I keep blowing LCD 1602 modules!!!

[MrRadish]

These are the cheap 1602 LCDs with the blue background that you can get from Amazon - 1602A V2.0

Any pointers to what I am potentially doing wrong or what the issue is would be gratefully received.

I am running a ATMEGA328P at 3.3V on a breadboard along with some other sensors DH11, RTC etc. I am using the LCD1602A as the display for the time, temp and humidity. The 1602 display is running from 5V suppied from a 7805 and the rest of the circuit is powered by a 3V3 reg. The 1602 data lines are driven directly from the ATMEGA at 3V3. There are numerous decoupling capacitors spread around the circuit

Issue: Over the course a day or so the LCD display starts to corrupt and eventually the 1602 blows going short circuit taking out the 7805 reg. This has now happened three times  Each time the LCD blowing has coincided with me prodding around with my oscilloscope looking at the output of the DH11 and RTC and after three times it is more than a coincidence. Could the Oscilliscope be introducing a spike big enough to blow the 1602?

[0xdeadbeef]

Missing resistor to limit the current for the backlight?
As far as I recall, something like a 10Ohm resistor is needed between 5V and the backlight supply input.

[MrRadish]

Resistor appears to be there (with a value other than 0R  ) and the back light still works after the 1602 has blown. The component on the 1602 that seems to be blowing is the chip labeled U2?

[PA0PBZ]

Are you sure the display is 5 Volt? I know there are also 3.3 Volt models...

[MrRadish]

Are you sure the display is 5 Volt? I know there are also 3.3 Volt models...

Fairly sure but it went through my mind and I did briefly test one at 3v3 without success

[0xdeadbeef]

I think I'm using the exact type in a project of mine without any sophisticated circuits on an LPC1768 Cortex-M3 (3.3V, but 5V tolerant pins).
The devices run several hours a days since a year or so...

[bktemp]

U2 is the expansion driver chip with no direct connection to the interface, except the power supply. Therefore my guess is overvoltage on the 5V rail. Is the LCD the only device connected to the 5V rail?

[MrRadish]

U2 is the expansion driver chip with no direct connection to the interface, except the power supply. Therefore my guess is overvoltage on the 5V rail. Is the LCD the only device connected to the 5V rail?

There is another component that is still plugged into the 5v rail that  I am not currently testing - it is a RF RX module

[MrRadish]

I think I'm using the exact type in a project of mine without any sophisticated circuits on an LPC1768 Cortex-M3 (3.3V, but 5V tolerant pins).
The devices run several hours a days since a year or so...

I note you have a 10 Ohm resistor on the 5V input - to limit current?

[ruffy91]

It's the Anode of the Backlight. 5V input is on pin 2 without resistor.

[bktemp]

The RX module is probably more tolerant to overvoltages than the LCD.
Try probing the 5V rail to see if it is properly regulated or maybe oscillating.

[MrRadish]

It's the Anode of the Backlight. 5V input is on pin 2 without resistor.

Yes - you are correct ;-)

[MrRadish]

The RX module is probably more tolerant to overvoltages than the LCD.
Try probing the 5V rail to see if it is properly regulated or maybe oscillating.

I have a fairly basic DSO (SDS 1102CML) and the 5v rail looks stable. There is a bit of noises when mains appliances are switched on or off. I detected a 5.5V peak but the actual could be greater. I suspected the 5V supply myself but I cannot see how this would coincide with me probing about with the DSO?

[Kjelt]

the 1602 blows going short circuit taking out the 7805 reg

That is also weird since the 7805 has short circuit protection AFAIK. Only thing these chips really don't like is a higher voltage on the output.
Do you run the input of the 3.3V reg at the same inputsource as the 7805? Could it be that there is 3.3V on the 7805 output while it has no input voltage?

[bktemp]

Maybe you are shorting out the 7805 while probing?
You could add a 5V TVS diode (SMAJ5.0A, SMBJ5.0A) to limit the 5V rail just in case something goes wrong and injects a higher voltage. This would at least protect the LCD.
If the 7805 really dies and not just gets hot because of the LCD shorting out the 5V rail, the problem is somewhere in your power supply chain.

[MrRadish]

the 1602 blows going short circuit taking out the 7805 reg

That is also weird since the 7805 has short circuit protection AFAIK. Only thing these chips really don't like is a higher voltage on the output.
Do you run the input of the 3.3V reg at the same inputsource as the 7805? Could it be that there is 3.3V on the 7805 output while it has no input voltage?

There is alot of heat so I assume that would be a factor

[Kjelt]

Ah well there you have it then, what is the input voltage to the 7805 and how many mA is your display consuming?
Above 1 W in the 7805 you should add a coolingplate, it never should have to become hotter than around 40^oC in normal roomtemperature.

So a guess it is probably not the LCD blowing the 7805 it is the other way around, the 7805 is overheating it should shutdown but perhaps it is destroyed putting the inputvoltage on its output blowing the LCD chip.

[MrRadish]

Ah well there you have it then, what is the input voltage to the 7805 and how many mA is your display consuming?
Above 1 W in the 7805 you should add a coolingplate, it never should have to become hotter than around 40^oC in normal roomtemperature.

So a guess it is probably not the LCD blowing the 7805 it is the other way around, the 7805 is overheating it should shutdown but perhaps it is destroyed putting the inputvoltage on its output blowing the LCD chip.

Sorry -  I should have specified heat when the issue occurs. normal running everything is cool to the touch. Your advice is all good though - thank you

[MrRadish]

Maybe you are shorting out the 7805 while probing?
You could add a 5V TVS diode (SMAJ5.0A, SMBJ5.0A) to limit the 5V rail just in case something goes wrong and injects a higher voltage. This would at least protect the LCD.
If the 7805 really dies and not just gets hot because of the LCD shorting out the 5V rail, the problem is somewhere in your power supply chain.

It is unlikely I am shorting 5V rail as the DSO leads are attached to fly leads I am then inserting into the breadboard. I will try the 5v limiting option :-D

[dannyf]

the 1602 blows going short circuit taking out the 7805 reg.

It is rare that those displays take more than a few ma + the led so something quite extraordinary must have happened for it to be able to go short and take out the regulator.

How did you establish that the 1602 went short?

The proper procedure in powering up any circuit is to put a power limiting resistor in the rail, something like 100ohm and then down to 10ohm would work. Once you have established that the current consumption is normal, take out that resistor.

[MrRadish]

the 1602 blows going short circuit taking out the 7805 reg.

It is rare that those displays take more than a few ma + the led so something quite extraordinary must have happened for it to be able to go short and take out the regulator.

How did you establish that the 1602 went short?

The proper procedure in powering up any circuit is to put a power limiting resistor in the rail, something like 100ohm and then down to 10ohm would work. Once you have established that the current consumption is normal, take out that resistor.

The module went short or very close to short - On two occasions the round black cap of U2 was blown off with a audible pop.

The census seems to be that these modules are not known to just blow up so the problem is with me. The corruption on the display and the eventual failure that seems to coincide with the DSO use is a mystery but the consensus is that it may be related to power supply.

Could it be a GND issue caused by a faulty connection either a wire or the breadboard itself?

[dannyf]

The module went short or very close to short

How do you know that?

Could it be a GND issue caused by a faulty connection either a wire or the breadboard itself?

Anything is possible. Without you showing a connection diagram, it is hard to tell.

Biggest struggle with getting any help is to figure out a way to help others help you.

[MrRadish]

The module went short or very close to short

How do you know that?

Well not for sure but IMHO S/C mainly because of the destruction and heat involved.

Could it be a GND issue caused by a faulty connection either a wire or the breadboard itself?

Anything is possible. Without you showing a connection diagram, it is hard to tell.

Biggest struggle with getting any help is to figure out a way to help others help you.

Great help has been offered so far and I appreciate you guys have to shoot into the dark due to the lack of information in areas you feel are important. That is part of the support process and part of my learning process I guess.

I am going to persevere and start a fresh with a new breadboard, wiring and without the other non LCD\ATMEGA components and see if I still have the same issue

[amyk]

Blowing the cap off the IC suggests a severe overvoltage event.

Is your scope ground really ground, or floating at some capacitively-coupled mains potential...?

[MrRadish]

Blowing the cap off the IC suggests a severe overvoltage event.

Is your scope ground really ground, or floating at some capacitively-coupled mains potential...?

eek - What is the best way to check this (and fix it if it is a problem)?

[katzohki]

Look at this video, it includes information regarding the ground of your oscilloscope.

[22swg]

You cannot run the LED backlight on 5v without a limiting resistor ! min 10 Ohm  this should  be >. 5 w   

[David_AVD]

I think a schematic or at least a decent photo is in order.  Maybe the 7805 is becoming unstable (oscillating) at certain times.

[suicidaleggroll]

Have you tried skipping the 7805 temporarily, and using a current-limiting bench supply to drive the LCD instead?  What's the input to the 7805?

[tautech]

I have a fairly basic DSO (SDS 1102CML)
I suspected the 5V supply myself but I cannot see how this would coincide with me probing about with the DSO?

Don't under estimate the ability of the CML range. They are good little DSO's.

As has been mentioned, I'd suspect probing technique error too.
Make sure you use 10x and meter for a GND loop between the probe reference lead and where you intend to reference from.

[sleemanj]

Where did you get the 7805 you are using?

I have seen chinese 5v regulators (1117 series in this case) that fail with more than 10v input and no current draw, worse the failure shorts input to output!

[bktemp]

I have seen chinese 5v regulators (1117 series in this case) that fail with more than 10v input and no current draw, worse the failure shorts input to output!

TS1117 are only rated for 7V, so nothing wrong here if they fail at >12V.
Some manufacturers allow higher input voltages, but not all are equal.

[dannyf]

For the record, I have taken Chinese-made 1117 to 20v, and encountered no problem.

[sleemanj]

TS1117 are only rated for 7V, so nothing wrong here if they fail at >12V.

Trouble is they were sold and marked as AMS1117 which supposed to do up to 15v (and legit ones do).

For the record, I have taken Chinese-made 1117 to 20v, and encountered no problem.

Yep, some are fine.  But there are either fakes and/or parts in an datasheet I eventually found with the same part number but 10v max input (the ones I bought didn't mention this datasheet, just the official AMS one, the IC markings are the same, although the fakes ones are left shifted a bit).

Anyway, my point is, testing the voltage regulator is a simple thing to do, set it up on your breadboard with a small load, and wind up the input voltage to see what it can handle.

[wraper]

Maybe the 7805 is becoming unstable (oscillating) at certain times.

And this possibility rises the question if 7805 input and output are properly decoupled with capacitors?

[MAS3]

Multiple capacitors to be sure.
We really need to see some pictures and/or schematics to be able to give some advice that makes sense.
Oscillation is a serious problem, unexpected to a lot of people and impossible to see using a multimeter (even though a scope was involved here).

[dannyf]

We really need to see some pictures and/or schematics to be able to give some advice that makes sense.

"That's TOP secrete",

[MrRadish]

Thank you for all the responses so far.

I setup the test again with a 20x4 LCD that I received yesterday from a different source  to the 1602s I have blown . I also connected the DSO earth directly to the GND on the 7805 (The input to this is 12v)  instead of an arbitrary GND rail point on the breadboard. So far after 24 hours - no screen corruption and fully working LCD.

As for CCT diagrams - My CCT is based on http://arduinoliquidcrystal.readthedocs.org/en/latest/_images/LCD_schem.png with the exception that the ATMEGA is running on 3v3 and the backlight is connected to 3v3 without an external resistor (a resistor appears to be in place on the board?). The LCD itself is running from 5v

One further thing is that the 1602 blowing event may have coincided with me changing the controls on the DSO (eg V/div) or even switching it on (~80% sure). Could these events have caused enough GND ripple to blow the 1602 if they were sensitive or am I just blowing smoke 

[bktemp]

Please post a photo of your complete setup.
Changing settings on your DSO does not produce any spikes. Maybe you have a bad ground connection somewhere and the smallest vibrations on the table lift the ground pin of the 7805 and therefore it increses the output voltage because of the missing ground path. But without more details of your circuit it is impossible to say.

[22swg]

  and the resistor at from +5v to  +LED ?   

[Monkeh]

You cannot run the LED backlight on 5v without a limiting resistor ! min 10 Ohm  this should  be >. 5 w   

and the resistor at from +5v to  +LED ?   

There's a 100R resistor already on the board, and the backlight has nothing to do with the IC.

[22swg]

There's a 100R resistor already on the board, and the backlight has nothing to do with the IC.

None of the 1602 LCD's I have had have a resistor for the LED and I have had loads. there is a zero Ohm link sometimes . If LED array cooks ,  over current could take out the 7805 and supply 12v to his IC ? UC

Correction    One of my stock has 2 x 51 Ohm provided

[Monkeh]

There's a 100R resistor already on the board, and the backlight has nothing to do with the IC.

None of the 1602 LCD's I have had have a resistor for the LED and I have had loads. there is a zero Ohm link sometimes . If LED array cooks ,  over current could take out the 7805 and supply 12v to his IC ? UC

Take a look at the photo on the first page.

If the LED fries it will typically go open circuit.

[22swg]

Ooops    Didn't open the thumb, perhaps his BB wiring is at fault.

[suicidaleggroll]

None of the 1602 LCD's I have had have a resistor for the LED and I have had loads.

All of the ones I've used have had a resistor installed for the LCD backlight.  Just went to the lab to check on one of the ones there...yep, 6R8 (it's a 3.3v unit from mouser).

[22swg]

I checked too  4 LCD only one (4x20 blue) had R8 and R9 as 51 Ohm and in BL circuit . 

[MrRadish]

Please post a photo of your complete setup.
Changing settings on your DSO does not produce any spikes. Maybe you have a bad ground connection somewhere and the smallest vibrations on the table lift the ground pin of the 7805 and therefore it increses the output voltage because of the missing ground path. But without more details of your circuit it is impossible to say.

Ok so lesson for me - I will photo the project and put together the CCT diagrams next time I post asking for help  .

I pretty much rewired the whole thing yesterday including an LCD of a different make and supplier (as mentioned above). I have added a more robust GND for my scope and I even replaced the breadboard I was using just in case that was an issue.  All is working well.

As I have changed a fair bit I dont think I will get to the bottom of the issue but I think I was either experiencing a GND fault that caused the 1602 to blow or it was a faulty batch of 1602s. Why the DSO seemed to be linked I do not know. I consider myself to be scope competent as an ex RF engineer but this 5v thing is new to me ;-) 

If I can stomach the pain I may order a few more of the 1602s and give them another go out of curiosity.

Thank you all for your help.

[RobertHolcombe]

Maybe you have a bad ground connection somewhere and the smallest vibrations on the table lift the ground pin of the 7805 and therefore it increses the output voltage because of the missing ground path.

Have personally had this happen too many times using breadboards, it gets expensive fast 

[electr_peter]

I am running a ATMEGA328P at 3.3V on a breadboard along with some other sensors DH11, RTC etc. I am using the LCD1602A as the display for the time, temp and humidity. The 1602 display is running from 5V suppied from a 7805 and the rest of the circuit is powered by a 3V3 reg. The 1602 data lines are driven directly from the ATMEGA at 3V3. There are numerous decoupling capacitors spread around the circuit

What is providing power to the circuit? Power source could be shifted up/down w.r.t. earth/DSO ground.