an idea i have is to use 1 shift register to drive all segments derectly or though a bcd to display driver chip (lession the data required) and to chose the display to use a shift register or shift counter of some kind controlled buy the pic and with this 2 wires could control as many displays as wanted could even add another shift register to make it 1 wire

if some1 wants to design this and give Weazel some credits im keen fo this circuit XD
want it to mesure voltage amperage frequncy and phase angle
to run off a 8 pin pic chip be cool and if theres a way to use this idea with 74 or 4000 series chips be cool to

 

Something like this perhaps?

Regards, Mike

 

View attachment weazel-plex idea 7 segment.bmp


the display chosing chip could be a flipflop controlled by the pic i hope u get the idea and maybe to save power and resistors only 1 segment runs at a time XD XD
if works run loads of displays with only 1 wire and very minamiul chips

 

Something like this perhaps?

Regards, Mike

Roman Black's Shift1, eh?

Clever. We used to use a very similar scheme back when the earth's crust was cooling, except we used two resistors, a cap and a 1N60P or similar Schottky germanium diode. The diode re-charges the cap quickly through the 100 ohm resistor. This provides for very fast recovery after the signal line goes high so that you don't have to wait so long after zeros or latches. Doing this you can more than double the data rate.




Kind of like this:

I guess what's old is new again, except not quite so good.

Sorry for the hi-jack!

 

Hi Bill,

It is an old technique but you have to give him credit for putting it all together in a simple cohesive 'system' package for newcomers. Also, please keep in mind that he uses the technique for a single 74HC595 driving an LCD display so the timing (high tau) isn't nearly as objectionable as it might be when driving a long string of 74HC595s.

Regards, Mike

 

... i hope u get the idea ...

Sorry, not quite following you. It seems you're still looking at two or three pins.

If you're trying to get it down to a single pin interface, why not just build a serial backpack? A simple circuit for 20 displays (expandable) with PWM brightness control is shown below. Then you'd only need a single pin on even the smallest host mcu, like the six pin 10F200 or 10F320 series devices, and they don't have the overhead associated with mux'ing the displays.

Regards...

 

This is a huge shift register specifically for driving 7-segment displays...though it may not be quite what you were looking for, you might find it useful to know about if you didn't already. Have fun

http://www.digikey.com/product-detail/en/MM5450YN/576-1341-ND/771810

 

Here's a 16-digit module for $7.60 from DealExtreme... I don't think I could buy the displays for that price.

 

im wanting to cut down the componet and pin count and cost its to run vfd and 7-segment display i dont want to have to buy big mcu or stacks of shiftregisters and for the vfd i dont want 100's of level switching transistors and if posible be nice to run 1 segment at a time sos esay to set brightness and keep currents low im not i know lots bout electronics but not everythink it was a ruff idea since the segment driver shift register has free ports could use 2 bcd to 7-seg chips and could use some kind of shifting flip/flop timed buy the pic to shift display lighted
im wanting a full circuit 2 run with 8 pin mcu want mcu to be as cheap as posible and run ur basic bcd input or volt amp frequcny etc stuff
and if posible to use the idea with out mcu useing only 74~4000 chips

 

i have tons of displays for free

 

I'm not sure if English is not your mother tongue, or you are trying to speak gansta, but, either way, it might be causing the communication to fail.

It sounds like you want to do this with a 2 pin MCU and a single LED. Being a designer and manufacturer of MCU development systems, I do understand your desire to build something of great value for nothing. You just need to know that that idea is not fresh. One also needs to understand the concept of feasibility.

Based on the obfuscated information you have given, good ideas have been presented here. You need to take them to the next level, or go hire someone that can. While, from time to time there are people on this site that will go to great lengths to help, it's not often we are willing to do everything.

What are you expecting? Would you like someone to design this for you, build the prototype and develop the code, then turn the completed project over to you? While all the time you tell us 'jus naw dan wid it, yo need tu spec id up'.

na y'all got sum booty tu spred rown yo? ma be u git sum foo down wid dat y'all goin to git dis
u said u is keen fo this circuit how keen ar u?

 

since its somethink others would use some1 mite make it and most the code is out there it just needs to be moded to the task
some1 may make it or not im not worreid eatheir way just wanted to put the idea out and see wat bites
its part of a bigger project of a home built universail 3 phase-brussless motor driver-1 ph start run winding driver psu want to be able to run many motors off it change phase angles voltages currents frequncy and drive brussless motors
i have a good start on that owning the smartdrive engineers manual there just need to get copy of there workshop software


mi english is poor because im part dislexic and have aspegers

 

...
Also, please keep in mind that he uses the technique for a single 74HC595 driving an LCD display so the timing (high tau) isn't nearly as objectionable as it might be when driving a long string of 74HC595s.
...

Is there a problem driving a string of shift registers? Only the last pulse of the entire string needs to be the long pulse.

...
We used to use a very similar scheme back when the earth's crust was cooling, except we used two resistors, a cap and a 1N60P or similar Schottky germanium diode.
...

Using an RC delay to derive both clock and data pulses is a very old system and I said that on my page. To my knowlege the system I publicised that uses two RC delays to derive clock data AND latch from just one line had never been done before in actual use, although someone had probably thought it up just like I did.

There wasn't much benefit using that type of system in the logic chip days, but now with microcontrollers easily being able to make any desired time periods on the output line it's quite a workable system for the low cost of 2 R and 2 C.

The big flaw of the system is that the last data pulse of the string is unusable, it's always a 0 because it has to be the latch pulse as well. For LCD use (as Mike said) the system is fine as the LCD only needs 6 pins so losing one of the 8 shift register pins is not a problem.

 

... Also, please keep in mind that he uses the technique for a single 74HC595 driving an LCD display so the timing (high tau) isn't nearly as objectionable as it might be when driving a long string of 74HC595s. ...

Is there a problem driving a string of shift registers? Only the last pulse of the entire string needs to be the long pulse.

I didn't mean to suggest there was a problem driving a long string of daisy-chained shift registers. Just that it might take a long time.

 

im using it to run a 11 digit vaccume fluresant that needs 80v so 11 shift registers is a little to much and 144 transistors is totaly overboard

 

Hi Bill,

It is an old technique but you have to give him credit for putting it all together in a simple cohesive 'system' package for newcomers. Also, please keep in mind that he uses the technique for a single 74HC595 driving an LCD display so the timing (high tau) isn't nearly as objectionable as it might be when driving a long string of 74HC595s.

Regards, Mike

I do. I said it was clever, and I meant it. I just think it could be improved a bit.

 

im using it to run a 11 digit vaccume fluresant that needs 80v so 11 shift registers is a little to much and 144 transistors is totaly overboard

billbo,

There is no other way to do this with a few wires. Sorry. Using a 1,2 or 3 wire interface the display data will have to be latched, or stored somehow by the display. If you are going to multiplex between 11 digits, you will need to get different data to the display as you enable each digit as well as data to select each digit in turn. You would need to be able to send interleaved data about what needs to be displayed and by which digit, many hundreds, if not thousands of times a second. You would need a lot of intelligence at the display to be able to separate the display data from the multiplex data and be able to frame it all.

What is the brightness of these displays at 9% duty cycle?
Do you really need 11 digits?
Why do you need to use vacuum fluorescent displays?
Why do you need to use such a small (8 pin) MCU?
How are you going to drive 80V displays with 4000 series logic?
Why do you need to use 4000 series logic?


If you were really keen on this I suggest you consider one of the previous suggestions, or use a modern display device, like a serial LCD, parallel LCD, or even and LED display. Even the LED display could be done with a single MCU with about 24 I/O and 8 resistors. Dirt cheap and easy as heck.

Bye now...

 

The big flaw of the system is that the last data pulse of the string is unusable, it's always a 0 because it has to be the latch pulse as well. For LCD use (as Mike said) the system is fine as the LCD only needs 6 pins so losing one of the 8 shift register pins is not a problem.

Not a real flaw if you load the display LSD first ending with the DP of the MSD, which would be easy to avoid needing. Or another idea, for a N digit display, use N+1 latches. The extra latch can be used for 7 bits of control data like brightness control, sign display (+/-), or other indicators, etc...

 

billbo,

There is one more solution to this I can think of. There is the purpose designed chip called the MAX7219. It handles all the multiplexing for a 8 digit display. I'm sure you could interface it to your ancient 80v fluorescent displays, but that would be up to you. It uses a 3 wire (+ ground) SPI interface.

They are cheap, all the work has been done already and you can read about them and how to use them right here.

 

I do. I said it was clever, and I meant it. I just think it could be improved a bit.

Sorry BillO I should have probably commented on your suggestion of improving the system.

Re using a diode for fast recovery I had actually considered that, but decided to stay with the very simplest of hardware ie just an R and C.

My reasons were that simplicity makes it more accessible to people who are inexperienced in hardware design, there are a lot of people clever with code but who start to get nervous with asymmetrical timing networks or adding multiple components etc. The other reason was that it's a relatively SLOW system anyway, and only suited for slowish tasks like display driving so I did not think speeding it up slightly was going to be worth the extra complexity.

Really, if speed is an issue it's best just to use the latch IC and 3 control wires, or in some cases you can just use the data and clock wires.