FreeEMS hardware feature wishlist (your suggestions here!)

[jharvey]

Here's another thought, most programming is done via PC w/ software. Would it be worth while to allow for some form of DIP programming? I'm wondering how often someone needs to do an on the fly software change, but doesn't have the PC (or PC with software) in hand. It's just a thought, and I suspect it won't stick, but is it worth while?

[Fred]

Here's another thought, most programming is done via PC w/ software. Would it be worth while to allow for some form of DIP programming? I'm wondering how often someone needs to do an on the fly software change, but doesn't have the PC (or PC with software) in hand. It's just a thought, and I suspect it won't stick, but is it worth while?

Can you explain what you mean? I don't understand.

[jharvey]

Fred I didn't see anyone kick your baby yet, so I figured I would

You like the standard size boards. I'm wondering why? Are there useful cards that might snap in? Are there PCB vendors that offer discounts? Standard housings? I'm not trying to trash your idea, just wondering why you like them. I think your from over the big drink, so I suspect Euro-cards are more common for you then they are here in the states. I've only seen them in communications devices, and the connectors have been quite expensive, not to mention they didn't seem to carry much current. Most of my experience with them has been CompactPCI. I might lean toward a board size of 2.5 in (63.5 mm) by 3.8 in (92.52 mm) because Express PCB offers a discount on this size.

You split one card input, one card output. Perhaps to cards each half input and half output might work better? For example if you have 4 injectors, one tach, one temp, one map and one O2. So 4 in and 4 out. There is a good chance the tach won't work for every possibility out there. So if you made the board up then discovered you needed a custom tach input, you have to toss the entire board, and start over. Could get costly. However if you split it 50% each board, you might not have to scrap the entire board.

About the power section, perhaps a chip with a switch mode pump might be handy. The upfront protection is good, but what if you have a power drop for any reason. Dead battery, resistance at the injector's high current, ect. A SMP could help prevent CPU hiccup's.

You like tech art's stuff. I'm curious what makes them different from the rabbit cores, and other prefab boards out there. Do they offer the brain boards as GNU or such? I've seen popularity with the http://www.arduino.cc/ I think these guys are GNU. Not so sure if the programming language thing, but perhaps other firware can be uploaded to it.

Any how, hope I didn't kick to hard

[jharvey]

Can you explain what you mean? I don't understand.

I guess I'll have to understand it first. I was remembering that way back in the days of the first DIY home micro computers, you had to enter the program by hand. Some did it with DIP switches, by setting up the DIPs then hitting a button to accept that line, then change the DIPs for the next line, ect, ect. Entering the program line by line. I was then thinking that someone might head off to the track, and find out the gas isn't quite the same grade as what they expected. If they wanted to change a value or two in the lookup tables, or change the program in any way, they would have to reprogram via PC. But if they simply needed to take off a couple of milli secs on the PWM to lean it out a bit, reset it to a known table, ect, they could make an adjustment via DIP switch entry, with out a PC.

Does that make more since?

I'm not sure how useful it is, but if you just dragged your vehicle a couple hours to the track only to find the PC is at home, it could be quite a handy as an emergency feature.

[Fred]

Fred I didn't see anyone kick your baby yet, so I figured I would ;)

You are in the hardware section. Yes we have decided on the processor and cpu board, but everything else is up in the air somewhat. There is no baby here to kick :-) I'm open to ideas (like you cable connector one for example!)

You like the standard size boards. I'm wondering why? Are there useful cards that might snap in? Standard housings!

and the form factor is appropriate to what we want to achieve. IE, similar physical size to MS, but metric.

I'm not trying to trash your idea

It's not so much that I like them, it's that I HATE imperial measurements. Unfortunately the electronics industry is a mix of the two with 0.1" pin spacings and metric to220 parts etc. The MS main boards are drilled for imperial screws (which you simply can't buy outside of the usa) forcing an m2.5 instead of m3 screw to be used and a loose fit because of it. Conversely you CAN buy m3 screws in the usa, AND the components are metric!

I've only seen them in communications devices, and the connectors have been quite expensive, not to mention they didn't seem to carry much current.

Connectors? I thought we were talking about boards?

Most of my experience with them has been CompactPCI. I might lean toward a board size of 2.5 in (63.5 mm) by 3.8 in (92.52 mm) because Express PCB offers a discount on this size.

See, I'd give you an extra 5 bucks to NOT have them that size any day :-)

You split one card input, one card output. Perhaps to cards each half input and half output might work better? For example if you have 4 injectors, one tach, one temp, one map and one O2. So 4 in and 4 out. There is a good chance the tach won't work for every possibility out there. So if you made the board up then discovered you needed a custom tach input, you have to toss the entire board, and start over. Could get costly. However if you split it 50% each board, you might not have to scrap the entire board.

Something along those lines is possible, yes. I've thought about that too. Except I think one board must have the inputs self contained (there are only 3 types, and we will only use 2 of those) and the other have auxiliary stuff on it instead.

About the power section, perhaps a chip with a switch mode pump might be handy. The upfront protection is good, but what if you have a power drop for any reason. Dead battery, resistance at the injector's high current, ect. A SMP could help prevent CPU hiccup's.

Tell me more.

You like tech art's stuff. I'm curious what makes them different from the rabbit cores, and other prefab boards out there.

Form factor. A plug in CPU is a good thing to have IMO. The tech art card was more expensive than other proto boards, but it was the only one that could be used like a module in a finished product. This is it's biggest value. For a single chip design, none of the other stuff I've found comes close to the XDP/XEP parts.

Any how, hope I didn't kick to hard ;)

No problem. I'm not sure where this post came from. It seemed out of left field and not related to the last one.

As for dip switches... I think if you tried to change anything significant blind with those you would just get headaches. Also, if you were trying to change something significant you would also be there for about a week doing it! :-) Not practical IMO, but an interesting idea none the less.

Fred.

[jharvey]

About the power section, perhaps a chip with a switch mode pump might be handy. The upfront protection is good, but what if you have a power drop for any reason. Dead battery, resistance at the injector's high current, ect. A SMP could help prevent CPU hiccup's.

Tell me more.

I've seen the SMP used in PSoC's. However I hadn't known about this one
http://www.eetindia.co.in/ARTICLES/2004 ... S=DOWNLOAD

I know there's an ap not for the normal SMP, I seem to recall it allowed the PSoC to run down to under 1V, but I can't seem to find it right now. Cypresses site sucks, and PSoCDeveloper's ap note tool seem to be down at the moment.

[Fred]

Thanks for that, very interesting :-)

I have to say though, if your battery voltage drops below about 6 (where an LDO can still function) when cranking then you need a bigger/newer battery or starter rebuild :-)

I'd probably err on the side of KISS for that.

I have code to detect and log low voltage instances. Ones where it goes low enough to stop working should be fairly obvious I would think.

Fred.

[jharvey]

Connectors? I thought we were talking about boards?

I honestly don't know much about the Eurocard, mostly what I found here
http://en.wikipedia.org/wiki/Eurocard_(PCB)

This article lead me to believe the connectors were standard with Eurocard, but it seems your only looking for the length and width standard. Great. Connectors are covered else where. About using metric vs imperial, I'm all for it. Go metric.

You like tech art's stuff. I'm curious what makes them different from the rabbit cores, and other prefab boards out there.

Form factor. A plug in CPU is a good thing to have IMO. The tech art card was more expensive than other proto boards, but it was the only one that could be used like a module in a finished product. This is it's biggest value. For a single chip design, none of the other stuff I've found comes close to the XDP/XEP parts.

So to recap, your vision the current have for the hardware platform is Techart's Adapt9S12X with two custom boards as wings. The custom boards would contain input, output, power regulation, and perhaps some other features.

I'm curious, can you get different 50 pin connectors? I know stacking has been talked about, if we stack, can we get straight instead of 90 degree? I'd bet they would be willing. Oops that that connector stuff again.

You like this techarts module, because they have had similar products in the past, and as time went on and new chips came out, they keep upgrading modules in a manner that allowed for upgrades using the same module pinout and form factor.

I also see the 50 pin connectors have both IO and AN, perhaps a single board will suffice for now. There are lost of extra IO and AN. I haven't looked at the data sheet for the MCU yet, are the AN path's only input? Or does the device have D/A?

I see the MCU is about $12 at digikey, the CAN devices are about $1.5, 232 chip is about $2, toss in $8 for misc components, and your up to around $25. Which means they produce the PCB and assemble it for about $100. That's not bad for a proto. The PCB in low quantity would likely cost $40 to $50 of that. So it's more like $50 to assemble.

Any how, hope I didn't kick to hard

No problem. I'm not sure where this post came from. It seemed out of left field and not related to the last one.

As I was reading this topic, I had a bunch of questions. When someone says this is what we are doing, and someone else questions why, it's kind of hard to get it to come out right. So I tried to help humorize it with the baby kick thing.

I know Fred has been leaning folks toward holding off on drafting up a layout, but I'm tempted to push for it. I like the techarts board as the heart, then I think we could layout a wing man, I mean wing board, that takes care of 6 inputs, ? outputs, and power conditioning. I think the other wing could be used for future expansion.

[Fred]

but it seems your only looking for the length and width standard. Great. Connectors are covered else where. About using metric vs imperial, I'm all for it. Go metric.

Yup, just the dimensions, glad you are all for metric :-) It poses a small problem with the 0.1" pitch pins etc (using both standards or funny decimals of metric) but I think it's worth it.

Here are the boards in bare form :

http://i260.photobucket.com/albums/ii15 ... 0_6790.jpg
http://i260.photobucket.com/albums/ii15 ... 0_7036.jpg
http://i260.photobucket.com/albums/ii15 ... 0_8206.jpg

So to recap, your vision the current have for the hardware platform is Techart's Adapt9S12X with two custom boards as wings. The custom boards would contain input, output, power regulation, and perhaps some other features.

I'm curious, can you get different 50 pin connectors? I know stacking has been talked about, if we stack, can we get straight instead of 90 degree? I'd bet they would be willing.

Close, not wings. You've been looking at their website, but not in enough detail, you can choose your connectors, and I already chose them :

http://i260.photobucket.com/albums/ii15 ... 0_7093.jpg
http://i260.photobucket.com/albums/ii15 ... 0_7092.jpg

This allows sandwich, not funny fragile (physically) wings.

The rest of what you said is spot on. Though the TA board needs some small mods to work well for us. No big deal though.

You like this techarts module, because they have had similar products in the past, and as time went on and new chips came out, they keep upgrading modules in a manner that allowed for upgrades using the same module pinout and form factor.

I hit them up about XEP upgrades and they said they had no plans to do one, but no plans to stop doing these. If they stop making them totally, we can make an XEP one that is pin compatible to reuse our board designs.

I also see the 50 pin connectors have both IO and AN, perhaps a single board will suffice for now. There are lost of extra IO and AN. I haven't looked at the data sheet for the MCU yet, are the AN path's only input? Or does the device have D/A?

A single board would probably suffice for now. 16ADC, no DAC.

As I was reading this topic, I had a bunch of questions. When someone says this is what we are doing, and someone else questions why, it's kind of hard to get it to come out right. So I tried to help humorize it with the baby kick thing.

It's all good :-)

I know Fred has been leaning folks toward holding off on drafting up a layout, but I'm tempted to push for it. I like the techarts board as the heart, then I think we could layout a sandwich man, I mean sandwich board, that takes care of 6 inputs, ? outputs, and power conditioning. I think the other sandwich could be used for future expansion.

Fixed it for ya ;-) If you want to, I am not stopping you. I'm just wary that I may arbitrarily move pins around on you. The chances of that are getting slim though, so it's starting to get reasonable to think about it now especially as the software starts to mature a little and get closer to being useful. You are not alone in your desire to build boards now. In fact, I may be alone in my hesitance to do it :-) It probably stems from not wanting to be responsible for messing up someones work by moving a feature to a different pin. I have to say though... I won't hesitate to move things IF I have to. So PCB designer beware. :-) All boards will have all pins though, so just a re route should fix any of my changes.

Fred.

[jbelanger]

And just to tease those who want to make boards a bit more, I've just discovered that the XEP is available in single quantity from Arrow: http://app.arrownac.com/aws/pg_webc/0,4 ... 2XEP100MAL (hopefully this link works).

Jean