One-Channel Control...
 
Notifications
Clear all

One-Channel Controller

99 Posts
21 Users
0 Reactions
38.8 K Views
Posts: 2541
(@lightup)
Famed Member
Joined: 17 years ago
dpetryk wrote:...In my design, the switch turn on at the zero cross at the beginning of a cycle and turns off at some phase angle somewhere between 0 and 120 degrees...

"Regular" triacs don't turn off until its current reaches around zero amps. Hence the use of the "phase on control". Perhaps you are using a IGBT?
Also how did you arrive at $2 per controller - if I may ask?


Reply
kingofkya
Posts: 2165
(@kingofkya)
Noble Member
Joined: 15 years ago
LightUp wrote:"Regular" triacs don't turn off until its current reaches around zero amps. Hence the use of the "phase on control". Perhaps you are using a IGBT?
Also how did you arrive at $2 per controller - if I may ask?

As i was mainly wanted ot know how the data rate was effected when other controllers (not your controlers) shared the same ac line. I think he meant he is doing leading edge dimming rather than falling edge.

In my testing it really did matter they produced around the same amount of noise, i had about an amp1 of resistave current, less when it was dimmed.


Reply
Posts: 3277
 Jrd
(@jrd)
Famed Member
Joined: 19 years ago
dpetryk wrote:Well I'm the guy that designed and built that 4 channel controller. I tried to get some interest in the Houston area with other like minded individuals but no one was interested.
SO I built 300 of the controllers and I use them in my show. A one of a kind of system. I use a string of LED lights that I had made in China custom ordered. I wanted to get the system simple and easy to use and sell it to the general public. But since I could not get any interest in the community I just quit trying to sell it to people that had not interest.

There just seemed to be a lack of interest overall and everyone has moved on to pixel lamps anyway.

I was interested in eliminating the communications cables. I did that by putting the communications on the AC power line. My strategy was a de-centralized or distribute the control system. It works great. I was working with LOR on using the power line communications but I dont think it worked too well for them primarily because the LOR controllers require two way communication for some things. My controllers do not because the communications is one-way. There is no need for two way communications. My system only uses abut 6 feeder cords that have plugs wherever there are controllers to be powered. All the lights on the house (150 channels) are on one cord. Controllers hung on the house near the lights. As you can see on the video the controllers are mounted on the arch. They are in a wall wart enclosure and potted for environmental protection.

What I refer to the transverter in the video simply modulates the serial data stream on the 120 vac power line. The controllers de-modulate the data stream and respond in the usual way. The key was finding a circuit that could extract the serial data stream off the power line that had a bandwidth of at least a megacycle because I wanted to pass DMX through it and DMX data stream clocks at quarter megacycle (250k baud rate). My circuit ended up to be ten times faster than I needed and it only cost about $2 to build. It was what made the power line communications (PLC) work.

I have a new design that eliminates the big blue transformer and maybe the other large inductor as well. I wanted to mold them on the light string and build them for really cheap and make a bizillion dollars. I dont think the market is there and It would take a truck load of money to get them on the shelves or retailers.

What you see is the first prototype of the system I wanted to be sure the concept worked in a real world deployment. It does. It saves a bunch of time and hassle setting up a display. It eliminates a bunch of wire. Controller addressing has to be managed properly. Cost high because I had to include the development costs but it could be made very inexpensive. Controllers remain on the light fixtures so setup and take down is faster.

I own an industrial control systems company and my time is limited so there would have to be a decent incentive to continue the development of the system.

So there you have it - mystery solved. - Video can be seen here - www.youtube.com/watch?v=TAh1gg0u8Xk

Your system is very cool and impressive. Would you consider making it available to the DIY community as a DIY project?

I noticed one of the boards appears to be mostly SMD components, Would it be possible to do this in through hole? Or would that make it too expensive/large?

Does you system inject the data in the upstream AC line also similar to X10 equipment or is the data only available on the outlet of the transverter?

I am looking for a low cost low channel count(at the endpoints) solution and this could do it depending on the answers to these questions.


Reply
Posts: 13
(@dpetryk)
Active Member
Joined: 17 years ago

Answers,

The $2.00 was for the demodulator circuit only. Not including all the other stuff for the controller.

I use a transistor not an IGBT

The transverter only sends data on the AC line going to the controllers. Some does leak back into the house circuit however. It is similar to X10 in concept only.

I have had 90 controllers on a single cord driven by 1 transverter. More than that the signal begins to get too weak for reliable operation.
The rule is that ONLY these controllers can be connected to the power line. NOTHING ELSE EVER. Because my controllers are designed
so they do not absorb much the signal. If you put a typical LOR controller across the line nothing will work because they and many
other devices have big caps across the line for noise and spike suppression which kills the signal. My controllers do not generate switching
noise so I dont need or want to put filters the AC line.

The controller is too big as it is. I am planning to cut the size to 1/3 to 1/4 of what it is now by eliminating the blue transformer and the other inductor
and making it even smaller so it can be molded onto the power cord. SMT is a must. I don't really have an incentive to re-design it because my system
works great now. Saving a few qubic inches on a new controller design does nothing for me. It would be and incentive if I had sales of a bunch of
systems and recover the $15k I put into the development.

ANother limitation is that it the controllers are designed for LED and resistive loads only. 120vdc series string. It will not drive large loads. 1 amp total for all 4 channels.

The only thing I might - I repeat might consider is selling the controller boards fully assembled with firmware installed. The volumes would have to justify
getting them spooled up for assembly. I am not in a position to support a lot of questions over the phone or email about issues that might come up.
I just dont have the time to take on a task like that. I'm sure there would be bunches of questions and it would probably make me regret the decision of selling
them in "KIT" form. Its not really in a form that is easily sold because there is little documentation and the utility program used to manage the addressing
and configuration of the controllers was written in VB6 / XP and needs to be brought up to date for the latest software tools and operating systems.

Not sure if it would be a viable DIY project or not. I am open for ideas.


Reply
Posts: 2541
(@lightup)
Famed Member
Joined: 17 years ago
dpetryk wrote:Answers,

The $2.00 was for the demodulator circuit only. Not including all the other stuff for the controller.

I use a transistor not an IGBT...

OK, that makes sense now that you are not using a triac.

...The transverter only sends data on the AC line going to the controllers. Some does leak back into the house circuit however. It is similar to X10 in concept only.....

Can you enlighten us on your transverter? FSK? RF frequency?


Reply
Page 12 / 20
Share: