I really like the specs on this pixel bridge, particularly the ability to have power coming from a 12 vac landscape power supply (I already have 12 gauge cables run to my deck from an existing 12 vac landscape ps).
But I'm having a bit of a hard time figuring where this fits into the overall architecture. Obviously this will run / control pixel strips or strings, but is this meant for real-time incoming wireless data feeds? I.e. this will receive a wireless data stream that is E1.31 or similar? If so, I don't understand why there are transmit capabilities (unless there are some kind of implied daisy chaining capabilities). I've read the "sACN to DMX, Wireless, Streaming FM & Local Audio Gateway" thread, but that doesn't make it any clearer.
If the wireless bandwidth is not enough for real-time streaming, I assume the wireless receiving provides a way to download some sort of control sequence, which is stored in the bridge and turned on and off (wirelessly, I assume, but that would take just a small amount of bandwidth).
If there are multiple of these bridges that are meant to be synchronized together (and I assume this is the case, since each bridge is limited to 3A which I assume will limit the available number of attached pixels), this would mean the transmitter would have to be configurable for multiple of these bridges.
I've tried to educate myself on the wireless protocols associated with the nRF51822, but what I've found doesn't match up (yet) with what I'm experienced with, regarding other 2.4 Ghz protocols (e.g. Wi-fi b and g).
Any help for someone that is a "mixed newbie" is appreciated!
Thanks,
Cliff
There is plenty of bandwidth to handle almost a full 8 universes over a 2Mbps channel. Now at that data rate the range will be reduced so there is some give and take depending on what you're wanting to do. This is not dependent or has any relationship with your normal WiFi channels or its protocol.
For lighting we work in the Master/Slave mode, just as DMX does, so the timing is totally under the control of the master. What does that mean? It means that the slaves only need to update the channels once for each new packet stream. If the slave has macro functions then it may have to do multiple updates but only for macros. For programming and other special control functions we can work using a different set of addresses much like the way RDM works with DMX. It goes over the same wireless channel but it can be identified due to it's different address again much like the way the start byte is used with DMX. This can even extend to changing the entire wireless frequency over the command channel or changing each element's pixel slot.
The design is such that a single PCB can cover any application. This does make the board a bit larger so that it can be powered from a landscaping transformer but it would be easy to leave that part off so you could get the board in a smaller case. For the TA-200 enclosure it does not matter but it does for the IP67/8 version it does. The board would have versions different parts installed. If however you install everything then you could not only drive a pixel string but also use the controller to send to three other slaves, i.e., one universe.
This was designed with the "sACN to DMX, Wireless, Streaming FM & Local Audio Gateway" in mind. That way you can control a very large number of wireless channels. How many totally depends on the speed of the wireless channel. At 250Kpbs it a single universe, 1Mbps it's 4 universes and at 2Mbps it's 8 universes.
The wireless protocols we're using are still in a state of flux. All of the work to date has been with the older nRF24L01+ modules which have a more limited 32 byte record length. The newer nRF51 series cannot only support that format but new format with up to 255 bytes of data per packet. This makes it much easier to create protocols to handle larger number of Pixels. In addition the wireless channel does not have to follow anything related to DMX. The master controller will handle the conversion from DMX universes to Wireless channels.
So at this point everything is still in the development stage and there is lots of work that has to be done before this sees the light of day. Right now it's presented so I can get feedback.
Thank, Timon! Your explanation is clear and the architecture is making much more sense to me. The "sACN ... Gateway" is the master controller, and the bridge (from this thread) is a slave (and obviously there could be multiple slaves). You (and others in the discussion I assume) are defining the higher level data protocols over the 2.4 channels (which makes sense - 802.11 / Wi-Fi doesn't fit for this). The gateway (master), I assume, does speak a higher level protocol such as E1.31 (over Ethernet) or one of the Wi-Fi protocols, for the input data stream (and whatever return data packets that are needed). The gateway and slave configuration / logic then knows how to allocate the universes / pixel addressing scheme to the one or more bridges / slaves.
This is interesting stuff, and I appreciate helping fill in my knowledge gaps (and correcting my mistaken assumptions).
Cliff
Latest update for the Wireless Pixel Bridge.
Changes:
- Change the chipset used for the power supply. It's still a Ti chip set but it now uses an external MOSFET instead of the internal one. This allows for higher power output so I can not get a full 6A which can supply up to 100 pixels. The parts rating may allow for a full 10A but I need to make sure I can get enough heatsink in the case.
- Bluetooth, I may be able to use the nRF51422 which has the builtin Bluetooth 4 stack. This would allow for bluetooth configuration when not in wireless mode. This is something to look at for future development.
- I single board design now allows for several different builds for power, DMX and wireless. Example, you can build to be powered from a low voltage source AC or DC source. Again this is a build option.
- If you leave off the processor you could just use the supply to make a power injector. Need to add a couple of wirepads for this to work.
- Stand alone mode for landscape lighting. When no DMX or wireless is active the controller can run your Pixel Based Landscape lighting. (Still need to add a light sensor for this to work)
- Fits a TA200 case.
Lookin' real good! Any idea when these might "see the light of day" 8^}
I've got to get ready for this years show so the $s and time have to be allocated. I supect this will be an early next year project along with the sACN to Wireless, DMX and FM Bridge since they will be working together.
Thanks for the update. The set of components working together makes sense.
timon wrote:I've got to get ready for this years show so the $s and time have to be allocated. I supect this will be an early next year project along with the sACN to Wireless, DMX and FM Bridge since they will be working together.
I'll support the current nRF24L01 protocols although I hope that we can all agree on a modified protocol that can support the larger packet sizes that the nRF51x22 can use. That way we'll all be on the same page. It's something that needs to be discussed after this season.
Most of my designs going forward are going to be ARM based primarily using the Cortex processors from NXP and Nordic. I'll be using the larger M3 and above for the Gateways and larger controllers and the M0 processors such as the nRF51x22 for the smaller ones. Using the nRF51x22 will allow every controller to be wireless or DMX based without doing anything special. Only thing I wish the Nordics processors had a boot ROM like the NPXs.
Already on the drawing board are AC and DC controllers, Servo controllers and Pixel controllers such as this one. All of the designs will have a common front end that allows for internal diagnostics so voltages, currents and temperatures can be monitored and reported back using RDM protocol. I'm also planning to run all controllers through FCC Part 15 testing since they contain transmitters, hopefully I can get this done somewhat inexpensively, since the last thing you want to do is deal with an angry FCC:(
timon wrote:I'll support the current nRF24L01 protocols although I hope that we can all agree on a modified protocol that can support the larger packet sizes that the nRF51x22 can use. That way we'll all be on the same page. It's something that needs to be discussed after this season.
(
Just because I have a thing against dead horses, Im curious about the larger packet format. Any idea on the change in throughput we could expect by going to the larger format? We are still taking about using the same data rates IE 250kbps vs 1MBPS etc. Do you think we will see more throughput with larger packets? Or just less chance to drop one?
Im also curious about FCC and open hardware. You are planning on releasing the hardware on the CC BY-NC-ND 3.0, if it undergoes FCC testing is that covered by the overall design of the product regardless of manufacturer? Or does it only pertain to a "built" version of the board?
Having been though FCC / Industrie Canada I can offer what I discovered. I was submitting for eval under the 13.56MHz intentional radiator for RFID products. I would guess you will also be under a similar guide, but at the frequencies of choice.
It's not cheap. The real cost is in the paperwork and prep of package for submittal. Not to mention the re-work of the PCB when the product fails the limits and then retest. In one case I learned why virtually all laptop PSs have a filter, the LV cord can act as an unintentional radiator. It can also feedback onto the actual AC cord. (which earned me another retest). Retests can add $2k for each.
I had two products and made a package deal and I think it was around $12K to start. But I had the company perform the complete package for me (with me submitting a lot of detail). Check out NWEMC. Or ask me specific questions and I'll try to help.
