DMX with dimming for Grinch??

[Deast]

I have been looking around for an answer to my question.... Is there an interface that allows you to control a grinch via DMX that also offers dimming? I saw the interface that RJ made for DMX to Grinch/Olsen, but it does not dim.

I guess what I am looking for is a DMX to Grinch with dimming, or a DMX to Renard converter.


Thanks

Dan

[Macrosill]

I believe a DMX-Renard interface is in development.

[mrpackethead]

Look for Renard 4p4, its a dmx renard..

[RPM]

Dan,

This is exactly what I'm going to be doing for this years display.
I had designed a 128 channel DMX controller that I was going to use, but then I decided to go a different route and came up with a DMX DIMMING interface for my grinch boards.

It takes a DMX signal and converts it to a dimmable signal for a grinch (or 595 board) and outputs 128 channels, so it's good for driving two grinch boards. It is powered by an external 12 volt AC power transformer and derives the zero cross signal for dimming from this.
I have the PIC code done and a prototype has been built and tested now and has been running 24/7 for several weeks now with good results.

I'll attach a preliminary schematic for anyone who is interested.


RPM



I have been looking around for an answer to my question.... Is there an interface that allows you to control a grinch via DMX that also offers dimming? I saw the interface that RJ made for DMX to Grinch/Olsen, but it does not dim.

I guess what I am looking for is a DMX to Grinch with dimming, or a DMX to Renard converter.


Thanks

Dan

[Deast]

Could you post or pm me the code and the BOM? I would like to build one for myself. I have a working grinch, and I decided to go the route of DMX.

Thanks

Dan

[dmcole]

RPM:

Shouldn't you have an output for DMX? Or, rather than a jumper for DMX termination, make it a solid connection?

\dmc

[RPM]

Dan,

I don't have a BOM together just yet, but when I'm done I'll release more info on it.
I'm thinking of getting some PC boards made up and if there's enough interest, maybe I can get some beta testers involved.
Right now the code only supports 60hz, but I'll be working on 50hz soon. Just keep in mind the schematic is preliminary, I just looked at it and it shows the Atmega8-16 PIC, while I am now using the Atmega48-20 and I have tweaked a few of the resistor values.

Robert



Could you post or pm me the code and the BOM? I would like to build one for myself. I have a working grinch, and I decided to go the route of DMX.

Thanks

Dan

[RPM]

The final version would probably have RJ45 jacks for DMX in/out. The schematic just shows a terminal block because that's what I like to use, since you can just attach both the in and out wires to it for daisy chaining, or install a jumper to terminate the line if it's the last device. I do this out of habit, since in my line of work I work with RS485 circuitry and that's how it's usually done for permament installations, but the RJ45's do make it easier for the DIY stuff.

Robert


RPM:

Shouldn't you have an output for DMX? Or, rather than a jumper for DMX termination, make it a solid connection?

\dmc

[dmcole]

I do this out of habit, since in my line of work I work with RS485 circuitry and that's how it's usually done for permament installations, but the RJ45's do make it easier for the DIY stuff.
That explains it.

\dmc

[Deast]

If you need a beta tester, let me know.

[NogginBoink]

Hey RPM, out of curiosity, how does it work? Does it just update the Grinch outputs 120 times a second and synch the data strobe with the zero crossing?

[RPM]

Actually, it updates the output 512 times per second (256 per half 60hz wave) for 128 channels, synced to the zero crossing. It uses the data/clock/strobe lines to update the outputs.

RPM

[NogginBoink]

Of course it uses all three lines! :)

But the strobe line is the one that should be synched to the zero crossing. You can fill the data into the shift registers, and blip STROBE on cue to transfer the data to the outputs in synch with the zero-crossing. Don't worry about shifting data into the chips in synch with anything; your concern should only be in synching STROBE to the zero-crossing.

Updating 512 times per second isn't going to be sufficient. There are 120 AC half-waves per second. Updating 256 times per half wave is 120 * 256 = 30,720 times per second. (512 updates/second is 256 updates each half-second. I don't know where I came up with 120 times/second in my post above.) You'll be shifting out 128 channels' worth of data which means you'll be shifting out about 30,720 * 128= ~4 megabits per second.

The Allegro chips are rated to a 25MHz data rate, so you should have ample headroom there. Be sure to choose a microcontroller that you can clock high enough to pump out those bits. You may also consider dropping the least significant bit of the dimming level, and refreshing only 128 times per half-cycle. This will halve your bandwidth requirement and probably have zero visible impact.

I'm really excited about this project. I'm looking at my requirements for this year and it looks like I could use a few of these in my show. I think this will be an excellent contribution to the community. Please let me know when it's finished.

[P. Short]

Quite frankly I think that there are going to be problems with this. People have been having difficulties with the Renard-Grinch (Ren-C) adapter, which is a similar design. The problems have to do with the signals going between the microcontroller and the Grinch through the RJ45 cable, and are probably signal integrity problems (possible ringing on the clock and latch signals, and perhaps ground bounce). In some cases people have had to reduce the cable length down to three or four inches because it didn't work with six inch cables. If I read this thread correctly, the plans are to double the number of channels, and hence double the clock frequency. I think that this will lead to difficulties.

--

Phil

[RPM]

Quite frankly I think that there are going to be problems with this. People have been having difficulties with the Renard-Grinch (Ren-C) adapter, which is a similar design. The problems have to do with the signals going between the microcontroller and the Grinch through the RJ45 cable, and are probably signal integrity problems (possible ringing on the clock and latch signals, and perhaps ground bounce). In some cases people have had to reduce the cable length down to three or four inches because it didn't work with six inch cables. If I read this thread correctly, the plans are to double the number of channels, and hence double the clock frequency. I think that this will lead to difficulties.

--

Phil

Phil,

I agree that there is the possibility of some problems due to the fact that there are many possibilities for wiring and power sources for the Grinch, but in my testing I have used my interface to run 1 Grinch with 12 feet of CAT5 cable (one 10 foot prefabricated generic cable plus a 2 foot home made "pigtail" off the controller board) using the on-board power supply to power the Grinch and have not seen any problems with timing.
I have tried two different Grinch combinations so far, one with the Maxim MAX6969 chips and one with the Allegro A6279EA-T chips (I don't have any MBI5027 chips to test) and they work equally well with "standard" Co-op SSR's.

I originally wanted 128 channels, but due to timing problems the code currently supports 64 channels.
I think it's possible to get 128 channels but the limiting factor is the clock speed of the AVR I'm using and more specifically, the SPI interface speed. I am using the ATMega8/48 AVR's hardware SPI interface to generate the required timing for the data/clock lines, but it only outputs at 1/2 the clock speed (16 mhz max clock speed for the ATMega8-16) in bursts of 8 MHZ frames. Basically, you send 8 bits and wait for the SPI port to finish sending, then next 8 bits...etc. The wait times in between burst are considerable, so the overall effective speed is about half of the 8MHZ SPI speed. This causes some timing problems at 128 channels, but it works well for 64 channels.

I'm working on a solution for this problem, but in it's current form the interface is very usable as the hardware SPI timing is very precise. It's because of this and the fact that I am buffering the output of the AVR with a 74HC14 that I believe makes the interconnect cabling requirements less critical.

I will be posting an updated schematic, AVR code and Bill Of Materials soon for anyone who want's to build one and give it a try.

Robert

[Deast]

I am definitely looking forward to the plans being posted!!!

[RPM]

Of course it uses all three lines! :)

But the strobe line is the one that should be synched to the zero crossing. You can fill the data into the shift registers, and blip STROBE on cue to transfer the data to the outputs in synch with the zero-crossing. Don't worry about shifting data into the chips in synch with anything; your concern should only be in synching STROBE to the zero-crossing.

Updating 512 times per second isn't going to be sufficient. There are 120 AC half-waves per second. Updating 256 times per half wave is 120 * 256 = 30,720 times per second. (512 updates/second is 256 updates each half-second. I don't know where I came up with 120 times/second in my post above.) You'll be shifting out 128 channels' worth of data which means you'll be shifting out about 30,720 * 128= ~4 megabits per second.

The Allegro chips are rated to a 25MHz data rate, so you should have ample headroom there. Be sure to choose a microcontroller that you can clock high enough to pump out those bits. You may also consider dropping the least significant bit of the dimming level, and refreshing only 128 times per half-cycle. This will halve your bandwidth requirement and probably have zero visible impact.

I'm really excited about this project. I'm looking at my requirements for this year and it looks like I could use a few of these in my show. I think this will be an excellent contribution to the community. Please let me know when it's finished.


NogginBoink,

Actually, in actuality all three lines are synched to the 60hz AC. First you cycle clock & data for all channels, then cycle the strobe line to output the data. This must be done 256 times for each half 60hz cycle, with the on/off timing per channel varied by the dimming value and then turn off all outputs at the zero cross and start over for the other half wave.
Not hard for microcontroller to do, but for dimming it's all in the timing.
I won't get into it here, but David Fansler has an excellent writeup here on AC dimming
http://www.dv-fansler.com/Computer%20Lights/dimmers.htm

Robert

[kostyun]

FYI RJ made this several months back - I have two completed ones sitting right here.

[scharbon]

I thought RJ's were for on/off only, I didn't think they dimmed. Or is there another board out there?

Steve

[NogginBoink]

NogginBoink,

Actually, in actuality all three lines are synched to the 60hz AC. First you cycle clock & data for all channels, then cycle the strobe line to output the data. This must be done 256 times for each half 60hz cycle, with the on/off timing per channel varied by the dimming value and then turn off all outputs at the zero cross and start over for the other half wave.
Not hard for microcontroller to do, but for dimming it's all in the timing.
I won't get into it here, but David Fansler has an excellent writeup here on AC dimming
http://www.dv-fansler.com/Computer%20Lights/dimmers.htm

Robert

There's no need to synch the data and clock lines to anything. The only requirement is that the output changes happen in synch to the AC signal. You can load data into the shift register at any point in time you'd like, only the output from the latch has to be timed accurately.

[RJ]

I think Frank is talking about the DMX converter that makes an On/Off Grinch run with DMX. I have a new controller that is 128 channels of DMX with dimming, power and ZC on one board called the Freestyle. It is like two grinches, a dmx convertor, and powersupply with ZC put together but it does dimming. It cost about what a 595 cost to build and works with all standard SSR's. There will be a coop starting this coming week for it.

RJ

[WWNF911]

I thought RJ's were for on/off only, I didn't think they dimmed. Or is there another board out there?

Steve


I think Frank is talking about the DMX converter that makes an On/Off Grinch run with DMX. I have a new controller that is 128 channels of DMX with dimming, power and ZC on one board called the Freestyle. It is like two grinches, a dmx convertor, and powersupply with ZC put together but it does dimming. It cost about what a 595 cost to build and works with all standard SSR's. There will be a coop starting this coming week for it.

RJ

To be clear,... the coop won't be here. You'll have to follow the link to the website at the bottom of RJ's signature in his post for that. :)

[RPM]

There's no need to synch the data and clock lines to anything. The only requirement is that the output changes happen in synch to the AC signal. You can load data into the shift register at any point in time you'd like, only the output from the latch has to be timed accurately.

The clock and data need to be synched to each other but not necessarily to the zero cross, you are correct in that the strobe is the critical line that needs to be synched to the zero cross.

Robert

[RPM]

I did a test yesterday and hooked up a Grinch to two SSR’s and eight 150 lamp strings and decided to video it. I also doubled up a C9 LED string to one of the channels just to see the difference between the LED's and incandescents. It’s the multi-colored string at the bottom of the video.

Although the quality isn’t the best, you can see the Grinch next to the two SSR’s with Cat5 cables laying over each other with no interference.
I was running Vixen 2 with an Enttec DMX Pro interface connected to my DMX dimmer interface and then to the Grinch.

I know this isn’t a full on test, but it’s a proof of concept and I plan on doing a full test with all SSR’s attached when I find the time to get more of the Xmas stuff out of storage to hook it up.

Robert

http://www.youtube.com/watch?v=0nPUxzt5pD4

[WWNF911]

RPM that looks awesome!

[scorpia]

Ok,

i have been sort of following this thread and sort of not. It looks to me like there is now a DMX-Grinch interface that dimm's. Cool really cool infact.

He's a question for you. have you considered modifying the design to work with the ledtricks?
You wouldnt need as many dimming levels as you use with the grinch I was thinking 16 levels of dimming and this should ease up the speed needed on the interface.

this would then give the Ledtricks a whole new interface and would make it a much eaiser device to control i think.

Anyway not sure if it can be done or not but would be interesting to see how it goes.

Peter

[dmcole]

RPM:

It's looking really, really good.

Thanks for all your hard work.

\dmc

[NogginBoink]

RPM, that looks fantastic! Please keep us posted on your progress!

[RPM]

Ok,

i have been sort of following this thread and sort of not. It looks to me like there is now a DMX-Grinch interface that dimm's. Cool really cool infact.

He's a question for you. have you considered modifying the design to work with the ledtricks?
You wouldnt need as many dimming levels as you use with the grinch I was thinking 16 levels of dimming and this should ease up the speed needed on the interface.

this would then give the Ledtricks a whole new interface and would make it a much eaiser device to control i think.

Anyway not sure if it can be done or not but would be interesting to see how it goes.

Peter


Peter,

I'll leave that one to RJ. I believe he has another version of LEDtricks that he's working on right now.

RPM

[Deast]

How is the progress on this?

[RPM]

How is the progress on this?

I've been working on getting the BOM together and a PC board layout. I'll attach a preliminary PC board layout.

The parts cost is about $26-27 right now, not including a PC board. I plan on getting a few prototype boards made up and if there's any interest, I'll probably get more boards made.


RPM

[dmcole]

... and if there's any interest, I'll probably get more boards made.

Oh, there's interest.

\dmc

[RPM]

Oh, there's interest.

\dmc


That's great to hear! I'll attach a picture of my perfboard prototype unit.


Here's some info about the design...

The PC board design is pretty close in layout to this one and hopefully I'll find the time to finish it up in the next week or so and get some prototype PC boards made.

The prototype in the picture shows a 6 pin programming port in the upper left corner, but I'm not sure if I should include that on the final PC board. It makes it easy for me to do programming changes to the PIC code, as I don't have to remove the chip every time I do an update, just hook up my programming cable and go.


I have used RJ45 sockets to make it easier to interface to the current DIY designs out there. The RJ45 socket on the left goes to the Grinch board input, and the two RJ45 sockets on the right are for the DMX IN/OUT connections. These two are in parallel so you can daisy chain several units and there is a termination jumper on the board used for terminating the line if this is the last unit (or only unit) in the signal chain. The on-board power supply outputs enough current to run 1 Grinch board with the current SSR's out there.

There are also two 2-pin terminal blocks on the board. The one on the bottom right is for 12VAC power source for the unit, which is supplied by a "wall-wart" power supply. The unit derives the AC Zero-cross reference signal from this, so it must be an AC power supply. It will operate with a transformer rated at 9VAC to 12VAC input with 1 ampere current rating. The transformer suggested in the bill of materials is a 12VAC unit rated at 1.66 amperes and will provide enough current to run two Grinch DMX Dimmer boards and two Grinch boards. The terminal block on the bottom left is currently used for testing and is connected directly to the +5vdc supply.
I had originally put this on the board to be able to get more current to the Grinch board, but the +5vdc is also hooked up to the RJ45 socket, so this may go away in the final version.

The DIP switch positions 1 through 9 are used for changing the unit's DMX address and switch position 10 is used for the loss of DMX signal "blanking" option. In the OFF position the unit will hold the last valid DMX data and in the ON position the unit will blank all outputs after 1 second.

In the upper left, the LED's are used for power/error status (red) and DMX signal (green).
The red LED will remain on solid when the unit is receiving power and DMX signal and it will flash when there is no DMX signal present.
The Green LED will flash only when the unit is receiving a valid DMX signal within the unit's programmed address.

The PC board currently measures 3.75" X 2.5" and is small enough to fit into a 4X4" J-box.

As you can see, quite a bit of thought and effort has gone into the design of this unit and I would like to share it with the DIY community.


RPM