HCL - what is the strategic concept

[ivan.ruban 3]

Great work! That's what I also try to achieve!

However, the HCL won't do it for me as it is impossible to do proper polling with an external script, so if you want advanced motion detection or anything more advanced than simple on/off/fast dim, you'll be out of luck. I'd spend the extra cash and get a HC2 if I had to buy again

Could you please give some examples. What kind of polling do you need? Cannot you use the event triggers in HCL instead of polling of HCL?

I have ditched the dirt cheap tablets (i bought some from china called Ampe at 450 SEK a piece), fortunately i had only paid for three of them) as they really didn't work that well and two of them burned out in less than a month. Now i've bought some Samsung Galaxy Tab Light (from Sweden actually ) they cost a bit more (1.000 sek pr piece roughly) but the experience is so much better anyway and they work much better. My advice is to spend a little extra on the tablets and buy a brand product like the samsung one.

Thanks for advice. I test this one:

Please login or register to see this link.

400 sek for 1.2 ghz 1gb RAM is not bad. And the device work fine. Nevertheless you never know when the device will die if you use it 24h/7d. Thanks for advice here.

Regarding mounting, good idea with sticky foam tape, I'll test it.

We are moving to our new house in a few months and then i'll hide the cables a bit better.

In case of new house you can of course plan everything in advance and hide everything in walls. In the case of existing house, I see you use "cable channels" to hide the power cable. That's what I thought to do as well.

Once again, great work!

Please login or register to see this image.

/emoticons/default_icon_wink.gif" alt=";-)" />

[Dalle1985 16]

Polling: I want to know when a motion detector triggers and what the room light level is. If i should do this with the HCL i would need to interogate the json every 100 ms which might make the hcl unstable. A work around is that i make the motion detector trigger something else for i.e. 2 sec and then poll for this trigger every second. Finally, I can check for last breach, but no matter how i do it, I'll induce a lack and an extra error source if something fails to trigger. Therefore, i prefer to have everything inside the box. But making the script for interogation and triggering won't be that hard, making it work smoothly is another issue. In my experience, if you want a single trigger to activate multiple outputs you need a delay in the php to make sure that fibaro catch all the http commands (i use 0,2 sec in the single script which does this, my heating scheduler).

Tablet: my experience is that the cheap tablets can't stand up to the continuous use.

House: well "new" is an overstatement, it is from 1952, but we need to paint, so i'll cut some grooves before we put up felt and paint so they'll be completely hidden.

[ivan.ruban 3]

OK. Now I got it. In case of motion sensor, yes – it is really important to know that fast. So yes, it will create a much load on a small ARM device in case of polling every 100ms.

In my case I plan to use REST API of OpenHAB application, so no polling is required for events. As soon as motion is triggered, HCL scene calls http request to OpenHAB which makes any complex script logic and does necessary callbacks to HCL.

But polling is of course needed for displaying the status. I can nevertheless poll every 1-2 seconds which hopefully won't affect HCL much. All critical events which requires faster reaction are done via event trigger in scene as I described just before.

As an alternative to polling in my case is to use OpenHab persistence DB which can store all "last values". And OpenHab will get the new statuses in HCL only in case when it's notified by a "SomeOfDevicesChanged" scene in HCL. Trigger in SomeOfDevicesChanged scene is "IF Dev1==ON or Dev1==OFF or DEV2==ON or ... press virt_dev-sendHttpNeedStatusRead"

As longer I think about big ugly scenes with graphical blocks as more I understand how good is to have LUA actually... But HCL has half of the price!!!

Please login or register to see this image.

/emoticons/default_icon_razz.gif" alt=":-P" />

[Dalle1985 16]

As longer I think about big ugly scenes with graphical blocks as more I understand how good is to have LUA actually... But HCL has half of the price!!!

Please login or register to see this image.

/emoticons/default_icon_razz.gif" alt=":-P" />

Well this really brings me back to my earlier point about the price: I'm already €4000+ into my project, and I still need to spend something like €2-3000 before I have all the stuff I need for our next house (which is 360 m2). Fortunately, I once got an electrical engineers certification during my education so I'm allowed to install everything myself, but had I needed an electrician for all the inwall mods, the price for that would probably be an extra €3000.

So in total, my project will run into €7000+, so I don't see why I should settle for an inferior product to save €300. If you are doing a small simple install in a holiday house or something similar I could understand, but not for a big full integration project.

[ivan.ruban 3]

Yes, I fully agree with you. ZWave = expensive, not as expensive as KNX, but it costs a lot anyway. For simple solutions it is cheaper and easier to think about open source + RfxTrx.

Another option for small projects is Vera Lite - UI sucks but otherwise it is a very good product. As soon as Miscaverde-guys make a good UI and apply a better marketing strategy - they will become a serious competitor to Fibaro, I believe.

Currently I think about quite a small project. And actually HCL meets my requirements in this particular case. But you never know what kind of devices and scenes you are going to think about later in the future. And then one day - you can face with limitations of HCL.

[CopyCatz 4]

Regardless of what the strategic concept is I would like to see a patch soon that fixes the basic issues with the HCL, most notably the broken sunset/sunrise functionality and the crashing. Both issues are not present on the HC, I would assume the HCL is targeted at people with a simpler HA setup, and I'd imagine this calls for an idiot proof working device.. I cannot use sunset scenes at the moment so my HCL has little advantage over a a timer switch for my lamp I can but at a DIY store...

[Cosworth32 0]

I have to admit I'm starting to warm up towards the HCL. My setup now includes a PHP server and that is were I put the intelligence in the landscape. So HCL has 2 primary functions for me:

1. Interface (pretty at times, responsive and "swiping" the lights on the mobile is just plain fun)

2. Z-wave controller.

Of course some scene logic is added but no heavy lifting is done here. My Synology NAS takes care of that

Good integration is only possible with the http API and I really hope it will be developed further to enable interaction of external devices with the Fibaro GUI elements.

I don't even miss Lua and don't use the "trick" either to run it on HCL.

So for me the "strategic concept" is clear:

HCL is for programmers who "don't need no stinkin'" Lua to do the job

Please login or register to see this image.

/emoticons/default_icon_wink.gif" alt=";-)" />

Might not be what the marketing guys intended but it works for me.

[KristianO 15]

Looking good!

Do you know how to pass attributes like temp, lux etc using a VD?

[CopyCatz 4]

Cosworth how are your experiences with the webapi? I'm building a similar external handler into Domoticz, the other day I was passing some incorrect url parameters to the homecenter and it promptly crashed. Had to reboot it. It seems you have to make sure the data sent is perfectly ok because there is no (or not enough) error trapping at the Fibaro side...

[A.Socha 133]

CopyCatz, can you give me list that url's?

[Cosworth32 0]

Looking good!

Do you know how to pass attributes like temp, lux etc using a VD?

They are not passed in the original request (not possible with HCL), but the PHP server simply polls the relevant values depending on the command.

My "Vacation" heating script just receives one simple command and the discovers the zones that need vacation mode set and sets it accordingly (toggles on/off based on current status).

Only requires one virtual device button.

[ Added: 2014-04-29, 17:07 ]

Cosworth how are your experiences with the webapi? I'm building a similar external handler into Domoticz, the other day I was passing some incorrect url parameters to the homecenter and it promptly crashed. Had to reboot it. It seems you have to make sure the data sent is perfectly ok because there is no (or not enough) error trapping at the Fibaro side...

Pretty good for calls to HCL. I've thrown lots of nonsense calls at it without problems.

The other way is less reliable and virtual devices sometimes just stop doing anything. That's why the status feedback is important to determine if the HCL has gone belly up again. These incidents don't have a clear pattern though, so not sure what is causing it.

[KristianO 15]

ok, im looking for a way to use google chars for all my devices.

- Polling device values using API

Please login or register to see this link.

- Retrive values and update the result into mysql database

- pull mysql database into google charts

Got any good examples in this area?

[CopyCatz 4]

Well I just installed a led strip + rgbw controller on my HCL, only to discover that it isn't supported in HCL scenes. That pretty much makes the led strip useless. If v4 isn't coming soon I would suggest that Fibaro removes the lua restriction from the HCL because I'm afraid I'm going to have to sell this thing before it loses too much market value! The way that the HCL is crippled now I just have a lot of things I can't do with my hardware without lua; minimotes not working, led strip not controllable from scenes, greenwave power consumption not detected...

Either make it work without Lua or give the HCL users a limited amount of Lua scripts to set up. I would be happy if could run 5 devices with Lua to get things working...

[Cosworth32 0]

Just to conclude this thread:

My HCL is now for sale as it is being replaced with a better product. I'm pretty sure I'll never look back and regret this decision. The biggest nails in that coffin were:

- Instability. HCL would just stop responding approximately once a week.

- No roadmap. The secrecy concerning new release and general lack of patching of issues from Fibaro caused me to loose hope.

- Censoring in the forums.

- Poor integration of virtual devices into existing device categories.

Bye!

[Guest Kuuno]

Just to conclude this thread:

My HCL is now for sale as it is being replaced with a better product. I'm pretty sure I'll never look back and regret this decision. The biggest nails in that coffin were:

- Instability. HCL would just stop responding approximately once a week.

- No roadmap. The secrecy concerning new release and general lack of patching of issues from Fibaro caused me to loose hope.

- Censoring in the forums.

- Poor integration of virtual devices into existing device categories.

Bye!

Hi

i understand You a little but still my thoughts after doing this for over 2 years...

-instability- 90% of the cases there has been a misconfiguration on the users side and rest can be some technical problem in any case the instability is not normal behavior for HCL and can be fixed. Dont know who supported you in this but pretty sure it could have been fixed

-what is the roadmap for apple? this is not an excuse but again from good partners you can get understanding what will be done and even make a list of what you need and how to solve it...

censoring - read about it and probably not good unless it is bad for the company in what case i would do the same thing

-agree but that is the case w HCL, going for HC2 would fix it and in terms of whole project the cost is not probably that big of an issue...

there is much room for growth but still, we never had to take a controller back because something not working...

Please login or register to see this image.

/emoticons/default_smile.png" alt=":)" srcset="https://forum.fibaro.com/uploads/emoticons/[email protected] 2x" width="20" height="20" />

but good luck with new controller and share your experience!

[Dalle1985 16]

For what it is, the HCL and HC2 are both good products. But I think people are expecting too much from the boxes.

I have come to the point where I've started fiddling with my own controller. And i'm not meaning a stick and a protocol from somewhere. I've started playing with the protocol. I've now got a basic understanding of the inner workings of the zwave protocol and is building my own library - look at it as a zwave library for php. So you can call zwave functions directly into php instead of having to use some sort of frontend. As I understand it, I need a license to distribute it (or to get official access to the protocol) and I won't spend 3000$ on it so don't expect it to go public, but the point is that these complete systems all come with limitations set out by the manufacturers based on their ideas for their products. So if you want more, you are on your own.

When you buy a packaged product, you get limitations. If you don't have the skills to develop your own setup, there are two solutions: pay someone to do it, or accept the limitations. Don't expect more from the product than it can provide. The HCL does not have the hardware to support lua scenes and customizing. Therefore it is not an option. If you hack it and circumvene the barriers to aboid people making their systems unstable and then wonder why the manufacturer discourages it, then i think you need to sit down and have a good rethink about that. That is not censorship, that is protecting your product. Apple does the exact same thing. You as the hacker might know what you are doing (or at least believe you do) but when you tell every other average joe how to do it, you are basically aiding them in destroying their box.

If you want lua and you want a more advanced control, cough up the cash for the HC2. If you want more than that, then you are on your own... But don't blame the products, because that isn't what you bought. I've looked into different controllers, Vera, Zipato, sticks + software. But none grants me the level of access you want.

Fibaro isn't an open source company, so don't expect them to have public roadmaps, full access to the box or any of that stuff.

In my perfect world, you get a protocol with a public Api and then you are free to develop your own frontend and interface. I have the interface ready, I have a raspberry talking to a single device, but I haven't got the full protocol down yet. So right now I'm relying on the HC2 for the integration between zwave and my controllers and it is actually pretty solid. But I just want something with a bit faster response and scaleability to allow for as many components as I want. I'm already at 60+ devices and counting and have a fairly advanced setup. I still make due with the HC2 but I make no illusions that the limitations are fibaros fault. I'm just expecting more from the HC2 than it can deliver.

[Guest mhn]

Sounds like you are a Razberry user type.

Please login or register to see this image.

/emoticons/default_icon_smile.gif" alt=":-)" />

Please login or register to see this link.

[Dalle1985 16]

Actually, no. I am using a rasberry for my experiments, but i am using a stick, as I want to make sure that I can run it on any setup, not just an RPi. The razberry has a decidedly nerdish appeal to it. My intention is to build something that anyone with a bit of programming skills could use and that integrates seamlessly with a similar library for other protocols i.e. Rfxcom

And haven't the razberry daughterboard been discontinued anyway?

[Cosworth32 0]

Thanks for the comments. I don't want to go into a tug of war here. If some poor soul reads this and other similar threads they may come to a different conclusion concerning what to buy. It surely would have helped me not to make a mistake to begin with.

[Guest mhn]

It doesn't look discontinued to me.

Please login or register to see this image.

/emoticons/default_icon_smile.gif" alt=":-)" />

Please login or register to see this link.

Maybe you should look at Openzwave?

But we better stop, we are way off topic.

Please login or register to see this image.

/emoticons/default_icon_smile.gif" alt=":-)" />

Regards

Morten