18x is not compatible with HD4480 LCD Module.

[Major_Mess]

18x is not compatible with HD4480 LCD Module?

Hi all.

I have an 18x connected to a HD4480 LCD and it has major problems. I have not found a single example on the web of anyone actually running an LCD module off an 18x, all search results return the use of the 18A. Some have updated their info to reflect the 18x being available but appear to have not actually used one. Most PICAXE users seem to be using one of the serial modules.

Has anyone successfully connected an 18x directly to an LCD module? If so how many different modules have you tried?

The circuit I have used is the same as the one here, which is the same as many others for the PICAXE 18 and 18A. The download and reset circuits are present. Originally the circuit was running on a 14M OK, but I required more memory for text so moved to an 18X I had lying around, which is Firmware 8.6.

After losing 1.5 days of my life I went and grabbed another 18x (Firmware 8.8) and it is even worse. = 2.5 days of life gone.

All power is 5v exactly and very clean.

I have breadboarded a 14M and 28x1 and used jumpers to the 18X socket, and using the same program they work perfectly.

Symptons
Unable to fully program 18X with LCD connected. Programmer starts and sometimes completes and inch or 2 of the progress bar than returns a memory error. Disconnecting the LCD enables the programming to complete and the program sometimes runs perfect straight away or will miss or have the wrong first few characters. Cycling the power will then occassionally see it continue to run OK and suddenly weirdness will happen and it will not run correctly again until a reprogram. This running OK is not consistent and generally it doesn't.

Nothing on display.

Weird characters.

Program does not stop at the same place. There is no consistency and very rarely the program does actually continue to run but with errors and it will eventually fall over.

Out of the blue the v8.6 Firmware 18x actually ran for about 6 power cycles, but this has only happened once.

----------------------

It appears to be a threshold of some sort and if it is an issue needs to be advertised as such.

From searching on the 18x it appears others have discovered "weirdness" issues in other applications with this Picaxe. Any ideas of what else it could be apart from a buggy model?

 

[moxhamj]

I've used lots of modules - 8x1 8x2 16x2 and 20x4. They all are interchangeable and work fine with an 18X. I've also programmed them using the original CPM mbasic and in machine code from a Z80.

The fact it won't download with the LCD connected suggests some sort of major wiring error. Eg power supply connected wrong on the module and it is overloading the supply.

Sometimes you can stare at an error for days and not see it. Could you pls take some photos of the board and upload those, and also post all the code.

Lots of people here have used these displays so collectively we should be able to help!

 

[Major_Mess]

The fact it won't download with the LCD connected suggests some sort of major wiring error. Eg power supply connected wrong on the module and it is overloading the supply.

Yep, hence the 2.5 days(and nights lol) at it. But the problem is inserting the 28x1 and 14M into the same physical socket used by the 18x (including power - everything) and they program and run fine.

Does the 18x require all inputs be tied to run correctly?

 

[Major_Mess]

Sometimes you can stare at an error for days and not see it. Could you pls take some photos of the board and upload those, and also post all the code.

For code you can use the full example here. Though that doesn't run on an 18x without replacing "byte" with another symbol name.

I will get some pics up, but the circuit is the same as the one on the above site with the addition of the download and reset circuit.

 

[Major_Mess]

Would'nt believe it, I just plugged my original veroboard and 18x in and it ran for a few power cycles before stopping. It then successfully downloaded the program OK with the LCD connected from that stopped state. Of course, next power cycle the program crashed again.

 

[moxhamj]

A photo would really help a lot because it doesn't make sense that you are putting the 18X, 28X and 14M in the same physical socket as they have different pinouts. Unless you are getting parasitic powering by input/output pins (which surprisingly won't usually damage the picaxe but it will give some strange outcomes)

The 18X needs the power pins correct, the reset pin tied high and the download pin with the 22k/10k resistors in place.

Sorry to go on about photos, but a schematic is one thing and a photo of the actual circuit sometimes is different.

This forum works 24/7 so the UK people should be able to contribute soon as well. Hopefully get an answer quickly for you.

 

[Major_Mess]

A photo would really help a lot because it doesn't make sense that you are putting the 18X, 28X and 14M in the same physical socket as they have different pinouts. Unless you are getting parasitic powering by input/output pins.

Thanks Dr. What I mean is I remove the 18x, put a 28x1 in a breadboard and jumper the pins from the 28x1 breadboard pins to the now empty 18x socket pins. Therefore the 28X1 is using the same circuit as the 18x would have. The 14M I tried was setup the same without the non applicable reset.

Anyway, has to be a power issue of some kind, and it maybe the LCD falling over and not the Picaxe. This woulld really only make sense if the 18x used a much higher current than the other Picaxe tested. I have some other LCD's here so will also try them.

I don't want anyone spending to much time on this until I have some more info, but was hoping someone may have come across the problem.

 

[moxhamj]

I did once zap an LCD display and it failed by drawing lots of current. So that is a possibility. Could you measure the current draw of the LCD with nothing connected and no current going to the backlight. Should be very low - a few mA at most. That would at least help isolate whether it is the LCD or the picaxe. Picaxes are all rather similar in current draw - about 4-5mA.

 

[Major_Mess]

Here is the front and back of the vero board. A 14M and 28x1 runs perfectly on this board and it will program with the LCD module connected.

I probably wont have a chance to try a different LCD module, and check the current until tomorrow.

 

[Major_Mess]

The LCD module with power only connected draws 1.2mA.

 

[Technical]

The symptoms imply a dodgy solder joint/component at either the 10k/22k download circuit or the 4k7 pullup. Our crystal ball guess would be the solder joint to the reset pin....

And in the future it's a good idea to put your IC socket around the correct way - inserting the chip even just once the wrong way around will mean reverse polarity situation which can permanently damage the chip!

 

[Major_Mess]

The symptoms imply a dodgy solder joint/component at either the 10k/22k download circuit or the 4k7 pullup. Our crystal ball guess would be the solder joint to the reset pin....

And in the future it's a good idea to put your IC socket around the correct way - inserting the chip even just once the wrong way around will mean reverse polarity situation which can permanently damage the chip!

As mentioned in the PM the socket went in upside down at 4am. The original board took the 14M and I modified it for the 18X, but after 20 hours straight of reprogramming and consistent desoldering it was time for a fresh board in case of unseen track issues etc, hence the one pictured.

Both boards behave the same with all chips. That is, breadboarding the 18x and jumpering results in the mentioned issues. Breadboard and jumpering the 14M and 18X runs and programs perfectly. This means it is extremely unlikely to be a soldering or board issue.

 

[Technical]

The output pin characteristics of all PICAXE chips - 14M, 18X, 28X1 are basically the same. So there is no reason why a LCD that connects correctyly to one chip outputs will not work with another chip.

So there is a hardware issue somewhere. The fact that the download fails still implies a voltage issue. Have you tried 4.5V form 3xAA cells instead of the 5V supply.

The serial input (download) pin characteristic of all 18 pin chips is different to the 14M. Therefore if right on a threshold on, for instance, supply voltage there may be a difference. We would like to see a 4.5V battery tested.

Also you appear to have connected a backlight (pins 15 and 16) directly across the PICAXE power supply, without any decoupling capcitance by the PICAXE. A backlight will take a very large current draw which could also affect the stabilityof the power supply.

 

[moxhamj]

I'm guessing but I think that is a 470R driving the backlight. Should be fine.

I don't see any capacitors on the board though. As technical says, you need decoupling. And it doesn't hurt to have a bit of capacitance anyway. My favourite values are a 0.1uF across the power supply pins, near the chip, and a 22uF anywhere on the board, across the power supply pins. You need both as they are providing different (and important) functions.

The other issue is that I can't see pins 7,8,9 and 10 connected to anything on the LCD display. These pins will be floating, and could be any value. So you won't be able to program the LCD reliably to work in 4 bit mode. Tie those 4 pins to ground.

 

[Major_Mess]

I'm guessing but I think that is a 470R driving the backlight. Should be fine.

I don't see any capacitors on the board though. As technical says, you need decoupling. And it doesn't hurt to have a bit of capacitance anyway. My favourite values are a 0.1uF across the power supply pins, near the chip, and a 22uF anywhere on the board, across the power supply pins. You need both as they are providing different (and important) functions.

The other issue is that I can't see pins 7,8,9 and 10 connected to anything on the LCD display. These pins will be floating, and could be any value. So you won't be able to program the LCD reliably to work in 4 bit mode. Tie those 4 pins to ground.

Thanks, I have had a .1uf tantalum directly across the power pins and the behavior was the same though this could have been before finding a faulty resistor, and I am not confident to say I covered this correctly. Power surely is the issue. Is a tantalum OK for the decoupling job? I will have to source the 22uf?

I had read about the tying of the unused data lines being required on some modules, but haven't tried that. I will give that a go too.

Anyway looks like I finally blew the 18x trying to bodgey up a harness to another display. Will most likely be a couple of weeks before I can replace it so will have another attempt then. This is not a finished board but a test section of a bigger project, driven by a 28x1 sending a continuous stream of values to be displayed. I have a couple of 28x1's spare so can use those in the mean time to continue.

Thanks.

 

[moxhamj]

22uF - value isn't that important. 10uF to 470uF will do.

Tying those pins low is important because the first couple of instructions you send to an LCD module is to tell it to work in 4 bit or 8 bit mode. If those instructions don't go through, it may think it is still in 8 bit mode and garbage will appear on the display.

 

[hippy]

@ Dr Acula : Not necessarily. The reset instruction sequences are designed to ignore the lower LSB's and some datasheets state the lines can be left floating. I've never had any problems leaving them floating with HD44780 drivers.

 

[moxhamj]

Hmm. So it must be something else then. Zapped device? Zapped pin(s) on a picaxe?