Trying to design an abnormal power up circuit

[Ian.M]

Not quite.  Reverse D3, then its good.

[EasyGoing1]

Yes, desert temperatures and small signal Schottky diodes aren't a good mix.  Try if it works OK with 1N4148 for D1 and for the diode clamp from ground to the C1-R1 junction.

All I had on hand were 1N4001 but the circuit works. However, with the 470nf cap in there, I could not get a voltage on the divider so I removed it, then ran the diode to ground before the 200k resistor. That was the only change I made, but it does work and it works well.

This is what worked:

[Ian.M]

The 470nF cap was to prevent 12IN keping EN active when the MCU tries to shut it off.  It only lets through a short pulse to start up the buck module so you shouldn't expect to measure a DC voltage on the divider.   The code in the MCU must takeover keeping EN active as soon as it starts.  However 470nF was a best guess for how long it needs to get going.  If the startup pulse is too short, increase C1 till the MCU starts successfully, then increase it another 50% to have a safety margin.

[EasyGoing1]

The 470nF cap was to prevent 12IN keping EN active when the MCU tries to shut it off.  It only lets through a short pulse to start up the buck module so you shouldn't expect to measure a DC voltage on the divider.   The code in the MCU must takeover keeping EN active as soon as it starts.  However 470nF was a best guess for how long it needs to get going.  If the startup pulse is too short, increase C1 till the MCU starts successfully, then increase it another 50% to have a safety margin.

OK, I bumped it to 680nF and it worked so the next size up that I had was a 1µF. I'd have to put the scope on it to measure the time that it stays on but its probably around 10ms or less based on what the meter shows me. I had to add a resistor to drain the cap... this is the final circuit:

[EasyGoing1]

increase C1 till the MCU starts successfully, then increase it another 50% to have a safety margin.

It's a nice circuit. It's simple yet effective and the short burst on is a nice touch. I can't foresee it ever causing a problem for the Pico since the voltage applied to EN is actually under 3 volts (like 2.6-ish). So much nicer than using the SS Relay.

OK, now I have questions:

1) Am I correct in assuming that once the cap charges, the full 12 volts will be applied to that 50k resistor feeding the base of the transistor?
2) The 1N4001 seems to be working (though I'm not sure what to expect if they weren't working). Do you foresee a problem sticking with this diode?
3) How did you find the datasheet for the buck converter based on only having the data from the Amazon page?
4) How did you know that the resistor on pins 2 and 6 was the one causing EN to default high?

Edit: I just confirmed that when in the off state, it only draws .3µA from the 12 Volt source feeding the buck. Life is good 

[Ian.M]

• The 50K resistor always gets the full 12IN voltage - the cap doesn't affect it.
• 1N4007 is slow and has a lot of junction area (its rated 1A continuous) so may let too big a reverse recovery spike through to the MCU.  Replace with 1N4148 or 1N914 or other fairly fast small signal silicon PN diode.
• Generally you google: SMD code "<code>" <package>
  where <code> is the code on the chip and <package> is what package its in e.g SOIC-8 or SOT-23-6 (not your chip) and if you are lucky you'll find a site that lists what the chip may be in the search results. If no result and the code is >3 characters, drop one character from the end of the code and try again.
  When you've got a candidate chip, google its datasheet see if the function and pinout make sense and and look at the datasheet's package markings section to check for a match.
• The datasheet has a sample circuit + explanation of its operation - see the section I quoted in Reply #9, then the pinout tells me which pins the EN pullup must connect to.

[EasyGoing1]

• The 50K resistor always gets the full 12IN voltage - the cap doesn't affect it.
• 1N4007 is slow and has a lot of junction area (its rated 1A continuous) so may let too big a reverse recovery spike through to the MCU.  Replace with 1N4148 or 1N914 or other fairly fast small signal silicon PN diode.
• Generally you google: SMD code "<code>" <package>
  where <code> is the code on the chip and <package> is what package its in e.g SOIC-8 or SOT-23-6 (not your chip) and if you are lucky you'll find a site that lists what the chip may be in the search results. If no result and the code is >3 characters, drop one character from the end of the code and try again.
  When you've got a candidate chip, google its datasheet see if the function and pinout make sense and and look at the datasheet's package markings section to check for a match.
• The datasheet has a sample circuit + explanation of its operation - see the section I quoted in Reply #9, then the pinout tells me which pins the EN pullup must connect to.

Thank you again for the time you took to help me figure this out. Your logical approach to the problem was clearly necessary. I'm really happy with how well it works. It's exactly what I wanted.

Mike

[EasyGoing1]

Replace with 1N4148 or 1N914 or other fairly fast small signal silicon PN diode.

This SMT component kit I got looks to have two different kinds of 1N4148 as you can see here ... do you have any idea what the difference is between these two?

[Ian.M]

The package!  LL-34 is a MELF style glass package, similar to the original thru-hole package but with flat disc ends.  SOD-123 is the usual rectangular epoxy 'brick' package and has the major advantage of not rolling away as you try to solder it!

[EasyGoing1]

The package!  LL-34 is a MELF style glass package, similar to the original thru-hole package but with flat disc ends.  SOD-123 is the usual rectangular epoxy 'brick' package and has the major advantage of not rolling away as you try to solder it!

What an odd thing to do ... make a cylinder into an SMD component.

[Ian.M]

Not so odd.  MELF is very much a technology of its era when the electronics industry was transitioning from thru-hole to SMD.  The component manufacturers knew how to package components in little tubes with wire ends, and switching to little tubes with disc ends wasn't too big a change.  Its hung on for parts where a true hermetic metal to glass seal is desirable, or for greater dissipation than epoxy packages if for whatever reason a ceramic package can not be used, as little externally square glass tubes cant be practically manufactured at a reasonable cost vs little round glass tubes.  *Nobody* likes it from an assembly point of view!