1

I made a mistake on a board recently when adding an optional PIC12F615 processor to supervise my ARM processor.

I didn't realize the PIC needed >4.5V to erase.

So, I've got this MIC2920A 3.3V regulator feeding the PIC (and only the PIC) and I need to let the programmer pod apply 4.5V to it.

Will it work? I don't see anything in the regulator datasheet that says it will or won't. The datasheet discusses how "bulletproof" this thing is, capable of handling -20 to 60V on the input pin. I don't see anything about high voltage on the output pin.

I've got 100 fairly expensive PCBs like this (20 stuffed). I can modify one board for development with a diode or reverse-power-protecting mosfet, but I'd like to be able to easily reprogram these boards later on. I'd planned on simply leaving the PIC off entirely if it didn't work... but it almost works.

EDIT

The regulator's input is a regulated 5V, so I wouldn't be going over that.

EDIT 2

It works!

I hooked up a bench supply, set it to 4.5V with a 30mA current limit, and then tried connecting it to the VDD pin of the ISP header (which is wired to the regulator's output pin). I didn't see any measurable current.

I was unable to force the programming pod (PICkit3) to supply 4.5V when there was already 3.3V present, though. The regulator had to be unpowered to program the PIC.

However, this regulator seems fine with it... the GND and VIN pins are 0V, and it draws no current or at least no measurable current.

I was able to erase and reprogram the PIC. Thanks!

darron
  • 3,491
  • 2
  • 29
  • 42

2 Answers2

2

Surmises only - hard to be sure for a specific part with no hard data.

Connecting one as stated and measuring current flow into regulator would tell you something.

A look at the block diagram at the bottom of page 5 of the datasheet suggests you will be trying to reverse bias junctions in Q24 & Q26 (top right of page).I'd expect the 4.5-3.3= 1.2V overvoltage to be low enough to have a good chance of being safe.

Is power applied to the regulator during programming? If not you may breakdown internal intrinsic body diodes (which we may not have got). After which anything can happen.

You mention a reverse protection diode. It's not obvious where you were going to put that. It's usual to place a diode from regulator output to regulator input - Schottky probably a good idea.

Russell McMahon
  • 147,325
  • 18
  • 210
  • 386
  • I plan on trying to apply 4.5V from my bench supply with a really low current limit to see what it does. I can apply power to the regulator or not, either way. I thought as you suggest that it would be safer to apply power to the regulator. – darron Nov 03 '11 at 20:22
  • I really like the reversed P-ch MOSFET as a reverse power protection device concept that endolith commented on in another answer: http://focus.ti.com/lit/an/slva139/slva139.pdf. Since that would prevent ANY voltage over 3.3V on the output pin, I can't think of why I shouldn't use that. 100 boards from now. – darron Nov 03 '11 at 20:48
  • I do like the reverse MOSFET polarity protector / reverse current preventer. I've got probably 100,000 of those in operation in a reversed battery protection scheme where a diode drop of voltage was unacceptable. An excellent solution. I mentioned the backward regulator bypass diode as its extremely easy to retrofit and very cheap. Using the MOSFET, at Vbat = 3.3V you want a Vgs_th of well below that so it's on hard at 3V3. [AO3423](http://www.aosmd.com/res/data_sheets/AO3423.pdf) is OKish and $US0.10 /1000 [at Dikgikey](http://search.digikey.com/us/en/products/AO3423/785-1016-1-ND/1855958) – Russell McMahon Nov 03 '11 at 22:01
  • I managed to get the reverse MOSFET trick working on a 36V input. Accidentally, actually. I placed the gate's voltage divider on the source side of the P channel MOSFET by accident in LTspice, and it works beautifully. I was experimenting with a divider on the input voltage side, and I couldn't get it to fully turn on that way. My analog skills are not yet advanced enough to be that clever on purpose... although that was an eye opener and it surely helped me along. – darron Nov 04 '11 at 00:04
  • Be sure chosen MOSFET is fully driven on by available voltage - as above. What MOSFET? What voltage? – Russell McMahon Nov 04 '11 at 02:24
  • For the 36V protection? (It's actually ~10-36V) IRF4905PBF. I'm not sure why I chose it (Rdson of 0.02 probably), but putting the gate voltage divider across the inputs but BEHIND the MOSFET works great. It wasn't turning all the way off in reverse power with the voltage divider across the inputs. With the divider on the source side, the MOSFET starts on but quickly turns off. A small diode across the back absorbs that ~20uSec blip until the MOSFET is off. – darron Nov 04 '11 at 05:12
  • the fact that the MOSFET partially turns off cuts the voltage divider difference from the negative rail even further, eventually bringing the gate fully to the source voltage. – darron Nov 04 '11 at 05:19
  • Here's what I'm talking about: http://i.imgur.com/BdtUD.png – darron Nov 04 '11 at 05:31
  • That hole on the right is where an optional diode goes to absorb the small kick until the MOSFET is off. The 30 ohm resistor is just simulating the device load. – darron Nov 04 '11 at 05:33
0

IIRC, I think the >4.5V is for bulk erase, but you can still perform a row erase at lower voltages.
If you can set your programmer to erase by row then you should be able to use the 3.3V.

Oli Glaser
  • 54,990
  • 3
  • 76
  • 147
  • This would be great. I'll try to figure out if my PICkit 3 can do that... this would be the best option. – darron Nov 03 '11 at 20:23
  • I don't appear to be able to do this with the PICkit3. I don't see a setting for that anywhere. I also tried marking an address range to preserve, hoping it would erase around it. That didn't work either. Oh well. – darron Nov 04 '11 at 01:47
  • @darron - Darn, sorry about that. At least it works with the 4.5V which is good. I think 1.2V over is unlikely to damage the regulator. I'm surprised the PicKit3 doesn't do row erase easily, I had a quick look at the manual and couldn't see anything either. – Oli Glaser Nov 04 '11 at 02:12