. Luckily we had a combo of new build at framing stage and old reno with plaster ripped off, so dragging cable only took a few hours and cost was minimal
I'm using common 12V coil, 250V / 10A rated relays fitted in DIN rail mounts. Yes, they could control outlets as well, but my plan is to move to SSRs because the relays make loud clicks when they turn on.
Very nice seeing something like that on a big scale! I have something like [] --> light switch --> socket-outlet --> wireless socket-outlet --> light bulb at the moment, because that was the simpliest way for me to get wireless controlled light in my room, but when I turn off the light switch on the wall, the wireless socket-outlet loses it's power. No problem, that it's no longer receiving, but it also loses it's state and it's default state is off. (1/3)
Awesome! I'm really excited about this channel! Maybe I'll be capable of making/planning my entire house circuit system! I do mechatronics. :D
I'll keep on watching! Really good!
Unless I misunderstand what you're saying, that is pretty much what I'm doing. The communication happens using either web or MQTT, depending on the particular node.
Lutron is no good for me because it's only 110V, won't work with Australian 240V mains. The big reason I'm going DIY, though, is to avoid the whole "closed-ecosystem proprietary package" integration problem. As soon as you start using classic commercial HA products, you have a problem linking it to anything else. I can't be bothered spending my time trying to reverse engineer some proprietary commercial system when I could make something myself from scratch in less time.
I get that at times, it's nice to have some additional logic available at the switch. But for turning lights on and off, I don't understand the need for an Arduino at every switch. Why not have the button just close the circuit and have a central Arduino handle that? Seems way overkill to me, but I'm not a pro here.
Same idea here. He could have cat5e to the switches and have room for the hot/neg pair and then three other pairs for buttons. Less logic at the switches and less heat.
Nice video. You may also want to checkout the review of home automation on my blog at *bryanreviews. com/home-automation-review/* Thanks, Harlan.
Yes it was. I implanted an RFID chip in my arm about 7 years ago using that implanter. I've done demos and talks about it a few times (and been interviewed on Sunrise, which is in one of the other vids on this channel) and I intend to do a future SuperHouse episode about it.
All nice but when the Linux box crashes - you have no lights. If I was refurbing a property, I would prefer physical switches that are also controlled through automation. That way when you have technical issues, your not plunged into darkness.
Richard Smith so true...
Richard Smithøå
Of course, I have yet to have a linux box crash on me. They get the most amazing uptimes.
Copper is expensive, surely it would have been cheaper to use a localised control at the light fitting? So have the original switched loop power supply as is usually installed, a small contactor/relay in a junction box per light fitting and a cheaper low voltage control cable back to your control box. You'd then still use your same switching panels etc, but the large mains cable is used more efficiently and saves you money. You could even go totally ethernet control to the fitting.
So glad I found your channel! I just wish you had a little more detail for the people starting, in your later videos. Like how to put the watchdog on that prototype board.
It's a tough balance between keeping the videos brief enough for people to watch, but providing enough information to do it yourself. It already takes days of work (up to a week) just to produce each 20-ish minute video so I can't afford to do much more. In recent episodes I'm trying to put more supporting information in the blog post for it, with more links, diagrams, source code, explanations, etc. If you only watch the video on CZcams, you're missing out on that extra information. There's a link in each video description going to the blog post for that episode.
Thanks!
Hi Jon. Have you considered doing an update now that you've got the MQTT stuff in place? Or perhaps an instructional on getting the MQTT infrastructure in place for all your devices to pub/sub to. How to handle redundancy, etc.
I'll be doing a new episode with an update on the architecture when I get a chance.
Hello, Like Doug Dale and Skulltronics said before, looking this system from a maintenance, ease of use, reliability and installation costs, automation systems should integrate with existing cabling infrastructure (atmel's zigbee modules, less cable, less installation costs), and a central console with something simpler like a small PLC (automation direct are really cheap and reliable) that would not require an expert to do some programing and HMI's.
Don't get me wrong, your system is excellent and I am no expert, but I am trying to understand the reason to chose that topology, keep up the good work, excellent videos. Sorry for my english.
Why not run a cluster of Raspberry Pi's one for each room of your house, prevents your single point of failure as well as distributing the network load. It also allows your home to easily scale up and down as you need as well as allowing integration to your televisions. It's also relativity lost cost and could incorporate into your current system with minimal effort.
Node Red would be a very good software to use for this, it is visual and node based, and has an awesome looking web UI available
Thanks for your replay.... I think you should look into it.. it seems to me that it might with the "too much cable" problem.... as all protocols, it’s not perfect and the main drawback that i see in it is that it may be a standard of wide spread use in Europe, it seems to be scarcely available in the US, which hiders it worldwide adoption... I don’t know how widely use it might be in Australia... Here in Latin America we are very far away from any type of standard or regulation on the field
If you just want to do a few simple things and plug in some commercial gear, yeah, maybe. It's far more limited though: it's a single-master, multiple-slave bus, rather than all-nodes-equal peer group like I have using Ethernet. I wouldn't gain anything by using DMX, but I'd lose a lot of flexibility.
Why not use the wireless switches that are already networked interfaced and can communicate with many devices and programs? Like the Lutron stuff or other systems? This seems way more complex then it has to be especially for existing homes. Basically you just replace the switches with the wireless devices and you have a hub of some sort.
That means I have to turn on the light switch and trigger my wireless socket-outlet to turn on the light. Everything would be much better if the default state would be "on". Is there any way to hack that wireless socket-outlet? The worst and craziest method I thought about is sending from another source every few ms an "on" signal to catch that unique moment I turn the light switch on.. :/ (2/3)
A contactor/triac would be cheap and reliable. Given the switch wire was probably already behind the wall plate. If access was an issue why not just mount the contactor/triac in the cavity with the already included arduino running the switch control panel? Easy to repair, no costly and large distribution boxes scattered around the house. I really cannot see any advantage in running 100's of metres of twin and earth for every light fitting unless for a commercial fitout and using large dimmers.
Wouldn't a system based on DMX standard be way more simplier to establish and to programm?
That totally depends what end result you're trying to achieve! Perhaps you could talk about your project in the Home Automation section of the Freetronics forum: I can't post URLs here in a CZcams comment, but if you search for "Freetronics forum" and go to the first result you'll see a section called "SuperHouseTV Home Automation"
I already have ordered a 433,92 MHz sender and receiver for my Arduino and the receiver could log the data emited from the wireless remote control anyway, so the current state would be known. But that is really not a good solution. Is there any way to alter the wireless socket-outlet, i.e. "solder an "invert" onto the PCB", to speak in a logical way? Haven't opened one of the outlets yet, so I just ask for "feasibility". :) Maybe you have a great idea. (3/3)
You said it, MQTT is machine to machine communication, but this is no CNC Lathe or PAC Automated Machine in an industrial plant with thousands of variables, where talking about less than 200 I/O points (maybe?), no need for extreme safety or quick response time, I am trying to understand why this over complicated topolgy with Linux and ethernet comunication.
What do you think about KNX??
you could use the same tiwsted pair for all the light switches, not to mention sensors, etc.
Can you use this setup for setting up dimmers for the circuits?
Either set it up hardwired, or using some type of logic to limit power output.
Very good and informative videos!
Oh please tell me that the hyperdermic syringe at 0:08 was for a hand implanted RFID transponder for area access control........please oh please oh please
Can you do a video on working as best as one can with existing home wiring? Maybe with use of wifi enabled chips and sonof...etc somewhere in the line in the attic crawlspace?
Awesome video, thanks, Jon. I liked the web server and web interface on the tablets. I was wondering how you can still provide the HTML/AJAX UI on the tablet once you migrate to MQTT.
now it's almost 9 years later, and i have to ask: would you still use the same architecture, or is there something better now?
Watching your video and your DIY home automation system, I really dream to do the same kind of thing one day... well done !
You're talking about latencies improvement with MQTT. Have you any measurement to compare H TTP vs MQTT ? Is the H TTP (TCP) latency really discernable in your local network ? Is MQTT a UDP based protocol ? Why do you need to get separate buses to segregate the messages sent ? Is it to specialize each program that will have to listen to one type of message ? Thanks
"U in English can never be pronounced as "oo"."
Ubuntu is not an English word. It originates with the Bantu languages in Africa.
"if you want to pronounce it as an African word, then it needs accents."
So you're saying that any language other than English that happens to use the same letterforms but with different pronunciation has to put accents on the letters? Try telling that to Spanish speakers who pronounce "j" as "h", etc.
Search for "wikipedia ubuntu philosophy" to hear pronunciation
I like your videos but I'm a little bit confused. I would like to ask you something. Why don't you use a client server based system and do the communication via sockets. The server then sends the command to the Arduino which then activates the corresponding rellay.
Great channel.
When you say PHP is heavy weight for this, what do you mean exactly? Server running it consumes lots of power compared to MQTT? What are the challenges met with PHP which are resolved with MQTT?
Would also like your opinion on technologies like zwave for this type of setup. Thx.
+vic ky By "heavyweight" I just meant that it uses a bit more disk space and CPU time than some other approaches. On a desktop class machine it's fine, but it can be slow on low power machines. Using MQTT it's very easy for small devices (such as microcontrollers with maybe a few kB of RAM and FLASH) to connect and communicate
Hi there.. I know this video is already old.. but I find it very informative to us beginners out there... can you please give us some more information for those led lighted switches... part numbers or links? Thank you cheers!
Could you make or send me instructions and stuff to get this running? I plan to do almost the same thing as you do.
would it be possible to do the same switch idea but instead of one on and one off botton just use the same button for the 2 actions and use a if statement inregaurds to the led for when powered on to change to power off command?
First of all: I love your YT series of SuperHouse
Do you think it is possible to use a Ethernet of Power way of communicating with arduino boards? That way you don't have to wire miles of ethernet cable, because all the communication goes through the power lines.
It would be nice to have 110 volts (or in my case 220 volts) running to the lights and switches. The arduino boards will than talk to the main switchboards through the existing power lines.
I haven't found a Ethernet-over-power shields or diagrams though...
+Edwin van den Akker or probably use Zigbee instead of Ethernet of Power, if it's a small home.
Nice work but i don't see any redundancy in the plan while everything is centralized.
Don't you feel unsafe, when your central linux machine goes out?
That's a very good point. I still have that problem, but my system has changed a bit since that video was uploaded. I'm doing an updated system architecture video now, it'll probably be the next one uploaded.
Absolutely stinging for the updated Architecture, 3/4 the way through a 1st floor addition, wiring is in progress.
A very complicated way to turn a light on. Publish post subscribe etc The logic is simple if the switch is on then turn on light, So run a cable to the switch test if the switch is open or closed then open or close the relay to turn on the light.
I'm currently working on a project where I can control my home devices using a kinect sensor and I'm really new to home automation, what devices do you recommend I use?
Instead of Apache, could you use Nginx? I know they both do the web serving aspect, I'm just wondering if there is a specific reason for it.
How is it different? It's either at one end of the switching wire or the other. Serious you want to argue about the semmantics of my posts and not the topic at hand?
I don't think you've really thought this project through. You have all the high voltage devices wired directly back to the central switchboard, the argument you made for this in a separate video was that you didn't like having lots of devices all over the place, but then you have an arduino control panel in each room...
You could have kept the high voltage cabling for the light switches and replaced the switches themselves with relays controlled by the arduino in each room.
Hi,
I am interested how you steer the relais, which are prob 24V by means of a 5V TTL signal of the arduino ?
I've never used KNX and all I know about it is what I've read on Wikipedia. Time for me to do some homework!
Did you get any of this inspected? Seriously, I am doing one of these systems and three sections of the US NFPA apply. I need a EE to stamp my plans and three separate inspectors and inspections. Nuts. What are you doing?
Hello. What about when your Linux server goes down ? How do you switch the lights on and off ? Isn't it better to control the relay directly from the light switch at once and from the other side from your Linux server and android tablets ?
+Cornel Ghisoiu If the Linux box goes down then you're right, things stop working. Luckily it hasn't failed in 5+ years of continuous use. I don't understand what you mean about controlling the relay directly from the light switch. If you do that, you may as well just have the switch and nothing else, because if the switch can control the relay to turn it on, other controls such as tablets couldn't turn it off anyway so you lose functionality
Could you set up a second server that is a clone of the first and set one to the primary and the other to the backup. The backup constantly checks that the primary is running and if it finds it isn't sends out an alert and takes over control until the primary comes back online?
As you said it's been very reliable, but all things good things must come to an end eventually? :)
Yes, that's certainly possible. Clustered controllers aren't common in home automation systems, but there are already tried and proven techniques for Linux server failover (I've done STONITH setups in the past) so that could easily be applied.
I'm only starting in computer and automation but why so much cables instead of running one serial bus (or 2 or 3 for different areas) around the house?
And why using a big Arduino board for each of the button and not a simple serial enabled button?
Good questions :-) Many home automation systems do use some form of serial bus (such as RS422 or similar) to allow devices to be daisy-chained. That saves cabling time but potentially makes each node more complex. Since I was dragging the cables myself and wasn't charging myself by the hour or by meter of cable, I opted for a star configuration so I'd have maximum flexibility with everything brought back to a central point. What do you mean about a simple serial enabled button? Do you have an example? Since this video was made I've replaced all the light switches with simple button panels that are managed by centralised Arduino based boards with many inputs, so I don't have Ethernet to every light switch anymore. The cable is still there, but it's used to sense the buttons directly. Each panel now has nothing on it but a wall plate with dumb momentary-action illuminated buttons in them. I'm going to do a video about the updated architecture soon.
+SuperHouseTV That makes sense if the cost of the cable is low for you.
I'm assuming the cables going to your wall switch sockets is shielded (6 wires eternet?) so you probably could have a pushbutton with a Arduino nano, (or check out the "digispark" board,) sending a coded signal to the control unit.
I'm myself digging through the heap of different serial protocols so couldn't tell which one is better but I've read that i2c and "can" can communicate on up to 100m of cable at a low baud rate.
+SuperHouseTV , could you give some details about the Arduino based switch input capturing board. a schematic perhaps or do you sell these at freetronics?
what happens if your network connection goes down?
That's a common (and very valid) concern. With many modern IoT devices they depend on an external cloud service to operate. I've specifically designed my system to be all locally hosted, so there's nothing external required for my house to operate. If my Internet connection goes down (which it does several times per day thanks to the crappy router provided by Telstra) the house keeps working fine. The only thing I can't do is control the house remotely.
awesome!
Hi again.5 years it's a long time. that means you should start backing up your system:), but now without joke could you not control same relay both from light switch and linux ? I mean that is single point of failure what you have done. it is great that is working, just it's a little risky. thanks again for reply.
+Cornel Ghisoiu Yes, it's backed up, and I maintain the configuration files in Git so I can recreate the machine any time I need to. Although the cooling fan is getting noisy, so I have a Raspberry Pi here on my desk as a replacement so I will retire the old machine soon. About controlling from both switch and Linux, I don't understand how. Which one would win? If the switch turns on the relay, and the automation controller wants to turn it off, should it override the switch? But then if the controller wants it off, and you turn on the switch, then it shouldn't override the switch anymore? It's a logic problem
You could have used Zigbee.
Steven Luzuriaga no way. 6lowpan over 802.15.4 if impossible to use cables. tcp/ip over ethernet otherwise.
+Steven Luzuriaga Isn't Wifi good enough?
what butttons you use ?
I've asked this question in another video of yours, but I would love to know whom to contact to have my desgins put on to a board. What software can I use on Ubuntu to design these boards?
For schematic capture and PCB layout there are 2 leading contenders in the cheap / hobby space: KiCAD and EAGLE. Personally I use EAGLE simply because my early work was based on the Arduino reference designs, which are published as EAGLE files. However, that's a proprietary package with a free-for-hobby-use limited version and paid pro versions. KiCAD is open source and very popular for open hardware designs, and we use it at Freetronics for many of our projects. There are excellent tutorials for both packages, but PCB design has a pretty steep learning curve so you'll need to have some patience when you get started.
Thanks, I am learning a lot, and sooner or later, I might have to miniaturize a prototype.
This guy was using KNX -like connections and architecture waaay before people knew about this standard
You should put XMPP IoT in the server to handle anything in and out of the server, so you can do higher level federation, the propper way. It has all the Authentification, Authorization etc already done and tested. You can have pub-push on many nodes at once. And you can have mash, and not only tree.
MQTT are good as long as you don't go out of your LAN, and are lower level.
Search IoT XMPP for more information.
archive.fosdem.org/2015/schedule/event/deviot12/
www.theinternetofthings.eu/joachim-lindborg-xmpp-internet-things
github.com/joachimlindborg/XMPP-IoT
Good idea, thanks
If the Ethernet off thin all house will off .What your solution???
+AW Yes, if the Ethernet switch fails then light switches etc will stop working. Ethernet is very reliable though. It hasn't been a problem even after many years
+SuperHouseTV when will you make more videos ?
+william roberts I want to start making more very soon, but I've been burnt out and too tired. It takes days of work just to make one video :-(
ok looking forward to seeing your next video
I'm interested to work in this field. I'm planning a carrier change. Can you kindly direct me on how to start my career? Thanks!
Use google, or any search engine,really..
I have to know what your power bill is every month. All of those arduinos and tablets running non-stop have to add up, no?
My power bill is higher than a normal household, but that's more because I usually work from home, not because of the home automation. The 6 monitors on my primary computer draw much more power than the home automation system, and I spend about 10 hours / day using those! I have a half-filmed episode about power monitoring, so soon I'll be able to answer your question in detail with specific measurements.
I work from home as well and my biggest spenders are definitely my workstations. However, I have run some rough numbers and a lot of my always on devices like a handful of poorly finished raspberry pi and arduino projects running through a Kill A Watt meter show that they would cost me between 5 and 10 dollars (each) a year to run. If you have one of those at every light switch that has to add up. Or are they in a low power state until a hardware interrupt (button press) wakes them up? I suppose they could all be asleep since they don't actually do the switching in your system.
Yes, it's surprising how a small usage can add up over a whole year! I've reduced the number of devices in my system dramatically by restructuring all the light switches. They're now just dumb buttons that use the existing Cat-5 cable to link back to 2 central Arduinos (one in each automation switchboard) that handle all the switches in the entire house between them. I need to do an updated video about the new system architecture, it's much more efficient than the way I used to do it.
Ah, so you've changed it. Less capability at the switch for sure but much lower TCO. That makes a lot of sense. Thank you the response!
I'm curious about which kind of communication used to read the dumb switch signal. Did you have to boost the signal?
kudos ! to ur video... plz try to upload the video for coding part too once again tysm
SuperHouseTV
Thank god for Z Wave, it's way less complicated and wireless.
Hopefully you never need/want to sell your house, I'm sure the custom built light switches, exotic wiring scheme and requirement of having a Linux server for the lights to work would be a turn off to 99% of buyers... Though if you found the right buyer you could get a premium....
This is very cool, but you better plan on living there your entire life because you're never going to be able to sell it. Nice video though.
+Harry M What have I been smoking? How many people are going to know how to fix this if there is a problem? There may be some but you have excluded 99.9% of the market. If there is a repair issue that needs to be done no electrician in the world is going to touch this. Let's say the main board fails not to mention that it is all run by a computer. Who's going to fix it?
+notta3d This is a very important point. It's something I should talk about in a future video. Home automation often *lowers* the resale value of a home instead of raising it, for this very reason. It's seen as a risk by potential buyers. DIY systems even more so, because you can't just find an electrician who is certified for that particular system. In my case, I made a judgement call that if (when) we sell the house, I can replace the contents of the switchboard with something like C-Bus modules very easily because the wiring is already set up for it. That way a buyer won't have to worry about something that's unsupportable.
Every car runs by a computer.....
lol it would suck to get lag.
"Ubuntu is not an English word. It originates with the Bantu languages in Africa."
Didn't I just say that, can't you read?
No, I'm saying that if you write in English, its English. If i was writing in Spanish,
French, German or even Youraba I'd use the accents, to inflect how it should be pronounced, I'm afraid that just how languages work.
I'm fully aware of what the Ubuntu community claims. However, that doesn't make it correct.
Awesome, no network -> no lights :)
Not really, you don't need network to turn the lights on. They are connected through cables, when he presses a button, that button sends a signal to the control box (through cables) and that box send a signal to Arduino (through cables) asking to turn on or off relays. No network is needed whatsoever.
He only would need network if he wanted to do it remotely, like turn on his house's lights while he is at work, so he would send a signal to his control box via internet, and the control box would send a signal to Arduino via cables.
His system is not optimal though, it was 8 years ago, it's waaay easier now in 2020 with a lot of information available and new sensors/devices.
man this musta cost a fair bit of cash..
looks expensive...
cercl
Your wasting money on those arduinos for light switches. Use 20 quid binary input module for your light switch and a controlling a relay in the panel. If you want to integrate linux, use bacnet4linux. en.wikipedia.org/wiki/BACnet
Nice project... But not very practical in the grand scheme....
Technology has skipped over this method. Too expensive. The cable is $$$$. Too time intensive.
Well, I think this whole MQTT nonsense is a poor design. You were closer to the better answer before if it was decentralized. I was a software architect that is now working at a different firm as a software analyst. It's a very bad concept. But am interested in some approaches but I don't want the computer to control the entire house. Outside things; ok.
flip me talk about over engineering. no fecking need for the half of that poo
Bad architecture, you need a low cost rs485 or canbus instead of ethernet. I am in home automation since 1997, ethernet is ok for touchscreens.Our installations usually have over 40 pushbutton locations, you will need an expensive ethernet infrastructure, and these besides been expensive also consume lots of energy ($$....)
Johan Sagaert everybody already has ethernet infrastructure, so it is cheaper than rs485. and it is already connected to internet(of things)
Johan Sagaert You're right about the power requirements and cost of Ethernet infrastructure (switch ports, etc) but it was a simple way for me to get started with something basic and increment it in small steps. Since this video came out (which was more than 3 years ago) I've changed all my light switches over to simple dumb buttons connected via the existing Ethernet cable back to Arduino-based input boxes that read dozens of buttons each. That's been working great for about the last year. I need to do another video about the updated architecture.
+SuperHouseTV An update would be great Jon.
+jok yhg It's high on my to-do list, but I've been suffering a bit of burnout (not from SuperHouse, just from work and life in general) over the last year so it's been hard to get onto some of the projects that I really want to do. Too much work, not enough hours.
+Johan Sagaert
Not a wonderful comment, not helpful. If you are a professional user/installer with a lot of resources, then maybe use something other than Ethernet, but for individual enthusiasts, Ethernet is well supported. Personally I use CAT5e everywhere for LAN and 1-Wire systems for simple interfacing with PC's etc. A few network switches are hardly going to break the bank.
you-bun-too not ew-bun-to :p
This is the most stupid thing I've ever seen. Run all new cables back to the switchboard?? Or multiple switchboards?! What a waste of time, effort and equipment. Just run 240 to the controller for the room, then run the cables of each light to the controller. Done.
I'm sorry, but U in English can never be pronounced as "oo". Now, if you want to pronounce it as an African word, then it needs accents.
:-/
wastefull