$20 LCR ESR Transistor checker project

[jellytot]

The "U4" MCU doesn't support reading the current firmware, only writing a new firmware is possible. It's designed that way.

Thanks madires. So no easy way to go back to original if i use the modified firmware unless I keep original u4 and either program a new one or modify the hardware ?

[indman]

jellytot,the easiest way is to buy some more clean U4 chips and keep the original one. Soldering them is easy, so you always have a choice.

[madires]

In the k-firmware documentation ttester-1.13k.Bohu.pdf at page 77 there is this:

For JFET transistors often the characteristic current Idss is specified, the current in the drainwhen the gate-source voltage is 0V.  Here, however, the current is given by a 680 Ω load resistance atthe source side of the JFET. The load resistor generates a reverse voltage Vgs, which is also shown.With a 470kΩ load resistor at the source side of the JFET the Source-Drain current will be nearlyzero.   With  this  circuit  we  can  get  the  Gate-Source  Cutoff  voltage  Vgs_off  exactly  enough,  if  thevoltage remain below 5V.  With this two operating points we can estimate the current Igss with thenearly quadratic characteristic curve of the current.  If the estimated current Idss stay below 40mA,a additional measurement is done without a additional resistor at the source pin.  With the measuredvoltage at the source pin we can compute a additional current value.  Now we can compute a betterestimated current Idss with this higher current valued, the gate-source voltage and with the knownquadratic current curve, if the value of 40mAis not exeeded.  Due to the symmetrical design of theJFET transistors, the drain and source can not be distinguished.

So I think it's feasible

Yep, as long as V_GS(off) doesn't exceed 4.7 or 4.8V. I'll add it to my to-do list.

[N9XYP]

4) avr8 toolchain 3.6.2 + Arduino-IDE 1.8.13 avr binaries (gcc 7.3.0)
like 3), but with binaries replaced by IDE ones

I am trying to set up my Win10 laptop  and was wondering which folder the binaries are located. These two have similar named folders. The second has more files.

1? C:\Program Files (x86)\Arduino\hardware\tools\avr\lib\gcc\avr\7.3.0
2? C:\Program Files (x86)\Arduino\hardware\tools\avr\avr\lib

[Chris02]

Hello,

I bought two testers from Aliexpress. I tried to change the firmware on the first one in november already and posted my problems here.


Cos I haven't so much time I bought a second one and tried to update this one. But it will also not work. What I have:

TC1 - 644PA inside - which I already tried in november to get working. MLCCs, diodes and measurement Rs (to 0,1%) changed, U4 successfully flashed.
Tester powers on, makes one measurement and powers off. Powers off so fast, that's not possible to read the result. As I described in my old postings. Same with v1.40

T7 - 344PA inside - HW ist china factory, no MLCCs or something else changed. U4 successfully flashed.
If there is no part in it, tester did not find any part (which is correctly) and restarts automatically a new measurement. This will be done 5 times and power off.
If there is any part in it (e.g. a C), tester starts to measure, shows result for some secons and will repeat measurment like a loop. It will never stop. Only if part is removed (then it will try only 5 times like no part at the beginning) or battery is empty.


For both testers config files are the same. Only change is the µC type in makefile. Other settings where copied from "clones".


Fuses on TC1 und T7 were different. But I can't remember how it was. I wrote the same fuses for both tester accidentally.
Now it is:

L: 0xFF
H: 0xD7
E: 0xFD (change like it is written in "clones")
LB: 0x3F

Any tips for me? 

Regards
Chris

[madires]

TC1 - 644PA inside - which I already tried in november to get working. MLCCs, diodes and measurement Rs (to 0,1%) changed, U4 successfully flashed.
Tester powers on, makes one measurement and powers off. Powers off so fast, that's not possible to read the result. As I described in my old postings. Same with v1.40

That's what the original firmware of U4 does. Presumably the reflashing of U4 went wrong.

T7 - 344PA inside - HW ist china factory, no MLCCs or something else changed. U4 successfully flashed.
If there is no part in it, tester did not find any part (which is correctly) and restarts automatically a new measurement. This will be done 5 times and power off.
If there is any part in it (e.g. a C), tester starts to measure, shows result for some secons and will repeat measurment like a loop. It will never stop. Only if part is removed (then it will try only 5 times like no part at the beginning) or battery is empty.

Either start the tester with a long key press (>0.3s) or enable UI_AUTOHOLD in config.h. It's explained in the README (and the German version README.de).

[Chris02]

TC1 - 644PA inside - which I already tried in november to get working. MLCCs, diodes and measurement Rs (to 0,1%) changed, U4 successfully flashed.
Tester powers on, makes one measurement and powers off. Powers off so fast, that's not possible to read the result. As I described in my old postings. Same with v1.40

That's what the original firmware of U4 does. Presumably the reflashing of U4 went wrong.

Thanks for your reply. If I start the tester with a longer press it will make just one measurement. Later I'll check the readme again.

TC1
I flashed U4 again. But still the same. See the attached logfile. For me everything is looking fine.

Code: [Select]

Checking target MCU ...
  MCU type: STC15L104W
  F/W version: 7.2.5Q

Current H/W Option:
  . Current system clock source is internal IRC oscillator
  . IRC frequency: 11.999MHz
  . Wakeup Timer frequency: 26.419KHz
  . Do not detect the level of P3.2 and P3.3 next download
  . Power-on reset, use the extra power-on delay
  . RESET pin behaves as I/O pin
  . Reset while detect a Low-Voltage
  . Thresh voltage level of the built-in LVD : 2.42 V
  . Inhibit EEPROM operation under Low-Voltage
  . Hardware do not enable Watch-Dog-Timer
  . Watch-Dog-Timer pre-scalar : 256
  . Watch-Dog-Timer stop count in idle mode
  . Program can modify the Watch-Dog-Timer scalar
  . Do not erase user EEPROM area at next download
  . Do not control 485 at next download
  . Do not check user password next download
  . TXD is independent IO
  . TXD pin as quasi-bidirectional mode after reset
  . P3.3 output HIGH level after reset
  . Testing time: 2018-6-28

  . MCU type: STC15L104W
  F/W version: 7.2.5Q

Adjusting frequency ... [0.750"]
Adjusted frequency: 11.999MHz (-0.010%)

Re-handshaking ... Successful [0.125"]
Current Baudrate: 38400
Erasing MCU flash ...  OK ! [0.343"]
MCU ID : F2D4246E1A6F8E
Programming user code ... OK ! [2.171"]
Programming OPTIONS ... OK ! [0.047"]

H/W Option upgrade to:
  . Current system clock source is internal IRC oscillator
  . IRC frequency: 11.999MHz
  . Do not detect the level of P3.2 and P3.3 next download
  . Power-on reset, use the extra power-on delay
  . RESET pin behaves as I/O pin
  . Reset while detect a Low-Voltage
  . Thresh voltage level of the built-in LVD : 2.42 V
  . Inhibit EEPROM operation under Low-Voltage
  . Hardware do not enable Watch-Dog-Timer
  . Watch-Dog-Timer pre-scalar : 256
  . Watch-Dog-Timer stop count in idle mode
  . Program can modify the Watch-Dog-Timer scalar
  . Do not erase user EEPROM area at next download
  . Do not control 485 at next download
  . Do not check user password next download
  . TXD is independent IO
  . TXD pin as quasi-bidirectional mode after reset
  . P3.3 output HIGH level after reset

  . MCU ID : F2D4246E1A6F8E
  . MCU type: STC15L104W
  F/W version: 7.2.5Q

  . Set frequency: 12.000MHz
  . Adjusted frequency: 11.999MHz
  . Trim error: -0.010%


  Complete !(2020-08-08 07:25:25)


[madires]

The TC-1 has two special pads for solder bridges to set the power-off time (P1/P2). If you select the maximum time does the TC-1 stay on for a few seconds more? If so, the original firmware is still running. Or if you have a scope you can probe PD2 (or was it PD1?). When powering on U4 sends some data to the ATmega. Again, only the original firmware does that.

[Chris02]

Thanks, I'll check it. To check signal on PD1/PD2 ist also no problem.

[Chris02]

I checked PD1/PD2 with a scope. No signals, only level changes.

PD2 changed from 2,8V to 4.8V after switching on. PD1 from 3.0V to 0V. And both back after power off.

And also no effect if P1 + P2 is connected to GND or VCC. Time is still the same. 

[indman]

TC1 - 644PA inside - which I already tried in november to get working. MLCCs, diodes and measurement Rs (to 0,1%) changed, U4 successfully flashed.
Tester powers on, makes one measurement and powers off. Powers off so fast, that's not possible to read the result. As I described in my old postings. Same with v1.40

Do not be offended, but you have not been able to solve your problem for almost a year.
Have you tried the alternative 2-transistor control scheme that madires suggested?
It is assembled from simple parts in almost 30-60 minutes.
Can you access the menu on this firmware? You need to run hardware tests to help identify potential problems.
Show a photo of your tester board.

[madires]

I checked PD1/PD2 with a scope. No signals, only level changes.

PD2 changed from 2,8V to 4.8V after switching on. PD1 from 3.0V to 0V. And both back after power off.

And also no effect if P1 + P2 is connected to GND or VCC. Time is still the same. 

In that case the new U4  firmware should be running. The m-firmware can be configured to run just one probing cycle and then power off by setting CYCLE_MAX to zero (default is 5). What are your settings for CYCLE_DELAY, CYCLE_MAX and POWER_OFF_TIMEOUT?

[aristarchus]

Hi,

Got a GM328A with 1.12k and wondering how can I upgrade its firmware.

Is there a compiled 1.40m firmware for this so I could flash it?
Or, is there any document on how to compile the sources and get the flash files from this repository?
https://github.com/madires/Transistortester-Warehouse/tree/master/Firmware/m-firmware

Alternatively I see this one but I am not sure if it is a 1.40m
https://www.mikrocontroller.net/svnbrowser/transistortester/Software/trunk/mega328_GM328/
Is this 1.40m?


P.S.
Never mind about how to compile, found this post that gives the steps
https://www.eevblog.com/forum/testgear/$20-lcr-esr-transistor-checker-project/msg1011926/#msg1011926
Spoke too soon, tried it but got errors so back to square one.

[Chris02]

I checked PD1/PD2 with a scope. No signals, only level changes.

PD2 changed from 2,8V to 4.8V after switching on. PD1 from 3.0V to 0V. And both back after power off.

And also no effect if P1 + P2 is connected to GND or VCC. Time is still the same. 

In that case the new U4  firmware should be running. The m-firmware can be configured to run just one probing cycle and then power off by setting CYCLE_MAX to zero (default is 5). What are your settings for CYCLE_DELAY, CYCLE_MAX and POWER_OFF_TIMEOUT?

Config is exactly the same as on T7 and T7 doesn't make any problems. I'll check HW again.

What do you think about replacing 78L05 to MCP1703? Will the measurement more accurate?
I already replaced 680R and 470k to 0,1%.

If MCP1703 is assembled just need to comment #define HW_REF25 or anything else?

[aristarchus]

Hi,

Got a GM328A with 1.12k and wondering how can I upgrade its firmware.

Is there a compiled 1.40m firmware for this so I could flash it?
Or, is there any document on how to compile the sources and get the flash files from this repository?
https://github.com/madires/Transistortester-Warehouse/tree/master/Firmware/m-firmware

Alternatively I see this one but I am not sure if it is a 1.40m
https://www.mikrocontroller.net/svnbrowser/transistortester/Software/trunk/mega328_GM328/
Is this 1.40m?


P.S.
Never mind about how to compile, found this post that gives the steps
https://www.eevblog.com/forum/testgear/$20-lcr-esr-transistor-checker-project/msg1011926/#msg1011926
Spoke too soon, tried it but got errors so back to square one.

Ok, finally managed to be able at least to compile the 1.40m from here, (used a virtualbox XP since in my W10 I could not)
https://github.com/madires/Transistortester-Warehouse/tree/master/Firmware/m-firmware

But now I'm faced with a more daunting task, to identify what settings are needed for the GM328A .

Are the exact settings for the GM328A (config.h, config_328.h) known or anyone done this before and can help here?

[Chris02]

Hi,

Got a GM328A with 1.12k and wondering how can I upgrade its firmware.

Is there a compiled 1.40m firmware for this so I could flash it?
Or, is there any document on how to compile the sources and get the flash files from this repository?
https://github.com/madires/Transistortester-Warehouse/tree/master/Firmware/m-firmware

Alternatively I see this one but I am not sure if it is a 1.40m
https://www.mikrocontroller.net/svnbrowser/transistortester/Software/trunk/mega328_GM328/
Is this 1.40m?


P.S.
Never mind about how to compile, found this post that gives the steps
https://www.eevblog.com/forum/testgear/$20-lcr-esr-transistor-checker-project/msg1011926/#msg1011926
Spoke too soon, tried it but got errors so back to square one.

Ok, finally managed to be able at least to compile the 1.40m from here, (used a virtualbox XP since in my W10 I could not)
https://github.com/madires/Transistortester-Warehouse/tree/master/Firmware/m-firmware

But now I'm faced with a more daunting task, to identify what settings are needed for the GM328A .

Are the exact settings for the GM328A (config.h, config_328.h) known or anyone done this before and can help here?

Check "clones" file for GM328 settings adn readme file also.
Compiling on win10 also works very well. I use WinAVR-20100110-install and need to replace msys-1.0.dll like on Windows Vista

[aristarchus]

Check "clones" file for GM328 settings adn readme file also.

Thanks,
I have already seen the clones text, it does not have the exact GM328A.
And it looks to me that to out the proper settings there the schematic is needed, does anyone got the exact GM328A schematic?

[Chris02]

Check "clones" file for GM328 settings adn readme file also.

Thanks,
I have already seen the clones text, it does not have the exact GM328A.
And it looks to me that to out the proper settings there the schematic is needed, does anyone got the exact GM328A schematic?

What's different?

Maybe this one fits to your tester?
https://dragaosemchama.com/en/2019/04/gm328a-reverse-engineering-new-firmware-and-tetris/

[madires]

What do you think about replacing 78L05 to MCP1703? Will the measurement more accurate?

The MCP1703 is a good choice because of it's 0.4% tolerance and the higher output current. What really matters is that the voltage regulator delivers a stable voltage. The exact output voltage isn't that important since you can measure it and set it in config.h.

If MCP1703 is assembled just need to comment #define HW_REF25 or anything else?

Yep, that will disable the external voltage reference.

[aristarchus]

Maybe this one fits to your tester?
https://dragaosemchama.com/en/2019/04/gm328a-reverse-engineering-new-firmware-and-tetris/

Yes, that is the one.
Have seen this before but it looks to me that the author has problems in the schematic, he has interchanged the lines FREQ with VOLTMETER, at least from what I've seen and I do not know if he has other errors in his reverse engineering.
(My error, after seeing it again looks it is ok)

Anyway, I hope that someone has the exact GM328A settings and give a help here and save me the trouble.

[Chris02]

Thanks, I'll order MCP1703. MCP1703 can be ordered in SOT-223 packe, which is pin compatible, but a littlebit bigger, to TO-89 LDO in TC1. TO-89 MCP1703 has got a different pinout.

Still one questions/sugesstion
I tried to test a brand new RGCL60TS60GC11. Only Diode is recordnized. I guess gate voltage is to low to switch the IGBT on?
What do you think about an small adapter PCB for testing IGBTs with higher gate voltage (even 15V ones). Higher voltage can be taken from boost converter and adjusted thru a e.g. LM317.

[Obelix2007]

@aristarchus

... the entry in clones-list is the first:

DIY Kit "AY-AT"

Regards Horst

[madires]

Still one questions/sugesstion
I tried to test a brand new RGCL60TS60GC11. Only Diode is recordnized. I guess gate voltage is to low to switch the IGBT on?
What do you think about an small adapter PCB for testing IGBTs with higher gate voltage (even 15V ones). Higher voltage can be taken from boost converter and adjusted thru a e.g. LM317.

Possibly, any semiconductor which needs more than 5V or 7mA can't be detected. Many users suggested similar ideas and the answer is that we like to keep the circuit simple. Adding support for higher voltages and currents would require an external switching matrix and would make the tester more complex and expensive.

[Chris02]

You, and also Karl-Heinz, make a great job and I like the transistortester very much. The TC1 is my second Tester. My first one was built in 2013, long time before Chinese built it in a thousand  and I still use it, like it and love it:)

IMHO over this long time electronic has changed. More and more power electronics come to our homes. Almost every new device you buy in your household has got inverters, many many Tools and many many more. So maybe the Tester needs an update for power electronics too ?!

Keep it simple is good. I like it. But what do you thing about an adapter PCB like it is for testing optocouplers. A small PCB with additional electronic which will be pluged in the texttool socket and even has got an own socket?

On this PCB is a selectable voltage source for different gate voltages, powered from the boost converter. Selection has to be done manually, e.g. with a DIP switch. The tester will not trigger the device under test directly, it will trigger an additional transistor (bosster/driver stage). This transistor will power the gate of the DUT and you can measure lot's of values. This will be possible with any tester which has got and boost for Z-Diode testing. You have to calibarate this PCB before to know the characteristics of the booster.

Of course there are also some disadvantages, e.g. you need to know which pin is Gate, etc.

But it is a step in the right direction

[perieanuo]

here you have some tested files
of course, recompile with your crystal frequency, I have 16Mhz version hex/eep files.
just change extension for ComponentTester - Copie.hex into ComponentTester.eep.
regards,pierre