PDA

View Full Version : New F16V2 differential expansion board in the works



dpitts
01-05-2016, 11:33 AM
New F16V2 differential expansion board in the works. This board will allow the F16V2 to drive 16 or 32 3-wire strings up to hundreds of feet away. The 16/32 strings are in addition to the 16 strings that can be driven from main board for a total of 48 strings max. When the expansion board is in 16 string mode each port can drive 340 pixels worth of data or 170 pixels in 32 string mode. The expansion board utilizes differential signals similar to those used on our Falcon hubs and Smart String controllers to drive inexpensive four string receivers. The receivers are in the PCB layout process and both expansion and receivers are expected to be tested in the next two weeks.

DrNeutron
01-05-2016, 11:40 AM
Dave you just keep making this lineup better and better!
My next board will definitely be a F16V2!

Daryl Hurd
01-05-2016, 01:35 PM
Awesome!

Lowepg
01-05-2016, 02:07 PM
Exciting!

I keep hearing about people driving pixels across longer-than-usual distances with f16v's (like 60+ feet).

Is there something unique to the "standard" f16 thats already more conducive to driving pixels long distance?

*** sorry if this is a dumb question, Im still trying to learn about this falcon!!

dpitts
01-05-2016, 02:39 PM
Mainly i think it i the 100ohms resisters in the F16V2 that allow for longer distance. The same is accomplished with SanDevice when you decrease the default 270ohm resistors.

Robo
01-05-2016, 03:06 PM
Amazing! Looking forward to these....

Wolfie
01-05-2016, 04:12 PM
New F16V2 differential expansion board in the works. This board will allow the F16V2 to drive 16 or 32 3-wire strings up to hundreds of feet away. The 16/32 strings are in addition to the 16 strings that can be driven from main board for a total of 48 strings max. When the expansion board is in 16 string mode each port can drive 340 pixels worth of data or 170 pixels in 32 string mode. The expansion board utilizes differential signals similar to those used on our Falcon hubs and Smart String controllers to drive inexpensive four string receivers. The receivers are in the PCB layout process and both expansion and receivers are expected to be tested in the next two weeks.

So let me wrap my brain around this....

This board outputs 16 or 32 differential signal outputs (either 2 or 4 out per port)?

And then you need the receivers then to convert back to TTL for sending direct to 2811 strings.

Right?

You are putting 4 differential signals down each TP in a Cat5/6 yes? Then the 4 channel receivers then split those 4 TP out to 4 TTL signals, yes?

Is the signal format RS485 on each TP?

So whats the price point here? This board and the receivers?

And will this board be compatible with V3 controllers?

I was thinking I wanted a couple V2s if the V3 is not coming soon and was going to throw down for 2 expander boards (standard TTL) until I saw this thread. Makes me want to wait for this baby to finalize. Can we combine orders later? Say I pony up for a couple V2 controllers, can I order a couple of these expanders and some receivers later and yall can roll the shipping together?

Lowepg
01-05-2016, 04:17 PM
Sigh.

This addition is making it harder and harder for me to stick to my story that "I need multiple 4-port controllers for remote elements and because of pixel control distance limitations"

smeighan
01-05-2016, 04:39 PM
Maybe this will clarify the falcon16v2. http://nutcracker123.com/nutcracker/images/f16v2.pdf

Wolfie
01-05-2016, 04:40 PM
Sigh.

This addition is making it harder and harder for me to stick to my story that "I need multiple 4-port controllers for remote elements and because of pixel control distance limitations"

You can do that for $8 or less a run. No controller needed. Just requires a bit of perf board and some soldering.

RogerH here turned me onto these chips...

This is all you need to run 2 four channel cables (and have 4 spare chips)(but 2 lots, that gives you 5 four channel cables):
http://www.ebay.com/itm/20PCS-MAXIM-MAX485CPA-MAX485-DIP-8-RS-485-RS-422-Transceiver-US-SHIPPING-M235-/401042528208?hash=item5d5fff57d0:g:TJAAAOSwHQ9WUTb 2

Some perf board and 8pin DIP sockets.

Each chip can be a receiver or transmitter. There are 4 twisted pairs in a cat 5. That means 4 channels of differential data. 4 of those chips on each end of the cable. One set to transmit, the other end to receive.

Feed each chip from a channel on your controller, opposite end each chip drives one string of lights.


Or you can pick up these which have all the impedance matching resistors and whatnot in place:
http://www.ebay.com/itm/10-PCS-MAX-485-MODULE-RS-485-MODULE-TTL-TO-RS-485-MODULE-COMMUNICATION-/261957769654?hash=item3cfde659b6:g:UdwAAOSwgQ9VpFN i

That will give you a 4 channel cable plus 2 modules backup. Buy 2 lots and that gives you 3 cables worth. Has blinky data lights, all components ready to go and even screw terminals for the cat 5 cable pair.

I picked up a lot of 5 for $5 (buck each) which gives me 2 channels down. I tested them by connecting them to opposite ends of a 500' coil of cat5. Fed one from a ECG-E2, other end went to length of 50 bullets. I used the other 4 wires in the cat5 to transmit 5v down for the controller and to power the pixels. Pretty sparkly lights. Took maybe 10 minutes to wire up.

Wolfie
01-05-2016, 04:56 PM
Maybe this will clarify the falcon16v2. http://nutcracker123.com/nutcracker/images/f16v2.pdf

Something is wrong with the math...

PDF says 32,768 channels total.

Most I can come up with is 32,640

16 x 680 = 10880 x 3 = 32640.

Thats 128 missing channels. Is that just the universe slop (512 is not divisible by 3) that is missing from the mix?


And what about the DMX ports? Is that included in the channel count? If you use the DMX outputs, does that mean you loose TTL outputs (one or more of the 16)?

K-State Fan
01-05-2016, 05:00 PM
Something is wrong with the math...

PDF says 32,768 channels total.

Most I can come up with is 32,640

16 x 680 = 10880 x 3 = 32640.

Thats 128 missing channels. Is that just the universe slop (512 is not divisible by 3) that is missing from the mix?


And what about the DMX ports? Is that included in the channel count? If you use the DMX outputs, does that mean you loose TTL outputs (one or more of the 16)?

64 universes times 512 = 32768 channels

Wolfie
01-05-2016, 05:02 PM
64 universes times 512 = 32768 channels

Ok, I thought it might be the universe slop that accounts for the missing channels when referencing pixel counts.

So how do the DMX channels factor in?

smeighan
01-05-2016, 05:05 PM
64 universes received = 32768
32 strings of 680 nodes for output = 32,640. Just means you can run some strings at more than 680. I really dont think those last channels will be used.
DMX channels do come out of the 32768. You cant output more than is received. I only have 32 AC channels, so i consider the DMX channels more like noise. You could take the extra 128 channels to drive DMX.

dpitts
01-05-2016, 05:05 PM
They are separate physical ports that run independent of pixel ports. They do however use the same 32768 bytes of data as there source.

smeighan
01-05-2016, 05:10 PM
You can do that for $8 or less a run. No controller needed. Just requires a bit of perf board and some soldering.

RogerH here turned me onto these chips...

This is all you need to run 2 four channel cables (and have 4 spare chips)(but 2 lots, that gives you 5 four channel cables):
http://www.ebay.com/itm/20PCS-MAXIM-MAX485CPA-MAX485-DIP-8-RS-485-RS-422-Transceiver-US-SHIPPING-M235-/401042528208?hash=item5d5fff57d0:g:TJAAAOSwHQ9WUTb 2

Some perf board and 8pin DIP sockets.

Each chip can be a receiver or transmitter. There are 4 twisted pairs in a cat 5. That means 4 channels of differential data. 4 of those chips on each end of the cable. One set to transmit, the other end to receive.

Feed each chip from a channel on your controller, opposite end each chip drives one string of lights.


Or you can pick up these which have all the impedance matching resistors and whatnot in place:
http://www.ebay.com/itm/10-PCS-MAX-485-MODULE-RS-485-MODULE-TTL-TO-RS-485-MODULE-COMMUNICATION-/261957769654?hash=item3cfde659b6:g:UdwAAOSwgQ9VpFN i

That will give you a 4 channel cable plus 2 modules backup. Buy 2 lots and that gives you 3 cables worth. Has blinky data lights, all components ready to go and even screw terminals for the cat 5 cable pair.

I picked up a lot of 5 for $5 (buck each) which gives me 2 channels down. I tested them by connecting them to opposite ends of a 500' coil of cat5. Fed one from a ECG-E2, other end went to length of 50 bullets. I used the other 4 wires in the cat5 to transmit 5v down for the controller and to power the pixels. Pretty sparkly lights. Took maybe 10 minutes to wire up.

you can create a line drivver, receiver for $8 and use it on your controller. This was describing adding this ability to the falcon16v2. the falcon16v2 has capabilities that i use. The OLED display, the bbuilt in e1.31 monitor, virtual strings (mix 2811, ink1003, rectangular 2811's all on one output), every string a different string type. I like the f16v2 and am happy that dave has added the differential board. BTW, his board looks to be around $5 per line to have the extended range so it is in the right price area.

For people who are happy with their existing controllers they should look to methods to extend distance , like what you describe. I like this solution using daves board.

thanks
sean

Gilrock
01-05-2016, 05:23 PM
Yeah I think people have gotten too used to the limitations in other controllers. With the F16v2 you don't worry about universe boundaries anymore. I had display elements starting on channel 511 in some spots.

Wolfie
01-05-2016, 05:30 PM
64 universes received = 32768
32 strings of 680 nodes for output = 32,640. Just means you can run some strings at more than 680. I really dont think those last channels will be used.
DMX channels do come out of the 32768. You cant output more than is received. I only have 32 AC channels, so i consider the DMX channels more like noise. You could take the extra 128 channels to drive DMX.

I do use DMX for floods. But only 12 channels (4 floods). So I can map the DMX ports to those slop channels in universes? Sorry, unfamiliar with the FC. I am looking for a better and more capable controller than the 2 ECG-P12s I have.


They are separate physical ports that run independent of pixel ports. They do however use the same 32768 bytes of data as there source.

Ok, thats good. Thank you.


you can create a line drivver, receiver for $8 and use it on your controller. This was describing adding this ability to the falcon16v2. the falcon16v2 has capabilities that i use. The OLED display, the bbuilt in e1.31 monitor, virtual strings (mix 2811, ink1003, rectangular 2811's all on one output), every string a different string type. I like the f16v2 and am happy that dave has added the differential board. BTW, his board looks to be around $5 per line to have the extended range so it is in the right price area.

For people who are happy with their existing controllers they should look to methods to extend distance , like what you describe. I like this solution using daves board.

thanks
sean

I too like the look of Dave's board. Its clean and clutter free and I like that. I am still waiting to hear what the price point on it and the receivers will be. I will likely be looking at at least one, maybe two depending on their price. And how did you figure the $5 a line price? Just guessing based on the other expansion board?

Anyway. my response was aimed at Lowepg who commented he was trying to justify 4 port controllers simply to control remote elements and I was offering him (her?) another solution instead of 4 port controllers. It was not directed as an alternative to Dave's board. I mean it could be, but you would be loosing money and ports off the controller by using the line drivers I linked. Its actually better to use Dave's expansion board, assuming you have a V2 to plug it into. I don't, so I used the Max chips off a J1Sys P12 to get some pixel distance. I don't have a huge yard. Its likely I won't need all 8 ports, not even close probably. But, its nice to have them ready and able :)

So is there an anticipated price point on the expansion board and 4 port receivers?

mharnish
01-05-2016, 05:30 PM
Sorry to semi-hijack the thread but aside from the new differential board can the F16v2 push power further than other controllers? I know when I go over 15 feet or so from my e682 with a 100 pixel string I start to get pinking and need to inject power. With the F16v2 can I run the 100 pixel string 25 - 30 feet from the controller without pinking? Is there any special gauge or type of wire needed for 25 - 30 feet from the controller?

This would definitely change my mind on what I am ordering for next year. I was planning a few e682s and a few e6804's but I could get by with 2 x F16v2 boards with expansions.

Thanks!

Wolfie
01-05-2016, 05:37 PM
That is one of the things I have read about the F16 boards is that it seems to be able to get more initial distance out to first pixel than many other controllers. But remember you can change the resistor packs on your e682 and get additional distance. I have a 682, I just didn't get it out this year. IN testing I got about 18' before I got unstable pixels out of the 682. I have not changed the resistor packs either. For reference, using the same wire and same pixels, I can only get 8-10' out of my J1sys P12 controllers before the pixels become unstable. That is why this expansion board is of interest to me. 10' ain't cutting it.

dpitts
01-05-2016, 05:39 PM
Usually the only way to solve the pink problem is to power inject or use better cable that has less voltage loss.

Wolfie
01-05-2016, 05:48 PM
Rofl! I read his question to say "blinking" not "pinking" <facepalm>.

Nevermind my last post then :blush2:

smeighan
01-05-2016, 06:20 PM
i would separate the discussion of distance to power drop and data signal.
People on australian christmas (ACL) tried using 3 wire cable, cat5 cable and coax cable on the data signal. the distance was in that order, shortest to longest, 3 wire cable, cat5, coax. I think cat5 goes longer because you have some noise immunity with a twisted cable for data instead of a single cable. coax went the farthes, 70 meteres (200 feet). cat5 was up to 60-70', 3 wire was around 30'.

in these data extenders that dave has , do note there is NO power being extended. the 300-400' of cat5 is carrying 4 pairs of data signals. Power must be supplied locally at the receiver boards.

Wolfie
01-05-2016, 06:30 PM
i would separate the discussion of distance to power drop and data signal.
People on australian christmas (ACL) tried using 3 wire cable, cat5 cable and coax cable on the data signal. the distance was in that order, shortest to longest, 3 wire cable, cat5, coax. I think cat5 goes longer because you have some noise immunity with a twisted cable for data instead of a single cable. coax went the farthes, 70 meteres (200 feet). cat5 was up to 60-70', 3 wire was around 30'.

in these data extenders that dave has , do note there is NO power being extended. the 300-400' of cat5 is carrying 4 pairs of data signals. Power must be supplied locally at the receiver boards.

I am using 4 conductor shielded alarm cable. It has a foil shield with ground wire. I found that by grounding the shield, it shortened my max distance by almost 2 feet. Counter intuitive that was let me tell you.

And yea, the pink issue is likely power, not data. As I said, I mis-read his statement to say "blinking" not "pinking". Pinking is typically caused by low voltage. Blinking or non-reactive pixels is typically caused by data corruption (too long wires).

As I said in my experiment using the ebay boards, I ran 5v down 4 of the cat 5 cables to power 50 pixels and the receiver. Realistically I would use 1 pair for 5v, and the other 3 for data. The 5v would only power the receivers as I plan to be all 12v pixels next year. I wouldn't expect to drive many pixels feeding power down 24awg wire inside cat5 so power supply at/near pixels was assumed, at least for me it was.

mharnish
01-05-2016, 10:08 PM
Ok, just to be clear there is NOT a significant change in the distance from the controller to the first pixel from a power perspective using the F16v2 verses other controllers, correct? I would still have to power inject if i am running more than 10 to 15 feet over 18 gauge 3 wire to my first pixel in a 100 pixel string.

If I could get 20 to 25 feet without needing power injection i could change my controller layout and use more dense controllers like the F16v2.

smeighan
01-06-2016, 06:13 AM
I run 40' from controller/receiver to first light. Power supply is next to controller, no power injection. 150 nodes


Sent from my iPhone using Tapatalk

Wolfie
01-06-2016, 12:09 PM
Ok, just to be clear there is NOT a significant change in the distance from the controller to the first pixel from a power perspective using the F16v2 verses other controllers, correct? I would still have to power inject if i am running more than 10 to 15 feet over 18 gauge 3 wire to my first pixel in a 100 pixel string.

If I could get 20 to 25 feet without needing power injection i could change my controller layout and use more dense controllers like the F16v2.

There is, AFAIK no difference in POWER from controllers. IE, the POWER from my P12, or your/my 682 or the F16V2 is all damn close to the same. Its the DATA driver circuits that have the most effect on controller>pixel distances. And thats what the resistor packs can change in your and my 682. Someone with a F16 can confirm, but I don't believe it has interchangeable resistor packs. That I have seen, very few controllers do.

Basically the resistor packs change the matching impedance (resistance if you will) of the output circuit to better match it with the cabling you choose to run. The better match you make, the less data corruption that occurs. Its sorta like using a 50 ohm coax to hook up your cable TV instead of the standard 75 ohm coax it expects. Will it work? Yes. But the quality of the signal is reduced. By matching the controller output circuit to the wire via resistor packs, the data can be driven further down the wire before losses corrupt it to the point the first pixel can't rebuilt the data packet. Baud rate also plays a part here too but thats a discussion for another thread I should think.

Do you have a multimeter? If not, go get one, even a cheap one (Harbor Freight, Home Depot, Menards). You really should have one on hand.

I am suspicious of your PSU(s). 20' should be nothing for power runs. Data, another story. I would recommend you measure the voltage at the PSU to verify its at 5 or 12v whichever is appropriate. If its wrong out of the PSU, there is usually a small adjustment dial (potentiometer or pot as they are called) on the PSU, typically near the screw terminals. Disconnect ALL cables except AC from the PSU. Then put your meter on the output terminals and adjust it for correct voltage. Doing this with everything disconnected, at least until you are sure of what you are doing, protects all your stuff from accidentally over adjusting the pot and toasting any of your equipment. Once you have a good voltage on the PSU and understand the function of the pot, reconnect everything and re-measure at the PSU terminals. Tiny tweaks to the pot to re-establish proper output voltage, but don't go to far as the voltage could raise too far unloaded.

After verifying the PSU voltage is correct, then measure it at the controller output. Then measure it at your first pixel. If you are not seeing the correct voltage at any location, you should investigate all connections, wires, terminals, solder joints between the last place you found good voltage and where you found low voltage.

dpitts
01-06-2016, 12:31 PM
F16V2 resistor networks can be removed from SIP headers and replaced with ease.

Wolfie
01-06-2016, 12:33 PM
There ya go! I figured the F16 would have them. I just hadn't seen one in person to confirm.

boarder3
01-06-2016, 04:18 PM
let me understand this correctly cause right now i use 682 boards and i can do i think maxed out is 16 strings of 63 pixels? you telling me i can 16 strings of 680 pixels? is that right and what protocols can it do since i cant find that in wiki. i use 3001 and gece and whatever else i buy this year.

dpitts
01-06-2016, 04:21 PM
The types supported are listed on this spreadsheet. http://nutcracker123.com/nutcracker/rgb_controllers/
GECE has a max of 64. That is the pixels limitation not controller.

jck46
01-06-2016, 11:22 PM
so you are saying the new boards will not work with the boards I have now but will work with the new ones on presale any idea of price yet

dpitts
01-06-2016, 11:37 PM
No. Just today i came up with a solution. They will work now.

Sent from my LGLS990 using Tapatalk

tonyhight
01-07-2016, 01:15 AM
Curious of thoughts. I have 3 e682's right now. My plans for next year would add 5 e6804's (smaller distant elements) and 2 more 682's. If I bought f16v2 and expansion to use for my megatree would there be any drawbacks or issues running with other controller and Pi? Are power capabilities the same on the board as controller? Tree has 24 strings of 100 nodes. Each string uses one output. So 12 outputs of controller and 12 of Expansion board. Would use 4 psu's. If it works it'll save me $150. Thanks!

--Tony Hight
http://www.hightfantasylights.com
http://www.facebook.com/hightfantasylights

smeighan
01-07-2016, 03:29 AM
Falcon16v2 can drive 340 nodes per output on v2, run 3 strings of 100 per output, you would use 8 outputs of the 16 . U would need to power inject. Add the new diff expansion board and u could have 8 receiver boards located. 300' away. Each receiver can drive four strings, each string can have 170 nodes


Sent from my iPhone using Tapatalk

tonyhight
01-07-2016, 08:05 AM
Falcon16v2 can drive 340 nodes per output on v2, run 3 strings of 100 per output, you would use 8 outputs of the 16 . U would need to power inject. Add the new diff expansion board and u could have 8 receiver boards located. 300' away. Each receiver can drive four strings, each string can have 170 nodes


Sent from my iPhone using Tapatalk
That's why I'm curious if it will work power wise the way I already have it set up. I try to power inject as little as possible. Most of my string are powered directly from the board to make plugging things in simpler along with less connection points and additional fuses to buy and wire to psu's for each injection run.

I know I can save outputs and cut down on controllers but in this case I'd ook at your example and see running no power off the board at all and injecting at the bottom of all 3 strings to ensure consistent brightness or powering the first string from the board and continuimh the data only to string 2 and 3 and injecting powered at the beginning and end of those two strings. Either way I've got two additional connections at the top (the strings are separated for storage) along with two to three additional injection points of power from psu at the bottom with another plug and fused wire for each.

--Tony Hight
http://www.hightfantasylights.com
http://www.facebook.com/hightfantasylights

jeffl
01-27-2016, 05:40 PM
I may have missed it but will the differential boards be connected to one of the 4 serial ouputs on the board and then be software configured like DMX, pixelnet or renard?

dpitts
01-27-2016, 07:51 PM
No. They are directly connected to FPGA/Processor board on the F16V2.

jeffl
01-27-2016, 08:05 PM
Will they connect to the existing F16V2 boards without additional hardware?

dpitts
01-27-2016, 08:09 PM
Will they connect to the existing F16V2 boards without additional hardware?

Yes they will.

thevikester
01-28-2016, 02:03 PM
let me understand this correctly cause right now i use 682 boards and i can do i think maxed out is 16 strings of 63 pixels? you telling me i can 16 strings of 680 pixels? is that right and what protocols can it do since i cant find that in wiki. i use 3001 and gece and whatever else i buy this year.

Boarder, you can run substantially more than that on 682s up to 12 universes in Unicast, 7 in Multicast. Your limitation per output would be to not blow the 5 amp fuse. Total nodes 2040...I ran 16 strings of 120 pixels per string last year...no issues. Multicast takes you down to 1190 nodes or 74 nodes per string if using all 16 strings.

Wolfie
04-27-2016, 02:45 PM
So whats the status here? Boards in dev? Ready to build? Need beta?

dpitts
04-27-2016, 07:47 PM
They have been tested. Will be released in next month.

Sent from my LGLS990 using Tapatalk

Wolfie
04-28-2016, 10:08 AM
Was there a sale on them already? I never saw notice of one.

dpitts
04-28-2016, 10:09 AM
No not yet

Sent from my LGLS990 using Tapatalk

Wolfie
04-28-2016, 10:12 AM
Whew! Thought I had missed it!

Missed you last night in the chat!