$20 LCR ESR Transistor checker project

[dazz]

dazz
1. Please make the resolution of your photos smaller, attach them to the message or hide them under the spoiler. See how other forum members attach photos.
2. Install m-firmware for your T4 clone and test your transistors again.Compare your results with k-firmware.
3. Carefully calibrate your tester according to the author's manual.

1. Sorry about that, fixed now.
2. I believe my T4 is at the latest version and it's calibrated. Got it from Kubi40's github repo a couple weeks ago. Is madires' the m-firmware you're referring to? Should I expect different results with that one?

[indman]

2. I believe my T4 is at the latest version and it's calibrated. Got it from Kubi40's github repo a couple weeks ago. Is madires' the m-firmware you're referring to? Should I expect different results with that one?

It is likely that there will be other results, which is why I suggested that you try the madires m-firmware.

[madires]

The TC1 shows just V_GS (and the corresponding current) with the JFET's gate pulled down via the 470k test resistor. Yes, the 680 Ohms resistor is used as current shunt. Can't say much more about the TC1's firmware as it is a modified k-firmware (fancy UI, less features, no source available). Your T4 shows I_DSS, V_GS with gate pulled down via 470k (plus current), and V_GSoff (current 0). The 'I=...@Vg=...' format is a compromise to have voltage and current in one text line.

[dazz]

2. I believe my T4 is at the latest version and it's calibrated. Got it from Kubi40's github repo a couple weeks ago. Is madires' the m-firmware you're referring to? Should I expect different results with that one?

It is likely that there will be other results, which is why I suggested that you try the madires m-firmware.

I'm looking for it. I need to recompile it to flip the display vertically, if you remember. But madire's repo is very different from kubi48's one. Is this it, please?
https://github.com/madires/Transistortester-Warehouse

Also, what about my friend's TC1? Will he get Idss and Vgsoff readings if he updates the firmware, or is that a limitation of that particular device?

[dazz]

The TC1 shows just V_GS (and the corresponding current) with the JFET's gate pulled down via the 470k test resistor. Yes, the 680 Ohms resistor is used as current shunt. Can't say much more about the TC1's firmware as it is a modified k-firmware (fancy UI, less features, no source available). Your T4 shows I_DSS, V_GS with gate pulled down via 470k (plus current), and V_GSoff (current 0). The 'I=...@Vg=...' format is a compromise to have voltage and current in one text line.

Oh, thanks for the info, that explains a lot. So no firmware upgrades for the TC1 then.
One thing I don't understand is, with a 470K source resistor, wouldn't that set the drain current very close to zero, hence measuring Vgsoff?

[indman]

Actually, I have two questions. I also have a few MMBFJ201 and MMBF5457, and they are both detected as BJT's by my T4, while his TC1 does spit out a reading, what's the deal there, please?

Some JFET-MOSFETs are not detected correctly by k-firmware, but are successfully detected by m-firmware.
I have experienced this effect many times. I also assume that Chinese firmware contains modified code not only for k-firmware but also partly for m-firmware. But I could be wrong about this.

[dazz]

Actually, I have two questions. I also have a few MMBFJ201 and MMBF5457, and they are both detected as BJT's by my T4, while his TC1 does spit out a reading, what's the deal there, please?

Some JFET-MOSFETs are not detected correctly by k-firmware, but are successfully detected by m-firmware.
I have experienced this effect many times. I also assume that Chinese firmware contains modified code not only for k-firmware but also partly for m-firmware. But I could be wrong about this.

Great, I'll try that.. I need to figure out how to flash madire's firmware with the display flipped vertically though.

[madires]

Oh, thanks for the info, that explains a lot. So no firmware upgrades for the TC1 then.
One thing I don't understand is, with a 470K source resistor, wouldn't that set the drain current very close to zero, hence measuring Vgsoff?

No, the TC1 can be updated, but needs a mod for the additional control MCU (either a different firmware or a two-transistor-circuit). And yes, kind of.

Edit: ... as long as it comes with a genuine ATmega.

[dazz]

So this is what I understand the meter should be doing, based on what's in the manual.
It initially measures Vgs an Id with a 680 ohm source resistor like this



These values represent an operating point somewhere in the middle of Vgsoff / Idss, and my guess is that's the info the TC1 is displaying, correct?

Then it adds a 470K source resistor to measure Vgsoff:



Finally removes the source resistor to calculate Idss:



Not even close, probably, just trying to get a better picture of what's going on and what values each meter is displaying

[indman]

No, the TC1 can be updated, but needs a mod for the additional control MCU (either a different firmware or a two-transistor-circuit).

Madires,as you know, now it is impossible to say with 100% certainty that the clones of the LCR-TC series can be upgraded, as there may be "surprises" from Chinese friends inside the package.

[madires]

So this is what I understand the meter should be doing, based on what's in the manual.
It initially measures Vgs an Id with a 680 ohm source resistor like this

All three circuits miss the 470k gate resistor. And in circuit #3 there's also a 680 Ohms source resistor. See CheckPins() in CheckPins.c for the details.

[dazz]

OK, thanks again, guys, really appreciate you help. So the TC1 is showing something "close" to Vgsoff and the small drain-source current the fet is drawing in that condition, but doesn't measure Idss at all. Hope I got it right now.

I'll see if I can upgrade my T4 with madires' firmware and that allows me to measure my fets. Is the procedure the same as with kubi48's firmware? Where is madires repo?

[madires]

Nope, TC1 doesn't show V_GSoff or anything close. For the m-firmware please see https://github.com/madires/Transistortester-Warehouse/tree/master/Firmware/m-firmware.

[indman]

Some JFET-MOSFETs are not detected correctly by k-firmware, but are successfully detected by m-firmware.

Here is an example of how JFET-N J201 detection differs on k-firmware and m-firmware 

[dazz]

I have madires' code compiled already with the added flag to flip the display, going to flash it now. Thanks, guys!

[dazz]

I have an issue with the display contrast, it's too high. I went to the menu and it was set to 22, but I don't know how to lower it, just kept increasing hoping it would cycle back to zero and up again, but it didn't. Also tried adding the CFLAGS += -DVOLUME_VALUE=10 flag to the makefile, but it seems to ignore it completely. Sorry to be such a pest, but can you guys please give me some pointers to fix this?

Using this: avrdude -p m328p -P usb -c usbasp -U flash:w:ComponentTester.hex -U eeprom:w:ComponentTester.eep

[madires]

From the README:

You can adjust the contrast for some graphic LCD modules. A short key press
increases the value and and a long key press decreases it. Two short key
presses will exit the tool. With a rotary encoder installed the value
can also be adjusted by turning the encoder.

To change the default value check out LCD_CONTRAST in config_328.h.

[dazz]

Got it working! It's now measuring my mmbfj201's, and the values for Idss and Vgsoff are almost identical to the ones I measured manually yesterday. Great stuff guys, vielen danke!

[dazz]

I'm still having a couple minor issues. The text in the display is shifted a little bit (see picture attached). I checked the flags in the documentation but couldn't find an obvious one to try, any ideas, please?

The second issue is, I added the POWER_OFF flag to the Makefile, but the tester keeps probing on a loop. Here's what I got in WinAVR, I added the last two compiler flags, and it compiles without error, not sure what I'm doing wrong.

Code: [Select]

# compiler flags
CC = avr-gcc
CPP = avr-g++
CFLAGS = -mmcu=${MCU} -Wall -I. -Ibitmaps
CFLAGS += -DF_CPU=${FREQ}000000UL
CFLAGS += -DOSC_STARTUP=${OSC_STARTUP}
CFLAGS += -gdwarf-2 -std=gnu99 -Os -mcall-prologues
CFLAGS += -funsigned-char -funsigned-bitfields -fpack-struct -fshort-enums
#CFLAGS += -flto
CFLAGS += -MD -MP -MT $(*F).o -MF dep/$(@F).d
CFLAGS += -DLCD_ST7565_V_FLIP=1
CFLAGS += -DPOWER_OFF
EDIT: Wait a minute, I commented out both those flags I added, compiled, flashed... and it made no difference. Looks like it's ignoring them completely.

[madires]

For the m-firmware edit config.h and config_328.h, as explained in the README. And don't add random flags to the Makefile! The shift of the display is caused by LCD_OFFSET_X (config_328.h). Simply comment it out. For probing cycles see CYCLE_DELAY, CYCLE_MAX and POWER_OFF_TIMEOUT (config.h).

[dazz]

For the m-firmware edit config.h and config_328.h, as explained in the README. And don't add random flags to the Makefile! The shift of the display is caused by LCD_OFFSET_X (config_328.h). Simply comment it out. For probing cycles see CYCLE_DELAY, CYCLE_MAX and POWER_OFF_TIMEOUT (config.h).

Oh, by README I thought you meant the pdf documentation. Sorry about that, got it now.

[indman]

This is an important difference for compiling k-firmware and m-firmware I think will be very useful for many people to know:
In the k-firmware all the basic settings before compiling are made by editing the file "Makefile". In the m-firmware the global settings are concentrated in 3 files: "Makefile", "config.h" and "config_<mcu>.h

[dazz]

All is good now, thanks guys. 

[madires]

The RTFM hint is really nice! I think I should add it as a default option, mabye UI_RTFM?

[indman]

The RTFM hint is really nice! I think I should add it as a default option, mabye UI_RTFM?

I also suggest that the RTFM be placed on the start page as well as in the "DATA" section