20x4 LCD Issues - CLOSED

[Gabriel]

Hi All,

I use the following screen:
https://www.sparkfun.com/products/9568

The schematic is based on this:
https://www.sparkfun.com/datasheets/LCD/Serial%20LCD%20Backpack%20v28.pdf


This screen is famous for at some point showing 2 black bars and not responding to serial commands.

The Serial part of it is handled by an on-board PIC, and i always assumed that the black bar issue was caused by poor/rushed serial firmware.

So as a test i decided to reprogram the on-board PIC16F88 with the Flex_LCD driver by PCM.
The problem persists, i need to cycle power multiple times and eventually, sometimes, it works.

I have a feeling this might be a timing issue when both the LCD Controller and the PIC are initializing, but I've tried disabling pin hardware, start up delays, etc.. with no success.

I'm hoping someone here has suggestions on what could be causing this.

Thanks
G.
_________________
CCS PCM 5.078 & CCS PCH 5.093

[temtronic]

That schematic doesn't show ANYTHING connected to _mclr....
I don't see any good VDD filtering....
Are all nonused I/O pins set low or high ?
Is the VDD supply rock stable and good for say 1 or 2 amps ?

I have to say I use the 20x4 LCDs all the time for years, using flex_LCD and have never had an issue (aside from flippin RS and E then NOT changing the code.....oopsy).

I think flex has a 35ms delay in the lcd_init function, though I always do.

[Ttelmah]

What is the part number of the resonator?.
What fuses are you using to set this up?.

My guess would be a processor startup issue.

As an ongoing comment you say:
"but I've tried disabling pin hardware, start up delays, etc.."

How long have you got for the startup delay?.
The original Hitachi chip required 90mSec after start, before any commands
could be accepted. The PIC however generally wakes before most LCD's
and many of the 'clone' drivers need up to perhaps twice this time. So
something like 250mSec before you try to send anything to the display
is a much safer delay.

[Gabriel]

Holy smokes... i never though of double checking the MCLR circuit, i assumed it was correct since it was an old Sparkfun design. I am a bit disappointed i must admit.

I checked the hardware and indeed MCLR goes straight to the breakout pin/header.

I have not had a chance to add the diode/resistor/pullup combo. But as soon as i get home tomorrow I'm on it and I'll report back with fuses and the rest of the necessary diagnosis info.
_________________
CCS PCM 5.078 & CCS PCH 5.093

[Gabriel]

So, turns out the MCLR circuit is pretty important. Who would have thought that! *sarcasm*

There's a lesson in trust for me. Even with the diagram on hand i could not see the error. It was just not possible the board didn't come with some kind of minimal MCLR circuit.

Thanks.
_________________
CCS PCM 5.078 & CCS PCH 5.093

[ezflyr]

Hello Gabriel,

I was bitten by a similar oversight by the Sparkfun 'engineering' staff a few years ago while working with one of their Cellular development boards. After several days, I could not get it to work as intended, so I dug out the manufacturers hardware development guide and found numerous departures from the recommended circuits. I modified the Sparkfun board to comply with the manufacturers recommendations, and voila!, everything began to work as expected!

Lesson learned!

In general, I like Sparkfun & their products, but this experience reinforced the notion that there is a BIG difference between a 'maker' and an 'engineer'!

As Ronald Reagan said "trust but verify"!
_________________
John

If it's worth doing, it's worth doing in real hardware!