LCD do display - but pins check out

[Scampy]

Guys, I'm debugging the mistakes I made when prototyping a PCB. I say I made, the schematic tested out OK in the DRC check, but the board failed to pick up some nets and thus I'm missing a few tracks to some components. However I've trouble shouted through the board and The PIC is running fine, I have comms through an FTDI 232RL chip, and it reads the 18B20 sensor and an corresponding monitoring LED flashes away to confirm the code is running, but the LCD simply lights up with the backlight and doesn't display anything at all.

I've traced the routing for the 4 bit data lines using a multi-meter for continuity, and using a scope I get negative pulses on enable pin, and positive pulses on the R/W pin - these are around 1.2s frequency which seem slow given that the pic is running at 40 Mhz with a 10 Mhz xtal. There are no solid blocks regardless if the contrast is grounded or taken high (or 4k7 placed between it and the 5v rail). If I replace the PIC back on the development board it works fine ! and the schematic for the LCD was taken from the manual for the EasyPIC5 board - any suggestions or further testing I can do.

[Tabsoft]

What are you getting on the RS pin?
Also have you checked the LCD's VDD pin to make sure it is at least 4.5V?

[towlerg]

sorry just reread your post. pulses on r/w doesn't sound correct, rs and r/w crossed?

[Tabsoft]

True.

Looking at the EasyPIC5 manual, R/W is tied to ground so you shouldn't read pulses on it.

RS -- RB4
R/W -- GND
E -- RB5
D4 -- RB0
D5 -- RB1
D6 -- RB2
D7 -- RB3

[Scampy]

Thanks for the replies,

OK here's confirmation on how it's wired.

LCD

1 GND
2 +5v
3 +5v via 4K7 resistor
4 RB4 (RS)
5 GND (RW)
6 RB5 (E)
7 - 10 GND (D0 - D3)
11 RB0 (D4)
12 RB1 (D5)
13 RB2 (D6)
14 RB3 (D7)
15 LED+ 5v via 100R resistor
16 LED- GND


I've attached traces of the pulses I get on RB4 and RB5 (sorry I made a mistake in my previous post) and on two of the data lines. I can confirm that the PIC is running as other functions such as RS232 coms and reading the DS18B20's all work fine. I will try a direct wiring option later (wires direct from the PIC pins), but as I can't find any shorts suspect that there is something else that's srewed on the board.

Moral is not to submit gerbers at 1am in the morning after 14 hrs of drafting the design, but to wait a day and check the design with fresh eyes in the morning ! - This board has so many mistakes in it I must of been totally brain dead !!

[Scampy]

Interestingly, I get the same trace when the LCD is returned to the EasyPIC5 board, but on RB5 the pulses are positive going not negative going !!!

[pedja089]

Pin 3 is contrast. And you need to put about 1-2K resistor to ground. Not to Vcc. Or for quick test, just short pin 3 to GND and you will get maximum contrast.
On EasyPIC boards you have trimmer connected to powers supply rails, and slider to LCD. When you just apply power to lcd with correct contrast, first line of display light up all pixels.

[Scampy]

Thanks for the suggestion, but I've tried taking pin 3 to grnd and +5 but it still makes no difference. I've tried wiring the LCd direct to the pins, but still no joy, so I'm at a loss, other than break all the tracks between the PIC and it's header for the LCD, but if there is an issue with the board's through hole plating or something similar then the fault could still be there

[Tabsoft]

What LCD are you using?

[andybarrett1]

What was the board designed in...?

Can you post the project artwork / gerbers at least ?

[Tabsoft]

The E line should pulse Positive.

If you were to send an "H" $48 to an LCD with a 4bit interface, you should see something like this.

[Scampy]

Thanks for the further info guys.

@Tabsoft - it's a 4 line 20 chr back lit display. I use them in all my projects and like I mentioned, works fine on the development board do something is screwed on the board, something that is taking the lines high or low and stopping the signals getting through.

I've traced out the schematic and compared it to the board and there are quite a few mistakes, not with the LCD part as that pings out but some nets were missed in the PCB layout module when the schematic was converted, but this was with a custom part, so that might of been human error on my part. I'm going to redraw the schematic, and re-work the PCB layout and check everything three times before sending the gerbers off. Still I have some posh new green and white coasters to put my cup of tea on so I guess all was not wasted

Thanks again for every ones contributions

[Tabsoft]

Good luck there, sir.
Nothing like PCB Coasters! I have a few myself.

BTW what Schematic/PCB software are you using?

[Scampy]

I use DipTrace. Up until now it's been excellent and I've produced several working boards from it.

To be honest I think it's 50:50 human error / software error The nets show up for the custom part in the schematic, but were missing in the layout... still having spent two days now trying to sort this I think it's passed debugging and time to start from scratch

[andywpg]

Good luck there, sir.
Nothing like PCB Coasters! I have a few myself.

I get some of that green felt from the fabric store and glue it on the back so they don't scratch the coffee table......

[Scampy]

Well after a few hours work I've redrawn the schematic and this time names each net, and give the pins on the connectors names so that everything can be traced and identified when it comes to converting to PCB layout.

It's nearly 2.30 am !! (struth don't time fly when having fun !) - let me know what you guys think

[Tabsoft]

Can't really read the schematic because the picture is so blurry.

[andybarrett1]

Same as TabSoft

If you could post the DipTrace files we might be able to help more.

Or a better res JPG.

BR

Andy

[Scampy]

Sorry guys - hope this helps

Seems the forum re-sizes the image - I'll try something else

Nope - does the same with URL's !!

Here's a link http://micro-heli.co.uk/New%20Schematic2.jpg

[Art]

IIRC correctly you only need the supply voltage and contrast voltage to see the square blocks on line 1.
So long as you aren’t driving the contrast with PWM (Melanie’s old trick), you should see the blocks with the chip unpopulated.

[Tabsoft]

Thanks, I can read this one.
Since you had issues with the LCD before, can you give us the model number?
Really do not want to guess.

[Scampy]

Thanks, I can read this one.
Since you had issues with the LCD before, can you give us the model number?
Really do not want to guess.

The only info I have is what's stamped on the back. RT204-1 (ver2.0) - uses the standard Hitachi HD44780 chipset. Closest datasheet is http://www.satistronics.com/myfiles/...CD/RT204-1.pdf

[andybarrett1]

Hi

Looking at the circuit all looks OK for me. LSB's grounded on display, good !... For me I would have decoupled (0.1uf) around all the chips including the display. I also would have used a trimmer across 5v and 0v (5k) on the contrast (Pin 3)so as I could meddle !!

Nothing wrong in checking all connections are right way around on the board D4-D7 also RS and E, so easy to assume... "I know" I have been burned so many times.

Could be worth checking the Xtal.... Maybe just try switching to internal clock (4Megs), I have had strange things happen around crystal circuits and displays, Some of the EBay types certainly :-) .... Not just with PIC's but other Micros also.

One last thing make sure you protect your AC SSR outputs ... I love MOV's ...At least do fuse them. Good practice to fuse the the low voltage AC before the bridge also.

Hope some of this helps

Andy

[andybarrett1]

Trimmer across 5v and 0v (5k) on the contrast (Pin 3)

Just noticed Pin3 LCD Vo is up high... Should be more Low (0v) if no trimmer ?

[Scampy]

Thanks for the suggestions. I will look at adding a trim pot for the contrast, and a few more decoupling caps in the mix.

Yes, I'll be fitting some chassis mounted fuse holders, or some of those C-13 type sockets with fuse built in to protect the heaters. I also slot the board under the SSR between the low voltage pins and AC pins as a safety feature too.

[Art]

[Scampy]

Just measured the resistance across the 10K pot on the EasyPic5 board. Between Vco and GND pins on the LCD connector I have 5.67K, with 3.12K between Vcc and Vco. I'll amend the schematic and put a 4.7K between GND and Vco

[Scampy]

OK - those ruddy gremlins must of been messing me about yesterday..

I disconnected the resistor on Vco and grounded the pin as it was in the original board design. Connected the LCD and powered up and got two lines of blocks which was something I never got previously and I had tried taking Vco both high and low !

Placed the PIC in the board and booted the board - bingo text displayed...

I'm at a loss as to why, what or how the thing failed to work previously - Oh the jpys of electronics !!!