Yes, see my last post, the silicon available would make excellent industrial computers with the adaptation you mention.

Regards
cww

 

I might just add that any of the devices that run Android can also be programmed down to the metal with the any of the GNU compilers. It wouldn't be at all difficult to use one board for both a real time executive for control and another of the same board for Android apps. The cost is such that there really wouldn't be a downside.

Regards
cww

 

> I might just add that any of the devices that run Android can also be programmed down to the metal with the
> any of the GNU compilers. It wouldn't be at all difficult to use one board for both a real time executive for control
> and another of the same board for Android apps.

That might be possible in theory, but what Android hardware are you talking about? I was thinking of regular consumer tablets. The problem with using them for the "down to the metal" stuff is that they don't have the right hardware. Initially they all came with a USB port, but it was a slave port, not a host port, so that the Android tablet could be a device to a host like a PC. Now finally we are getting some tablets with host USB ports. But the host support for the wide range of industrial USB devices is lacking. Those Android host ports mostly just support keyboards and serial ports. Device drivers have not been written for the vast majority of industrial I/O.

Robert Scott
Real-Time Specialties
Hopkins, MN

 

Hi Robert

Check these out
https://www.olimex.com/Products/OLinuXino/A13/A13-OLinuXino-WIFI/
https://www.olimex.com/Products/OLinuXino/iMX233/

There are many others building boards using the mobile class silicon to make inexpensive SBCs They address your issues with USB host ports and various power options. The driver thing is a chicken and egg problem as usual. Since everything in Linux is abstracted as a file, if you have knowledge of the protocol you can write support in simple cases.

Here's one more:
http://apc.io/order/

Regards
cww

 

> The goal of writing source code only once is rarely achievable in practice.

WRONG !!!

Please search for Applications that are based on Qt.
http://en.wikipedia.org/wiki/Category:Software_that_uses_Qt
For example "Google Earth".

I talk about http://pvbrowser.org in special.

 

I would think it more accurate to say that it requires care from the outset to achieve it and it is much more easily achieved if you start from the OSS side than from the proprietary side. If you make portability a requirement, it will be portable. If you code for one environment and expect it to be portable to another you may be less fortunate. It's fairly easy to paint yourself into a corner if you use libraries and add-ins without checking. The most frequent problem is feature X is not available across all platforms, but you really like feature X.

Regards
cww

 

Yes, it's happening.

Check out the "CuBox" that is 2"x2"x2" and runs on 3W. It's a powerful little device.

I like to think of WYSIWYG applications that could run on top of Java/Linux/Android embedded applications, particularly through a web browser. Once you get to the point of regular, commercial software/games, etc you could envision the device hosting it's own slick GUI and command line for programming. Heck, they could host the HMI portion, integrate with the enterprise, etc, etc. Pretty slick!

BTW - I sprung for a Nexus 7. It's sweet!

----
Nathan Boeger
http://notanotherindustrialblog.blogspot.com

 

Yeah, there's a lot of potential in this cutting edge silicon. I packs a lot into a single chip and the horsepower vs cost is exceptional. Even if it doesn't penetrate the automation market I think Android is a good thing to be involved with. The Nexus would be kind of an indulgence right now, but it's not a bad investment.

Regards
cww

 

For the Doubters:

http://www.zdnet.com/windows-has-fallen-behind-apple-ios-and-google-android-7000008699/

Can you say: Disruptive Technology?
I knew you could.

Regards
cww

 

And your point Curt?

What does this have to do with an office / corporate / plant floor environment. By far and away this has not changed inside of company business walls.

In the consumer world this is very very true, but on the plant floor Wintel is still 90% or better. Will it change (who knows) my guess is business corporate applications and office desktops and IT departments and big time Control Vendors have a long way to go to see your dream of a all tools are free and the Blob will keep you running, just ask for help and it will come freely running, in other words I feel safe in BETTING at least another 12 years just like the last 12 this rant has gone on......in other words long beyond MY working career.

Dave Ferguson
Control Systems Engineer
Still just a tool user who doesn't think the hammer matters, it is what you do with it.

PS: See where this was sent from while not in the office

Sent from my iPad

 

I forgot and if it makes you feel better Curt.....my entire control network infrastructure is built on VMWare Esxi which is a Linux kernel......then all of my virtual server infrastructure, mostly Server2008 R2 boxes (roughly 30) are run on top of this.

This IS becoming fairly common in industrial datacenters and control system.

Dave Ferguson
Control Systems Engineer

Sent from my iPad

 

Hi Dave

Nothing to see here, move along! Even if you choose to ignore it, change is afoot. Even Microsoft is reacting drastically to becoming a bit player in the future of computing. Their problem is that the masses are discovering that you can do everything you need to do without Windows. And even worse, given a choice, people are, en masse. And while they still are afraid of Linux, they are familiar with Android and that it's just as easy to use and the world didn't end when they made the switch and it's already at a price point that simply will not support Microsoft's business model or the expensive hardware to support Windows. And I don't doubt that some will think spending ten times what they need to, to use windows on an expensive, fragile, power hungry PC is worth it. Given the choice between that, however, and a low cost, low power, fanless, bulletproof device that does the same things will be a no-brainer. The only obstacle is the software, but suddenly there are a _lot_ of Android programmers around. Your support point might have held water, but also quite remarkably, there are 100 times more companies doing Android including most of the largest. And I don't think they are going to quit. Big automation may circle the wagons around their idol, but doing that too long will simply put them in the same position _as_ Microsoft, they will simply be bypassed on the road to progress. What they do is simply not all that hard to do anymore. And if there are more attractive choices............

All that remains is for someone to package the existing state of the art for industry, which is far easier than making an industrial PC or any traditional PC for that matter. And to port some software to Android. I'm betting someone will the opportunity in that . A little late for my career as well.

Regards
cww

 

Hi Curt,

i agree with you that such small mobile devices will be used in automation and that a sort of Linux is the favorite OS for these devices.

But mobile devices and thus Android are optimized for saving energy. They try to switch off as much as possible because of the battery. This is in contrast to what we need in automation. In automation the devices must operate all the time and even a HMI screen should be on all the time.

I think we will use the same embedded hardware but pemanently connect it to a power supply.

A week ago i ordered a
http://www.raspberrypi.org/
and will do experiments with our pvbrowser on that device.

 

Curt

You seam to think I am against this change. Not at all.......never have been. What I have said for 12 years is that plain and simply, I don't care what hammer I get, I make stuff with the tools. At the moment, the only tool good enough for me to safely, supported, and available time tested ability, on a day in and day out basis......are the ones supplied to me by my employer and happen to be from the big players and happen to be Windows based.

I live in an Apple world, dabble in Linux and agree things are moving. I agree Microsoft is attempting to change although it appears that Windows 8 and 12 are miserable failures and probably cost Steve Sinofski his job a few weeks ago. But until it happens and is testing big time on a full scale plant running 24/7/365 year in year out, it will not be adapted by real companies. Don't read that wrong, it will be dabbled with, but until full scale big time adaption happens, it won't.

Again I have learned to effect what I can effect and since I do not buy the tools I use, I care very little, if the tools required change, I will change but I do not foresee it being soon, nor do I care.

I will effect what I can and after all solving plant production, quality and money making are what I effect. Someone else decides my hammers.....I used to preach tools and after 25 years learned I can work with any and all of them, and they pay for me to learn them.....so who cares what they are, I am after bigger fish.....core, deeper issues and solving control problems....I have enough trouble keeping up with that.

Dave Ferguson
Real Live Control Systems Engineer

Sent from my iPad

 

> A week ago i ordered a http://www.raspberrypi.org/
> and will do experiments with our pvbrowser on that device.

The RPi is a nice inexpensive Linux system, but one thing still missing is an inexpensive NMEA touchscreen to use with it. If you (or anyone) comes across something please let me know!

I have a 3 MB zip (ARM Linux) on this page that can be used to compile GUI programs (directly on the RPi):

http://www.turbocontrol.com/easyfpgui.htm

 

> i agree with you that such small mobile devices will be used in automation and that a sort of Linux is the favorite OS for these devices.

Android is just an application framework for mobile devices on top of a stripped down Linux kernel.
I can't see how phone related applications should be very useful for automation solutions ... ?

The way how Android is scheduling threads may be OK for phone apps ...
but is it also OK for hard or soft real-time requirements in the automation industry?

> But mobile devices and thus Android are optimized for saving energy. They try to switch off as much as possible because of the battery.

Energy is saved mainly be the usage of ARM SoCs ... not by switching devices off, IMHO

> This is in contrast to what we need in automation.

I can see a remarkable trend for ARM based devices in the automation industry.

A real nice board for industrial applications is the "Beaglebone": http://beagleboard.org/bone ($70-80)

Its CPU (or SoC) offers a lot of interfaces neeeded for automation applications.

Best Regards
Armin Steinhoff

 

I've had a Raspberry Pi running for a couple months now. It's not too appealing as a development platform with it's limited RAM, but it would make a fine thin client. It was optimized for cost, but is still powerful enough for many uses and it uses very little power. The power issue is a trade off, from what I've read, most of the power management on the ARM chips can be switched off if you aren't concerned about saving the last milliwatt. Most designs seem to use a power controller of some sort to manage this. I've got a cross compiling setup for a normal PC but haven't had the time to mess with it much. I'm thinking that even optimized for continuous operation, you might be looking at 20 watts, or less if you turn the clock down. Two big things in the ARM world are that Linux 3.7 has unified the various ARM architectural differences and the 64 bit ARMs will be showing up soon.

Regards
cww

 

No, it's nothing personal Dave, I just call em as I see em. And it's a good thing to have several viewpoints. From my viewpoint , it's so glaringly obvious that other platforms are technically and esthetically far better suited for the work we do, that I am driven to speak out. And it's true that one person can have only limited effect, but if you can cause a community to see what you see, you can have a much greater effect. If you can cause an executive somewhere to see how much silliness and chaos endured for Microsoft's benefit would go away with a platform that can be tailored for automation use, and most importantly that at this moment, at least technical people would now consider an Android proposal seriously, you can have a larger effect. There will come the oft mentioned tipping point. sooner would be better, after which people will wonder "what were we thinking" to cling to MS so long. It shouldn't be that hard a thing to see, the mobile/other embedded people have dropped MS like a rock and are experiencing a meteoric rise in their fortunes and a return to vigorous competition and innovation. And the money goes into _their_ pockets rather than to Redmond. That last point is the coffin nail, it's going to be hard to justify the Microsoft tax as device prices drop. the momentum is making the alternatives better at a rate that no single company can match. That's how I see it. If big automation holds out, they will simply be replaced by the new big automation. Android is not the ultimate for our uses either, but it realistically can be, under the brand, it's Linux. And Google can kick open the door.

Regards
cww

 

I agree Armin, but Android is really just a brand. There is no reason an "Automation Edition" couldn't include all the best parts of Linux for automation. And that would be a formidable package with a name that gets people to jump on the bandwagon. And if they switch to plain ARM Linux, that's fine too:^). I haven't been following the ARM RT progress, but you can bet there is interest. What Google has done is make Linux cool. with it's little green robot. It can even make an old C programmer look at Java:^).

Regards
cww

 

> I agree Armin, but Android is really just a brand. There is no reason an "Automation Edition" couldn't include all the best parts of Linux for automation. And that would be a formidable package with a name that gets people to jump on the bandwagon. And if they switch to plain ARM Linux, that's fine too:^). I haven't been following the ARM RT progress, but you can bet there is interest.

Try to apply the PREEMPT_RT patch to the stripped down kernel of an "rooted" Android distro 4.x and install additionally Ubuntu or Debian on a "loop device".

VNC will help you to show the e.g. LXE graphics on a host PC. Works great also without the PREEMPT_RT patch ...

Regards
Armin Steinhoff