0
\$\begingroup\$

If I wave solder in the battery, it begins draining rapidly until the RTC chip (clock) is set. So I installed a jumper. Now I manufacture without the jumper, and then in the bed of nails testing, I set the clock and install the jumper before I package it and put it in stock. I never EVER want the jumper removed, but it is happening. Is there a way to make a lockable jumper or other ideas for a solution?

\$\endgroup\$
  • 3
    \$\begingroup\$ Why is the jumper being removed? \$\endgroup\$ – Barry Dec 4 '19 at 16:23
  • 3
    \$\begingroup\$ What is the RTC chip? Is this a known behavior the manufacturer will admit exists and has no workaround? What about socketing the battery so that it can be installed later (and eventually replaced)? What about soldering a shunt? \$\endgroup\$ – Chris Stratton Dec 4 '19 at 16:25
  • \$\begingroup\$ It is being removed because they don't know what they are doing, and end up trying everything. They shouldn't \$\endgroup\$ – Robert Farley Dec 4 '19 at 16:40
  • 1
    \$\begingroup\$ few != none, there's nothing here to suggest that an exotic battery type is being used, and such solutions often bring associated product challenges. \$\endgroup\$ – Chris Stratton Dec 4 '19 at 20:17
  • 1
    \$\begingroup\$ @MarcusMüller I went to say that LiSOCl2 typically claims 30 years. I see Tadiran have upped the ante :-). || IF they can main tain the 0.7% loss / year that would give 50% capacity at 100 years, and 25% at 200 years. IF - which seems unlikely, but. || They say ". Due to very low self-discharge, Tadiran cells can survive accelerated testing for 90 months at 72°C (the equivalent of hundreds of years of continuous operation)." \$\endgroup\$ – Russell McMahon Dec 6 '19 at 1:00
1
\$\begingroup\$

You could have a solder jumper land pattern that does not have solder on it normally (opposite side of board from wave soldering). Only apply a solder bridge when required. There are many land pattern designs out there if you google them.

\$\endgroup\$
  • 1
    \$\begingroup\$ I like this idea, too bad the ICT bed of nails can't solder. \$\endgroup\$ – Robert Farley Dec 4 '19 at 18:06
1
\$\begingroup\$

Could you wave solder in a battery holder instead, and install the battery later in your manufacturing process?

Here's an app note from TI about this sort of thing using their parts: https://e2e.ti.com/blogs_/b/powerhouse/archive/2014/06/05/maximize-shelf-life-with-a-one-time-push-button-switch

\$\endgroup\$
1
\$\begingroup\$

More ideas:

  • conductive paint dabbed onto a SMD resistor footprint
  • use a simple graphite pencil to shorten two pads (your RTC doesn't draw enough current, hopefully, to make these couple 100 Ω a problem)
  • make through holes in the right distance. Use an office stapler to close the circuit
  • big fat REMOVING THIS JUMPER DAMAGES THE DEVICE AND IS DETECTABLE silkscreen or neon sticker
  • through-hole rivet
\$\endgroup\$
1
\$\begingroup\$

You could add an inexpensive SOT-23 or SC-70 MOSFET and a (say) 10M resistor and put the jumper in place to turn the RTC off.

There would be a drain of a few hundred nA, probably less than the battery self-discharge, until the jumper is removed.

schematic

simulate this circuit – Schematic created using CircuitLab

\$\endgroup\$
  • \$\begingroup\$ I suppose nobody make a discrete antifuse. This idea however seems like something that could be opened automatically with a very low current conventional fuse, or else a through hole resistor or loop of wire that gets manually clipped off. \$\endgroup\$ – Chris Stratton Dec 4 '19 at 20:19
  • \$\begingroup\$ @ChrisStratton As far as I know, nobody makes one. The "zener zap" method is used in some chips for trimming, but I doubt OP wants to characterize ordinary parts for that. \$\endgroup\$ – Spehro Pefhany Dec 4 '19 at 20:40
  • \$\begingroup\$ An antifuse (great word) is what I was thinking. I was wondering if there existed something like that that would maybe arc over and make a connection when hit with a strong pulse of electricity. Something like you can buy to repair a string of Christmas lights; you put it into the string of lights, pull the trigger and if must fuse something in the light string to "repair" the string of lights. \$\endgroup\$ – Robert Farley Dec 5 '19 at 19:29
  • \$\begingroup\$ I don't know of any discrete antifuse product. A quick test verifies that an LL4148 subjected to a 5A/5V (power supply settings) for about 0.5s will perform this function (also with some visible light indication- red LED?) but characterizing it for a production application, including reliability, would be non-trivial, to say the least, and you would have to ensure the "programming pulse" didn't screw up the other circuitry. \$\endgroup\$ – Spehro Pefhany Dec 5 '19 at 19:40
0
\$\begingroup\$

If you can find a wire wrap gun somewhere, you can make a permanent no-solder jumper (assuming the current jumper is a standard 0.1" square pin).

\$\endgroup\$
0
\$\begingroup\$

The RTC datasheet says that after it has been totally unpowered, the device only needs the main supply shortly after connecting battery. Apply the main power to board just after manufacturing so it stops consuming battery, even if it is being tested later.

\$\endgroup\$
  • \$\begingroup\$ Might be tricky, because the datasheet says that Vcc initialization needs to be done every time main power is lost. This seems painful enough on its own that I'd probably be shopping for a new RTC, myself... \$\endgroup\$ – spuck Dec 4 '19 at 20:48
  • 1
    \$\begingroup\$ No, it does not say that. VCC is tne main power. It says if all power has been lost, VCC initialization needs to happen by applying VCC for at least 1 millisecond after battery has been connected. \$\endgroup\$ – Justme Dec 4 '19 at 21:22
  • \$\begingroup\$ Section 8.2.2.3 of the datasheet says after "after every total power loss situation". I misspoke in my previous comment when I said every time "main power" was lost. I apologize. \$\endgroup\$ – spuck Dec 4 '19 at 21:26

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.