Page 3 of 5

Re: Thermo couple input conditioning

Posted: Tue Sep 08, 2009 1:00 am
by MotoFab
jharvey wrote:The availability might concern me a pinch. I see it from Newark, Farnell, and one other supplier. Farnell and Newark are really the same, so it's basically a 2 vendor item. I bet there has to be others just like this that are more commonly available. On a good note, both Maxim and Newark are fairly good sources. So I'd bet this chip will be obtainable in a couple years.
Digikey stocks them too. I'm pretty sure the AD595 part is a ceramic through-hole dip, and that sort of package may not be around much longer. I get what you mean about continued supply though, Maxim sometime discontinues parts, but they give plenty of warning. It's like any other component in a design, while the product is shipping you must check for scheduled end-of-life at the beginning of every quarter.

Also the $13 vs $5, that's an increase of $7 per cyl, which is kind of annoying. I'd bet you're on the right track, but I'd also bet there is a more common chip out there. Lets keep looking for a bit more, and see what happens. National semi seems to have quite a selection of chips, not sure if they have something quite like this, but they did pop up alot with octopart.com.

I need to go but was looking at the digitals listed here.
http://www.national.com/analog/tempsensors/products
National has quite a selection! Maybe they, or other companies, make an NTC thermistor converter with a serial comm, that accommodates a remote mountable thermistor.

Hey there Fred :wave

Re: Thermo couple input conditioning

Posted: Tue Sep 08, 2009 1:45 am
by mk e
jharvey wrote:Looks like the two above mentioned chips are basically the chips that every one uses. As mentioned above, the AD chip is fast, and the max chip is digital. How about I draw it up with both, but we plan for the MAX chip to be the primary one. That way if someone wants a faster response, or is trying it on the cheap cheap, they can, but we'll focus on digital support first.

Any objections or thoughts? Some times I wonder if the PCB will fit under the hood. Hi hi. I'm sure it will fit just fine, perhaps with a bump in the hood, but it will fit.
The people who will want to build a 5554 will want the best if it's available....so the AD chip but would live with the MAX chip

How big do you think the board is going to be?

Re: Thermo couple input conditioning

Posted: Tue Sep 08, 2009 2:52 am
by MotoFab
Best is a squirrely term here. What do you expect the maximum high-end use? Perhaps a thermocouple in every exhaust port, and watch the temperature graph follow the exhaust valve as it opens and closes?

Re: Thermo couple input conditioning

Posted: Tue Sep 08, 2009 3:01 am
by jharvey
I think I'll plan for both sensors, but as separate circuits, so say 24 thermo-couplers, and you choose where / if you want the fast sensors. It adds a feature of being able to monitor the temperature of misc devices to ensure they don't over heat. A quick story, I once was working with a low end motor controller, it has 12 FETs parallel driven. It had some supporting circuitry, but we buggered it. One end of the controller was closer to the motor, which caused that end get slightly hotter than the other end. The temperature compensation sensor was on the other end of the board. Once the FET closest to the motor overheated, it popped, which then caused the load to shift, which caused overheating, which caused a pop, ect down the line of FETs until no little FET's were left standing in a line. An expensive mistake because it only had one sensor. On this board I'd say a likely long term mode of failure is from overheating. One end of the case is visible and therefore exposed to the sun, what have you. It might be nice if we had some sensors telling us the case was getting two hot, or that some misc device was getting to hot. Thoughts?
mk e wrote:How big do you think the board is going to be?
About the board size, good question. I've stopped to ponder that a little bit, but have been putting it off until we had the schematics ironed out a bit more. By putting the discretes on top and bottom and breaking it into two separate boards with the brain stacked on top as a third. One board containing the power drivers, with injection, and ignition ect. The other with the lower power and lower voltages like signals in and signals out. Done that way I think the real space constraints won't be the PCB layout, but the heat dissipation capabilities. PCB wise, I think the 1 euro size we used for FreeEMS 1.0 would be about right. So that's around 7"x4"x4" ish. The key is can we get rid of the heat.

What's an acceptable case size? Perhaps we could have a remote case for staged injectors and some of the less common extras?

If we decide to go with a larger X,Y of the case, I might copy FreeEMS 1.0. FreeEMS was designed with SMT pads under thru hole components. This allowed easy bulk asm, while allowing blind DIYers. Takes a fair bit of PCB real estate, and is a 2 layer design. I like 2 layer designs when possible, you can get a scope on it, see burn or broken traces, ect. If I straight up copy that design, I would keep these modules. I think that FreeEMS has almost all of the most common features you want, except for the 6 cyl vs 12 cyl thing. Hmmm, I have to stew this over. Any objections about breaking this into two cases. Case 2 including thermocouples, staged injector drivers, and extra IO's. Case 1 being a 2 board layout, with 12 injector, 12 ignition, 2 RPM, 2 MAP, all the basic inputs including IAC and some of the extras we haven't drafted up yet. Basically case 1 is an OEM 12 cyl, and case 2 is the performance add-on. With this design, I think we can keep each case at 1 euro 7"x4"x4"ish. Perhaps case size is something we should bring up in it's own thread. Lots of possibilities here.

Um speaking of heat, I don't recall peak and hold injectors as a software bullet. Theory being, that you drive the injector hard for say 10ms, then you duty cycle the FET such that the injector solenoid field collapses and prevents the injector from heating up, ect. This allows for a more accurate turn off time, and it prevents pre-heating the fuel. We should probably add that to the bullets if it's not already buried in there.

Re: Thermo couple input conditioning

Posted: Tue Sep 08, 2009 1:44 pm
by mk e
I think there are 2 groups who will want TCs.

The first group is the turbo crowd. In this application 4 hz is plenty fast as it takes quite some time to burn a turbine. For this group 1 TC would cover over 95% of the users and 2 would get you 2 99.9%. I would see the TC signal triggering a warning light on the dash in this application.

The other group is the roadracers and maybe dragracers. They will want a TC in each cylinder to find problems before the engine goes bang. In this application I would expect to light a warning light at some low set point and drop the cylinder at a high set point. A failing or plugged injector or low fuel pressure can lean a cylinder/engine and cause damage in just a couple cycles so I don’t think a 4 hz response would be of much value to these guys.

I would like to see the fast chips on the 5554 board and not bother with the low cost option. At $100 to read 12 TCs as well or better than any other option on the market it’s already priced at less than 10% of what any other option would cost and a great value. I see no reason try to cost cut any more at the expense of response time and I don’t want to think about 2 options in the software.

Re: Thermo couple input conditioning

Posted: Tue Sep 08, 2009 5:22 pm
by jharvey
The slow sensor is the one that costs more. I think the real expence is the PCB. You can always leave off the chips if you don't want them. Or populate them down the road if it's a feature you do want. I'm tempted to draft it up with both, and expect for blank pads.

Re: Thermo couple input conditioning

Posted: Tue Sep 08, 2009 5:27 pm
by mk e
jharvey wrote:The slow sensor is the one that costs more.
Then why are we even talking about it?

Re: Thermo couple input conditioning

Posted: Tue Sep 08, 2009 9:25 pm
by jharvey
The digital may be less prone to stray RF, and other such analog noise(s). So it may be a bit more robust.

Re: Thermo couple input conditioning

Posted: Tue Sep 08, 2009 11:07 pm
by mk e
jharvey wrote:The digital may be less prone to stray RF, and other such analog noise(s). So it may be a bit more robust.
The TC wires are running all though the engine bay analog, the I/O board should be away form that.

I say don't bother with the digital chip for 5554 hardware and will try it and see what we get. We don't need to clutter the board up with multiple options.

Re: Thermo couple input conditioning

Posted: Wed Sep 09, 2009 12:32 am
by MotoFab
mk e wrote:At $100 to read 12 TCs as well or better than any other option on the market it’s already priced at less than 10% of what any other option would cost and a great value.
A 12 TC conditioner using the AD595 will likely add $300 all in.