$20 LCR ESR Transistor checker project

[mrmox]

CORRECTED SCHEMATIC FOR CHINESE CLONE LCR-T4
from 91make (and others)

It took days of searching for me to FIND the revised corrected schematic.
The errored one is very widely distributed and there are lots of fake russian sites trying to get
you to sign up for some service complete with fake testimonials about how great the service is.

Here I found Transistortester_T4_1d_100dpi.png by Roland dated 29.08.2016 with the corrections
I had proposed (and perhaps more).

https://www.mikrocontroller.net/topic/248078?page=23

https://www.mikrocontroller.net/attachment/304138/Transistortester_T4_1d_100dpi.png

Best wishes to all.

[hobbyist]

What is the right voltage input for the frequency counter on an AY-AT tester with the M firmware?  The PDF for the K firmware seems to say that the input should be 2.4 volts.  Is it OK to have a sine wave input?  Are negative voltages a problem?  If not, then does the tester expect a frequency that is 0 volts low and 2.4 volts peak?  Do I need to use a Schmitt trigger for low frequencies?  What voltages would damage the tester?   

[madires]

The frequency counter (pin T0) is triggered by a low signal and the AY-AT has a voltage divider between Gnd and Vcc plus a 100 nF DC blocking cap as input circuitry. Any voltage lower than Gnd or higher than 5 V at T0 can damage the ATmega. If you limit the current the clamping diodes are able to protect the input to some degree. Sine wave is fine, but I haven't tested very low frequencies.

[DAIRVINE]

Hello Madires,
To compile: Download source from Madires current website.
Download and run AVRSTUDIO 7. Do not choose to install the latest version. Then follow the instructions in:
https://microchipsupport.force.com/s/article/Creating-a-project-from-Makefile-in-Atmel-Studio
Steps for importing makefile project into Atmel Studio :-
Step 1:- Go to Atmel Studio->Tools->Extension and updates->Available Downloads
Step 2:- Install "Create Project from Makefile" extension.
Step 3:- Restart Atmel Studio
Step 4:- Go to Atmel Studio 7->Tools->Create Project from Makefile
Step 5:- Choose the makefile and project name according to the instructions given in the window.
Step 6:- Click "Create Project"
Step 7:- After the project is created it can be built within Atmel Studio itself.
Then add an empty dep subdirectory to remove build errors.
Please can I recommend that you add a blank dep subdirectory to avoid this AVRstudio 7 compile error?

Also, please can you explain how to compile for a particular known model of tester? In my case an ordered T4, I cannot figure it out.
Thanks.

[madires]

Please can I recommend that you add a blank dep subdirectory to avoid this AVRstudio 7 compile error?

I'll update the Makefile.

Also, please can you explain how to compile for a particular known model of tester? In my case an ordered T4, I cannot figure it out.
Thanks.

Please see the file "Clones" for the settings for your clone and edit config.h and config_328.h accordingly.

[DAIRVINE]

Re compile options for T4,
OK, thanks. I will follow that.
Which IDE are you using?

[madires]

A text editor and make - quite spartan.

[carrascoso]

Hi.
They gave me a GM328 that they broke when measuring a charged electrolytic capacitor. The SVR05-4 was shorted, and the Atmega was bad too.
I removed the SVR05-4 and the P6KE6V8A since they are not necessary for the proper functioning of the meter, and I programmed another Atmega at 20Mhz (and tested with several) and also changed the Xtal to 20Mhz.
Now it turns on smoothly, but it doesn't work properly.

When I turn on the meter without any components in the measurement socket, it should indicate that there are no components.
Well, it indicates that there are high value resistors or high value capacitors connected.
With normal use of the meter I put components in the socket and except for the capacitors all other components seem to measure well.

The same happens with firm M and firm K, so it is a meter problem.

E measured continuity from the 3 points where the component to be measured is placed to pins 23-24-25 of the Atmega.
From those 3 same points to pins 14-15-16-17-18-19, checking that the resistors are good.

7550 output voltage when measuring component, stable 4.97V.
TL431 output voltage, 2.48V stable.

I would appreciate a help. Thank you

   

[Cliff Matthews]

I wonder if we'll see a Raspberry Pi RP2040-based transistor tester? $4 for dual Arms, on-board slave MCU's, and 12-bit A/D's 

[indman]

TL431 output voltage, 2.98V stable.
I would appreciate a help. Thank you

2.98V or 2.48V?
Show us the T1-T6 hardware test results

[carrascoso]

TL431 output voltage, 2.98V stable.
I would appreciate a help. Thank you

2.98V or 2.48V?
Show us the T1-T6 hardware test results

Sorry 2.48v
I do not understand the hardware results T1-T6
thanks

[indman]

K-Firmware manual carefully read section 5.5 Selftest Function. M-Firmware has the same section + Self Test

[carrascoso]

Performed test 5.5 in K-Firmware, it is carried out resulting between the test terminals 0ohm.
Performed Self Test with M-Firmware, it performs it with the same result.

But after I test without connecting anything, the problem continues.         

[indman]

Full Selfest is performed after entering the main menu of the tester

[carrascoso]

The self-evaluation and calibration is carried out within the menu. It does both without problem but the problem remains.
Firm M only asks for the step of shorting the three terminals, it does not ask for the >100nf and <10nf capacitors. The firm K does ask for the three steps.
I have other GM328 meters, one with the firm K and one with the M and they work fine.

[indman]

carrascoso, don't you understand me? I ask you to write the results that show the T1-T6 hardware tests! Then the reason why phantom parts are detected with empty test contacts will be clearer. Most likely the reason is a leak on the board or damaged processor ports.

[mrmox]

carrascoso does not seem to appreciate that "the test runs" is not enough!  The test never says "ERROR", it is up to the user to READ THE NUMBERS and compare them to the description of the test.  Often they are reported in mv and the answer is supposed to be a low number for all 3 (like less than 20mv or so...the code author has subtracted off the expected value and is displaying the difference for your convenience).  Other times a value is reported directly.  Best Wishes.

[DAIRVINE]

Madires
AVRstudio 7 has the advantage of a simulator and, if you have a debugwire adaptor, debug.
AVRstudio 7 requires you to create the Madires  project from makefile.
AVRstudio 7 requires ComponentTester to be renamed ComponentTester.elf for the simulator and debug to work.
I can do this by changing the object file extension in the makefile, see changes below.

Is it possible to change the makefile to generate ComponentTester and ComponentTester.elf for everyone? Makefile is arcane and I checked google and tried CP and XCOPY but failed to replicate the file with a .elf extension.

# link firmware
$(NAME): ${OBJECTS}
    ${CC} ${LDFLAGS} ${OBJECTS} ${LIBDIRS} ${LIBS}  -o ${NAME}.elf


# create hex file of firmware
%.hex: ${NAME}.elf
   avr-objcopy -O ihex ${HEX_FLASH_FLAGS} $< $@

# create hex file of EEPROM data
%.eep: ${NAME}.elf
   -avr-objcopy ${HEX_EEPROM_FLAGS} -O ihex $< $@ || exit 0

# create dump of firmware
%.lss: ${NAME}.elf
   avr-objdump -h -S $< > $@

# output size of firmware and stuff
size: ${NAME}.elf
   @echo
   @avr-size -C --mcu=${MCU} $<

[madires]

I could add a new target "elf" to the Makefile. Does a manual "cp ComponentTester ComponentTester.elf" work?

[DAIRVINE]

Madires,
"I could add a new target "elf" to the Makefile. Does a manual "cp ComponentTester ComponentTester.elf" work?"
Exactly. I tried that but failed to get it to work.

I tried a number of variants suggested by google using CP and XCOPY and COPY and none that I tried resulted in a .elf file with AVRstudio. I am sure that I am making a trivial mistake because I am making random guesses without understanding it.
 
You probably would benefit from using an IDE and AVRstudio has the benefit of a simulator so you can run to a breakpoint and step through your programs without any hardware and, I assume, if you acquire a debugwire interface you can debug the hardware.

[madires]

Have you tried the "cp ..." in a shell (not via the Makefile)? AVR studio doesn't run on my PC natively.

[carrascoso]

Hi.
After doing different tests I had not been able to find the problem. She was even going to make a video calibrating the meter to see if it could be fixed.
I decided to change the socket where the component to be tested is placed.
Due to the tin with which it was soldered (this equipment was given to me already assembled) it has been impossible for me to remove it, but I have been able to remove a lot of tin from the holes.
And in order not to damage the PCB by doing some force, I decided to leave the same socket, but using good quality tin.
And from this moment on, the meter works correctly.
Possibly the problem was the tin with which the assembly was made. If it fails again, I would comment on it again.

Thank you very much for your help.

[mrmox]

A while after I posted this I realized that the Makefile scheme for the K-firmware and the M-Firmware are vastly different.
I now believe the K-Firmware uses specialized Makefiles for each PCB permutation, while with the M-Firmware
you make #define changes in the config.h and config_328.h files by editing them yourself.
MY COMMENT APPLIES ONLY TO THE  K-Firmware!  (But let me know if I am wrong!)
Sorry if I contributed confusion.


This is working for me to get to breakpoint debugging on hardware but it goes in the finish.mk file

%.lss: $(TARGET)
   $(AVR_TOOL/PATH)avr-objdump -h -S $< > $@
   ls $<
   cp $< ../mega328_T4_v2_st7565/JaysAS7_LCR-T4v2.elf
   @echo JAY thinks he can COPY THE ELF.

size: ${TARGET}

(I included some context) (you will need to adjust it for name)

IIRC, you have to copy it before it gets deleted!

Best All.

[DAIRVINE]

Madires,
Re AVRstudio is not compatible with your computer, that is a shame.
I have used and can recommend Eclipse as a good IDE, but it is a lot more tricky to set up than AVRstudio. Which effort is probably not justified unless it was your job, given that what you do now is working. I find the eclipse forums are far more helpful than I ever found IAR  or Keil. Also I have never used eclipse with debugwire so am not certain of the debug part.

I look back with some nostalgia at AVR MCUs. Having looked afresh I now see limitations of the old AVR MCUs.

AVR studio works fine for me so far with the object file renamed to .elf in the makefile and the empty dep folder. Anyone using avrstudio with your code will find the solution in this forum. So there is no need to solve the copy issue.
Thank you for looking at it.

[madires]

I wonder if we'll see a Raspberry Pi RP2040-based transistor tester? $4 for dual Arms, on-board slave MCU's, and 12-bit A/D's 

Nope, the Transistortester needs a 5V MCU.