Welcome along :-)
Like it or hate it, typically I don't go easy on anyone ;-)
I think you have failed to let go of one of the MS things :
Single source hardware.
Because this is truly open you can/could/will? implement your idea and give users another option, a modular option :-)
It won't be like MS where there is one board and you are stuck with that. There is one design right now, but once that works, I bet you a dozen more spring up in varying forms. It should be very entertaining at some point.
Thanks for the suggestions anyway!
Fred.
Letting go of the "MegaSquirt mindset" (hardware)
Re: Letting go of the "MegaSquirt mindset" (hardware)
DIYEFI.org - where Open Source means Open Source, and Free means Freedom
FreeEMS.org - the open source engine management system
FreeEMS dev diary and its comments thread and my turbo truck!
n00bs, do NOT PM or email tech questions! Use the forum!
The ever growing list of FreeEMS success stories!
FreeEMS.org - the open source engine management system
FreeEMS dev diary and its comments thread and my turbo truck!
n00bs, do NOT PM or email tech questions! Use the forum!
The ever growing list of FreeEMS success stories!
-
- TO220 - Visibile
- Posts: 13
- Joined: Mon Sep 15, 2008 3:53 am
Re: Letting go of the "MegaSquirt mindset" (hardware)
I take that back Bruce has been very helpful to me. I just wanted to diffuse that before anything happened.
Re: Letting go of the "MegaSquirt mindset" (hardware)
Also welcome along, I also won't hold back.
I know we haven't gotten to do much with the general docs yet, freeEMS is still very young, however I think many of the things you mentioned have been implemented or are being implemented.
Right now, the goal is a board with the extreme basics, it only does naturally aspiration, can do up to 6 sequential spark and fuel and includes all the required inputs and outputs to function. The "brain" is detachable. So as CPU's get smarter/cheaper, they can be changed, but for now we are using a specific card. Also other features can be added with the use of add-on modules, much like a PC104 stack. So lots of room for growth.
Keep reading up on it, and I think you'll like what you see. Enjoy.
I know we haven't gotten to do much with the general docs yet, freeEMS is still very young, however I think many of the things you mentioned have been implemented or are being implemented.
Right now, the goal is a board with the extreme basics, it only does naturally aspiration, can do up to 6 sequential spark and fuel and includes all the required inputs and outputs to function. The "brain" is detachable. So as CPU's get smarter/cheaper, they can be changed, but for now we are using a specific card. Also other features can be added with the use of add-on modules, much like a PC104 stack. So lots of room for growth.
Keep reading up on it, and I think you'll like what you see. Enjoy.
Re: Letting go of the "MegaSquirt mindset" (hardware)
The basic idea is there and more or less correct. Certainly there's money to be saved having a small number of widely applicable boards, but.... You can have a 5 cyl board by not populating one of the injector and coil slots. Certainly that's what I will be doing for my four banger.
It wouldn't be hard to remove them from the drawing and remake the board.
Right now, we're trying to concentrate on getting something standard working. Certainly adding in more things isn't too bad, and I think the idea of a modular board stack is what we have in mind. replacing boards is supported, and designing better boards to add on is encouraged.
At least, that's how I see it so far.
It wouldn't be hard to remove them from the drawing and remake the board.
Right now, we're trying to concentrate on getting something standard working. Certainly adding in more things isn't too bad, and I think the idea of a modular board stack is what we have in mind. replacing boards is supported, and designing better boards to add on is encouraged.
At least, that's how I see it so far.
Re: Letting go of the "MegaSquirt mindset" (hardware)
Obviously I am late to the party, but here are my thoughts on this:
SMD as much as possible. TO-92 parts that need to be heatsinked should be standing up so they can be bolted to the enclosure. I've got a toaster oven and have 'hand soldered' dozens of full SMD boards with my stopwatch 'reflow curve' now, and it's way, way easier than soldering up a through-hole board. It takes me maybe 1/4 of the time. You can even get solder paste stencils for cheap, so you put the stencil on the board, smear the solder paste on like a silkscreen, remove the stencil, place all the components (which don't have to be exactly accurate even), and then put it in the toaster oven. I turn it to about 100C for about 4 minutes to bake out the moisture, then up to about 180 for 2 minutes to let the flux etch the pads, then crank the oven up as high as it goes (about 240C) and watch it closely. In about 60 seconds, the solder goes shiny and all the parts settle into place, surface tension pulls the multi-lead packages perfectly into position! Turn off the oven, carefully open the door, and let cool. The entire board is now soldered.
Sure, I usually have a few places to touch up, but it's super easy to do. Then if needed, you can do conductive soldering rework for bad parts. But really.. a stencil, some paste, and a $20 toaster oven and anyone can turn out pro SMD boards.
Further, as far as board design, I would design a basic board that does a simple application, and then have a small expansion bus to uplink to another board to expand it. You won't have many pins to work with, so working with an actual digital bus vs bringing in every signal is the way to go. A CAN interface can run up to 1 mbps and should be fast enough to handle most events across boards if the protocol is designed well. This way you also are able to support extra boards that you haven't even thought of. Plus, CAN controller chips are cheap, and they offload all the bus processing from your main CPU. Lastly, CAN is tolerant of lots of noise and can be expanded out to many feet from point to point.
Anyway, just trawling through the forums getting acquainted with the project and thought I would offer some ideas...
SMD as much as possible. TO-92 parts that need to be heatsinked should be standing up so they can be bolted to the enclosure. I've got a toaster oven and have 'hand soldered' dozens of full SMD boards with my stopwatch 'reflow curve' now, and it's way, way easier than soldering up a through-hole board. It takes me maybe 1/4 of the time. You can even get solder paste stencils for cheap, so you put the stencil on the board, smear the solder paste on like a silkscreen, remove the stencil, place all the components (which don't have to be exactly accurate even), and then put it in the toaster oven. I turn it to about 100C for about 4 minutes to bake out the moisture, then up to about 180 for 2 minutes to let the flux etch the pads, then crank the oven up as high as it goes (about 240C) and watch it closely. In about 60 seconds, the solder goes shiny and all the parts settle into place, surface tension pulls the multi-lead packages perfectly into position! Turn off the oven, carefully open the door, and let cool. The entire board is now soldered.
Sure, I usually have a few places to touch up, but it's super easy to do. Then if needed, you can do conductive soldering rework for bad parts. But really.. a stencil, some paste, and a $20 toaster oven and anyone can turn out pro SMD boards.
Further, as far as board design, I would design a basic board that does a simple application, and then have a small expansion bus to uplink to another board to expand it. You won't have many pins to work with, so working with an actual digital bus vs bringing in every signal is the way to go. A CAN interface can run up to 1 mbps and should be fast enough to handle most events across boards if the protocol is designed well. This way you also are able to support extra boards that you haven't even thought of. Plus, CAN controller chips are cheap, and they offload all the bus processing from your main CPU. Lastly, CAN is tolerant of lots of noise and can be expanded out to many feet from point to point.
Anyway, just trawling through the forums getting acquainted with the project and thought I would offer some ideas...
Re: Letting go of the "MegaSquirt mindset" (hardware)
CAN support is on the cards, but one of the design philosophies was that MOST users would be satisfied with a single chip solution. The CPU we're using has 90+ IO pins, and should keep all but the hardest core people happy. I'd definitely like the ability to chain stuff like extra ADC units off of it, though. Log EVERYTHING :-)
SMD/not is covered by the same comments that I made earlier about the SMP reg. I'd quite like the first version to be through hole and linear and KISS. After that, though, whatever goes! :-)
Fred.
SMD/not is covered by the same comments that I made earlier about the SMP reg. I'd quite like the first version to be through hole and linear and KISS. After that, though, whatever goes! :-)
Fred.
DIYEFI.org - where Open Source means Open Source, and Free means Freedom
FreeEMS.org - the open source engine management system
FreeEMS dev diary and its comments thread and my turbo truck!
n00bs, do NOT PM or email tech questions! Use the forum!
The ever growing list of FreeEMS success stories!
FreeEMS.org - the open source engine management system
FreeEMS dev diary and its comments thread and my turbo truck!
n00bs, do NOT PM or email tech questions! Use the forum!
The ever growing list of FreeEMS success stories!