285 replies to this topic

[mjr]

Thanks for the update, but now you are just scaring me!  ZOINKS! that's a lot of parts and it's not even all of them!  Not sure I'm up to this diy task.

 

It's not for everyone, for sure - I don't think the LedWiz people have to worry about it cutting into their business.  But I should point out that I'll eventually put together some real documentation on how to build it - I don't want anyone to think that the preview page is meant to be the whole user's manual.  That's just to share information on the project in these early stages.  It's never going to be a turnkey solution like an LedWiz, but hopefully I can eventually make it less intimidating than it is at the moment.

[mjr]

Just got email that my prototype batch is fresh out of the factory and ready to ship.  I should have them in a week or so.

 

[Rappelbox]

I can't wait to get my hands on it!

[jreising326]

I'm interested as well.  I'm building a cab and just got the Pinscape Nudge installed and will try and get the plunger working over the weekend.  I really appreciate your effort on this.  Great Work!

 

Joe

[GeCh]

that looks amazing.

if you have enouth boards what would be the price for  shipping one set to france? 

I already got the tlc5940 chips ready 

[mjr]

that looks amazing.

if you have enouth boards what would be the price for  shipping one set to france? 

I already got the tlc5940 chips ready 

 

I'd rather not ship outside the US (it'll be expensive), but happily there's another member here who's doing a run in Europe.  mk47 is in Germany, and he has a batch being made right now.  I'd suggest PM'ing him here and see if he can ship some over to you - should be way cheaper than from the US.

[Rappelbox]

Hi mjr, GeCh

 

I'm in talks with mk47, meaning that i will possibly organize a group buy/Soldering Service etc. like i did for PIN2DMD

 

Let's see what this awesome project will lead us to... Atm I have the naked KL25Z running in my Cab, works like a charm.

Only nudging is a bit "sensible" @mjr: do you have any suggestion for that?

 

Greets,

Dom

[GeCh]

Pm send to mk47..

I would like to have the boards as fast as possible.

Edited by GeCh, 04 December 2015 - 09:45 AM.

[AndyMB]

It's a me too for a "Europe" prototype board order.  PMs sent.

 

Thanks for the great work on these boards, it solves all my problems in one foul swoop!  Design objective #1 DONE :-)

[vampirolatino2]

guys just a heads up. it will not be "cheap", the components alone that is needed adds up. But I have a lot of spare parts that if any of you guys need one in USA I could do requests for selling them at cost + shipping, just because we are a community and this is mjr masterpiece, not mine.

Let me first finish and test them first before group buying

[wshadow]

I've just been talking to mk47 about possible signal fading, propagation delay and interference when daisy-chaining up to 8 TLCs (or 4 boards including the KL25Z interface board).

I might be a little late on that since you guys already ordered your prototype boards, but as a fallback idea, I thought about using the 3 remaining pins (PTC0,3,4 according to schematics ) on the KL25Z board for talking to the TLCs.

 

The idea is only to chain the 2 TLCs per board.

Assuming that PWM_GSCLK, PWM_XLAT, PWM_SLCK and PWM_BLANK can be connected in parallel to all TLCs throughout the system, only one additional PWM_SIN pin per board (or chain of 2 TLCs) is required.

 

The resulting 4 available chains of 2 TLCs each (or 4 boards, including the "main" board) would be able to provide the projected 128 outputs (not even including the chime outputs).

You can stick to bus wiring as it is and put 2 more pins on the header or create additional I/O headers (2 more would be required - for 3 expansion boards besides the KL25Z interface board), with one PWM_SIN line on each.

When sticking to bus wiring, one would be required to jumper the expansion board so that the correct PWM_SIN line in the bus is used. Worst case when jumpered wrong: 2 banks of outputs fire at the same time - go cry

 

Not an expert on your source code, but I assume only minimal effort is required to let the software know how to use 4 different PWM_SIN lines on the output assigments (should even be calculatable on-the-fly with MOD dividing )

[mjr]

I've just been talking to mk47 about possible signal fading, propagation delay and interference when daisy-chaining up to 8 TLCs (or 4 boards including the KL25Z interface board).

I might be a little late on that since you guys already ordered your prototype boards, but as a fallback idea, I thought about using the 3 remaining pins (PTC0,3,4 according to schematics ) on the KL25Z board for talking to the TLCs.

 

The idea is only to chain the 2 TLCs per board.

Assuming that PWM_GSCLK, PWM_XLAT, PWM_SLCK and PWM_BLANK can be connected in parallel to all TLCs throughout the system, only one additional PWM_SIN pin per board (or chain of 2 TLCs) is required.

 

The resulting 4 available chains of 2 TLCs each (or 4 boards, including the "main" board) would be able to provide the projected 128 outputs (not even including the chime outputs).

You can stick to bus wiring as it is and put 2 more pins on the header or create additional I/O headers (2 more would be required - for 3 expansion boards besides the KL25Z interface board), with one PWM_SIN line on each.

When sticking to bus wiring, one would be required to jumper the expansion board so that the correct PWM_SIN line in the bus is used. Worst case when jumpered wrong: 2 banks of outputs fire at the same time - go cry

 

Not an expert on your source code, but I assume only minimal effort is required to let the software know how to use 4 different PWM_SIN lines on the output assigments (should even be calculatable on-the-fly with MOD dividing )

 

It might be possible, but I'd recommend trying it first with the current daisy-chained setup, since it's so much simpler in both hardware and software.  This is exactly how the chips were designed to work, after all.  Reports I've seen indicate that the practical limit should be at least 10 chips, although some Arduino people do seem to have problems with chains of just 2 or 3.  I was able to get 4 working reliably in a breadboard environment, which is quite poor in terms of the physical signal path properties; hopefully the PCB version will be quite a lot cleaner and deliver clear enough signals for a decent length chain.

 

The big difficulty on the software side with what you have in mind is that I'm using DMA through SPI to clock out the data bits to the chips.  That limits which pins you can use, to the point that really only one SIN channel is possible.  You could replace the DMA/SPI code with direct bit-banging to GPIO pins, but the DMA/SPI approach takes a ton of load off the CPU for the chip data transmissions.  The load would be significant doing it with direct GPIO writes, especially doing several sets of chips in parallel.
 

And I must add, I really want to hear more about your cabinet and what all you're hooking up to this many outputs!  I've kind of run out of ideas at around 64.

Edited by mjr, 07 December 2015 - 06:07 AM.

[mjr]

I got the boards back from the fabricator, and got my Mouser order with most of the parts.  I'm just waiting for a few additional parts now, then I'll be able to assemble and test the boards.

 

[vampirolatino2]

Excellent news! ..... can't wait 

[mk47]

I'm one step ahead ;-)

 

But it's completely untested ...

 

[roar]

I got the boards back from the fabricator, and got my Mouser order with most of the parts.  I'm just waiting for a few additional parts now, then I'll be able to assemble and test the boards.

 

 

Hello mjr, if there is a group by are there any plans to ship these boards to Canada?

Edited by roar, 23 December 2015 - 02:29 PM.

[mameman23]

Cool project, I'm certainly down for getting inline for a batch board purchase.


Sent from my iPhone using Tapatalk

[mjr]

Hello mjr, if there is a group by are there any plans to ship these boards to Canada?

 

I think the best bet would be to find a couple of other people in Canada and do a group order from there, so that the final distribution shipping is all domestic.  Re-shipping them internationally probably won't work very well cost-wise.  (I'm in the US.)

[NobodyYouKnow]

Just a thought on the per-channel current limit on the MOSFET board. The SIP header proposed in the BOM has a current limit of 3 amps per pin. I have not yet found a higher current alternative that fits in the same form factor limits enforced by the board.

[mjr]

Just a thought on the per-channel current limit on the MOSFET board. The SIP header proposed in the BOM has a current limit of 3 amps per pin. I have not yet found a higher current alternative that fits in the same form factor limits enforced by the board.

 

 

That does seem to be as high as that size of connector will go.  To get a 4A connector you have to next size up, which is about twice the spacing, so only half the outputs would have fit in the same space.   I figured it was worth the trade-off to pack more onto the board, since (a) practically everything in a virtual cab is under 3A, and (b) everything runs intermittently.  The only thing in my cab that draws more than 3A is the gear motor, which is about 4A.  When I get mine set up I'll try running the gear motor on it continuously for a while and see how long it takes for the pin to get hot.  I think the specs on these parts tend to be fairly conservative, so expect it'll be okay despite pushing the nominal limits, but I'll see what happens.

 

If it does turn out to need something beefier, or if you have something even higher current that you're concerned about, one hackish workaround would be to run some wire from the solder pad meant for the pin header and wire up a heavier duty connector off-board.  I could also try coming up with a board variation that has a few special heavy-duty outputs that go to a bigger connector, possibly sacrificing a couple of outputs to make room.   Do you have an idea of how many such outputs you might eventually need?

Edited by mjr, 25 December 2015 - 10:53 PM.