RGBW controller 0-10V input, strange measurement

[jorritf 1]

Recently I've been trying to get a 0-10V brightness sensor working with the RGBW Controller (v2.5).

 

However, when connecting the sensor to GND and IN1, the RGBW controller doesn't report anything lower than 1,7V. The brightness sensor itself appears to be fine, and happily reports values in the entire 0-10V range, according to my volt meter.

 

Also, when nothing is connected to the controller, my volt meter reports 10,6V when measuring between GND and IN1, IN2, IN3 and IN4. That doesn't seem right?

This is the case in every mode I put the controller in, so also in the "ANALOG 0-10V" mode (config option 14 set to 0x8888)

 

What can I do to get a correct 0-10V reading from the controller?

 

Power supply: 24V din rail 0,8A

RGBW controller: FGRGWBM-441 v2.5

Sensor: Loxone 0-10V brighness sensor 0-100.000 lux (needs 24V DC)

 

Please login or register to see this attachment.

 

Please login or register to see this attachment.

[petergebruers 1,268]

I am not at home to give you an exact value, but I am pretty sure there is a pull up in the module, I tested it. That's why you measure 10,6 volt on the open input. The sensor should drive the output voltage, aka the output should be 'active'. Is it possible to link to the data sheet so I can have a look? Do you have a silicon diode to connect between in and gnd, to check if the module reports around 0,6 volt?

[jorritf 1]

On an open IN1 input the controller reports "99".

On a diode between IN1 and GND it reports "5", so that's close enough to 0,6V. Volt meter reads 0,56V.

 

According to the Loxone brightness sensor spec:

"The sensor has a built in 0-10V transducer. You can therefore connect it to all devices which use the standard 0-10V signal."

Spec sheet:

Please login or register to see this link.

 

I have it connected to the same 24V power supply as the RGBW controller. When not connected to the RGBW controller, volt meter readings on the sensors output are as low as 0,04V in relative darkness and around 7.9V if I shine my phones flashlight on it.

[petergebruers 1,268]

The "diode" plus "open" test give the expected results. Good. The manual isn't very specific, I had hoped to find out source and sink capability. I'm still not at home, so I am unsure if you would get better usable range by adding a pull down resistor... you could try 1 k just to see where that gets you.

[petergebruers 1,268]

I've checked a few things and I have to say... it's odd.

First... the input of a RGBW module looks like a 18k resistor to 10.5 Volt (take that with 10% accuracy). Otherwise put: if you want the IN pins at zero, you'll have to sink 0.6 mA. So it really depends on the capabilities of the output of the sensor (current and resistance) how low you can go.

You mention that 1.7 Volt is the lowest you can get on the input. That seems a lot indeed! If it were due to output resistance, that would be 3.5 kOhm. That's... high!

So I assumed that maybe this module needs more than 18 kOhm. But that doesn't seem to be the case. I checked the specs of a Loxone controller and I find: "4 analogue inputs 0 - 10VDC, 10 bit resolution. Input resistance: 10kΩ.". The documentation doesn't specify what reference level that resistance connects to. Anyway, the input resistance is comparable to that of the RGBW.

The only ideas I have right now:

- Maybe it can only source current (well), and can't sink current (enough). That would make the device incompatible unless a you connect a buffer between sensor and RGBW (theoretical possibility, no ready made solution).

- Maybe the device has separate ground connections, and they are not at the same level? Not likely, but I don't have a layout of the connections of this sensor. Would be nice if it were as simple as that.

- Maybe it is a bad device. But... If you connect it to a multimeter (only), you say it produces reasonable values. So not a very likely explanation either.

I'm stuck, and I *might* buy one of these devices to find out. I did a quick check and it's rather pricey, listed at slightly less than 100 EUR. I know its an outdoor device, and has a wide range. But I've been thinking that I want a device with logarithmic response.... Do you know where I can buy it a little bit cheaper?

[jorritf 1]

Some measurements with extra resistors:

 

HIGH lux scenario:

• sensor alone: 10,1V
• at RGBW without resitor: 10,1V
• at RGBW plus 3kOhm: 10,1V
• at RGBW plus 2kOhm: 8,6V
• at RGBW plus 1kOhm: 5,6V

LOW lux scenario:

• sensor alone: 0,04V
• at RGBW without resistor: 1,7V
• at RGBW plus 3kOhm: 1,0V
• at RGBW plus 2kOhm: 0,8V
• at RGBW plus 1kOhm: 0,5V

It seems no resistor option "restores" the entire 0-10V range.

I'll have to check how these new measurements relate to actual lux readings, because I'm not sure if clipping or range compression occurs in the scenarios above.

 

To date I haven't found a cheaper way to source a sensor like this. It seems the only option available in a ready made package. I've seen people Arduino their way to a lux sensor, so imagine my surprise to see a small atmel microcontroller on this sensor.

 

The sensor and RGBW controller share the same GND.

 

[petergebruers 1,268]

I was indeed afraid that none of the resistor options would really restore the full range. Maybe 2k2 or 3k3 can get you a more useful range... but that depends on what range you want to measure and what range you set the device to. But you knew that already, of course, I'm repeating this for myself!

Please login or register to see this image.

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

From your experiments I deduce the sensor can source current, but isn't good at sinking. To solve that, you'd have to add a buffer. I don't have a ready made solution in mind, but I am thinking "CMOS opamp with output capability to ground".

In the meantime... I have tried to build a sensor myself... with only one component: a LDR (CdS cell). I had a few of these LDR and I found one 7,5 mm diameter LDR that gives me a useful range from 1 to 10.000 lux. This device is logarithmic, so it has fine resolution at the low and and course at the high end. The difficulty is finding a good housing + diffuser for it, though some suggest coating it with some silicone rubber is all it takes.

If you want to now more about this, or yous want ta talk about solving this incompatibility, send me a private message.

[petergebruers 1,268]

Some time ago, I built a little circuit to adapt the output of the Loxone sensor to the RGBW. With this contraption, the input of the RGBW can get as low as the sensor. That's around 0,06 V. That will give you a much improved range. This circuit should work for other sensors too.

Please login or register to see this attachment.

[emielstroeve 22]

any change this problem is solved in HC2/RGBW? I've set up an RGBW switch to input-lux. It gives me the option to set a voltage level for 0 Lux and the same for the high end lux measurements. 

 

I've orderd a Loxone 0-100.000lux/0-10V sensor, I'll see what happens.

 

cheers

emiel

[petergebruers 1,268]

I don't think the HC2 has anything to do with it, it's purely a matter of the input characteristics (versus the output capabilities of the sensor). I predict the dark level reported will be too high. Unless there was a hardware change...

[emielstroeve 22]

I don't think the HC2 has anything to do with it, it's purely a matter of the input characteristics (versus the output capabilities of the sensor). I predict the dark level reported will be too high. Unless there was a hardware change...

but I can set the input levels in HC2.... let's see what happens.

 

cheers

emiel

[petergebruers 1,268]

but I can set the input levels in HC2.... let's see what happens.

 

cheers

emiel

Yes, but the sensor will never report below 1,7 V, so you all readings below that voltage will be the same (So that's 17% of full scale, for example 3400 Lux on the 20.000 Lux scale). If you're interested in daylight, that probably won't bother you. But if you want to detect dusk/dawn/dark, then I'm afraid you'll need a hardware solution (see my post above) or select the 100 or 500 Lux range.

A word of advice if you'r using this outdoors : set DIP switch #1 to 1 and #2 to 0, to obtain a 60 sec smoothing filter. Otherwise, you're RGBW will report changes like crazy.

[emielstroeve 22]

but I can set the RGBW to read 1.7V als 0lux. 

 

thanks for the dipswitch settings! will do that.

 

cheers

emiel

[petergebruers 1,268]

but I can set the RGBW to read 1.7V als 0lx. 

 

thanks for the dipswitch settings! will do that.

 

cheers

emiel

You can do that, but I doubt that setting the RGBW like that will make you happy...

Let me give an example, then please tell me if this is what you want to do.

Loxone set to 1000 lx full scale, so 1 Volt per 100 lx.

You set 1,7 V to be 0 lx and 10 V to be 1000 lx. To get from 1,7 V to 10 V the software interpolates, so it's uses 120 lx per volt for its display.

Mathematically said: lx displayed = (volts measured - 1.7) * 120

Now get some real measurements:

Bright light outdoors = 10000 lx -> Loxon outputs 10 V, RGBW measures 10 V, so RGBW displays 996 lx. No surprise here...

Inside, good light = 500 lx -> Loxon outputs 5 V, RGBW measures 5 V, so RGBW displays 396 lx. Maybe some surprise here...

Dim the lights, so we get 170 lx -> Loxon outputs 1,7 V, RGBW measures 1,7 V, so RGBW displays 0 lx. That's what you've configured it to do...

Dim the lights, so we get 50 lx -> Loxon outputs 0,5 V, RGBW measures 1,7 V, so RGBW still displays 0 lx.

Complete darkness, 0 lx, as expected.

So actually, the midrange will be inaccurate, and anything below 170 lx reports as 0.

If you're playing with that "1,7 offset" idea, then I'd rather set 1,7 V -> 170 lx (instead of 0 lx).

To get from 1,7 V to 10 V the software interpolates, so it's uses 100 lx per volt for its display (instead of 120 in the previous example).

Mathematically said: lx displayed = (volts measured - 1.7) * 100

Now get some real measurements:

Bright light outdoors = 10000 lx -> Loxon outputs 10 V, RGBW measures 10 V, so RGBW displays 1000 lx. No surprise here...

Inside, good light = 500 lx -> Loxon outputs 5 V, RGBW measures 5 V, so RGBW displays 500 lx. Good!

Dim the lights, so we get 170 lx -> Loxon outputs 1,7 V, RGBW measures 1,7 V, so RGBW displays 170 lx. Still good!

Dim the lights, so we get 50 lx -> Loxon outputs 0,5 V, RGBW measures 1,7 V, so RGBW displays 170 lx. Uh oh! Not good.

Complete darkness, still displays 170 lx.

I think the second example demonstrates best what happens: you can't use the lower 17% of your measurement range. No matter what you try, the information never enters the RGBW, so no math can retrieve it. But I think the second setup is more accurate. With a hardware buffer, you can get the lower limit to be 0,1 or 0,2 instead of 1,7. Edit: setup 2 is identical to setting nothing.

[emielstroeve 22]

you're right, it doesn't work.... 

 

Conrad order is on its way

Please login or register to see this image.

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

 

ajuus

emiel

[emielstroeve 22]

Some time ago, I built a little circuit to adapt the output of the Loxone sensor to the RGBW. With this contraption, the input of the RGBW can get as low as the sensor. That's around 0,06 V. That will give you a much improved range. This circuit should work for other sensors too.

Please login or register to see this image.

/public/style_images/master/attachicon.gif" alt="attachicon.gif" />

Please login or register to see this link.

working like a charme!!

 

cheers

emiel

[petergebruers 1,268]

working like a charme!!

 

cheers

emiel

Thank you for your feedback! Have fun!

[henzel 0]

On 27-9-2015 at 10:17 PM, petergebruers said:

Some time ago, I built a little circuit to adapt the output of the Loxone sensor to the RGBW. With this contraption, the input of the RGBW can get as low as the sensor. That's around 0,06 V. That will give you a much improved range. This circuit should work for other sensors too.

Please login or register to see this attachment.

Dear Peter,

Can you help me with the correct items for the Conrad order?
On the basis of your small drawing? Customer service at Conrad could not help me.
Thank you very much.

 

[petergebruers 1,268]

16 hours ago, henzel said:

Dear Peter,

Can you help me with the correct items for the Conrad order?
On the basis of your small drawing? Customer service at Conrad could not help me.
Thank you very much.

 

 

The device is not a (single) module, but an electronics d.i.y. project... Do you have any experience with electronics? I am assuming no so, because finding the "CA3140" is as simple as typing "CA3140" in the Conrad "search" box", I can't make it any easier than that. But maybe that's not what you are asking?

[henzel 0]

12 hours ago, petergebruers said:

 

The device is not a (single) module, but an electronics d.i.y. project... Do you have any experience with electronics? I am assuming no so, because finding the "CA3140" is as simple as typing "CA3140" in the Conrad "search" box", I can't make it any easier than that. But maybe that's not what you are asking?

Hi Peter, I have little experience with electronics. But I do have an electrical engineering background. I do understand that the board is made up of different components. Conrad could not help me with the CA3140 (EZ). the other components is not a problem. But I understand that I need the CA3140.

1 x CA3140
2 x 1N4148
2 x 100nF
2 x 10Kohm
1 x 1M ohm

Connecting components is not a problem but the right components I can use some help.
Thanks for your comment and help.

[petergebruers 1,268]

Well... When I type "CA3140" in the search box on conrad.be or conrad.de, I get one result. This is the URL of the search

 

Please login or register to see this link.

 

Don't you get a result?

 

You can also buy it from Reichelt:

 

Please login or register to see this link.