Propeller (Prop) Dimmer Future

[DynamoBen]

A while back I designed a Propeller based dimmer. http://doityourselfchristmas.com/forums/showthread.php?t=7863&highlight=Propeller

When I started this project it was my intention to make a fully featured dimmer. However I also wanted to make a development platform which could be used to create smaller (lesser featured) dimmers based on the propeller.

Today I noticed that the DIP version of the Propeller is back in stock, which got me thinking. Is there any interest in a dimmer that has a subset of features or should I keep working on the bigger version? If I were to make a lesser featured dimmer what is most important cost, capabilities, SMT/DIP, ability to customize?

[budude]

That would be great - having a Prop driving a number of constant current driver chips sounds like a great way to get a high density RGB/LED controller. I don't see a real need for a new AC based design but that's just my own opinion.

The Prop IDE looks like a nice way to create code as well - I've been looking at the DMX code snippets on the Parallax site and it looks so easy to do a DMX RGB dimmer!

[DynamoBen]

That would be great - having a Prop driving a number of constant current driver chips sounds like a great way to get a high density RGB/LED controller. I don't see a real need for a new AC based design but that's just my own opinion.


Interesting idea. In regards to the code its object based so its easy to swap things in and out. At present the AC dimming code is separate from everything else so it would be fairly easy to swap it out with PWM code.

In regards to the DMX code, I've taken the objects from the parallax site and customized them. As we speak I'm looking into implementing RDM.

Bear in mind we don't have to do DMX, we could do any serial protocol (like renard) or ethernet. My end goal has always been to make the Prop a building block for projects in this community.

[DynamoBen]

With some mods this object could work. http://obex.parallax.com/objects/437/

[budude]

Interesting - so if I read the documentation right this would support 800 channels with PWM - but of course you'd need a pile of daughter boards to hold the 100 '595 chips! So basically this would be a Super Grinch with dimming and DMX? I'm not quite sure on their math to get to 100 though.

[DynamoBen]

...so if I read the documentation right this would support 800 channels with PWM - but of course you'd need a pile of daughter boards to hold the 100 '595 chips! So basically this would be a Super Grinch with dimming and DMX? I'm not quite sure on their math to get to 100 though.

Yep thats the way I understand it. The only real limit is DMX which is 512, which is only 64 ICs. ;)

What refresh rate is needed?

[DynamoBen]

The 4094s I'm using should work for this. By removing the "advanced" features it frees up enough pins to allow 224 channels of control (7 banks of 32). However if I were to build this I wouldn't do more than 192, most likely less. I have eight processors, each bank of 32 uses 1 processor, DMX uses 1 processor, and finally processor 0 is used to launch the code to those processors.

[budude]

I'm liking it! For my current project I just expanded on the current RenSS designs with 6 PICs for 48 channels - but a Prop is way cheaper than that let alone accounting for having 3x as many channels... The layout would be way easier also - a single chip (albeit a giant 40 pinner) instead of 24!. You've got the juices flowing now - I think I'll go shopping for some Props!

[DynamoBen]

Okay so if we are going to do this then we need to not only break this thread out, but more importantly figure out the final channel count. I would really rather be around 128, it keeps cog use down, makes the board more compact, and is a nice round number. Do you need to be higher than that?

[budude]

Well - the more the merrier for my needs specifically - but 128 is probably a good choice - perhaps the extra cogs can be used for other purposes down the road. Are there any tools that can measure the load/performance of the Prop while it's running? It might be good to know how much headroom there is left after loading up the 128 channels.

It might be nice if it could be designed in a modular approach where you add on output boards to the main board - kind of like the Helix or Firegod controllers but with other options like AC, lowDC, highDC, etc.

[DynamoBen]

Well - the more the merrier for my needs specifically - but 128 is probably a good choice - perhaps the extra cogs can be used for other purposes down the road. Are there any tools that can measure the load/performance of the Prop while it's running? It might be good to know how much headroom there is left after loading up the 128 channels.

It might be nice if it could be designed in a modular approach where you add on output boards to the main board - kind of like the Helix or Firegod controllers but with other options like AC, lowDC, highDC, etc.

We can run more if you like performance is not going to be a problem. Think of the prop like having 8 separate micro-controllers that all share the same RAM. The DMX processor grabs the data and stuffs it into RAM, each of the output processors (32 channels each) watch their part of RAM and set PWM rates.

When you say AC do you mean standard dimming?

[budude]

We can run more if you like performance is not going to be a problem. Think of the prop like having 8 separate micro-controllers that all share the same RAM. The DMX processor grabs the data and stuffs it into RAM, each of the output processors (32 channels each) watch their part of RAM and set PWM rates.

When you say AC do you mean standard dimming?

Yeah - standard dimming, ZC, etc... really more of a reference though to a daughter board approach where you break out XX channels out to a board that can have different AC or DC configurations like 16/24/32 channels at a shot. Just thinking out loud I guess on the best way to terminate the outputs...

[DynamoBen]

Yeah - standard dimming, ZC, etc... really more of a reference though to a daughter board approach where you break out XX channels out to a board that can have different AC or DC configurations like 16/24/32 channels at a shot. Just thinking out loud I guess on the best way to terminate the outputs...

Technically I already have all of this, at present its 64 channels. In essence what you are asking for is a higher channel count, DMX only version of what I have now. Correct?

[budude]

Technically I already have all of this, at present its 64 channels. In essence what you are asking for is a higher channel count, DMX only version of what I have now. Correct?

I followed your other dimmer - great project - but probably a little high end for the every-day DIYC'er (and I think you acknowledged that while you made it). I think DMX is the way to go myself. It looks like you went with standard SSR/RJ45 outputs but I'm thinking more like Greg did with Helix where you have a single run out to a daughter board with X channels on it. Just going by the group buys here and at DLA, it's pretty clear that off-board SSRs have their place but aren't nearly as popular as the Ren8/16/24 or Lynx Express.

Again - just thinking out loud and trying to think what would be most useful to the "collective".

[DynamoBen]

I followed your other dimmer - great project - but probably a little high end for the every-day DIYC'er (and I think you acknowledged that while you made it). I think DMX is the way to go myself.

First off that dimmer is using DMX, and has been running that way for months. Second I don't think its beyond the "every-day DIYC'er" because I am an every-day DIYC'er. I'm an an Electrical Engineer (I'm really a software tester) or an electronics expert by any means, I'm just a guy with a bench and soldering iron. Everything on that board was soldered by hand. I agree that the surface mount stuff may scare some and frankly that was me before I soldered up this board.



It looks like you went with standard SSR/RJ45 outputs but I'm thinking more like Greg did with Helix where you have a single run out to a daughter board with X channels on it. Just going by the group buys here and at DLA, it's pretty clear that off-board SSRs have their place but aren't nearly as popular as the Ren8/16/24 or Lynx Express.

Again - just thinking out loud and trying to think what would be most useful to the "collective".

My project is modeled after and in some cases a direct copy of what Greg did. The dimmer code to drive the outputs is the Helix dimmer code (prop users stick together).

What you are asking for is RS-485 outputs that go to channel drivers. Here is the deal when Greg designed the Prop128 and then Helix he didn't use support DMX to them. By using DMX you are functionally making each dimmer/controller its own daughter board since data, not dimmer outputs, is transmitted via RS-485.

If the limiting factor is the RJ45 connectors that can easily be changed to any style connector, that happens to be the convention here. In its current configuration the first bank of 4 outputs can be SSRs, next one can be relays, and the next can be PWM. If however the goal is to be pin compatible with the Helix, that could be done by adding RS485 transceivers and removing the on-board 4094s.

Maybe the compromise is to create a "Propeller DMX (RS-485) base board" and just have headers for the outputs. Then people can can connect any output (or input) boards they want, could be ULNs, 595s, 4094s. This would allow mixing and matching of output methods which would also allow people to customize the controller for their purposes. The only caution here is with flexibility comes complexity.

[budude]

First off that dimmer is using DMX, and has been running that way for months. Second I don't think its beyond the "every-day DIYC'er" because I am an every-day DIYC'er. I'm an an Electrical Engineer (I'm really a software tester) or an electronics expert by any means, I'm just a guy with a bench and soldering iron. Everything on that board was soldered by hand. I agree that the surface mount stuff may scare some and frankly that was me before I soldered up this board.

Yep - I know it was DMX. I'm really in the same boat - HW EE originally but am a Software Tester by trade currently... Just saying adding some of the bells/whistles might add cost that might not be a necessary addition in relation to the current Renard designs - no disrespect meant of course.



My project is modeled after and in some cases a direct copy of what Greg did. The dimmer code to drive the outputs is the Helix dimmer code (prop users stick together).

What you are asking for is RS-485 outputs that go to channel drivers. Here is the deal when Greg designed the Prop128 and then Helix he didn't use support DMX to them. By using DMX you are functionally making each dimmer/controller its own daughter board since data, not dimmer outputs, is transmitted via RS-485.

If the limiting factor is the RJ45 connectors that can easily be changed to any style connector, that happens to be the convention here. In its current configuration the first bank of 4 outputs can be SSRs, next one can be relays, and the next can be PWM. If however the goal is to be pin compatible with the Helix, that could be done by adding RS485 transceivers and removing the on-board 4094s.

Maybe the compromise is to create a "Propeller DMX (RS-485) base board" and just have headers for the outputs. Then people can can connect any output (or input) boards they want, could be ULNs, 595s, 4094s. This would allow mixing and matching of output methods which would also allow people to customize the controller for their purposes. The only caution here is with flexibility comes complexity.

Agreed - again my point was to raise the discussion specifically around this exact point - basically - where do the outputs end? I am all for flexibility and if I was laying out something for myself I would go in some particular direction but it may not be useful for anyone else. Just looking at if a compromise between flexibility and reducing the cable count a bit should be discussed. How that's accomplished - I'm completely open to.

[DynamoBen]

Agreed - again my point was to raise the discussion specifically around this exact point - basically - where do the outputs end? I am all for flexibility and if I was laying out something for myself I would go in some particular direction but it may not be useful for anyone else. Just looking at if a compromise between flexibility and reducing the cable count a bit should be discussed. How that's accomplished - I'm completely open to.

I'm stripping all the fun stuff out of my other schematic to create a "base board." I will post something once I'm done. The thought is to make a building block and let folks like yourself take it where you want. I'm thinking this might be similar to the arduino shiled model but the base board would have DMX, zero-cross, and programming hardware.

[Matt_Edwards]

I would be interested in joining the coop if it does get off the ground, so I can learn about Propellers. they seem to be really feature rich, the environment and support are excellent.

[DynamoBen]

I would be interested in joining the coop if it does get off the ground, so I can learn about Propellers. they seem to be really feature rich, the environment and support are excellent.

Sweet! I'm drawing as quickly as I can, I'm keeping cost over features on the forefront this time.

[budude]

I like the "shield" idea a lot - I know it's been touched on in other projects but this really makes sense. A few different break-out boards for various output configurations and you're good to go. What are you planning to pass through the pins (besides the outputs obviously)?

I ordered a couple Props and a "plug" today - looking forward to joining the "party"! The timing is good - I'm at the crossroads for a couple projects and this may factor into it now.

[DynamoBen]

Here is the first pass of the base board. I still have a direct AC connection to power it the board (transformer and what not). However I'm thinking about removing all of that and requiring users to either attach an external DC or AC (standard dimming) power supply. Thoughts?

Also I could breakout the addressing portion to a header or DIP, for those that like the switches can use them for those that don't its just I/O. Thoughts?

Finally I swapped out the high speed USB programming port for low speed serial. The high speed chip was Surface Mount and a touch pricey, serial is cheaper. In either case it avoids needing to buy a prop plug and allows for serial to program and diagnose. I could ditch the whole thing and make ppl buy a prop plug or use an assembled FT232 board (sparkfun or DLP design). Thoughts?

OLD FILES DELETED

[budude]

Here is the first pass of the base board. I still have a direct AC connection to power it the board (transformer and what not). However I'm thinking about removing all of that and requiring users to either attach an external DC or AC (standard dimming) power supply. Thoughts?

Also I could breakout the addressing portion to a header or DIP, for those that like the switches can use them for those that don't its just I/O. Thoughts?

Finally I swapped out the high speed USB programming port for low speed serial. The high speed chip was Surface Mount and a touch pricey, serial is cheaper. In either case it avoids needing to buy a prop plug and allows for serial to program and diagnose. I could ditch the whole thing and make ppl buy a prop plug or use an assembled FT232 board (sparkfun or DLP design). Thoughts?

Are you planning to keep the QF package or would it be the PDIP40? I'm OK with either but it may scare off some.

I know the power supply section adds bulk/cost but I think it makes things less error prone to keep it with standardized components. There always seems to be a lot of discussion on the controllers that do not have onboard PS about what to use, etc whereas there's no questions (obviously) on the ones that do. Can it be made to support 240/50Hz? I guess that would affect the ZC bits and associated code as well but I think Greg's code allows 50Hz operation.

Addressing - I think the header would work fine - really whatever is the cheapest since you don't mess with that too much. Would this be to program it initially or used during power up each time?

I like the low-speed serial idea - super cheap and again, it isn't used much/often so why have a 'plug' sitting in your drawer for 99% of the time. I haven't read up on the plug interface - this SM device - it does the USB and level translation?

[Matt_Edwards]

Keep the uC SMD.

I would leave the AC Transformer off the PWB. That allows a lot more global options. Also for development purposes I tend to use a smallish transformer purely because they are always easier to handle. Production would use a correctly dimensioned/spec'd transformer.

Leave the switches in, but having a break out/expansion port is a good option.

As for the USB / serial port, I have never been a fan of the discrete component serial conversion ccts. I would prefer either a USB "Shield" or just a 4pin SIP header.

[DynamoBen]

Are you planning to keep the QF package or would it be the PDIP40? I'm OK with either but it may scare off some.

The QF is smaller and is in abundant supply however I could see how it could scare ppl off. I'm indifferent it comes down to board size and layout but I suppose we should design for DIP.



I know the power supply section adds bulk/cost but I think it makes things less error prone to keep it with standardized components. There always seems to be a lot of discussion on the controllers that do not have onboard PS about what to use, etc whereas there's no questions (obviously) on the ones that do. Can it be made to support 240/50Hz? I guess that would affect the ZC bits and associated code as well but I think Greg's code allows 50Hz operation.

You are correct in that the code will support 50 and 60Hz. However my current power supply design is for 120 only. I ditched the ability to go 240 largely because of trace clearances and the fact I couldn't test it. I'm open to suggestions from the 240 folks.

If we keep the power supply it makes the board bigger and taller (tallest part is the transformer), and adds cost. Again I'm indifferent here.



Addressing - I think the header would work fine - really whatever is the cheapest since you don't mess with that too much. Would this be to program it initially or used during power up each time?

It would be for addressing only and I would write the code in a way that it would be checked periodically.



I like the low-speed serial idea - super cheap and again, it isn't used much/often so why have a 'plug' sitting in your drawer for 99% of the time. I haven't read up on the plug interface - this SM device - it does the USB and level translation?

The prop plug is just an FTDI FT232, it converts USB data to TTL serial...no biggy. I think the serial port should still have a DB9, I think the Halloween crowd might use it.

Finally do we want data isolation, I removed it for cost reasons but could put it back.

[DynamoBen]

Keep the uC SMD.

It really wasn't hard to solder by hand, and keeps things small.



I would leave the AC Transformer off the PWB. That allows a lot more global options. Also for development purposes I tend to use a smallish transformer purely because they are always easier to handle. Production would use a correctly dimensioned/spec'd transformer.

So 1 vote to keep the transformer, and now 1 to ditch it. I kind of think ditching it might be the best bet, and maybe even safer. But it does open up us to lots of questions from more novice users.


Leave the switches in, but having a break out/expansion port is a good option.

Done, if someone hates them then they can skip populating it and use the header for expansion.



As for the USB / serial port, I have never been a fan of the discrete component serial conversion ccts. I would prefer either a USB "Shield" or just a 4pin SIP header.

If the FTDI chip weren't so small I would just keep it. I did hand solder it but I used the flood and wick method, not for the faint of heart. I can do a SIP for those that have a prop plug and do a "shield" for those that don't. Does that work? (I might even include the FTDI part, for myself)

[Matt_Edwards]

If the FTDI chip weren't so small I would just keep it. I did hand solder it but I used the flood and wick method, not for the faint of heart. I can do a SIP for those that have a prop plug and do a "shield" for those that don't. Does that work? (I might even include the FTDI part, for myself)

I am happy soldering the FTDI chips. I still have a few FT232 from an old project. I don't flood and de-solder any more. Grab some solder paste and hot air re-flow them. It is a snack. If this is an option, I'm cool. Another thing to consider is use a Mini USB connector, they save a lot of space.

[DynamoBen]

Another thing to consider is use a Mini USB connector, they save a lot of space.

They also tend to break off...Been there done that.

[budude]

wrt to the transformer - yes it adds cost but it is for 128 channels worth of controller and you're probably going to end up buying some type of transformer anyway - why not get the one that everyone will be using and be sure it works...

wrt to SM - dunno - the Prop is not too scarey and is probably doable so again - don't care either way but I think the PDIP would have more acceptance. The FTDI chip - that one I'm not sure on - I understand how it would make it simpler to use but I'd stick with what could be built with a good soldering iron and not resort to having to buy a reflow oven/tool. If you're going to go the USB route I would go with the DIP module - it's easy to install.

[DynamoBen]

wrt to the transformer - yes it adds cost but it is for 128 channels worth of controller and you're probably going to end up buying some type of transformer anyway - why not get the one that everyone will be using and be sure it works...

Matt_Edwards if you were to go off board what would you use for power? What global options are you speaking of?


wrt to SM - dunno - the Prop is not too scarey and is probably doable so again - don't care either way but I think the PDIP would have more acceptance. The FTDI chip - that one I'm not sure on - I understand how it would make it simpler to use but I'd stick with what could be built with a good soldering iron and not resort to having to buy a reflow oven/tool. If you're going to go the USB route I would go with the DIP module - it's easy to install.

This would be the comprimise but coudl be a pain for the international folks.
http://www.sparkfun.com/commerce/product_info.php?products_id=718

[Matt_Edwards]

Matt_Edwards if you were to go off board what would you use for power? What global options are you speaking of?

Any E-Core or Toroid transformer. Altronics has good ones. Locking in a PCB transformer virtually means either you add multiple component parts (like RPM's redo of his Grinch Dimmer) or we just don't fit the trany, mount one off board any way and waste extra space.
The other advantage to off board is the use of AC wall wart to beginners. there is no mains wiring in the controllers are all.



This would be the comprimise but coudl be a pain for the international folks.
http://www.sparkfun.com/commerce/product_info.php?products_id=718
They will be OK to. just choose one and the will us it. You could of course add both, the pads for those that want to SM Solder, and pins for those that will buy in the break out board. I would prefer the FTDI Quad packs (selfishly have those and not the SOICs) but will use what ever you put in

[DynamoBen]

Any E-Core or Toroid transformer. Altronics has good ones. Locking in a PCB transformer virtually means either you add multiple component parts (like RPM's redo of his Grinch Dimmer) or we just don't fit the trany, mount one off board any way and waste extra space.
The other advantage to off board is the use of AC wall wart to beginners. there is no mains wiring in the controllers are all.

So I guess the compromise is to not include the AC side. We could however create a "standard" 120V AC board for the US folks, and if the 240V folks want to do the same they can.



They will be OK to. just choose one and the will us it. You could of course add both, the pads for those that want to SM Solder, and pins for those that will buy in the break out board. I would prefer the FTDI Quad packs (selfishly have those and not the SOICs) but will use what ever you put in

If I add pads it will be for the SOIC, largely because the Quad have a bunch of passives and a crystal whereas the SOIC is all inclusive.

[DynamoBen]

In summary:

Surface Mount Prop
no AC onboard (transformer)
USB interface (onboard with pads and "shield"), USB mini connector
SIP connection to TTL (inline with USB)
Addressing via dip-switches with header for general I/O

What about DMX isolation? By using the Surface Mount chip it makes replacing a fried one a bit more difficult.

[Matt_Edwards]

In summary:

Surface Mount Prop
no AC onboard (transformer)
USB interface (onboard with pads and "shield"), USB mini connector
SIP connection to TTL (inline with USB)
Addressing via dip-switches with header for general I/O

Sounds good.


What about DMX isolation? By using the Surface Mount chip it makes replacing a fried one a bit more difficult.
Isolation won't help replacing a fried chip;)
Isolation is a good benefit if we can maintain cost control.
I take it you are suggesting if we don't go SMD for the 485 driver, we won't need isolation?

[DynamoBen]

Isolation won't help replacing a fried chip;)

True just fewer and lesser expensive chips.



Isolation is a good benefit if we can maintain cost control.

The expensive part is the DC/DC transformer. Oh, and of course board space.



I take it you are suggesting if we don't go SMD for the 485 driver, we won't need isolation?

Personally I like isolation regardless, however its easier to replace a DIP. With that being said a prop does cost around $8, its worth protecting.

I guess now is the time to start running some numbers. *sigh* I hate making BOMs.

[DynamoBen]

As I'm researching and updating the schematic I'm thinking about going with the DLP Design USB interface. Its a few dollars more but has the same footprint as a wide DIP package. Also it should allow most ppl to single source the parts.

http://www.dlpdesign.com/usb/usb232r.shtml

The only downside to a shield is cost, it ends up being the most expensive part on the board.

[DynamoBen]

I worked up a quick BOM. This has no transformer (but does have rectifier), on-board USB (no shield), isolation for DMX, and addressing. This BOM includes all the part costs except for I/O connectors.

Looks like its going to be near the $50 mark for parts alone, is this a problem?

For USB shield:
Rremove FTDI232RL, SN74LVC2G07, and mini USB connector.
Add shield at $20

Non-Isolated:
Remove NMR100C DC/DC, 6N137, and a few resistors

[budude]

I say keep the isolation - if I'm driving 128 channels I want to make sure it stays up. 16 channels I could afford to lose in a show (maybe) but not 128 so it would be good to make it more bulletproof.

I've already voted for the shield - a bit more cost but proably a lot less frustration in the end - easy to replace if necessary also...

You'll want to have per-channel (or at least 4-channel blocks) sink/source settings to support active low (e.g. standard SSRs) or active high (e.g. ULN2803) outputs - not sure if you mentioned this already...

[DynamoBen]

I say keep the isolation - if I'm driving 128 channels I want to make sure it stays up. 16 channels I could afford to lose in a show (maybe) but not 128 so it would be good to make it more bulletproof.

No problem. The reason I mentioned it is because besides the shield this is the next expensive portion of the circuit ($8 for DC/DC).



I've already voted for the shield - a bit more cost but proably a lot less frustration in the end - easy to replace if necessary also...

It doesn't cost anything to add a location for the shield, can do.



You'll want to have per-channel (or at least 4-channel blocks) sink/source settings to support active low (e.g. standard SSRs) or active high (e.g. ULN2803) outputs - not sure if you mentioned this already...

At present everything is broken out with the intention of shift registers, or similar, to be added. So really its all up to firmware and the I/O boards at that point for functionality.

The only direct I/O access is the address pins.

[DynamoBen]

I don't have the USB shield yet but should soon. I wanted to post what I had so far.

OLD FILES DELETED

[n1ist]

Looks good. In ControllerDMX, DMX_RO should be pulled up to 3.3V, not 5V. Also, J1 should be labelled Terminate instead of PROG, and there's something strange with the connection dots between pins 7 and 8 of the RJ45s.

/mike

[DynamoBen]

Looks good. In ControllerDMX, DMX_RO should be pulled up to 3.3V, not 5V. Also, J1 should be labelled Terminate instead of PROG, and there's something strange with the connection dots between pins 7 and 8 of the RJ45s.

/mike

All items have been corrected on my local copy. BTW the 3.3 vs 5 mistake exists on my other controller, I just I'll be cutting some traces. ;)

[n1ist]

You are probably OK with the series resistors limiting current into the ESD diodes on the processor. No need to cut traces, just load R26 sideways one one pad (if it fits) and flywire the other end - the output of the 6N137 is open-collector.
/mike

[DynamoBen]

You are probably OK with the series resistors limiting current into the ESD diodes on the processor. No need to cut traces, just load R26 sideways one one pad (if it fits) and flywire the other end - the output of the 6N137 is open-collector.
/mike

Now that I look at it again, I did that on purpose no fix is needed. According to the datasheet Vcc must be driven by 4.5 to 5.5V, hence the 5V rail and 1K resistor in series (keeps voltage in safe range for uC).

[n1ist]

I have just looked a bit more at the 6N137, and the data sheets are a bit muddy... Vcc is 5V. The output, however, is open collector (or open drain in one version), so you should be able to pull it up to 3.3v and have it work fine. All of the data sheet examples have the output pulled to 5V, but that may just be because the data sheets were written in the old TTL days, so nobody thought of level shifting to 3.3v.

/mike

[DynamoBen]

I have just looked a bit more at the 6N137, and the data sheets are a bit muddy... Vcc is 5V. The output, however, is open collector (or open drain in one version), so you should be able to pull it up to 3.3v and have it work fine. All of the data sheet examples have the output pulled to 5V, but that may just be because the data sheets were written in the old TTL days, so nobody thought of level shifting to 3.3v.

Well maybe its best to leave it a 5V, its working on my current project w/o fail.

[j1sys]

Ben -

I've been following (as best I can) your thread. Not being a prop-head makes it a little hard.

A couple of questions about your design for the DMX section:

1) can i copy it ?;);) Seriously I am working on a parallel project that will look very similar, don't think i copied it exactly. I like your isolation design.

2) I'm assuming that R15, R16, R21 are for biasing the line for RDM specs. But according to the wiki (i've not yet paid for the official spec)
First, a controller’s (console’s) output is terminated. Second, this termination must provide a bias to keep the line in the ‘marking state’ when no driver is enabled.
The reason for the additional termination is that a network segment will be driven at many points along its length. Hence, either end of the segment, if unterminated, will cause reflections.
A DMX console’s output drivers are always enabled. The RDM protocol is designed so that except during discovery, there should never be data collisions. To assure this lack of collisions, while making possible implementation on different platforms, there are times when all line drivers are required to be disabled. If nothing more than the termination was done, the line would float to some unknown level. In that case one or more random changes might be read on the line. These random changes greatly decrease system accuracy. So the biasing of the line is required
To assure this, section 2.4.1 (Line Bias Networks) of the standard says; “The command port shall provide a means to bias the termination of the data link to a value of at least 245 mV and verified by using the test circuit described in Appendix F.” Are you planning on this being the controller (console) on the bus? Can/should multiple devices be supplying the bias?

Just trying to understand DMX RDM interfacing better.

-Ed

[DynamoBen]

1) can i copy it ?;);) Seriously I am working on a parallel project that will look very similar, don't think i copied it exactly. I like your isolation design.

That would make yours a copy of a copy. I copied it from another users DMX dongle project, so I have no issues with that.



2) I'm assuming that R15, R16, R21 are for biasing the line for RDM specs. But according to the wiki (i've not yet paid for the official spec) Are you planning on this being the controller (console) on the bus? Can/should multiple devices be supplying the bias?

From the standard:
"This standard does not mandate a specific bias circuit."
(What I have mimics their command port reference circuit)

However you make a good point, this project is not a command port so I need to remove the biasing. Updated my local copy of the schematic.

[j1sys]

Thanx, I didn't know if I was missing something on the biasing.

[DynamoBen]

Thanx, I didn't know if I was missing something on the biasing.

My other project can be run as a command port, I just copy pasted. :rolleyes:

[DynamoBen]

Question:
At present I have ZC sense tied in prior to an on-board bridge rectifier circuit. Would it make more sense to ditch the rectifier circuit and run the ZC to its own header like the renard?

By doing this the rest of the board would just need a standard DC wall wart or equiv. If I don't do this the board would require an AC power supply for phase angle dimming.

[n1ist]

You could leave the rectifier (then the input is AC or DC, and not polarity sensitive), add the ZC connector and a 2-pos jumper block to select whether you want the ZC to come from the main input or ZC connector. That way you could set the board up to work either way, and don't need to use a regulated brick if you are running DC. The only drawback would be the 1.4V drop from the bridge when feeding in DC.
/mike

[DynamoBen]

The only drawback would be the 1.4V drop from the bridge when feeding in DC.


...and the additional parts and board space. I'm also a bit concerned about confusion. The renard already exists as a model. Also don't forget by removing the bridge I would remove the large cap on the incoming power.

Are there any advantages one way or the other internationally?

I see this board typically acting as a sister board sitting on top of a PCB with I/O and power on the motherboard. Or it can act as a building block tethered to other boards, which is how you are thinking of it.

[DynamoBen]

Well - the more the merrier for my needs specifically - but 128 is probably a good choice - perhaps the extra cogs can be used for other purposes down the road.


n1ist posted in another thread about an LED matrix which got me thinking.

May counts assumed that each I/O pin when to one LED. What if instead these pins went a matrix then an LED? Instead of 32 LEDs you could drive 256, it would take some crafty programming to do it but in theory is should work.

http://www.tigoe.net/pcomp/code/wp-content/uploads/2009/03/ledm88g_circuit.gif

n1ist project:
http://diylightanimation.com/index.php?topic=375.0

[DynamoBen]

Design Change:

I've made it about 1/2 through the RDM code and was considering the interaction between the dip switches addressing and RDM. At this point I'm thinking about getting rid of the dip-switches and using the serial interface (or RDM) as the sole method for addressing and configuring this controller.

Since we have to have a serial interface to load firmware we might as well use it for configuration. I will obviously need to come up with a simple serial protocol. My thought at this point is to make it easy enough to configure with hyperterm and a basic standalone app (I may need some assistance with the software part).

Any objections or concerns?

[BH]

HI, Don't know which RDM interafce you are using to test with but would like to recomend the DMX-TRI by JESE. Been using it myself to help develop some PIC based DMX stuff and have found it to be very reliable, not done any RDM myself yet except try a bit of off the shelf kit to see what its all about. Look foward to seeing the dimmer in action.

[DynamoBen]

HI, Don't know which RDM interafce you are using to test with but would like to recomend the DMX-TRI by JESE. Been using it myself to help develop some PIC based DMX stuff and have found it to be very reliable, not done any RDM myself yet except try a bit of off the shelf kit to see what its all about. Look foward to seeing the dimmer in action.

I have a homebrew method for testing at the moment but will add it to my resources. Thanks.

[DynamoBen]

Sorry folks I've been out of town for 10 days so I've not done much with this. I have the schematic mostly complete, I need to add the shield and header for serial. I also have a big start on the PCB.

Is there anything you would like to have me add or consider at this point?

[DynamoBen]

Here is the latest schematic, which should have everything on it. Comments or changes? BTW pinouts may change a little as I lay out the board.

NOTE: If you look at the USB schematic you will notice a SN74LVC2G07 which happens to be a SMD part. This exists to stop a known issue with the serial interface that prevents the prop from booting properly. If you look closely you will noticed that it is required even when you use the shield. Not sure if this is a deal breaker.

If this is a deal breaker a pin-through part will need to be sourced (of which I'm not aware) or you have to live with the issue if you use the shield.

OLD FILES DELETED

[n1ist]

Ben-
I just noticed that you are chaining the 3.3v regulator from the 5v rail. I seem to remember that there was some issues doing that on the Lynx. The 3.3v reg needs 4.75v min to regulate, and the 5v specs 4.8v as its min output, so it looks OK, but they had some low-temperature issues where the 3.3v dropped out of regulation and the processor didn't like it. If you aren't pulling much 3.3v, maybe you should run that regulator from Vin instead.
/mike

[DynamoBen]

Ben-
I just noticed that you are chaining the 3.3v regulator from the 5v rail. I seem to remember that there was some issues doing that on the Lynx. The 3.3v reg needs 4.75v min to regulate, and the 5v specs 4.8v as its min output, so it looks OK, but they had some low-temperature issues where the 3.3v dropped out of regulation and the processor didn't like it. If you aren't pulling much 3.3v, maybe you should run that regulator from Vin instead.
/mike

That's how this projects big brother is wired; change has been made. Thanks.

[j1sys]

Ben -

I've been programming all day and my eyes are bleary but is it my imagination or do the links to the schematics not match the content?

like main has dmx, and dmx has usb, and usb has main or something?

-Ed

[DynamoBen]

Ben -

I've been programming all day and my eyes are bleary but is it my imagination or do the links to the schematics not match the content?

like main has dmx, and dmx has usb, and usb has main or something?

-Ed

You're right, issue has be corrected. Sorry.

[DynamoBen]

You could leave the rectifier (then the input is AC or DC, and not polarity sensitive), add the ZC connector and a 2-pos jumper block to select whether you want the ZC to come from the main input or ZC connector. That way you could set the board up to work either way, and don't need to use a regulated brick if you are running DC. The only drawback would be the 1.4V drop from the bridge when feeding in DC.
/mike

You will notice in the latest drawing I forgot to add this. I will try to do that this week. However as I'm thinking about it, if a user wants low-voltage ZC they would not only need to move the jumper but also select the correct resistors. This is all well and good until you use a different voltage power supply then you are desoldering. With high-voltage ZC only this issue is mitigated.

I'm now leaning toward high-voltage ZC and ditching the rectifier. Thoughts/comments?

[DynamoBen]

Hmm not hearing a lot of feedback on this as of late. I have a few open items that I need info on.

1. ZC should it be high voltage, low voltage, or both?
2. What style connector do we want to power the board? Headers, screw terminal, barrel (or some combo)
3. Do we like the pin layout for I/O or would it make more sense to combine all the I/O into one or a few connectors (blocks of 8 for example)?
4. What style connectors do we want for I/O? headers, screw terminals, etc.

In regards to I/O I have them laid out in blocks of 3 (see drawing) I assume that a shift register is attached to each. However for the Halloween crowd this could be limiting, it might make more sense to use headers this would in turn make the board more compact. Thoughts other ideas? In the end I want this as flexible as possible but not complicated to use.

[Matt_Edwards]

Hmm not hearing a lot of feedback on this as of late. I have a few open items that I need info on. Sorry I have been a bit busy too.

1. ZC should it be high voltage, low voltage, or both?
I would use low voltage (and add a second set of resistor to a 2 pin connector for High voltage only of space allows)


2. What style connector do we want to power the board? Headers, screw terminal, barrel (or some combo)

I am partial to a Molex 2 pin KK and Terminal block combo. (I made a Component with 4 pads to suit for my SPI SSRs)

3. Do we like the pin layout for I/O or would it make more sense to combine all the I/O into one or a few connectors (blocks of 8 for example)?
Keep them on separate connectors


4. What style connectors do we want for I/O? headers, screw terminals, etc.
Your call. I use Molex KK range, as you can use both Crimp and IDT versions
IDC headers are cheaper and more readily available. For Shift Register - NOT TBs


In regards to I/O I have them laid out in blocks of 3 (see drawing) I assume that a shift register is attached to each. However for the Halloween crowd this could be limiting, it might make more sense to use headers this would in turn make the board more compact. Thoughts other ideas? In the end I want this as flexible as possible but not complicated to use.

A header connector will only work for all is you repeat GND, 3.3v and 5v in every group so the Ribbon cable can be separated into 6 cores for send to the Shift registers.

[DynamoBen]

I am partial to a Molex 2 pin KK and Terminal block combo. (I made a Component with 4 pads to suit for my SPI SSRs)

Keep them on separate connectors



I want to clear up a design assumption I've been working from. I've working with the assumption that this board could act as a daughter board since it has no real I/O OR be wired to satellite boards via some sort of connector(s).

Should I be accommodating both or is it too much work for too little effect?

[Matt_Edwards]

I want to clear up a design assumption I've been working from. I've working with the assumption that this board could act as a daughter board since it has no real I/O OR be wired to satellite boards via some sort of connector(s).

Should I be accommodating both or is it too much work for too little effect?

I think we are on the same page. I am assuming board board will be a self contained processor with no Real I/O. (The closest analogy is the Grinch Dimmer) SSRs will require an additional interface card like a Grinch.

[DynamoBen]

I was at a regional propeller convention yesterday and was talking about my project(s) with the owner of Gadget Gangster (http://gadgetgangster.com/). I mentioned that folks here are a bit timid about SMT parts but really interested in DIY and these project(s). He suggested doing either full or partial kits through his website. A partial kit could have the SMT parts already soldered but you purchase the balance, or a full kit comes with everything. By doing this the overall expense goes up but might be a workable solution for those intimidated by SMT or that want to purchase a kit year round.

He and I are going over a number of other ideas and the details. This is all still a germ of and idea but I wanted to get some feedback from the group. Thoughts?

[dmcole]

He suggested doing either full or partial kits through his website. A partial kit could have the SMT parts already soldered but you purchase the balance, or a full kit comes with everything. By doing this the overall expense goes up but might be a workable solution for those intimidated by SMT or that want to purchase a kit year round.

Last year I had suggested doing a co-op for an FM transmitter based on the NS73M which comes in only an SMT part (see http://www.mikeyancey.com/FM-Stereo-Broadcaster.php), using just this idea: find a company that could do the PCB and attach the SMT part(s), and doing a kit for the through-hole parts. My guess was that in quantity it would end up being between $40-$50, which is a little more than an Vastelec, but that we could provide an easy RDS interface and would have a little more control over availability.

Alas, my skillz are so meager that I could never have put this all together, so nothing further came from it.

But I still think the idea of designing SMT circuits and finding a supplier who would make the PCBs and attach the SMT parts is a good one while we as a hobby get over the hump of doing our own SMT soldering. It will help create demand for boards that might otherwise either not be built or end up with too small of an audience to end up with good coop prices.

Ben, you might consider breaking this out as a new, stand-alone thread in order to get sufficient input.

\dmc

[DynamoBen]

Ben, you might consider breaking this out as a new, stand-alone thread in order to get sufficient input.


Actually I want to leave it here because Gadget Gangster really only deals with a limited number of parts. I agree this is a bigger topic (which I'm discussing elsewhere) but for now I want to limit it to this project specifically.

[dmcole]

Actually I want to leave it here because Gadget Gangster really only deals with a limited number of parts. I agree this is a bigger topic (which I'm discussing elsewhere) but for now I want to limit it to this project specifically.

As always, I didn't understand ... :( ...

\dmc

[DynamoBen]

As always, I didn't understand ... :( ...

\dmc

It was a good point. :) No worries.

[n1ist]

Sparkfun has an NS73M breakout board.
/mike

[dmcole]

Sparkfun has an NS73M breakout board.


Yes, I know that. My devious plan was to cut Sparkfun out of the equation entirely, which I hoped would drive down the overall cost.

But, it's not to be.

\dmc

[DynamoBen]

BTW I worked on the PCB some more over the weekend, which required some schematic changes. Its about 85% complete, I also added the headers for a network interface. While I know we aren't using it now it only increased the board size by 1/2 inch, and it doesn't cost anything to add it. So at the moment you can connect via USB, DMX, or Ethernet...depending on the code of course.

I've also been doing some research on port expansion the current design assumed the use of shift registers. I've been reading up on the MCP23016 which seems to be a nice alternative if you want the ability to set inputs or outputs on a pin-by-pin basis. Just thinking ahead.

[DynamoBen]

Here is the latest DMX Prop Controller Schematic (rev D).

Updated to rev E

[DynamoBen]

Here is the latest Ethernet Prop Controller Schematic (rev A).

Updated to rev E

[dmcole]

Here is the latest DMX Prop Controller Schematic (rev D).

FWIW, the DC/DC converter you spec'd -- the NMR100C from Murata -- seems to be in short supply these days. Mouser is quoting 16 weeks out and Octopart says only Farnell and Avnet (?) have them in stock.

\dmc

[DynamoBen]

FWIW, the DC/DC converter you spec'd -- the NMR100C from Murata -- seems to be in short supply these days. Mouser is quoting 16 weeks out and Octopart says only Farnell and Avnet (?) have them in stock.

\dmc

Hmmm thats no good. Anyone know a good sub?

[budude]

How about http://www.mouser.com/ProductDetail/Murata-Power-Solutions/NKE0505SC/?qs=sGAEpiMZZMtwaiKVUtQsNSXDIqixAheByS2Ji%2fT%2fg5 Y%3d?

[DynamoBen]

How about http://www.mouser.com/ProductDetail/Murata-Power-Solutions/NKE0505SC/?qs=sGAEpiMZZMtwaiKVUtQsNSXDIqixAheByS2Ji%2fT%2fg5 Y%3d?

That should do it, now I have to change the PCB.

[budude]

That should do it, now I have to change the PCB.

I didn't look at the pinout close but can the PCB accomodate either? Might be handy if one or the other comes and goes...

side note - are you using DipTrace? It looks like it from the schematic - it's what I'm using also...

[DynamoBen]

I didn't look at the pinout close but can the PCB accomodate either? Might be handy if one or the other comes and goes...

side note - are you using DipTrace? It looks like it from the schematic - it's what I'm using also...

Looks like the big difference is pin spacing and size. The one you suggested is smaller so I will just go with that.

Yes I'm using DipTrace, love it! Tons easier to use than Eagle.

[DynamoBen]

That should do it, now I have to change the PCB.

Digikey has a Recom Power part that looks to be a drop in replacement.

RE-0505S (http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=945-1141-ND)

[DynamoBen]

The schematics are being peer reviewed in the developers forum to make sure the design is sound.

[DynamoBen]

I'm starting to make the I/O boards for the prop controller. Here is the first one which is a 32 channel dimmer board with external SSRs. I've got more on the way to do motor control, servo control, mixed voltage output, contact closure inputs, I2C general I/O, and LED control.

[budude]

I'm starting to make the I/O boards for the prop controller. Here is the first one which is a 32 channel dimmer board with external SSRs. I've got more on the way to do motor control, servo control, mixed voltage output, contact closure inputs, I2C general I/O, and LED control.

Remind me - - what is pin 5 on the SSR used for? I noticed they are all tied together back to the Prop connector. Other than that - pretty straight forward - looks good.

Visualizing this - is this meant to be a "shield" type of board directly connecting into the Prop module? The reason I ask is if it's not - what is the maximum distance between the Prop and this module? I'm thinking something like the original external Helix boards (no SSRs - just bits to drive external SSRs)...

[DynamoBen]

Remind me - - what is pin 5 on the SSR used for? I noticed they are all tied together back to the Prop connector. Other than that - pretty straight forward - looks good.

Its open for whatever you like. On my other dimmer it was for some 1-wire temperature sensors. Think of it as a bonus I/O.



Visualizing this - is this meant to be a "shield" type of board directly connecting into the Prop module? The reason I ask is if it's not - what is the maximum distance between the Prop and this module? I'm thinking something like the original external Helix boards (no SSRs - just bits to drive external SSRs)...

Its not a shield in the traditional sense because it doesn't stack. There is a jumper between the I/O board and the baseboard.

This dimmer I/O isn't meant to go too far. The Helix has a RS485 interface between the baseboard and the dimmer module thats how he gets the distance. With that said you could make an I/O board to do this.

[budude]

This dimmer I/O isn't meant to go too far. The Helix has a RS485 interface between the baseboard and the dimmer module thats how he gets the distance. With that said you could make an I/O board to do this.

Right - didn't mean to imply I would be able to get any real distance as the Helix but it would be good to know what this distance should be so we can start thinking about how you would put this into an enclosure. Obviously something like "a couple feet" would be more flexible than something like "as short as possible" or 6" for example. Some kind of range would be good - I know it's not critical for the schematic layout but again - just thinking how it would all go together...

[DynamoBen]

Right - didn't mean to imply I would be able to get any real distance as the Helix but it would be good to know what this distance should be so we can start thinking about how you would put this into an enclosure. Obviously something like "a couple feet" would be more flexible than something like "as short as possible" or 6" for example. Some kind of range would be good - I know it's not critical for the schematic layout but again - just thinking how it would all go together...

A couple of feet doesn't seem unreasonable, it really depends on how quickly data is being sent. Because I'm using the Helix code I haven't really looked to see how quick things are being sent.

[DynamoBen]

Here is the next I/O power, the ULN2803 with 32 outputs.

There are two propeller drivers for this board. On is a PWM driver, the other is a simple state driver. This I/O board can be used for motors, LEDs, valves, relays, and countless other things. Each bank of 8 outputs can have a different voltage as long as they share the same DC common.

[budude]

Here is the next I/O power, the ULN2803 with 32 outputs.

There are two propeller drivers for this board. On is a PWM driver, the other is a simple state driver. This I/O board can be used for motors, LEDs, valves, relays, and countless other things. Each bank of 8 outputs can have a different voltage as long as they share the same DC common.

Are the ULN2803A's 3.3V tolerant/compatible? I can't quite tell from the data sheet I have.

Minor stuff - fix run from U8.1/pin 13 (OE output) from going through the chips for readability. Also the connection to U8.1/pin 11 pin 5 of J1 - it looks like it comes from the unused pin 7 of J1.

[DynamoBen]

Are the ULN2803A's 3.3V tolerant/compatible? I can't quite tell from the data sheet I have.

Sure is.


Minor stuff - fix run from U8.1/pin 13 (OE output) from going through the chips for readability. Also the connection to U8.1/pin 11 pin 5 of J1 - it looks like it comes from the unused pin 7 of J1.

Fixed.

[n1ist]

Would it make sense to add ground to the I/O connectors too? The 2803 switches to its ground pin, and a local ground on the cable will keep the switching currents local.

Also, the '595 doesn't specify that it has a power-on reset, and can come up in any random state (I have had that before with bad results). I would add a POR to the ~SRCLR pin.
/mike

[DynamoBen]

Would it make sense to add ground to the I/O connectors too? The 2803 switches to its ground pin, and a local ground on the cable will keep the switching currents local.

Whoops, while adding connectors I deleted the common/ground. I will put the grounds back in.


Also, the '595 doesn't specify that it has a power-on reset, and can come up in any random state (I have had that before with bad results). I would add a POR to the ~SRCLR pin.


Power on Reset to SRCLR. Is it just a matter of toggling SRCLR?

[n1ist]

A pulse to ~SRCLR will clear them. I tied ~SRCLR from an I/O on the micro and added a resistor to ground so the '595 would clear at power up. I then drove it high in my init routine. In my case, I couldn't let the outputs come up random since that would cause the downstream devices to do nasty things (similar to driving an H-bridge from a '595). I could have used a resistor/cap/diode combo to do the same, but I had spare I/Os...

/mike

[DynamoBen]

A pulse to ~SRCLR will clear them. I tied ~SRCLR from an I/O on the micro and added a resistor to ground so the '595 would clear at power up. I then drove it high in my init routine. In my case, I couldn't let the outputs come up random since that would cause the downstream devices to do nasty things (similar to driving an H-bridge from a '595). I could have used a resistor/cap/diode combo to do the same, but I had spare I/Os...

Got it. I will tie SRCLR to an I/O pin and add some code to toggle the pin high at boot. Thanks for the tip.

[DynamoBen]

Got it. I will tie SRCLR to an I/O pin and add some code to toggle the pin high at boot. Thanks for the tip.

Done.

[DynamoBen]

I also posted the lasted schematics here:

DMX version:
http://doityourselfchristmas.com/forums/showpost.php?p=116496&postcount=76

Ethernet version:
http://doityourselfchristmas.com/forums/showpost.php?p=116496&postcount=77

I have one pin that I'm not sure what to do with. Thoughts?

[DynamoBen]

On the ULN board would it be better to limit the board to 8 outputs and make it "cascade-able"? Basically is 32 overkill?

[DynamoBen]

Update: Sorry folks my life recently got busy and I haven't had much time to finish this. However in the next two weeks I will be jumping back into it feet first. I'm still laying out the boards so I need to wrap that up. My tentative plan was to have 5 boards of each type (ethernet and DMX) made and to get them to the board house no later than the first week in Sept.

So in short this project is paused for two weeks then back to play. Sorry for the delay.

[DynamoBen]

First PCB complete, just needs standoffs.

This the DMX version, Ethernet is next.

[budude]

First PCB complete, just needs standoffs.

This the DMX version, Ethernet is next.

Good to see this is still moving forward! Is this a single layer board? If so perhaps it would be better to move to double sided? The ground plane has a lot of holes in it especially around the Prop chip. Other than that it's a nice compact layout. You mention standoffs - I assume you mean there are still mounting holes to be placed?

[DynamoBen]

Good to see this is still moving forward! Is this a single layer board? If so perhaps it would be better to move to double sided? The ground plane has a lot of holes in it especially around the Prop chip.

This is currently a dual layer board. The vias you see are linking the grounded floods between the two sides.


You mention standoffs - I assume you mean there are still mounting holes to be placed?

Yes I did mean mounting holes, I just added a 1/4 inch to each side an threw some holes on the board.

[DynamoBen]

I want to get the DMX board out this week. I'm waffling on one item, the capacitor immediately following the rectifier. It's currently 2200uF/16V however my last project used at 4700uF/16V. These two parts have dramatically different foot prints and I see merits for both. Anyone have any strong opinions about which one to go with?

Please speak up ASAP, these boards need to go this week :!:

Boards get uploaded tomorrow (Sept 3), speak now.

[DynamoBen]

I want to get the DMX board out this week. I'm waffling on one item, the capacitor immediately following the rectifier. It's currently 2200uF/16V however my last project used at 4700uF/16V. These two parts have dramatically different foot prints and I see merits for both. Anyone have any strong opinions about which one to go with?

Please speak up ASAP, these boards need to go this week :!:

Boards get uploaded tomorrow (Sept 3), speak now.

Order placed, should see boards in 8 days, now I have to order the parts. Assuming there are no issues with these boards I will make the Ethernet version next.

BTW I may start a new thread once the boards are complete to discussion next steps like daughter boards and code. Good idea?

[DynamoBen]

New thread about the PropController: http://www.doityourselfchristmas.com/forums/showthread.php?t=12334