How can you help me help you?

[mrpackethead]

Friends,

Litterally a few minutes after first hearing that a long anticipated co-op for controlled RGB pixel leds had been abandoned, i was contacted by a long standing member of the "blinky flashy" community to see if there was something i could do to "fill the gap" for the many people who now are feeling stranded. Others have echoed that same sentiment to me by phone, email and instant message. For a Sunday Afternoon my email was really busy. At heart i am a DIY type of guy, and i know the satisfaction of building something yourself and getting to to run. The reason i do this stuff for a living now is that it was a way that i could build even bigger things than i could ever afford to do by myself, or in the time scales that a hobby would allow. So right now i feel for you all, who had developed plans around something that now isn't going to happen. I have been planning for some time the introduction of a "consumer" range of RGB pixel lights, that are better than your average "cheap" pixel, but are more cost effective than our commercial/industrial product line. This was planned for release in March this year, but events seem to have overtaken me, and i'm left wondering what i can do to help out my DIY friends. Its not practical nor economic to produce a rgb pixel light at the price point that has been bandied about. There is real costs in doing this properly, including the cost of licensing!

How can i help you guys? or how can you help me to help you.. If you've got a sensible suggestions then i'd be all ears.


Kind Regards

Andrew

[LabRat]

Uh... what coop? (I'm a "pixel lurker": reading and tracking but not having committed to anything) I'm curious as to what hardware was being considered, but isn't anymore.

[maffeirw]

As far as I can tell, RJ's controller coop is still on and he is arranging for us to buy the pixel strings, flexible led strips and pixel modules directly from the manufacturer. We'll just have to wait and see.

[bcstuff]

There will be no coop because RJ was unable to come to a fair agreement with Phillips due to their unreasonable licensing fees.

The information on where to buy the RGB pixels will be available to anyone after the people who have already purchased the controllers have time to place their orders.

So yes, there will be no Coop but you will be able to buy the RGB pixels anyways.

Source:
http://diylightanimation.com/index.php?topic=4723.msg79039#msg79039

If you are not referring to the coop mentioned above, my apologies.

Brian

[mrpackethead]

Folks, lets try to keep this on topic, THe fact is that the co-op failed, this is'nt about talking about why RJ's attempted Co-op failed, and to guess why it did, is'nt helpful for anyone.. I'm offering our resources to fill the gap if i can help any of the blinky flashy community who have been caught out by this.

[RJ]

Man if you can get them cheaper than the $.58 the users are going to pay for them as it stands and they are compatible with The Smart String controller that would be great! I am a little lost on the stranded comment since they have them avaliable cheaper than we were going to be able to do it on the coop?

I can not get them a better price for the user so this would help them greatly! let me know and instead of sending the users to the supplier I will send them to you to order, everyone likes to save money. I need to order a bunch myself.

RJ

[eroberts]

MPH and RJ,

Way to keep it civil.

I've been bouncing around DIYC, DIYLA, and other sites for the last few months - lurking and learning about RGB. If I've learned anything, there is some deep divide between some of the bigger players in this field that get aired on all of the major sites. (RJ this, MPH that, P something else...and more).

If anything came out of this whole initiave, it would be that, in the Christmas spirit, whatever issues everyone has with each other would get healed over. Healthy coopetition is great, but there have been some downright distateful things being said amongst the factions.

[griffixdc]

not to sound of topic.....i just would like to know which forums or forum have the best information to Learn rgb and pixeles/nodes i am having trouble getting it :(

[n8huntsman]

If you havn't seen these, I highly recommend them: http://www.holidaycoro.com/rgb/
From what I've seen the info is very segmented amongst the different sites. Being that RGB is relatively new and still in the works, answers are tough to come by. There isn't yet a nice document that lists all the options with pros and cons. I tried to put something together, but it was too diffucult to get all the info on the different controllers. I am by no means an expert but I have read a lot, just PM me or start a new thread if you have any questions and I'll try to answer or point you in the right direction. If someone wants my excel file and can help fill it out let me know. It might be useful in the wiki if we can get it populated, but right now, I just havnt been able to gather enough info.

[griffixdc]

duplicate sorry:(

[griffixdc]

If you havn't seen these, I highly recommend them: http://www.holidaycoro.com/rgb/
From what I've seen the info is very segmented amongst the different sites. Being that RGB is relatively new and still in the works, answers are tough to come by. There isn't yet a nice document that lists all the options with pros and cons. I tried to put something together, but it was too diffucult to get all the info on the different controllers. I am by no means an expert but I have read a lot, just PM me or start a new thread if you have any questions and I'll try to answer or point you in the right direction. If someone wants my excel file and can help fill it out let me know. It might be useful in the wiki if we can get it populated, but right now, I just havnt been able to gather enough info.

thanks for the assistance and answer to my question i will keep that in mind when i get more involved with RGB lights :)

[pmscientist]

Correct me if I'm wrong, but doesn't the issue w/the Phillips patent only apply when both the controller and pixels are sold by the same entity?

If so, clearly anyone who is interested can find links to inexpensive pixels in various DIY lighting forums. However, it seems like getting the pixels is one of the things that causes some people a bit of trepidation.

In the short term, I can think of two things that might help out. First, a buyers guide that familiarizes people with the process of purchasing pixels directly might help. Second, organizing a group buy would simplify the buying process, save some on shipping, and possibly get another cent or two off per unit if quantities were large enough. Heck, maybe it's a blessing in disguise. There are multiple form factors of the pixels available, which might get more people interested if they have multiple mounting options.

Unfortunately, those might not truly answer your question, MPH, as both of those might be hard for you given you sell licensed pixels and controllers.

I understand that BAM (Bit Angle Modulation) may provide an alternate and public domain method of performing a function similar to PWM, just without the bee ess. Certainly that would have to be implemented in a chip that replaces the 2801/1804/6803/etc., as that's where the PWM occurs. Since you probably keep a closer eye than most on what's available in that area, maybe you could find out if any of the manufacturers have plans to implement such a method, and if so, when they might become available?

[mrpackethead]

Correct me if I'm wrong, but doesn't the issue w/the Phillips patent only apply when both the controller and pixels are sold by the same entity?


I'm not going to attempt to make any comment on the ins and outs of what Philips Patents cover and dont cover, and what kind of thigns you can do to avoid the. I'll leave that to the professionals who we pay to keep ourselves out of trouble. We've done what I think we need to do.

[Entropy]

I think the issue is: To be sure of guaranteed legality, the price is going to be a lot higher, and MPH can offer a competitive solution in the 100% legal arena.

However, many people have successfully gone "under the radar" with small direct imports of assembled strands. This is of questionable legality but it has worked for some of the small noncommercial endusers here. It's just not an option for commercial entities like MPH or RJ.

[pmscientist]

I understand not commenting on what Phillips does or does not claim their patents cover.

I wasn't suggesting that anyone not in the chip design or manufacturing business work to actually create a chip free of patent concerns for those designing complete systems. I was suggesting that we as a community keep an eye out for solutions that may use a non-infringing technology such as BAM. It may also be worth making requests to Ray and other distributors for such products, and maybe those with manufacturer relationships could make similar requests on that end of the chain.

[dpitts]

Now I am confused. when I read RJ's post he stated he had already set up the ability for individuals to buy the Smart strings from the manufacturer and that he was going to post a link. He did not mention patents and all this legal stuff. So I am led to believe we can buy the strings. Is this not correct.

[Entropy]

You can buy the strings, but in some countries, using them may be infringing on Philips patents.

Companies tend to go after importers and resellers when asserting patents, so anyone in these positions has to be REALLY careful. In theory, if you ordered infringing products directly from Asia, you could wind up having them seized by customs, but it is highly unlikely the patent holders will go directly after you. (They could legally do so, but it generally isn't worth the lawyer fees on their end to attack end users.) I have not seen this happen to any endusers purchasing assembled strands, but MPH has indicated that some "bare chip" imports of WS2801s have been seized in the past.

In short - these strands will be obtainable directly from Asia, but doing so entails operating in a legal grey area. (Some people believe that the Philips patents don't cover these strings and thus they are fully legal, but most resellers are doing a "better safe than sorry" approach.)

[mrpackethead]

You can buy the strings, but in some countries, using them may be infringing on Philips patents.

Companies tend to go after importers and resellers when asserting patents, so anyone in these positions has to be REALLY careful. In theory, if you ordered infringing products directly from Asia, you could wind up having them seized by customs, but it is highly unlikely the patent holders will go directly after you. (They could legally do so, but it generally isn't worth the lawyer fees on their end to attack end users.) I have not seen this happen to any endusers purchasing assembled strands, but MPH has indicated that some "bare chip" imports of WS2801s have been seized in the past.

In short - these strands will be obtainable directly from Asia, but doing so entails operating in a legal grey area. (Some people believe that the Philips patents don't cover these strings and thus they are fully legal, but most resellers are doing a "better safe than sorry" approach.)

http://www.copynot.com has an interesting definition;

A patent is an intellectual property right, granted by a country’s government as a territorial right for a limited period. Patent rights make it illegal for anyone except the owner or someone with the owner’s permission to make, use, import or sell the invention in the country where the patent was granted Patents protect the features and processes that make things work. This lets inventors profit from their inventions. Patents generally cover products or processes that contain ‘new’ functional or technical aspects.

If i did'nt think I needed to get a licence from phillips to import and sell product i wouldn't have. Its not like i like sending them money. I've just accepted that this is something i need to do to be compliant with the law.

By not licencing them, i woudl be subject to the possiblity of substatial civil litigation and thats not somethign i want to do, and its certainly not somethign that i would want to potentially pass on liablity to my end users to either, and i certainly would'nt be telling my customers to do this either..

The questions that everyone needs to ask,

does the product fall under a valid patent?
do i need to get permission for it?
can i get away with it?
what would be the consequences if i got taken to task.

For me that was, Yes, Yes, No, Dire!

I aggree with Entropy, that they are unlikely to chase someone for several hundred dollars of imports, unless they decided that they wanted to make an example of someone..

[Entropy]

In the case of "does it fall under a valid patent" - Sometimes a person might think it does not, but the patent holder thinks it does. In this case, even someone who thinks they are legal might pay the fees just in case. (Easier than winding up in court, even if it makes you feel dirty.)

[djulien]

I think there are 3 categories of patents:
- genuine patents where the idea is unique and should be protected/licensed because the inventor spent considerable resources to reduce the idea into practice
- over-reaching patents, where the core idea may be a valid patent, but it has been over-generalized to engulf other ideas that should legitimately be separate
- noise patents, where the idea is so obvious or general that it cannot seriously be treated as a unique invention

For example, software patents often fall into the second or third categories.

I'm not familiar with the Phillips patent to know which category it falls under, but from:
http://diylightanimation.com/index.php?topic=4723.msg79039#msg79039


Lawyer wanted me to pay royalties for anything that might ever come close to being hooked to a Smart String including the ATX power supply you would purchase to power you units.

... it sounds like the second (or possibly third) category, but I guess the lawyers get to decide.

(sorry if that strayed off-topic)

It seems like there is enough expertise and talent, even fabrication resources, represented in this forum to be able to come up with an alternative, non-infringing design. This approach is typically used within the open source software community when a patent is found - they seem to be quite successful at reworking a design and separating out the patented technique from the non-patented stuff surrounding it.

What would be the desirable features for a DIYC-designed RGB pixel string? (such as physical size or shape, max string length, color depth, max updates per second, type of controller, price point, etc)?

Or is the main issue that this patented chip has such a low price point that, even with a group buy, we could not come close to that same price using alternate parts for the pixels? For example, isn't there another component or way to turn LEDs on and off rapidly so that we don't need to rely on a proprietary, patented chip? Maybe we can all come up with something non-infringing that would do the job. BAM was mentioned, but it seems like practically any on/off technique that happened at at least 65 Hz or so and added up to the correct on-vs-off ratio would work.

(sorry if I over-simplified; just trying to understand the issues)

don

[Entropy]

Unfortunately, there are enough patents within Category 2 to likely kill any attempt at making things non-infringing enough to escape any patents out there.

Anything not using one of the Asian chips is likely to explode in terms of cost per node. For example, my I2C-controlled project - it would be lucky to ever get that below $1/node, and I effectively gave up on the project for 2 main reasons:
1) I couldn't come close to the Asian Nodes in CPN, even without patent licensing fees
2) Mine were most likely infringing on two patents (not even counting the fairly broad-scope Philips and Color Kinetics patents on general RGB control) specific to I2C control - the patent inventor even stopped by to say hello. He effectively said he had no interest in lawsuits against DIYers, but it's the sort of thing that could never go commercial and I would have been nervous about even doing group buys of the design.
3) I had technical difficulties with I2C and Atmel's USI approach to I2C on lower-end AVRs. AVRs with hardware I2C would have killed my CPN.

[djulien]

Would averaging the cost of the uC over 2 or 3 nodes be acceptable? (ie, an Active/Passive arrangement where only every-other node has a uC, or passive-active-passive, where it is only on every third node) If so, then it might be possible to meet the target CPN approximately.

don

[Entropy]

Depends on the desired target display. Some people need strands, and having a controller per 2-3 nodes is a NIGHTMARE of wiring and weight for strands. But for some other display types (such as, for example, MPH's tubes), semicentralized control isn't so bad. At that point you find cheap chips domestically that are in a fairly improved situation as far as patents go. The NXP PCA9635, for example, is already licensed to use the I2C patents I mentioned above and hence readily available commercially in the USA. It permits pretty good CPN as the chip is a bit over $1 and can drive 5 RGB nodes with one free channel. It's a bit more questionable as to whether additional licensing of the Philips/CK "system level" patents is needed, but it puts you in a far better situation than the Asian nodes.

I'm still considering 1-2 years down the line doing some "tubes" using the PCA9635. (3-chips per 16 nodes config) However I'm massively behind on everything these days, and already have a major project in mind that will consume huge amounts of time if I ever get around to starting it. :) (Involves a huge learning curve, but that is one of the primary reasons for doing it - I want to get proficient with ARM Cortex-M3 micros and FPGAs. I've been wanting to get started with FPGAs for years, but last year was when affordable dev boards like the Papilio One started really hitting the market.)

[djulien]

I've been wanting to get started with FPGAs for years, but last year was when affordable dev boards like the Papilio One started really hitting the market.)

Are there any FPGAs that are inexpensive enough to use for an RGB node? Each node would just need a shift register, a self-running clock, and a few count-down counters, which could be done with the gates in a FPGA. I saw some listed on Mouser but they were up around $4 - $5 each, which is too much for an RGB node.

don

[P. Short]

I doubt that there are any FPGA's in the right price range, although there might be some CPLDs in the range. However, none of these are likely to be in a desireable package (i.e. small size, low pin count).

Another thing to look at might be some of the small Cypress PSOC parts, they have hardware blocks built in might fit the bill.

[Entropy]

My comment on a major project involving FPGAs wasn't in relation to their use in a node, so getting a bit offtopic in regards to this thread. I agree - FPGAs or CPLDs are NOT the choice for individual nodes, and I don't believe even a good choice for multichannel dimming.

They are, however excellent for problems that people here have been targeting with the Propeller (large-scale multichannel bitbanging), which is what I have in mind.

Basic idea: Construct a core for a Spartan-3E that exposes some BlockRAM banks inside the chip in a similar manner to SPI EEPROMs. Then construct cores that take segments of BlockRAM contents and output them in various protocols (DMX, WS2801, etc). That way the FPGA acts as a "hardware accelerator" for any chip that has a high speed SPI output. For this, I have a Luminary LM3S6965 in mind (50 MHz ARM Cortex-M3, integrated Ethernet MAC/PHY). It's a MAJOR project and I'm not sure if I even have the attention span to see it through. However, it covers the aspect of learning two things I've been wanting to learn lately. I know I'm much farther behind on starting this due to a variety of things, including the hard drive from hell (curse you and your wonky firmware WD) and a foray into Android porting. (There's an Android port for the Windows Mobile phone I use that has some issues, I've gotten involved in solving some of them until AT&T gets in gear and stops releasing Android phones crippled by either lowend hardware or backlevel software so I can get a native-Android phone.)

Getting back on topic:
For individual nodes, the various small 8-bit micros are your best bet. It's just not possible now for them to touch dedicated dimmer ASICs such as the WS2801 in price/performance.

[ags0000]

...
They are, however excellent for problems that people here have been targeting with the Propeller (large-scale multichannel bitbanging), which is what I have in mind.

Basic idea: Construct a core for a Spartan-3E that exposes some BlockRAM banks inside the chip in a similar manner to SPI EEPROMs. Then construct cores that take segments of BlockRAM contents and output them in various protocols (DMX, WS2801, etc). That way the FPGA acts as a "hardware accelerator" for any chip that has a high speed SPI output. For this, I have a Luminary LM3S6965 in mind (50 MHz ARM Cortex-M3, integrated Ethernet MAC/PHY). It's a MAJOR project and I'm not sure if I even have the attention span to see it through. However, it covers the aspect of learning two things I've been wanting to learn lately...

Now that is an intriguing idea. I've not had the opportunity to work with ARM cores or FPGAs, but have always wanted to. That sounds like fun. I've been going down the Propeller-based path (and have had fun learning along the way - which for me is at least as important as the end result) so far. The dedicated hardware approach would probably be easier to provide the maximum refresh rate possible (limited by the actual node signal timing/protocol) for an unlimited number of nodes (using parallel output blocks as needed for channels). Of course, putting Propellers in parallel (parallel Parallax Propellers?) would do the same thing. The Prop is pretty inexpensive (as a controller, not as a pixel device) at around $9. What would you expect the cost to be for the FPGA approach?

Oh, and on AT&T and Android: I've given up. I was an early adopter of one of their earliest Android native phones, the Motorola Backflip. It's really targeting a younger social-networking crowd, but regardless, it is a steaming heap of... junk. Without a doubt, the thing it is best at is testing me on a daily basis on my self-improvement efforts regarding patience, anger-management, and as Frank Costanza would say (actually, scream) "Serenity Now!".

[Entropy]

I think now that VZW has the iPhone, AT&T may finally shape up. (The Atrix 4G and Inspire 4G looks very promising, just waiting for a good phone with a hardware keyboard.)

The Backflip is a classic example of AT&T Android badness from 2010. While the entire year Verizon was machinegunning Droids all running Eclair or Froyo on high-end hardware, AT&T released crap like the Backflip (Backlevel Android - 1.6 I think?), Aria (HVGA screen in 2010???? You're joking!), and other crappy Motorola devices. (Between them sticking with MOTOBLUR and making it as hard as possible to install third-party ROMs, I will NOT be purchasing a Motorola Android phone in the near future.)

As to FPGAs - 100-pin XC3S250Es are $13, 144-pin ones are $15. Main limit on channel count is the number of BlockRAMs available - I think the 250E is going to be limited to around 8k or 16k channels. I need to fire up ISE and check when I get home. I can't remember if it was a 2k or 4k BlockRAM entity that ate up two of the 18 BRAM slices available. Some sort of flash memory is needed for configuration ($2-4 or so), but I think the Props also rely on external configuration memory? Biggest issue is that VHDL has a steeper learning curve than Props, and also the Props are available in DIP.

It may be in theory possible to use a MicroBlaze or similar CPU to avoid having the ARM to handle Ethernet stuff, but at $15 for an LM3S6965 (remember, the Prop Ethernet solutions need the fairly expensive Wiznet module), it's probably easier to just have a nice beefy ARM to handle the Ethernet side.

Main bottlenecks for lighting use would be:
1) Not enough BRAMs to support all that I/O - probably even a 100-pin package would have tons of unused I/O before you hit the BRAM limit
2) How fast you can get data into the BRAMs - Over short distances SPI seems to be capable of 10-20 Mbits/sec though. Haven't tested this yet though.

It looks like a project was just released that covers the first phase of my intended project - http://excamera.com/sphinx/gameduino/hardware.html#low-level-spi-interface - 32KB of BRAM is exposed via an SPI interface, exactly what I was planning on doing.

[djulien]

For individual nodes, the various small 8-bit micros are your best bet. It's just not possible now for them to touch dedicated dimmer ASICs such as the WS2801 in price/performance.

If the price of a low-end microcontroller is averaged over 2 - 3 nodes, it seems like it could hit that price. For example, a $0.50 PIC driving 2 nodes would be $0.25 per node, which is pretty close to the $0.20 that I saw mentioned for the WS2801. Are there other components in there that I overlooked that would push the price higher? (I assumed that the PCB, regulator, or other parts would be comparable cost).

don

[n8huntsman]

Dude, are you bumping all the threads, or just the ones I'm watching?

[tonyjmartin]

bump


Dude, are you bumping all the threads, or just the ones I'm watching?
Not unless you're watching the same twenty-four threads. Most of which were only stale by an hour or two (link (http://doityourselfchristmas.com/forums/search.php?searchid=21163)). :confused:

[n8huntsman]

LOL, I was actually watching about 6 of them. My email alerts go to my phone and blew up all at once.

[chilloutdocdoc]

If the price of a low-end microcontroller is averaged over 2 - 3 nodes, it seems like it could hit that price. For example, a $0.50 PIC driving 2 nodes would be $0.25 per node, which is pretty close to the $0.20 that I saw mentioned for the WS2801. Are there other components in there that I overlooked that would push the price higher? (I assumed that the PCB, regulator, or other parts would be comparable cost).

don

It may or may not work for smaller designs, and I'm not sure on the length, but what about a mid-range processor in the beginning of the chain, with a string of shift registers (or led drivers like the tl5940?) I've got a project on the back burner with somewhat this idea. I'm not sure the max distance or anything for them, but I know the arduino community likes them.

edit: what about something like the dumb (as he calls it, i think it's ingenious) matrix that djulien implemented?

[Materdaddy]

Dude, are you bumping all the threads, or just the ones I'm watching?

Glad I'm not the only one annoyed. Not that I want to see other's annoyed, but after seeing about 10 of the threads with nobody even mention the bumps, I was thinking it was only me.


Not unless you're watching the same twenty-four threads. Most of which were only stale by an hour or two (link (http://doityourselfchristmas.com/forums/search.php?searchid=21163)). :confused:

Your link is bad, the search re-directs to an "ID" that is cached for a short time, here's the link you want: http://doityourselfchristmas.com/forums/search.php?do=finduser&userid=2977&contenttype=vBForum_Post&showposts=1

[Entropy]

If the price of a low-end microcontroller is averaged over 2 - 3 nodes, it seems like it could hit that price. For example, a $0.50 PIC driving 2 nodes would be $0.25 per node, which is pretty close to the $0.20 that I saw mentioned for the WS2801. Are there other components in there that I overlooked that would push the price higher? (I assumed that the PCB, regulator, or other parts would be comparable cost).

don

While someone did have some success making a dimmer from a $0.50 PIC, they had serious issues with dimming consistency (IIRC, the dimming was in a free-running software loop, not triggered by a timer interrupt because that PIC didn't have one). Also the $0.50 PICs don't have enough I/O for more than one RGB note.

Once you start moving to centralized control, there are some good fairly inexpensive solutions such as the PCA9635. But you've now lost all of the advantages of nodes like the WS2801 in terms of wiring convenience.

[djulien]

While someone did have some success making a dimmer from a $0.50 PIC, they had serious issues with dimming consistency (IIRC, the dimming was in a free-running software loop, not triggered by a timer interrupt because that PIC didn't have one). Also the $0.50 PICs don't have enough I/O for more than one RGB note.


I hadn't considered timing drift (I assumed it would be small or could be synced to the clock signal).

I think that an 8-pin PIC actually could drive 2 RGB nodes (but not at 100% duty cycle all the time). Attached is an example of 2 nodes, using a 50% max duty cycle if both nodes are showing the same color.

The idea is that each I/O pin can turn on one of the colors in each node, but not at the same time (output low to turn on the active node, output high to turn on the passive node, or tri-state to turn them both off). The active node would use the leading portion of the refresh cycle, and the passive node would use the trailing portion. So each I/O pin goes through the following states during each refresh cycle: low (for active node's time), tri-state (while neither node wants that color), high (for passive node's time) - sort of a back-to-back or push-pull type of PWM. The 3 time slots would always add up to the total refresh cycle time.

As long as the active time + passive time for any individual color is <= 100%, the nodes get the colors they want. If the total on-time for an individual color adds up to more than 100%, the nodes need to compromise - the plug-in could enforce this, resulting in a slight loss of brightness in that particular case. However, LEDs like the PLCC6 5050s seem bright enough that if 2 side-by-side nodes are on then they really don't need to be on at full brightness anyway (imho), so a 50% duty cycle in that case is probably okay.

don

[djulien]

While someone did have some success making a dimmer from a $0.50 PIC, they had serious issues with dimming consistency

Do you remember who/what/where? I'd like to read up on what they did.

don

[Entropy]

I think it was Phil Short (not sure), but he never released anything because he never got it working consistently. Look through the archives in Development for my (at this point dead) I2C node project - he mentions it somewhere in that thread I think.

[dirknerkle]

I think it was Phil Short (not sure), but he never released anything because he never got it working consistently. Look through the archives in Development for my (at this point dead) I2C node project - he mentions it somewhere in that thread I think.

I think part of his trepidation with releasing it was that there were many exposed areas where high voltage/current existed and he didn't feel it was as safe as it should be if released to the general public.

[djulien]

I think it was Phil Short (not sure), but he never released anything because he never got it working consistently. Look through the archives in Development for my (at this point dead) I2C node project - he mentions it somewhere in that thread I think.

Thanks for the pointer. There's so many different ideas and parts in there that it's hard to decide which one is better. oh well.

don