Navitrino - An open source controller for Renewable Energy Applications

[KenB]

List,

In a quest for a common starting point or building block for renewable energy monitoring, control and automation systems, I propose that we design, build and market our own controller - the Navitrino.

(I've googled it, and the word does not match any documents - so you have practically a unique product name).

Based on the Arduino, open source platform, the Navitrino would be aimed at DIY home automation and renewable energy monitoring projects.  It would serve as a common platform and a good starting point to get a project up and running quickly.

Features above those provided by existing Arduino:

1. Removable bulk data storage provided by SD Card (either standard or microSD, up to 4GB)
2. On board dataflash memory 8Mbit for non volatile storage and used when no SD fitted
3. Temperature sensing channels - up to 8 channels for low cost thermistors or PT1000 with extra plug in shield (keeping costs down).
4. Dallas One wire interface Real Time Clock with back-up battery or supercap
5. RS485 interface - cheap way of communicating down long cables with other devices
6. Compatible with existing Arduino shields - so that Ethernet can be added with a low cost (£13) shield
7. 4 channels of 10A Mains relays for direct switching of ac loads (pumps, valves etc)
8.  8 channels of opto-isolated  inputs allowing pipestats, thermostats etc a direct connection.
9.  I/O and relay outputs brought to sensible size screw terminals (eg 5.08mm pitch, 10A 240V ac)
10. Communications to PC via removable FTDI USB to serial lead
11. Wide input voltage switching dc power supply - higher efficiency and run it from 6V to 60V making compatible with most battery systems
12. PCB size designed to fit in a standard low cost enclosure - such as a Honeywell (or other) wiring centre box.
13. Based on ATmega328 or ATmega1280 - if you want the extra capabilities of the ATmega1280 - buy one and plug it into the Navitrino board as an upgrade shield *
14. Easy I/O connections for standard LCD modules and Keypad
15. Fibre-optic interfaces for safe sensing of gas meter reflective digits or electricity meter



*  This would allow some users to buy the Navitrino and use it as a specialist shield for the Arduino Mega  - just by removing the ATmega328 from its socket on the Navitrino.


Any other suggestions?



Ken

[Flamethrower_]

Ken,

My electronics knowledge is zilch..! and when you start talking C-coding I stand there baffled thinking I am in the wrong room!

But I am a practical guy and a tinkerer if what I understand in laymans terms is that you wish to create a method of whole house energy consumption monitoring and recording system + extras, then I am all for something that can be made to be cost effective, plug and play and a useful tool.

You can count me in as an interested party.

Regards,

Rob

[guydewdney]

In a similar way - I would like modules I can plug-in-and-play. Not everyone is happy with scary-looking coding and wires everywhere.

[Ted]

Any other suggestions?

Having played around with some weather and data logging in the recent past my experience was that the Real Time Clock was a major weak-point - with time drift making it near impossible to leave a box alone for more than a few days before the sync to real time was too far off to be acceptable.

Not sure what the solution is (from either a practical or economic viewpoint - atomic, network, etc) but reliability of this would be crucial for me.

[KenB]

Rob, Guy,

It's certainly true, we need to produce a product that has sufficient user friendliness to be usable by the interested layman, yet provide the means for the more  adventurous to roll their own.  Getting this balance is going to be the hard part.

The Arduino has a proven track record and has been adopted as a development platform in the most unlikeliest of camps including artists, lighting and stage designers and other groups who need to embrace technology but until now have not found an easy entry point.

For the layman, there needs to be a layer of coding above C, a further absraction above the machine language into a far more human readable form which provides a simple scripting tool.  For example:   "If tempsensor1>40 then start pump2 until tempsensor1 <20"

If enough people want to do the same thing - i.e. program in human readable language, then the tools will appear to do this.



Ken

[KenB]

Ted,

Real Time Clock drift should be in the order of +/-1 second per day.   Good crystals should be around 10 parts per million.  In order to save money, they may have been replaced with cheaper  20,30 or even 50 parts per million  crystals - resultaing in 2,3 and 5 seconds drift per day. 

If it is more than this then the solution is to get a better crystal, or use a module that gives access to the Rugby or similar atomic clock, or synchronise to the web, and get the time off the net.

These problems are not insurmountable.  I bought a weather station from Lidl's that accesses the Rugby clock for about £6 - showing that the building blocks for such a system are fairly cheap.


Ken

[Pat_]

Ken,

Having recently installed my first EV solar hot water system, I've been thinking around how to approach a monitoring system. This suggestion of yours seems to cover most of the obvious requirements.

It seems that a 1-wire, plus a Pt1000 interface meets the sensor requirements I had considered.

Is the compiler required freely available? Or assembler?

I might well be interested in contributing. Does the FTDI approach allow a dedicated driver access or is it subject to the vaguries of simulating, and trying to find, a virtual serial port in the PC? All the USB interfaces I have developed have had dedicated drivers, and a unique product identifier, which makes PC interfacing a lot more straightforward and reliable.

[KenB]

Pat,

The FTDI cable is based around their FT232 chip and is supplied with dedicated drivers.  I have had no problems in just plugging in a FTDI cable and having the PC nummerate the USB port.

The C compiler is distributed free of charge and regularly updated.  This is true open-source computing. Have a look on the main Arduino site for details of the Arduino philosophy.

http://www.arduino.cc/

There is an Arduino users forum and no end of information and code examples available on the net.  The Arduino project has attracted folks from all sorts of backgrounds and provided a rich library of useful code routines.

At it's simplest an Arduino is just a ATmega328 microcontroller with a clock oscillator and power supply.  Running an on chip bootloader which allows it to communicate with the Arduino IDE to receive code updates and send serial data back to the host.  Whilst the intellectual step to achieve this could have been done with any number of different microcontrollers, Arduino seems to have cornered the market and is rapidly achieving critical mass.




Ken

[Pat_]

The FTDI cable is based around their FT232 chip and is supplied with dedicated drivers.  I have had no problems in just plugging in a FTDI cable and having the PC nummerate the USB port.

Ken,

I do have a particular interest in a USB to a PC (as well as an ethernet connection), so forgive my questions. Presumably the required upgradability from Arduino rules out using a USB capable AVR microcontroller?

I believe you that enumeration of an FT232 is not a problem, but does this not involve the assignment of a random virtual serial port which:
a) your application then has to find, by user input
b) may well be different the next time you enumerate the device

If this is still the case then it is a very unsatisfactory interface compared to having a fixed known product identifier. That was my point.

[KenB]

Pat,

There are a lot of things unsatisfactory about USB - however, I am a user and not a programmer, so there is little that I can do about it.  The computer industry has a long history of leaving things half-done, and seems to think that a half baked product is acceptable leaving the user to sort himself out.  I have often heard that if Microsoft made cars, they would take 5 minutes to start in the morning, would go out of control and crash for no reason, and should they break down at the bottom of a hill, the best way to diagnose the fault would be to push them back up the hill and see if they do it again. We've all heard this before, but in the minds of the general user there is clearly something lacking.

Common platform, reliable open source tools are what we need.  The Arduino goes part way to achieving this goal.


Ken

[Pat_]

Ken,

Ah, I assumed when you mentioned C that you *were* a programmer. The reason I spoke up was because I would have been in a position to sort out a sensible USB interface which does work reliably, and could easily also be open source, unlike the FTDI. I have written the applications and firmware for around 5 USB products, plus written the drivers, both for 32-bit and 64-bit systems up to Windows 7. However, you didn't answer the question concerning USB versions of the AVR, I presume this is because they would not be compatible with this Arduino concept.

I guess I'll have to pursue my own solution, then. Or possibly make my own personal serial/usb add-on board for your board, using a PIC. Since PCs tend not to include legacy interfaces these days there isn't a lot of choice but to use USB, and I would need a hassle-free interface if I went to the trouble of writing applications.

Pat

[stephend]

Ken, I think this is a great idea.  The arduino is low power and easy to work with... the only sticking point for my skillset is the additional electronics and soldering needed to interface to sensors and relays.  so it would be smashing if these could be more plug and play.  Phidgets has already taken this type of approach with a more powerful platform and a host of sensors: http://www.phidgets.com/products.php?product_id=1070
Would be nice if the Navitrino could be compatible with phidget sensors (they're 0-5V, so it would just mean choosing the same type of connectors)

[KenB]

Stephen, List,

At $265 the Phidgets board is about ten times the cost of what I was proposing 

Attached is something a bit more reasonable which came from Canada for about $30.  Although this one used a 28 pin PIC, a similar design could be produced around the ATmega328 following most of the Arduino architecture.

The board in the photo has 4 x 10A mains rated relays with change-over contacts, 4 x opto-isolated inputs,  8 high current (darlington) transistor outputs for driving additional relays etc and a further 4 digital inputs. Most of the connections to the board are via chunky screw terminal blocks.

By replacing the serial connection with a low cost USB to serial port, it would make the board more usable. The power-supply would need to be redesigned to work from a wide range of battery voltages.

It is impossible to design a board which fulfils the requirements of every application - but it's worth having a go and getting most of the really useful stuff on there.

Ken



 

[wookey]

Ken, I agree there is real need for an energy monitoring system, and I've been pondering the issue for a while. Your interfaces list matches mine pretty closely (except I'd think zigbee and 868Mhz are important too). As you mentioned elsewhere I'd love to see you at the Cambridge energy-monitoring hack day where we will be having a serious go at all this stuff.  

However we have one point of disgreement: I am not convinced that the right way to do this is with arduino-alikes. It is one practical solution, but I favour a more capable system that can run embedded linux. I think in the long term this will provide much better connectivity and interfacing to external logging sites and of course it provides a huge software base. The hardware is slightly more expensive, but only a tiny bit and that will matter much less that the cost of the IO components. The hardest bit (particularly if you want to make it both expandable and consumer-acceptable) is actually mechanical design. How big is the base system - which things are standard, which are extras, how many extra 'shields' can you plug in, etc. I'm not sure that we'd actually manage to agree on this forum, but if we could it would be great.

A big part of making this work is actually working out protocols and interfaces to other devices (SMA, fronius, Victron etc inverters, boilers from loads of places, resol and TDC controllers, efergy, currentcost meters and so on). NaturalWatt have done some work on talking to all these devices, but are currently keeping the software proprietary. I think that's the wrong approach and the first person/company/group that achieves critical mass on talking to 'most' devices with a sensible design could be very successful. This group has a good chunk of the relevant expertise (and we're not the only people working on this stuff - we need to talk to them too).

Pat_. I don't understand why you think that FTDI interface modules are Free Software-incompatible. libftdi is free software and interfaces with the FT23xx series.

[Paulh_Boats]

Ken,

Your design idea is great....just like my Eco Controller I was working on a few years ago  . I put the project on the back burner because the Resol did all I needed and although I designed an I/O board and PCB and looked at all the costings but it did not seem a commercially viable way of spending my time. It also had a web interface that Ivan thought was a winning feature.

But open-source and teamwork could make things worthwhile. However there will be hundreds of hours of work in design, programming, testing and hardware construction. If we can get around 10 people things might work - for the 2 main disciplines, hardware and software, we'll need one or two professional engineers to steer things in the right direction. For anybody that can use a soldering iron there will plenty of less demanding practical work, likewise those that like wiring up sensors and tinkering plenty of testing to be done.

I have a degree in electronics, but C/C++software is my key skill these days. Also I'm not too shabby at doing technical presentations in front of a group of people, if it comes to that...convincing the chosen few will be easy, but convincing Government and Industry that the project might just save the planet is, in my opinion, a critical success factor.

cheers
Paul