Go Down

Topic: 24V DMX (Read 703 times) previous topic - next topic

Freakazoid

Feb 29, 2012, 08:24 pm Last Edit: Feb 29, 2012, 08:29 pm by Freakazoid Reason: 1
Hi,
i' ve got some old LED fixtures, which accept DMX (Color Kinetics). My problem is, that they need it at 24V.
I tried to amplify the voltage with a n-channel mosfet. My setup worked well, but now when I try to recreate it, it doesn't.
In my first successfull setup there was just one fixture and now there are 11, but I dont think thats the problem.
I attached my schematics so maybe someone comes up with an idea that solves my problem or one that leads to a solution.

Thank you.

DVDdoug

I don't imagine you have an oscilloscope to look at the 24V DMX data signal... ;)  If you have a multimeter, you can write a little test program to slowly toggle the data line and check if it's going between (approx) +24V and zero.   

I'm not really familiar with the DMX specs/requirements, so I'm not sure about any of this...  But, it looks like you are supposed to have a balanced ("push pull') connection.   Is that right?    And, I think you are not getting enough current through the 10k resistor.   Do you have any idea about the current requirements or the input impedance of your DMX devices?  (RS-485?) The more devices you hook-up, the more of a problem that 10K resistor becomes.

A lower-value resistor might "work", but you probably need a better driver circuit to drive the signal high, rather than relying on a pull-up resistor.   Unbalanced drive might "work" too, but I don't think it's compliant to the spec.   You should be able to get an RS-485 driver chip.

Freakazoid

#2
Feb 29, 2012, 10:28 pm Last Edit: Mar 01, 2012, 12:26 pm by Freakazoid Reason: 1

I don't imagine you have an oscilloscope to look at the 24V DMX data signal... ;)  If you have a multimeter, you can write a little test program to slowly toggle the data line and check if it's going between (approx) +24V and zero.  

That's exactly what i did. It doesnt give me voltages of 24V an 0V but I just didn't know why until now.


I'm not really familiar with the DMX specs/requirements, so I'm not sure about any of this...  But, it looks like you are supposed to have a balanced ("push pull') connection.   Is that right?    And, I think you are not getting enough current through the 10k resistor.   Do you have any idea about the current requirements or the input impedance of your DMX devices?  (RS-485?) The more devices you hook-up, the more of a problem that 10K resistor becomes.

A lower-value resistor might "work", but you probably need a better driver circuit to drive the signal high, rather than relying on a pull-up resistor.   Unbalanced drive might "work" too, but I don't think it's compliant to the spec.   You should be able to get an RS-485 driver chip.

How do I know what kind of resistor I need. Would adding a potentiometer, so i can just try, be an idea?
What do you mean with a better driver circuit? Is an RS485 driver chip capable of going up to 24V?

The coves do use RS485 (http://www.colorkinetics.com/support/datasheets/iColorCoveLT.pdf).

Thank you so much for your answer.

DVDdoug

Quote
What do you mean with a better driver circuit? Is an RS485 driver chip capable of going up to 24V?
Hmmm...  I think the light operates off of 24V power, but according to this page, the RS-485 control signal does not go that high.   The receiver needs to see a minimum differential of 400mV (+/- 200mV), and has maximum limits of -7 to +12V.  The transmitter is supposed to put-out a minimum differential of 4V (+/- 2V).

I Googled "Arduino DMX" and I found this circuit, which runs off of 5V and ses a 75176 chip.

tjbaudio

You are misinterpreting the info.  Color Kinetics uses a 3 wire power/data distribution.
Red +24V DC
White data
Black Common

They are NOT DMX.  To use with DMX they require a controller/PSU unit.  They also use a proprietary data format.  Not sure if that format is just DMX ran as an unbalanced signal or if it does some data translation. I also have no idea what voltage the data line runs at, my gut says 24V on the data line could fry it.

Freakazoid


You are misinterpreting the info.  Color Kinetics uses a 3 wire power/data distribution.
Red +24V DC
White data
Black Common

They are NOT DMX.  To use with DMX they require a controller/PSU unit.  They also use a proprietary data format.  Not sure if that format is just DMX ran as an unbalanced signal or if it does some data translation. I also have no idea what voltage the data line runs at, my gut says 24V on the data line could fry it.


I try to built the controller myself. It definitely is DMX, just on 24 an 0 volts. As is said, it worked before with one fixture. I was sending DMX data to the single fixture.
And it worked for him, too: http://mit.edu/loganw/www/color_kinetics.html

jfenwick

#6
Mar 01, 2012, 03:01 am Last Edit: Mar 01, 2012, 03:16 am by jfenwick Reason: 1
If you look at the PDS-150 board there's a pair of transistors, one of which is FZT790A, and the other I can't completely read on my board.
There's also an SK14 Schottky diode. I think you probably need these two things to change the signal... maybe increase the current more than a microcontroller is capable of doing?

tjbaudio

OK, based on the given info and your schematic it does look like it should work.  Have you tried it with just one light again?  perhaps your mosfet is damaged?

Freakazoid

OK it was true the MOSFET was dead. Probably due to random testing  :smiley-roll-blue:. I replaced it with a new one. Now there are the following results:
5V @ Gate -> 24V @ Drain  :)
0V @ Gate -> 21,1V @ Drain  =(

So what can I do to get the results I need?

jfenwick

Here's an alternate schematic trying to do what you're doing from a known CK hacker, not sure if it helps at all...
http://web.mit.edu/carboxyl/Public/ck/ckschem.jpg

Go Up