Puma board for FreeEMS

[nitrousnrg]

I just pushed a footprint module titled usb-B_u-mini-standard. Has the three common USB B connectors in it.

Splendid!

Also, last night I put a led in the RPM circuitry, just like this


The output behavior was ok, loud and clear 0v and 5v

[jharvey]

A quick look at Newark didn't produce a 10k, 1/2 watt 0805 resistor, however it did produce a 5k 0805 from Vishay with this part number CRCW08055K00FKTA

Can I encourage an 0805 be used. A 1206 may have problems meeting a large temperature expansion tolerances. Hmm, can a 10k 1/2 watt smt be obtained? In my 10 second search on Newark, I didn't find one.

[nitrousnrg]

A 1206 may have problems meeting a large temperature expansion tolerances.

Good point. And using 0805 would be easier for the pick&place. My tradeoff was only between component cost and assembly cost; now expansion made a difference.

Hmm, can a 10k 1/2 watt smt be obtained? In my 10 second search on Newark, I didn't find one

In digikey they're mostly 1206/1210.

A 0805 can work up to 0.25W (there are 1/3W versions, but I want to use only one type of resistor), so with 2 of those we're fine.

Oh forgot to reply to you, Fred. Adding now the cap... btw, you don't know my hippie version yet.

Just to write it down: One of the fields of each component should say "optional", to automate the assembly process.

[Fred]

Marcos, if that RPM_OUT goes STRAIGHT to the CPU pin, it's good, and if you add more shit as optional in front of the inputs, then its even better :-)

I have a good feel for what you're like man, it's evident in everything you do, whether you are aware of it or not ;-)

Fred.

[jharvey]

I did a QUCS circuit simulation on the AN protect circuit, such that we can predict the warp-age and attenuation of certain signals. I'm still working on the diodes for OV protection and such, but the filter parts are becoming decent. Here's a draft snap shot.



I just pushed this to my git repo.

[nitrousnrg]

good!

That seems like a band reject filter; is there an upper cut-off frequency?

What about a weak pullup on those inputs? Just to have some diagnosis in case of an open circuit.

whether you are aware of it or not ;-)

haha, lately I'm not aware of lots of things happening around me :p

[jharvey]

You are correct. All caps set as a low pass filter are in fact a ban pass. Basically they all have inductance that overrides the capacitance when they Hz goes way high.

The inductor in series with the cap is to simulate the small amount of inductance of the cap. However, that doesn't seem to be working the way it should, as it currently simulates as low pass, but not high. If you click the picture, it will show you a larger copy of that picture.

[nitrousnrg]

Oh, so its a high freq parasitic inductance, good. I know SMD stuff has less inductance than TH, and the inductance should appear in the datasheet as a resonant frequency (<- uni contents, not real in-field experience)

I saw the enlarged pic, and was expecting to see the band somewhere. If I get the time I'll give it a try in the uni, I have the qucs files with me.

[Fred]

What about a weak pullup on those inputs? Just to have some diagnosis in case of an open circuit.

I strongly disagree with this, it can easily be done with a meter or external resistor on the input wire while setting up, and an NC ADC will show floating/noise. I guess it's as easy as leaving it off, but it shouldn't be on there as it will skew whatever signal you apply, even if just a little bit. No one likes inaccuracies... especially when they are completely preventable/unnecessary. Think about the user with a very weak high impedance voltage source, they'll be screwed by such a thing... and in the process you save someone with shit wiring? Doesn't make any sense! Which reminds me, is "removing the pull resistors from the thermistor circuits" item on the todo list for spin2? I hope so.

The above is analogous to putting "flood clear mode" in the code, just plain wrong, it serves no part of normal operation, and can easily be done in other ways without complicating the design further than is necessary.

Fred.

[Fred]

Just to expand upon that, I guess adding the pads for them is fine, provided there is space and it doesn't detract from some other aspect, BUT, the DEFAULT instructions/build/assembly/manufacture should definitely not have them. Make sense?