0
\$\begingroup\$

I have a simple circuit I built using a 7-segment LCD glass display and an ATTiny (just a simple timer for now). I was hoping to use the setup in a low power, single coin cell device (3.3V - 2.7V), but then I noticed that the recommended voltage on the LCD is 5V. I looked for lower voltage LCD panels (they use them in the little freebie calculators that give away, and those have little coin cells) but either I'm looking in the wrong place or they don't exist.

Never mind, I just tried using lower voltages, and I can almost get it usable at around 3V by putting a large (\$>10M\$) resistor from the signal pins to GND (somewhat legible at 2.9V and 50Hz PWM at 50% duty). I thought this might work because I think the segments are capacitive so the resistor increases RC and causes the segment to stay lit longer.

So my question is, does anyone know if there is a secret to getting these LCD panels to work at lower voltages? Ideally I'd like to get down around 2.5V and still somewhat readable, and I imagine it's possible since I've seen simliar devices that must be working from small 3V batteries. Or am I barking up the wrong tree and there exist LCDs somewhere that already use a lower voltage?

And before I get "there are 3V panels listed on Mouser, Digikey, etc.", I see that some of them come up as such in the product databases, but they are not <3 digits (I only need 2 digits) and the datasheets/technical drawings/whatever all still list 5V anyway, so I'm guessing they are all the same.

\$\endgroup\$
  • 2
    \$\begingroup\$ There's a lot missing from your question. But given that these are AC devices what you should probably be doing is driving the common segment with one output pin, and then driving the segments you want to activate with a signal of the opposite phase, and the segments you do not want to activate with a signal of the same phase. If what you have is an effectively raw element (as the AC specification hints, though the voltage specification counter indicates) a DC voltage on them won't really work unless it is pulsed, and unipolar pulses are bad in the long run. \$\endgroup\$ – Chris Stratton Oct 29 '18 at 14:06
  • \$\begingroup\$ Sourcing questions are off topic, but if you want a low voltage raw element with fewer digits you could try salvaging something from a clock. You'll probably have to deal with multiple common selects there. And either make a board compatible with an effectively iron-on ribbon (get it off the old one with a hair dryer) or with the pads perfectly placed for the zebra strips (careful not to break them). \$\endgroup\$ – Chris Stratton Oct 29 '18 at 14:08
  • \$\begingroup\$ @ChrisStratton That is exactly what I am doing, opposite phase between signal and COM. Then a resistor between signal and GND gives me very much increased segment darkness. \$\endgroup\$ – TrivialCase Oct 29 '18 at 14:42
  • \$\begingroup\$ @ChrisStratton If sourcing information is off-topic here, where can I go to discover this? I might be interested in obtaining these screens in medium quantity but can't seem to find them (and other types of LCD seem to have killed the ability to google effectively). Of course this only applies if sourcing a new part is what I need, it could be that there is some other combination of pulse width and RC(L) that people use to get the same result. \$\endgroup\$ – TrivialCase Oct 29 '18 at 14:44
  • 1
    \$\begingroup\$ If a resistor to ground is changing things, you don't have a proper push-pull driver configured - but rather probably have a pin configuration mistake. Have you probed both the common and segment lines with a scope while nothing is connected, and ideally seen opposite phases on a dual trace scope?. In terms of sourcing questions, it has never been the intention of the Stack Exchange network to cover everything that someone might be interested in, rather it is an intentional decision to only cover topics which fit this particular model well; there's a whole other Internet for the rest. \$\endgroup\$ – Chris Stratton Oct 29 '18 at 14:57
1
\$\begingroup\$

You should be driving the common in sync with the segments, either in phase or 180° out of phase (depending on the on/off state of the segment).

The DC content of your drive waveform should not exceed some tens of mV or you will damage the display permanently. If you are driving your (static) display with single ended port pins you will get less contrast and will damage the display eventually. To minimize the DC content you want the out-of-phase waveform to be very close to 50% on off ratio. You can work out the permissible error in time from the frequency (which should be in the range 32-100Hz for this display).

5V is required for Lumex displays that operate up to 80°C, 3V is okay for 60°C. With your 3V supply (and properly driving the common) you will be driving the display with 6V RMS so it should be fine.


It's quite possible to source LCD displays that fit your requirement, in fact custom displays are more common than standard ones. There is an NRE cost and usually a minimum order quantity. You get to specify the type of drive (static or the more common multiplexed), pick a fluid temperature range, a viewing angle, glass size, type of termination (zebra strip, FPC or pins) and so on. Figure on a starting cost in the low thousands of dollars for an ordinary display.

\$\endgroup\$

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.