I am new in the AVR world. I use Atmel Studio 7, I program with USBAsp an ATmega328p.
I upload my program to the device by ISP and the program runs ok.
BUT ! when I disconect the programmer and connect an external power supply the microcontroller seems not to be programmed. It doesn't do anything.
It only works when the USBAsp is connected and the program is downloaded in the device.
What am I doing wrong ?
Thanks in advance, Folks !
MOST programmers require the target board to be separately powered. According to the schematic I just saw, if you use a 10-pin programming connector, the target CAN be power by USBAsp.
What do you have connected to the reset pin on your target?
If your USBAsp has a jumper to connect power to the Vtarget pin so that the programmer provides power, what happens if you disconnect that jumper and provide your own target power? Does it still program?
Jim
Do you have capacitors on your chip power pins? Is voltage connected to all of the supply pins & all gnds connected? Show your schematic. What is your clock source---internal RC?
Is your power supply supplying clean 5V or 3.3V? Did you measure that pwr is actually getting to the pins?
Why do you say the chip is "doing nothing"?
I have powerd up the circuit with an externay power supply and programmed the chip, in that case when I took off the USBAsp the program still works, BUT when I take off the external 5V for a moment and repower . The chip seems to have nothing programmed, "and do nothing".
It acts as if the chip was new, without any program.
The only way it works is uploading the program by the USBAsp and keeping it connected.
I think is a bad configuration of the FUSES, but I'm new and have no clue !
Thanks again !
A 328 has two power pins, VCC and AVCC, both must be connected to power and have 100nf caps to adjacent GND pins.
Jim
Could be an uninitialized variable. During the programming operation the SRAM is cleared and 0 is acceptable value for the variable to allow the program to run. After a power cycle the SRAM can contain any values and just so happens the power on value causes your program to crash.
Thanks for the answers. I have power in both VCC and AVCC. Even all the variables are not initialized. I'm not using SRAM positions, just "PUSH" and "POP" (which I think uses SRAM, I do not know if I'm wrong with this. Excuse me, my unknowledge).
I use the ISP with USBAsp programmer as I said. I upload the program to the 328p and the program runs ok, no bugs, it shows a counter in the LCD module and it's ok.
The problem is when I turn off the circuit and turn it on again, it act as if it loose the programming and it doesn't run at all. The LCD module doesn't initialize, etc.
I am not used to work with AVR's perhaps there are something wrong in the uploading method using the USBAsp, or something that I have to inizialize with fuses, so the chip can store the program.
Even all this, It works when I programm it, there are something that makes the device, seems to loose the program when it is turning off.
It is very unlikely that you "loose the program". What leads you to the conclusion that it is not working?
One of the problems that has been reported a few times has been that there is no power supply path once the programmer is unplugged. It has seemed to be a missing ground connection. What happens if you leave the programmer plugged into the target but the programmer unplugged from the PC (e.g. unpowered)?
Another test: unplug your programmer, as you have described and try to restart it. Use a voltmeter to measure the voltage between MCU ground pins and the MCU Vcc & AVcc pins. Also between ground and the MCU reset pin; it should be close to your supply voltage.
Jim
First Jim, Thank you very much for your time answering me ! ! !
I have done a small vid 1min for you to see the exact problem.
I think it is not about a missing ground or bad power, because the circuit works and the programm also works. The problem is, it seems not to stay in the flash memory.
https://photos.app.goo.gl/amXPqT...
If I disconect de USBAsp its the same (it is not powering the circuit). In this moment I am using an external power supply.
I argue that "lost program" is almost never true. People have claimed seeing it, but there have always been other causes, in the end.
I wonder if you are really using the simulator instead of actually programming the target?
Try this:
Program it outside of debug mode using the Tools > Device Programming menu. There, you can program it straight via ISP and you can VERIFY it (that is, read it back to the programmer and compare with the program it tried to load). That will tell you if the program is really in there. Also, there is no debug mode to get in the way. THEN try to disconnect and restart.
Jim
Beside the lcd, what parts are not working?
Do you have some pins, or leds that are working?
The problem is the LCD may take a long time to get ready after power is applied...do you have a second of delay before the program starts to send things to the LCD?
Of course all of the other parts & pins should be fine. Does even an LED light up?
Usually I'd look at power or the reset pin, but I'm guessing avrcandies is right on this one. When you load the program, the thing has already been turned on for a few seconds or more when the program starts. When you apply power, the program starts immediately and perhaps initializes the display before the display is ready to be initialized.
Try adding _delay_ms(1000) at the start of the program before you initialize the display. You likely can get away with something shorter than a second, but let's get it working first.
The largest known prime number: 282589933-1
It's easy to stop breaking the 10th commandment! Break the 8th instead. 
Protest agenst sensorship.
FIRST thank you all for your help !
I made some other tests to figure out what the problem is. I tried changing the chip. Using the chip in some other app. As I said, VCC and AVCC are powered. both GND pins are at 0V (PIN 7 and PIN 22) the AREF pin is also to 5V. The RESET pin has a 100k resistor and a 1N4148 to power and a 104 to ground. I have programmed the chip by the USBAsp program (outside the Atmel Studio). I am using an old PC power supply to work with the 5 volts.
I don't want to give it up, because when I upload the programm it works ok and it do what it have to do. I really don't think it loose the programm, but there is something that it doesn't work in stand alone mode. What is the difference or what changes in between charging the programm in the flash memmory or leting it start alone?
The chip is 99.999% NOT bad, or it wouldn't program/run in the first place.
Do you try blinking an led and see if that keeps working?
You forgot to post your code for any extra help.
FIRST thank you all for your help !
I made some other tests to figure out what the problem is. I tried changing the chip. Using the chip in some other app. As I said, VCC and AVCC are powered. both GND pins are at 0V (PIN 7 and PIN 22) the AREF pin is also to 5V. The RESET pin has a 100k resistor and a 1N4148 to power and a 104 to ground. I have programmed the chip by the USBAsp program (outside the Atmel Studio). I am using an old PC power supply to work with the 5 volts.
I don't want to give it up, because when I upload the programm it works ok and it do what it have to do. I really don't think it loose the programm, but there is something that it doesn't work in stand alone mode. What is the difference or what changes in between charging the programm in the flash memmory or leting it start alone?
AREF should never be connected to VCC, it should have 100nf to GND, you risk internal damage to the chip.
Connecting AREF pin to Vcc (Avcc) is kinda ok if AVCC is selected as the reference, otherwise.... maybe it makes a nice software destruct technique....10, 9, 8,7,5, 4 ...
I wonder how many freaks have purposely built in a destruct setting of some sort (intentionally!)..
I am here again and again thank you all very much !
I add a beep while it is counting and this told me that the program is running. The beep sounds, but the LCD module doesn't initialize, it doesn not show the characters, anyway when I upload the programm IT WORKS ! !
I took off the 5v to the AREF pin, I put a 104 cap to ground instead, as you told me and I add 100k pull-ups resistors in the C port PC1-PC2-PC3-PC4-PC5. where I have the strobe, data and clock pins of the CD4094 and the RS and E pins of the LCD module.
I ask again, what is the difference betwen upload the program or let it start alone ? What register shoud I have to beware of ?
I show you the program if it is worth. I know the program is ok and works but now when I turn the circuit off and turn it on again the LCD module doesnt inicialize`.
|
;*************************************************************** ;Control lines definition ;LCD commands ;Var definitions ;INPUT [0] OUTPUT [1] CONFIGURATION ;Reposicionar el Stack Pointer ;*************************************************************** LDI BUFFER, line2
;The counter STARTS here.
LDI RES1, 0x00 ;StayHERE: ; 0123456789ABCDEF ShowNUM: LDI TEMP, 0x30 LDI TEMP, 0x30 ;************************************************************** ;************************************************************** ;*************************************************************** ;*************************************************************** ;*************************************************************** |
Before you begin writing anything to the LCD, make sure you include a 1 second delay, that will be plenty. Just put it in after you set up your ports.
LCD_init:
;Power-Up delay
CALL Dly_15mS ;DELAY 15mS <==== Make this delay longer!!! Try 100ms
;Reset the LCD controller
LDI BUFFER, Set_LCD
CALL LCD_iw
Jim
I HAVE SOLVED THE PROBLEM !!! I changed the Y register by the Z register, the problem was I din't know the way of pointing to the data with Z+.
Now I can took off the power of the circuit and when I power it agani, it keeps running.
LDI ZH, HIGH(MSJE_01<<1)
LDI ZL, LOW(MSJE_01<<1)
MSJE_01: .DB "- TEST machine -"
SendMSJE:
LDI TEMP, 0x00
Loop_MSJE:
MOV ZL, YL
ADD ZL, TEMP
LPM BUFFER, ZL
CPI BUFFER, 0x24 ;'$' stop sending character
BREQ End_MSJE
CALL LCD_dw
INC TEMP
CPI TEMP, 0x10
BRNE Loop_MSJE
End_MSJE:
RET
=======================
================CHANGED :
=======================
SendMSJE:
LDI TEMP, 0x00
Loop_MSJE:
LPM BUFFER, Z+
CPI BUFFER, 0x24 ;'$' stop sending character
BREQ End_MSJE
CALL LCD_dw
INC TEMP
CPI TEMP, 0x10
BRNE Loop_MSJE
End_MSJE:
RET
I changed the Y register by the Z register, the problem was I din't know the way of pointing to the data with Z+.
You still haven't said what the failure was.... it worked when you used the programmer?
I really don't know why using the Y register produce loss of data in the stand alone mode. When I changed by the Z register, it works. The LCD initialization have the same timing, the power supply is the same I was working with. I only change the Y reg, as you see in the uploaded program part.
CPI BUFFER, 0x24 ;'$' stop sending character
How does that work then? The text does not have a terminating '$' ??
MSJE_01: .DB "- TEST machine -"
I guess you are simply relying on the test against TEMP, 0x10 ?
(BTW $ terminator? Really? Are you old enough to remember CP/M then?)
