DFH - Defacto FreeEMS Hardware in KICAD

[baldur]

Regarding hall effect current sensing, Allegro Micro make integrated circuits that do this in an SOIC package.

[jharvey]

You are correct, I've even got one that works nice. I seem to recall it cost like $10 indicated up to 100A and requires you to make pads for it to be soldered to. I'm sure that lower cost ones can be obtained. It's nice because it take calibration out of the equation. However, the hall I've chosen works well for ignition pickup, and costs around $1, and they are common for RC stuff as well and appear to be quite rugged.

[AbeFM]

Ignition? I thought it was for injection? Ah well.

Anyway, thanks for all the comments! To me all diodes are the same, so I'm glad to see someone with a bit more to say about it on here. Don't ask me about charging the caps, or I'll say throw inductors at it, that's my solution to everything.

Two things from what you wrote:
1) The Thermo stuff - for reading air temps in a turbo application, even once a second would be pushing it, temps change QUICK going from flow under vacuum to 2 bars compressed. My turbo spools in a fraction of a second (a large fraction admittedly) but certainly waiting 10 seconds, that's a 1/4 mile in a quick car. Using the idle temps will be an issue.

I'm actually wondering why we don't see more thermocouples? Certainly they can be made small and low thermal mass, but I imagine the thermistors are right.

2) On the diode protection, can't you put both in parallel? The small ones to clamp small errors to < -0.2, and the zeners for the high energy stuff? Same for the 5v6 zeners, could you use 10V ones to keep obscene spikes out, and the smaller stuff to take the edge off?


About the opamp inputs, they work fine on 5 volts, though I think the part number might be wrong, I used radioshack jfet based ones I believe, I'll find out though it's posted here. Using the wrong ones indeed does work poorly! The basic advantage is that it's highly tunable, has a lot of hysteresis, and takes advantage of the gate delay to ignore very short spikes. My sensors are about as bad as you can imagine, and this circuit works very well on it. I believe it to be broadly applicable, but that's no garuntee.

[jharvey]

1) The Thermo stuff - for reading air temps in a turbo application, even once a second would be pushing it, temps change QUICK going from flow under vacuum to 2 bars compressed. My turbo spools in a fraction of a second (a large fraction admittedly) but certainly waiting 10 seconds, that's a 1/4 mile in a quick car. Using the idle temps will be an issue.

Having the thermos switched, allows duty cycling if required / preferred. However, to use the switch, you have to break a PCB jumper. Right now they are intended to be 100% duty cycle, if we develope the soft to back it, we can duty cycle sensors, but only if you really want to.

About the diodes, I'll have to think about that some. Perhaps we do want some more holes for such a setup.

[Fred]

If you have a low voltage clamp and a low impedance source (the cap) then the higher voltage zener will never see any action unless the over voltage is extremely long. In that case various other things will be cooked already anyway I would think.

Fred.

[gearhead]

Jharvey,

Do you have this any farther than 1.03. Do you have board layouts, etc? I opened up the files in the 1.03 attachment and there appears to be things missing. I am trying to create a simple device to be able to start dev work on this and want to leverage others' work if possible.

Keith

[Fred]

Jared, is there any chance of you creating a git repo for this and filling it with history? This would involve the following steps :

• sign up to github
• create project on github
• find old versions of design
• place each old version into the git project (over writing the previous one) and commit it
• when all are placed and committed push the result up to github

This will allow proper tracking of change over time and distribute the files nicely such that we have built in backups :-)

Fred.

[jharvey]

I can try the git thing, but I'm still in a fog about it. I currently don't have a gut feel for what happens and when. This will be a learning curve for me. I can see what I can do over this next week.

About the missing things, I'm a bit confused about the 1.03 thing. My last release was A.20. I've got some small changes, but haven't posted them, because I wanted A.21 to be a significant change into maturity. That's the drinking age in the US, and the time in your life when you're considered a man.

So far I removed a trace, and added some text or something like that. Not all the much of a change. I was going to do the next release when we had the pinout for the TA card established, and I can run those traces to the pins on the brain board.

[gearhead]

Well, where is a complete archive of all of A20. That is what I want to start from, then. I cannot find it.

Gearhead

[Fred]

Jared, if you want me to get the ball rolling on the git stuff, I don't mind. In fact, more than happy to. If you can just go through the thread and collate all the links to all the versions into one post below this then it will be pretty easy for me to do.

I've just duplicated this forum to the new server and the entire forum setup is under git so that no f*ckups like that last outage can occur again. It will also mean I can easily back it up from a remote machine and share the setup with others to help on it etc etc.

Fred.