$20 LCR ESR Transistor checker project

[tom666]

@ryanna
Hi,
mirror of the SVN archive is at:
https://github.com/svn2github/transistortester

Attached zip file contains only the contents of the appropriate directory "mega328_2X16_menu".

[madires]

Are you sure is MENU_ITEMS 5->6 and not 5->8?
As i can see in user.c says define SW_ESR MenuID[Item] = 8; and next:

Code: [Select]

#ifdef SW_ESR
    case 8:              /* ESR tool */
      ESR_Tool();
      break;
    #endif


6 is correct. 8 is just an internal ID to identify the menu item.

i have compiled it and hex is 45.9 kB (45909 bytes) is ok right?

The compiler gives you the actual size of the binary firmware which shouldn't exceed 16k for ATmega 168. If it does you'd need to disable something else. The hex file is always larger because it's the firmware encoded in Intel format.

These are my settings in config.h

Code: [Select]

//#define HW_ENCODER
//#define ENCODER_PULSES   2
//#define HW_REF25
#define UREF_25           2495
//#define HW_RELAY
//#define HW_ZENER
//#define HW_FREQ_COUNTER
and finally:
#define SW_ESR

Is here anything that needs to be enable/disable and its not?
Thank you for the help.

That looks fine!

[micro88]

Hello again,

I have been trying to compare the performance of two different testers purchased from ebay China.

The MK328 which I managed to repair as mentioned earlier and the other is GM328A purchased from
http://www.ebay.com/itm/Digital-GM328A-Transistor-Tester-ESR-Meter-Frequency-Square-Wave-Generator-V2PP-/201479185670?hash=item2ee9186506:g:U24AAOSwbdpWYnJc

I tested a 2N3055 power transistor on both with emitter on 1, base on 2 and collector on 3
The results differed . The GM328A detected the pinout correctly while the MK328 shows interchanged collector and emitter.
Also hfe shown differs by a factor of 2.

I am trying to figure out why the MK328 is showing wrong pinout.

I have attached the results displayed.

Any comments are welcome with thanks in advance.

Cheers.

 

[madires]

I am trying to figure out why the MK328 is showing wrong pinout.

Which firmware versions do the testers run?

[micro88]

@madires

Both testers show 1.12k after selftest.

Thanks.

[madires]

Both testers show 1.12k after selftest.

The version with the symbol at the top left is a little bit more recent (1.12k is the current version under development) but I have no idea why C and E are swapped.

[micro88]

I tested a few other transistors and found that some of them like BC547 and SL100 show similar behavior on both the testers. While for some transistors like 2N3055 the MK328 swaps the collector and emitter. Will post here if I am able to resolve it.

Thanks once again.
Best regards.

[upsss]

Are you sure the battery on both testers is good?  In any case, I would verify the input and output voltage on the 5V regulator on both units while testing.  If the differential voltage is less than 2V, I would replace the battery.

[madires]

I tested a few other transistors and found that some of them like BC547 and SL100 show similar behavior on both the testers. While for some transistors like 2N3055 the MK328 swaps the collector and emitter. Will post here if I am able to resolve it.

Have you run the self-adjustment on both testers? One possible cause could be detection process for BJTs. It's run for C-E and C-E while the correct pinout is determined by the higher hFE.

[micro88]

@madires and @upsss

I have run the self test on both the testers. In fact there is no problem with the GM328A. Also the batteries used are new and 5V measures 4.98V. The only hardware change which I made to MK328 is that I had to remove the connection of the TL431 regulator to pin PC4 of the processor because I found that if I keep it connected the output of TL431 goes to zero. So I concluded that there is something wrong with PC4 which is loading the TL431 .

May be I should replace the Atmega328. I shall try that once I get a good hot air gun and a new smd chip.

Thanks.

[speedy29]

Help, I need to figure out which os install on ATMEL 328, since I was spoiled the processor.
Attaching photos of the device

[tom666]

@speedy29
In the attached file you can find the necessary

[speedy29]

Grazie

[tom666]

Di niente

[timofonic]

@madires

Hola.

I'm sorry a lot for my spasm!

Thanks a lot to you and the rest of the Component (can we call it this way?) Tester Team.

How much people collaborate on in this great project? i Are you all the from Germany?

You can ignore my stupid questions, I'm sure they can be foolish for people enough skilled in electronics. I would love to do versatile abs and smart projects like this someday! Perfect to make the life easier ofr both professionals and I aficionados and economical. Electronics needs more great DIY stuff like this!

I appreciate your incredible effort if all of you!

I'm sure you are busy to reply stupid things.

Did you see that component tester that showed waveforms?

About the (transistor) component tester :

If having a bigger uC:

- Do transistors and certain components have known characteristics that could be identified to know the model despite the code might be erased?

- Could the data sheets be transcribed and redone in saved in a database and compare the results to show the possible model?
* Of course, by crowdsourcing this daunting job of collecting information and using a converter to the very low overhead file format.

 I know it's a crazy and probably impossible idea, but I'm curious about how difficult would be to achieve that by community contributions.

Do you think the IC tester from Chinese would be easy to replicate? I suppose the software would be more difficult than the hardware.


Danke!

[mauroh]

Just received and assembled my DIY M12864 from Banggood!!!!
It is great and I'm really impressed by its capabilities, hats off to everyone who contributed to this amazing project 

It came with and old k firmware and to upgrade it I've added the ICSP connector (see pictures).

I've noticed a difference in the measurement using different firmware:

The 1.12k 523 and 1.12k 555 seems to be fine and the 220nF that came with the kit measure 217.2nF ESR=.29?
The 1.12k 629 (latest) it measure all the capacitors way low and the 220nF became 165.7nF ESR=.11?

I'm performing the calibration each time I load a new firmware but i think I'm doing something wrong.

It could be the fuses?? please see the attached picture.

I've noticed also that it fails to identify TRIACs is this a known bug??

Thanks
Mauro

[madires]

How much people collaborate on in this great project? i Are you all the from Germany?

Karl-Heinz (k-firmware) and me (m-firmware), plus a few helping with support (like tom666), donating hardware (PCBs or clones) and providing some code ideas. Karl-Heinz and I are Germans.

If having a bigger uC:

- Do transistors and certain components have known characteristics that could be identified to know the model despite the code might be erased?

- Could the data sheets be transcribed and redone in saved in a database and compare the results to show the possible model?
* Of course, by crowdsourcing this daunting job of collecting information and using a converter to the very low overhead file format.

 I know it's a crazy and probably impossible idea, but I'm curious about how difficult would be to achieve that by community contributions.

Do you think the IC tester from Chinese would be easy to replicate? I suppose the software would be more difficult than the hardware.

It might be possible to determine a class like small signal or power transistor, but nothing specific. The tester can't distinguish a BJT built for RF applications from a general purpose small signal one, for example. Detecting or testing ICs is a completely different story. You would need several voltages for power supply and signals. Some universal programmers can do that. Additional I/O pins would be used more likely for a frequency counter prescaler and some other hardware options, but not for testing ICs. And it's essential to have a 5V MCU.

[madires]

I've noticed a difference in the measurement using different firmware:

The 1.12k 523 and 1.12k 555 seems to be fine and the 220nF that came with the kit measure 217.2nF ESR=.29?
The 1.12k 629 (latest) it measure all the capacitors way low and the 220nF became 165.7nF ESR=.11?

I don't see any possible cause for that. Maybe Karl-Heinz got an idea.

I've noticed also that it fails to identify TRIACs is this a known bug??

None at all or a few specific ones? Some TRIACs can't be detected because the tester's trigger voltage/current is too low.

[timofonic]

How much people collaborate on in this great project? i Are you all the from Germany?

Karl-Heinz (k-firmware) and me (m-firmware), plus a few helping with support (like tom666), donating hardware (PCBs or clones) and providing some code ideas. Karl-Heinz and I are Germans.

If having a bigger uC:

- Do transistors and certain components have known characteristics that could be identified to know the model despite the code might be erased?

- Could the data sheets be transcribed and redone in saved in a database and compare the results to show the possible model?
* Of course, by crowdsourcing this daunting job of collecting information and using a converter to the very low overhead file format.

 I know it's a crazy and probably impossible idea, but I'm curious about how difficult would be to achieve that by community contributions.

Do you think the IC tester from Chinese would be easy to replicate? I suppose the software would be more difficult than the hardware.

It might be possible to determine a class like small signal or power transistor, but nothing specific. The tester can't distinguish a BJT built for RF applications from a general purpose small signal one, for example. Detecting or testing ICs is a completely different story. You would need several voltages for power supply and signals. Some universal programmers can do that. Additional I/O pins would be used more likely for a frequency counter prescaler and some other hardware options, but not for testing ICs. And it's essential to have a 5V MCU.

Thanks a lot for your charm and learning message! I don't deserve it...

Only two are the main brains of this project? I'm impressed!

Would you both consider in promoting it to attract newcomers? Are you using German in the project?

Would you consider to have a dedicated project site with collaborators summarizing three project in blog posts?

So making that would be a very complex and big beast, very probably lots of research, brainstorming, design, prototyping and an expensive BoM. Would you a good filtered SMPS simplify it or there are lots more to take into account?

Am I right in something? Feel free to comment it if you want, I want to learn

What can be the reason of not adding IC testing? Too much complexity? Too much pins? Do you consider it out of the scope of the project and would be better to be a separate project? More interest in checking analog components?

I sorry, my ignorance made me unable to understand it! And I just learned a tiny amount of it! I wasn't enough aware of the complexity of specialized electronics, I'm starting to understand this is a current universe!

Thanks for your reasonings: They are inspirational, motivational and didactic. I must study a lot harder!

This reminds me to Inductive Chain Learning, Constructivism and Active Learning

[mauroh]

I've noticed a difference in the measurement using different firmware:

The 1.12k 523 and 1.12k 555 seems to be fine and the 220nF that came with the kit measure 217.2nF ESR=.29?
The 1.12k 629 (latest) it measure all the capacitors way low and the 220nF became 165.7nF ESR=.11?

I don't see any possible cause for that. Maybe Karl-Heinz got an idea.

I've noticed also that it fails to identify TRIACs is this a known bug??

None at all or a few specific ones? Some TRIACs can't be detected because the tester's trigger voltage/current is too low.

Hi madires, thank you for the feedback
I made few tests flashing different firmware but I'm getting the same strange values with the latest build.
Tomorrow I'll make some other attempts…

Note the usual repository
https://www.mikrocontroller.net/svnbrowser/transistortester/Software/trunk/mega328_st7565_kit/
is no longer available and now I'm using the mirror on
https://github.com/svn2github/transistortester

Regarding the Triacs, i tested 4 similar variants and got the same result:

Model   ohm
BTA06  75.5
BTA12 100.3
BTB12  69.6
BTB16 83.6

Most probably current/voltage are too low.

Edit:
Do you happen to have the m-firmware compiled for the M12864 Banggood kit??
In the next days I'll put together the toolchain to built it.

Mauro

[madires]

Regarding the Triacs, i tested 4 similar variants and got the same result:
Model   ohm
BTA06  75.5
BTA12 100.3
BTB12  69.6
BTB16 83.6

Most probably current/voltage are too low.

Yes, I can confirm that BTA12 or BTA16 aren't detected. TIC206 or Z0402 are fine.

Edit:
Do you happen to have the m-firmware compiled for the M12864 Banggood kit??
In the next days I'll put together the toolchain to built it.

I've put the display setup into config.h, but I have no feedback yet. You also have to change the I/O pins for the rotary encoder in config.h (ENCODER_A, ENCODER_B) from PD2/PD3 to PD1/PD3.

[mauroh]

Hi madires,
    I modified the config.h uncommenting the M12864 DIY display section, commenting the previously selected one and changing the ENCODER_A from PD2 to PD1.

The result is strange, the eep file is too big..., please see files below(I had to rename the ComponentTester.eep to ComponentTester_eep.hex, forum restriction file format...).

Out of curiosity, is it possible to get the results of the tests on the UART??

Thanks
Mauro

[mauroh]

On the documentation I found the option to enable the UART!!
Tomorrow I'll have few minutes to play with it.
Will let you know

Mauro

[madires]

Hi madires,
    I modified the config.h uncommenting the M12864 DIY display section, commenting the previously selected one and changing the ENCODER_A from PD2 to PD1.

The result is strange, the eep file is too big..., please see files below(I had to rename the ComponentTester.eep to ComponentTester_eep.hex, forum restriction file format...).

Out of curiosity, is it possible to get the results of the tests on the UART??

What's strange exactly? The hex files are always larger than the binaries because they are encoded in Intel format. The compiler's ouput gives you the real size:

Code: [Select]

AVR Memory Usage
----------------
Device: atmega328

Program:   23944 bytes (73.1% Full)
(.text + .data + .bootloader)

Data:        184 bytes (9.0% Full)
(.data + .bss + .noinit)

EEPROM:      739 bytes (72.2% Full)
(.eeprom)

The m-firmware doesn't support the soft UART (yet?).

[i_cant_read]

Hi everyone!
Someone might have asked that already in this threat but what is in your opinion the best one out of those many versions?
Thanks for your replies!
Greetings,
Jannis