Problem with 16x2 display, blocks (as if 8x H is dispalyed.)

Hi all,
I have an issue with a 16x2 lcd display from Gleichman.
After connecting all and trying the “hello world” I get Blocks like HHHH HHHH displayed at pos 9-12 on both lines.

Anyone have a comment on that?

Did I blow the display or is it a software / programming / library issue?

Or even reset (if possible) of the lcd registers?

When I only connect the display power no blocks are shown, but after initialising Arduino (due) the blocks appear. The rest is ok, showing the expected texts and everything.

You're probably not going to want to hear this but it is probably just a bad connection between the lcd and the pcb. If you look closely you should see small metal tabs that are twisted on one side of the pcb. Those secure the lcd to the pcb. We have had other posts here where people claim they were able to fix the problem you have by straightening out those tabs, removing the lcd and cleaning the contacts and remounting it. I've never tried it but everyone who has said they had nothing to lose because the lcd was useless the way you see it now.
I would have to agree with that assessment. It is s no lose proposition. Do some more searching and maybe you can find the forum post where they walk you through the whole procedure.

allright...

Thanks, now I at least have something to go on, whether it is good or bad. Being a new on this I did not have a clue on where to look for.

Thanks again and I will search the forum on this.

BR,
Herman

Look at this post just in case it is related:
http://forum.arduino.cc/index.php/topic,128635.0.html

Just took the lcd display apart, no anomalies found on the pcb, everything was clean and I cleaned it again. (Even my soldering was clean....).

To test it, I turned the screen 180degr. around and pushed it on the pcb. With the result that the "blocks appeared in the same location as before.
Conclusion: The lcd is driven to show the blocks as they appear, which tells me that the chip is probably broken.

I probably messed it up when testing to get some life in it, by over powering it because of a wrong wiring for the lcd contrast on pin 3.

So in anyway, thanks for the comments and links, those were usefull.
BR Herman

Wow! A well-focused picture for once! Full marks for focus but frankly, way too big!

Herman01:
To test it, I turned the screen 180degr. around and pushed it on the PCB. With the result that the "blocks appeared in the same location as before.
Conclusion: The LCD is driven to show the blocks as they appear, which tells me that the chip is probably broken.

Double Wow! I am fascinated that in turning the display around, it actually worked. Clearly the design is symmetric which is quite sensible, but I really would not have anticipated that.

You will note there are two "COB" - Chip On Board - chips; the master driving the left hand side of the display and the slave, the right. The blocks in question are the first eight to the right of the centre line; indicating a break in the data to the slave chip, not the drive to the display.

haha, yeah sorry for that picture size....
Thanks for the clarification on the lcd functioning part. Makes sense to me now.

Finally I realized which contacts were affected.

... The blocks in question are the first eight to the right of the centre line; indicating a break in the data to the slave chip, not the drive to the display. ...

All eight columns on the right side of the display are controlled by the auxiliary controller and four of those columns are displaying the correct data. Since the data from the main controller to the auxiliary controller is undoubtedly multiplexed I can't see how the problem could be in the data connections between the two chips if half of the data is getting through.

It looks to me as if the problem is somewhere between the auxiliary controller and the display.

Don

floresta:
Since the data from the main controller to the auxiliary controller is undoubtedly multiplexed I can't see how the problem could be in the data connections between the two chips if half of the data is getting through.

Yes, it presumably is multiplexed, but multiplexing involves "select" lines for columns or groups of columns - and that is exactly what we are seeing here - data is failing for one group of columns.

And note - it is data, not display bits. When bits are missing, the display goes dark for those bits as it does when there is no excitation. In this case however, the faulty blocks are light, corresponding to a logic failure either within the IC, or an input to it. It is defaulting HIGH.