FreeEMS hardware feature wishlist (your suggestions here!)

[Fred]

Data logging information :

SD card pinout and info :
http://www.interfacebus.com/Secure_Digi ... inout.html
http://en.wikipedia.org/wiki/Secure_Digital_card

GPL code for the spi interface :
http://www.mil.ufl.edu/~chrisarnold/com ... __mmc.html

An idea to avoid complications from FS structuring :
http://osdir.com/ml/hardware.microcontr ... 00031.html

A way of getting proper functionality without having to write extensive code :
http://www.sparkfun.com/commerce/produc ... ts_id=7955

That sparkfun website has a LOT of interesting stuff on it.

Imagine for track use data logging all engine parameters, GPS position, 3axis accelerometer, 3 single axis yaw rate sensors, 4 shock pots, 4 wheel speed sensors, one driveshaft speed sensor, and any other goodies you can think of :-) You could spend hours analysing a single lap!

Admin.

[Fred]

I propose that for ignition duties, sufficient places for to220 packages are put aside for on board drivers. If these are not desired, smaller self protected fets could be safely used to drive external peripherals larger fets for larger items with more current demand.

If the pin out of the chosen ign device does not match the common fet layout, a set of pads with jumpers could be included to re route the pins for the different part.

Having examined many options, the IRGB14C40L seems like a good choice. At high temperature it offers 14 amps continuous duty. Most coils require around 10 amps, thus the VB921 unit is not acceptable, and I was unable to find the remaining types for supply around the world.

Looking at the front of the unit, the pin order from left to right is 1 gate 2 drain 3 source for the IRGB14C40L

Having checked, the chosen FET pin layout turns out to be the same, so thats good. It will be best to provide space for as many heatsinked to220 devices as possible.

Given that we have 6 fuel injector channels and 12 ignition channels and that someone with such an engine would want to have the usual PWM idle, fuel pump relay, etc, 6 spares seems good, which will leave board space for many spare O channels in a 4 cylinder application. That leaves us with a total of 24 to220 sites. Having thought about this, you could put them side by side and get 10 across the euro card (10mm each) but there would be no room for other tracks. Instead if they are in two banks, with 6 per side of two ally bus bars, then there will be 40mm of spare board for other outputs to run through. Using both sides of the board that is plenty of space for a bunch of tracks to exit through.

Admin.

[gearhead]

A couple comments.

I like the to-220, but I have recently been converted to Dpak. TO220 seems to be hard to find in some of these protected FET devices and even some IGBTs. Interestingly, if these are placed in line on one end of the board, a bar could be clamped across them as a head sink. These are usually cheap and available in many flavors. The Ipak is a decent substitute for to92 but and is the same mechanically as a Dpak, but with longer leads. It gives options. If this is not clear, I'll sketch something.

to92. Please give us at least 1.5mm or 2mm spacing if inline, or optimally, the pads should be in a triangle. Old eyes and sloppy irons make tight spacing hard and it is not necessary.

Gearhead

[Fred]

I like the to-220, but I have recently been converted to Dpak. TO220 seems to be hard to find in some of these protected FET devices and even some IGBTs. Interestingly, if these are placed in line on one end of the board, a bar could be clamped across them as a head sink. These are usually cheap and available in many flavors. The Ipak is a decent substitute for to92 but and is the same mechanically as a Dpak, but with longer leads. It gives options. If this is not clear, I'll sketch something.

I understand what you mean by that, but I'm unconvinced that all units would be at exactly the same height. Being firmly attached to the board like that means that if they are not exactly the same height, the contact to that strip will be poor. Yes, you could put some paste in between, but paste is not meant to be a filler and even the best stuff leaks out/moves around over time. If their is a significant gap there, you will end up with an air gap in between and overheating parts.

I had no trouble sourcing a large variety of suitable FETs for cheap in little old NZ, I simply have trouble believing that the to220 parts are hard to find somewhere larger like the USA. I even found these IGBT's locally : http://nz.farnell.com/jsp/Semiconductor ... ku=8650829

If we go with to220, simple non limited FETs could be picked up at a retail place and a flyback circuit added to support them for those without access to a more complete catalog. You can't walk into a place like "Maplin" (UK) or "Jaycar" or "DSE" (OZ/NZ) and buy SMD FETs. You also can't get the limited variety of either type anyway, so if you have to order you will be able to get whatever you choose.

Additionally, (and I could easily be wrong) I really think we can pack in higher power density with the stand up devices.

to92. Please give us at least 1.5mm or 2mm spacing if inline, or optimally, the pads should be in a triangle. Old eyes and sloppy irons make tight spacing hard and it is not necessary.

100% agree, inline to92 parts are difficult for the best of us and should be avoided. There is absolutely no reason to populate them with that tight a footprint.

Also, as mentioned before, If the initial board layout doesnt suit everyone, additional ones can be designed for the DPAK packages etc for those that desire it. I think for now KISS is best.

Keep the comments coming! Good work!

Admin.

[gearhead]

with regard to the smd fets...

one example is the vnd5n07 (my favorites, actually). I have a sleeve of these in ipak format. These are the ones that are 'omnifets'. Current limited, thermal protected, short circuit protected, etc. The through hole ipak version is not available any more, though the dpak version is. These will work great for injector outputs w/o any trap as well as driving just about any load except an ignition coil. If all devices are the same, a bar should clamp across them, though with the low rds on, I do not know how much they will need to dissipate... Lets calculate... at full load of 5A and P=i^2*R... we should be dissipating 5^2* 0.2=5w With a case that is 3.75 degC/W, the case will only get an 18C temp rise so, 40C, maybe. And this is assuming constant current. I do not think we will even need anything. Just food or thought.

The Philips fets you mention do not seem to have any availability over here in the USA. One thing we should do is give this availability deal a bit of thought as well. One possibility would be to have an alternate BOM for US vs. EUR vs. Australasia. One thing is that Japanese transistors are not too available over here. Ditto the European types.

[Fred]

Hi, you raise a valid point, and because of that, I've done something about it :

http://www.diyefi.org/forum/viewtopic.php?f=9&t=39

Everyone, let us know what you have at your disposal.

The only reason that I chose that Philips FET is that it was half the price of the VNP5N07 and VNP20N07 units with similar and better specs. Any are totally fine, and all were readily available. There must be a supplier that can send you some of one of them to the US of A.

You have neglected the tricky property of power dissipation during switch on/off. That depends on current sink/sourcable from the cpu supply pin, slew rate of that pin, slew rate of any intermediate drivers, gate capacitance etc. I agree that the little fellows will handle the jandle just fine, and you could always use vnd20n07 parts instead which definitely would.

The packaging is my primary concern, and although I do realise that they are easy enough to solder, they still aren't as easy as to220 parts and they don't have the heat capacity, and to the average kid on the street, they aren't as available IMO. I mean, if a non protected FET was used from the likes of Dick Smith, it would work OK, but you couldn't pick it up locally from a retail place if you needed to.

Let us know who your suppliers are and we'll go from there :-)

Admin.

[gearhead]

well, you can pretty much forget about 'local shops' in the US. In Columbus, OH there was one shop that sold NTE transistors and Radio Shack has pretty much nothing useful. Here in Chicago, ditto, though I have not made it to the one shop in Niles, yet, to see what they have. I am pretty much set with online purchasing. I do all my electronic shopping at Digikey.com, primarily. If they do not have it, then I try Newark, Mouser, MCM, Allied and Avnet (in that order). All online. I have been able to secure some 'samples' from manufacturers on occasion, but picking ans choosing are not that good. One example is the Freescale 'portless' 4250 pressure sensor. I do not think you can get them any more as samples, but I got 5 as freebies to use as baro sensors.

$0.02
Gearhead

[Fred]

Yeah, RadioShack is pretty much the equivalent to Maplin, DSE and Jaycar. Not all that much, but the basics are there usually. I didn't say they were good, but they exist.

Farnell claim to supply to the US, but then again, Digikey claim to supply NZ in the same way... Farnell here doesn't carry those vnp parts either, but it does carry the Philips ones.

I suspect this will be a slippery one to deal with. Provided there are equivalents available, its not crucial what parts are used. So long as the equivalent drops in place we are good to go.

Admin.

[Fred]

Bump.

I've completed an initial list of hardware spec. It was done fairly quickly, so comments, suggestions, corrections, additions are welcome.

Consider this list a maximal list that we may end up with a subset of. (i.e. we'll include everything we can)

Admin.

[Fred]

With regards using the multi layered approach, some have suggested that its not up to the task of a rugged environment. I'd like to put this in perspective somewhat.

PC104 style boards have been used stacked in fairly severe environments for some time. Here are some examples :



This one is so rugged I included the article that went with it :



http://www.armadilloaerospace.com/n.x/A ... ews_id=175

The sandwich layering is quite common and when using the proper standoffs with metric bolts etc they will not give any trouble while kept dry :



There are more modern version available too :



and these are REALLY nice :





One of these little PCs could be integrated into a system like this and provide massive capabilities with regards data logging, on board tuning and of course display or real time variables.

No chance of your laptop being stolen and security could be implemented easily with serial comms to the PC and some interface that required a password/PIN and/or physical device to be present before starting could occur.

Just a thought.

Admin.