Electronics > Projects, Designs, and Technical Stuff

Low Voltage Detection & Power Cut IC

<< < (10/14) > >>

Ian.M:
I did recommend a screened cable many posts back.   Nobody wants a car you cant start if you parked under a power line or too close to a high power radio transmitting station.

However it does seem over-sensitive to ambient EMI.  What's the actual circuit, both schematic and photos of the layout, zoomed in on the board, and also zoomed right out showing how the leads to all your test gear are routed. 

Also what's your bench lighting?  If its low voltage please give specific details of its transformer or PSU.  It's quite common nowadays for cheaper lighting (and some badly designed expensive stuff) to radiate excessive EMI.   The only types of lighting (apart from actual flames, chemlights etc.) that you can be reasonably confident *WONT* radiate EMI, is non-dimmed incandescents (but not low voltage with a switching PSU), fluorescents with line frequency magnetic ballasts and low voltage DC LEDs with a linear power supply (or battery) and linear LED current control.  Unfortunately incandescents are being phased out in many countries and magnetic ballasts for fluorescents are becoming difficult to find.  If you've got low EMI bench lighting don't throw it out if you upgrade to modern lighting, as you may need to go back to it when testing sensitive circuits!

1Meg||560K=359K, which is quite a high impedance and may be making it excessively sensitive to noise.  As Forrestc just suggested, try reducing the resistor values an order of magnitude or two.  You may well need an external feedback resistor to get enough hysteresis.

Forestc has pointed out how the PIC's various peripheral modules can be used to help de-glitch the sensor low voltage lockout signal.  However IMHO its important to first get the signal as clean as possible, otherwise you risk designing a circuit that works on the bench, but is the weak link, failing in the field due to localised high EMI, that is not strong enough to disable an unmodified vehicle.

JDW:
IanM. & forrestc,

My humble and sincere thanks for kind help and advice you gentlemen are providing.  I made a new video this morning showing my room lighting for Ian.  The same video shows the result of using a smaller 10k & 5.6k resistor divider (with PIC Hysteresis ON):

https://youtu.be/5Xk1dtJ5m4Y

Circuit is on a breadboard using discrete components because it's still in the early design stage.

I am now about to try the CLC suggestion to see if I can program my PIC satisfactorily to address the chatter, ensuring that the fingerprint sensor is never toggled ON & OFF repeatedly by a slow decrease in voltage across the chattering voltage range.  I will report back on that after I've coded and tested.

Thanks.

Ian.M:
Your circular fluorescent probably has a magnetic ballast and therefore would be low EMI.  The give-away clue is the presence of a gas discharge/bimetallic strip starter, as they are usually associated with magnetic ballasts as most electronic ballasts have integrated starter circuits.  I haven't got a clue whether or not the fluorescent strip lighting has HF electronic ballasts.  If it has the characteristic flickering startup of a gas discharge/bimetallic strip starter, (or you can find one on each strip), probably not.   Otherwise check the label and see if you can Google its specs.  Most manufacturers are proud enough of their energy efficient electronic ballasts to boast of them in the specs.

Still no photo of your breadboard or of the experiment setup on your bench.  >:(  Without seeing the actual circuit its very difficult to comment usefully on noise issues.

From the soundtrack of your video I get the impression you have two DMMs with long UNSHIELDED leads connected direct to your PIC's comparator pins.   Those leads will be acting as antennae and picking up EMI and taking it right to where you don't want it.  Try isolating the meter leads from the pins with 100K resistors (which will only cause a 1% readout error on a 10Meg input impedance DMM), and at the meter end of the resistor adding 100nF to ground to reduce any EMI the leads picked up.  Also, reduce the loop area available to pick up H-field EMI by twisting the leads together - 5 to 10 turns per meter is plenty, it just needs to keep the positive and negative leads together for as much of their length as is practical.

JDW:
Ian, thank you for sharing your thoughts on room lighting and noise.  I actually discovered my stupidity about an hour ago with regard to my Fluke 8845A's RED unshielded wire being the source of induced noise into my voltage divider resistors.  When I removed that red probe wire, all the chatter you saw in my videos vanished.  Not a single blip across the voltage threshold from low to high.  But I think the experience was good in that it is a reminded that if noise is present, chatter could still occur even with the PIC's Hysteresis turned ON.  And for that reason, I still need to code the CLC as a preventative measure.

For now, I've connected up a PNP and NPN in place of the FETs we've been talking about, since I don't have any FETs in my parts stock.  I split my Tx line (Tx from my PIC, Rx at the sensor end) across E-C of an NPN BJT (to make it switchable), with 10k base resistor of that NPN tied to the collector of my PNP (2SB772).  The base of my PNP connects to a 560-ohm resistor which connects to the Comparator Output of my PIC (the same Output pin I was probing on my scope).  So basically for a rough test I am using the PNP to power the fingerprint sensor and cut the Tx line too.  In my testing thus far it actually works and I only measured a scant 1mV voltage drop across the PNP to the output.  For testing, I've got a Power Supply that triggers the fingerprint sensor's flash memory wipe code every time I kill power and start power to that Power Supply.  But with the NPN and PNP attached, the memory no longer gets wiped at all.  And again, what all this is doing is ensuring that the sensor's power gets cut when the voltage drops below 2.8v and keeps it cut all the way down to zero.  So far, it's working even with my PIC's brown-out circuit enabled (LO setting).

Of course the PNP base resistor requires a few milliamps of current which isn't desirable, so it's best to use a P-MOSFET instead.  But again, for this rough test, the goal I set out to achieve (preventing flash memory wipes) seems to be working well.  I will create a schematic that graphically explains what I just said later today.

Thanks again for your kind help thus far!

JDW:
Apologies for my delayed reply.  Below is the simplified schematic of what I am testing on a breadboard -- "simplified" in that it leaves off my 3V relay, LEDs and other less relevant devices for greater clarity.



I don't have any MOSFETS in my parts stock, so I breadboard-tested with a PNP.  The PNP works great to kill power to the F.Sensor when Vdd dips below 2.81v, but the PNP's base resistor results allows a continuous 4.2mA current flow as long as the Base is LOW (fed by the Comparator1 OUT), which is most of the time.  If using a P-MOSFET will eliminate that 4.2mA then I should use a FET instead.

Using an NPN to break Tx in the manner shown in the schematic actually works, and signal integrity looks fine on the scope, but I am unsure if that's a sound approach -- it is a bit different than what forrestc recommended in (1) of his earlier post.

My switching power supply has internal protection against the output (3.3v@250mA) being shorted to ground, so I think the PNP's collector (which powers the F.Sensor) should be OK as shown on the schematic.  Leaving Ground as a straight-through connection between boards is probably okay as well for the same reason.  And the 3 signal wires have basic diode ESD protection with resistor as shown (for the PIC side circuit board).

I've not had time to do the CLC or the PIC internal voltage division (to eliminate R1 & R2) that forrestc suggested, but I intend to do that soon and will report back at that time.

Thanks to IanM. and forrestc for your kind help to date.

Navigation

[0] Message Index

[#] Next page

[*] Previous page

There was an error while thanking
Thanking...
Go to full version
Powered by SMFPacks Advanced Attachments Uploader Mod