1 wire LCD driving

L

lanternfish

Senior Member
MPep said:
Hi all,

Whilst browsing the 'net, found this http://www.romanblack.com/shift1.htm. Should be PICAXEable.
He also has plenty of other interesting ideas/circuits that are well worth browsing his website for.

MPep.
Click to expand...
Thanks for posting that MPep. I haven't seen that clock/latch delay circuit for a long time and an acquaintance was asking me about something similar the other day - I couldn't for the life of me remember how it was done.

Simple, elegant and highly functional.
 
M

moxhamj

New Member
Goodness me, that is highly cunning. I like it! Very useful for an 08/08M where it frees up pins for inputs.
 
W

westaust55

Moderator
Agree with Dr Acula that it is certainly a cunning/novel approach.
But as the Roman Black article mentions, it is slow.

By comparison with test I have done recently on msec/char data transfer:
0.006ms for a 6-wire direct –PICAXE to LCD module
0.140ms for a 3-wire 1 x 74HC595 shift register in 4-bit data transfer mode
0.060ms for a 3-wire 2 x 74HC595 shift register in 8-bit data transfer mode
0.055ms for a 2 wire 2 x 74HC4015 shift register in 8-bit data mode WITH back light control
4.00ms for the Roman Black 1-wire 1 x 74HC595 method with/without back light control

Note that the 2 x 74HC4015 method incurs some delay as it checks for the state of a backlight flag with every character sent.

There are certainly times when a 1 wire serial interface is needed, particularly with smaller PICAXE having limited IO. But reduced data transfer speed also means reduced speed for the remainder of the program as well.

I am currently working towards construction of the 2 x 74HC4015 shift register method – awaiting the IC’s to test. A proto bard and code is ready for the testing.
 
I

inglewoodpete

Senior Member
Isn't a serial LCD like the AXE033 a 1-wire device? At 4800baud, thats about 2mS/character and saves heaps of precious programme space.
 
M

MPep

Senior Member
I only posted it as an alternative idea. Yes I did read that it wasn't the fastest, but what is the point in following the already trodden path when the alternative path may be more fun? And educational!
 
W

westaust55

Moderator
Not saying that ROMAN BLACKs 1-wire LCD should not be built.
Just comparing the speed of various methods that I have looked at.

Yes, the AXE033 is a 1-wire device. Rev Ed literature does indicate that it is only 2400 baud (2400,N,8,1) ==> 4ms/char.

Certainly go ahead and have a go builting the Roman Black interface.
I for one cannot at this time offer a quick and cheap alternative.
The 2-wire method I am looking at is already derived from what others have done before me.
 
Last edited:
hippy

hippy

Technical Support
Staff member
If one can set the Data line RC timing so 10us = 1, 170us = 0, then this should be quite fast ...

SendB0:
b0 = b0 ^ $FF
b1 = 1
bit12 = bit0 : PulsOut TX_PIN, b1
bit12 = bit1 : PulsOut TX_PIN, b1
bit12 = bit2 : PulsOut TX_PIN, b1
bit12 = bit3 : PulsOut TX_PIN, b1
bit12 = bit4 : PulsOut TX_PIN, b1
bit12 = bit5 : PulsOut TX_PIN, b1
bit12 = bit6 : PulsOut TX_PIN, b1
bit12 = bit7 : PulsOut TX_PIN, b1
PulsOut TX_PIN, ? ' Load pulse
Return
 
Last edited:
You must log in or register to reply here.
Top