Puma board for FreeEMS

[jharvey]

Also an idea for spin doctor, some kind of inboard temp sensor to know the aprox ambient temp for logging purposes and such.

[Fred]

1. In my opinion, 56 pages is way, way, too many sheets for a design of this complexity. I do understand why some people may choose to use a hierarchal schematic design approach, but I really see no need to have page after page of a single resistor with a single diode (or a few that have no components at all). It really makes it somewhat difficult to navigate the design. I tend to use a "Main" sheet (shows connector(s) assignments) and then full sheets of sub-circuits (Input(s), Output(s),...).

BIG plus 1!

2. You will have to assume that some users will not use KiCad, and will be relying on pdf versions of the schematic (or printouts) while building / troubleshooting. Because of this, I would suggest using the smallest standard sheet size for all of your sheets (we use "letter" size, but I think "A4" is what the civilized world uses). Also, I was not able to search for a component (i.e. "R101") using Acrobat Reader on the pdf version of the existing schematic. I am able to conduct searches in schematic pdf's that I generate, by "printing" to a free pdf creating tool called "FreePDF" from within my design tool. I don't know though, if it is my schematic capture program or FreePDF that allows this.

It's not unfair to just say "install KICAD" because it's free software, freely available, only disk space is used to do it. However I agree that many wont want to, just because it's super easy to navigate a webpage or pdf. I wonder if we can export to latex and go from that to PDF and HTML for full docs. I have seen various websites that offer downloads in either format, that is probably how they do it.

Any sort of internal temp sensor should be built and connected through a production spec header. Putting it on the schem and PCB for Beta spins and then pulling it out afterward and therefore changing the design in the process could introduce hw bugs and require another phase or two of testing. In software, once a release is "frozen" only bug fixes and critical feature omissions should be made. The same applies here. We have enough to review and change without introducing more variables like that.

Fred.

[jharvey]

1. In my opinion, 56 pages is way, way, too many sheets for a design of this complexity.

I agree, but I don't know an easy way to populate parts in a bulk fashion. I'm working on a python script will hopefully help with some of the issues relative to bulk processing tools, but that's still a bit off. My programming skills are reasonably weak. For now it's drawn that way to ease the development steps. With luck, we'll find a way to make it easier for end users.

I would suggest using the smallest standard sheet size for all of your sheets

This may be another mild limitation of KICAD. There isn't a scale feature so the symbols are always 100% scale. If you want to fit more parts on a page, such that you can keep the page count down, you have to use a larger sheet. When you print a PDF, you typically are given a choice to scale to page. That's probably the best way to scale it to an A sized (or B sized) piece of paper. Also I believe an A sized piece of paper is different for you, than it is for me. That's an ANSI vs ISO thing. Scale is a bit of a bugger.

Thanks for the feed back.

[Fred]

JHarvey, I'm not sure if Marcos mentioned this elsewhere, but the thermocouple chip outputs chip/board/ambient temp when nothing is attached to it! So, you have your temporary temperature in this design, at least, anyway :-)

Fred.

[TonyS]

1. In my opinion, 56 pages is way, way, too many sheets for a design of this complexity.

I agree, but I don't know an easy way to populate parts in a bulk fashion.

I'm not sure what you mean by "bulk fashion"?

- Huff

[jbelanger]

JHarvey, I'm not sure if Marcos mentioned this elsewhere, but the thermocouple chip outputs chip/board/ambient temp when nothing is attached to it! So, you have your temporary temperature in this design, at least, anyway

Fred.

No, it doesn't. The AD597 has a 0V output with nothing connected to it, if it is configured to read a thermocouple. I just had a quick look at the pdf and the configuration is all wrong. It's a mix of different modes and some wrong connections. And it's connected to 5V instead of 12V which would limit the temperature read to 500ºC which is not very useful for an EGT measurement.

Jean

[jbelanger]

Actually, let me correct this: it's not a mix of modes but it a mix of pins between the package shown in the data sheet and the SOIC package. And you can use the thermocouple configuration (if it was correctly done) as a thermometer if you ground both inputs not if you leave them floating.

Jean

[jharvey]

the thermocouple chip outputs chip/board/ambient temp when nothing is attached to it!

That's the first I've heard of it. I'll have to look over it a bit to see how that might work. I was thinking it might be handy because of the talk and concern about the regulator heating up. I know the atmosphere ambient shouldn't hit that 125C, but if under the bonnet, things can get hotter than ambient. I'm not sure if an onboard sensor is really all that important or not. I kind of feel like it is because many devices have it, but eh, I'm not sure it's all that handy, and I know it's not critical. Just wanted to toss it out there. Perhaps if it gets to hot, it can switch into energy cool mode or something. Bringing this up now, it probably a bit OT for the current goals.

[edit added the below]

Actually, let me correct this: it's not a mix of modes but it a mix of pins between the package shown in the data sheet and the SOIC package. And you can use the thermocouple configuration (if it was correctly done) as a thermometer if you ground both inputs not if you leave them floating.

I'll have to look at this closer. Thanks for looking it over and give us a heads up. Perhaps we can find a work around for this version and correct it on the next spin.

[jharvey]

I'm not sure what you mean by "bulk fashion"?

On DFH, I drafted up one sub circuit, where I populated the component values and ironed out most of the details. I put that original sub schematic in a folder called templates. I then wrote a simple script that would copy that template file into the higher level directory. Theory being that if some one wanted to copy the design, for more or less cyl, ect, they could simply change that script, and add their own sub circuits. In my case, this allowed me to populate the symbol properties like MFG, and vendor information, once, and copy them 5 times. This greatly reduced my work load by at least 5.

When Macro's did his copy of DFH, he didn't catch that early on. I believe this was his intro to KICAD, and it's simply an item he didn't happen to catch in the early stages. To make it more interesting, he also found a different technique than my template theory. He used one file that automatically turns into the several sub files. An interesting approach. Any how, we now have hundreds of components that need at least 6 properties addressed and probably modified. Which is at best several hours of labor if done manually. It would be nice to change these properties in bulk, by saying a 1k resistor should be xxx and they all get updated. This is what I mean by bulk fashion. It would be nice to have a script or some kind of tool that would allow me to change all 1k resistors to mfg part #XXXXXX. Right now I know how to do it with KICAD via hitting the E key, then typing the 6 fields in. This lends it's self to copy paste errors, and other human issues. My script will rake through KICAD .sch files, and when it finds the line for a 1K resistor, it inserts the correct properties on the next several lines, such that you can change all the properties of a 1K resistor in a bulk fashion.

Does that help clear it up? Also any chance you have python skills, or would like to learn python? My script is almost working, but I've still got a ways to go.

[Fred]

Thanks Jean!