We all familiar with driving common character LCD using a few wires via some kind of latch or shift register.

The following shows how someone in China can do that *Without* any additional ICs.

The LCD shown is NOT a serial LCD.

The source of the pictures are from www.ourdev.cn . The webpage http://www.ourdev.cn/bbs/bbs_content_all.jsp?bbs_sn=4301955 shows how it is done, along with C source code.

As a challenge, the poster has also shown the following picture, using only one wire. It doesn't get any better than that.

For normal 7-segment display, he offers this picture, driving three 7-segment LEDs using a single 74HC595. Can you figure it out?

 

I've used a 4-pin interface before for a 3-digit 7-segment display....

Don't really care much for the 'trick' RC time constant type interfaces. For example, the Roman Black 1-pin 74HC595 LCD interface takes over 1-msec to load the shift register (yuch!).

Regards, Mike

 

I suspect he's using some kind of multiplexing for the segments. The cathodes of the LED display are driven by one pin so perhaps this pin is a clock and by loading the register fast enough the display appears to persist. Perhaps he is using two pins on the shift register to extract the LED display to light. Either way it's quite clever.

 

I cant read any of it (except whats on the LCD).. But this looks quite cool:

Wirelessly POWERING AND TRANSMITTING THE DATA signal! Awesome.

 

It is getting interesting with each day passes.

The author/designer has not revealed the actual circuit diagram or control codes yet. He only said the wireless setup took him about a day from concept to being a working prototype.

Here two more photos with additional info added by myself. The source of photos is the same as mentioned in post#1.

 

I have an idea.

For a 7 seg LED only displaying only numbers, certain groups of segments always light at the same time.

For example the B and C segments are lit during the numbers 1, 3, 4, 7, 8, 9 and 0. And the A, G and D segments are lit during the numbers 2, 3, 5, 6 and 8.

With some clever multiplexing this could be used to control the displays. I think only 5 pins on the shift register would be needed to control each display, and the remaining 3 could be used to multiplex the displays individually.

 

Actually, no, that idea wouldn't work.

I think I've figured it out. It's really quite clever.

The capacitors near the serial input form an RC circuit which delays the clock. By adjusting the RC constants to match the data rate, the clocks can be phase-shifted, providing clocks to each display.

I'll try and get a little demo circuit up.

While it's clever it is nowhere near being practical. It simply would be too susceptible to clock variations and capacitor tolerances.