$20 LCR ESR Transistor checker project

[indman]

Currently, I could buy either one (of the two), to be pass to a friend.

Is that a real good friend? If so, then why make such a disgusting gift that has a lot of flaws from the start?
Is it such a big problem to find a GM328A c DIP ATMega328 variant or TQFP factor now?
If you still have to pick one shit out of two possible, it better be LGT8F328.

[Feliciano]

I'm exploring the options, the costs, and the wait time. Thanks for the answer.

[indman]

I'm exploring the options, the costs, and the wait time.

I'm sure you'll make the right choice, good luck!

[Feliciano]

Another question: what is the maximum Symbol size that we could use for a 160x128 display? According to the config is 32x32, and I read here that 32x39 is not meant for that resolution, so I wonder whether it would worth to develop something in the middle, like 32x35
(note: my current font size is 8x16. Has one to be an exact multiple of the other?).

[madires]

For practical reasons the symbol size should be a multiple of the font size to match the probe numbers and their positioning. Otherwise you would get unpleasing gaps between symbol and probe numbers. Another important point is that some display controllers address the Y axis via 8-bit pages, i.e. you can only address the page and not specific rows. And we can't use a display buffer because the ATmega328's RAM is too small for this. So the height of fonts and symbols needs to be a multiple of 8 for those controllers.

[madires]

After surviving the April's Fool Day here's the real deal. Get your programmer ready and enjoy!

v1.52m
- Added quick-check tool for diodes and LEDs (SW_DIODE_LED).
- Support for BH1750VFI ambient light sensor (HW_BH1750).
- Fixed protocol issue for addressing chips in read mode in function I2C_WriteByte() for bit-bang I2C.
- Fixed scaling error in the voltage divider calculation in T2 of the Self Test function.
- Option for packed output format of font display for test purposes (FONT_PACKED).
- Added new font font_8x16_alt_hf.h (FONT_8x16_ALT_HF, thanks to Feliciano@EEVblog).
- Corrected bitmap for O umlaut in font_8x16_iso8859-2_hf.h (thanks to Feliciano@EEVblog).
- Fixed data byte issue in I2C_ReadByte() for hardware TWI (reported by fail@mikrocontroller.net).
- Corrected the colors for 0.1 and 0.01 multipliers in Display_ColorCode() (reported by chaoslegion@VRTP).
- Option to switch temporarily to auto-hold mode after a component is found (in continuous mode, UI_AUTOHOLD_FOUND, suggested by Maniaxx@EEVBlog).
- Added option to optimize the interrupt vector table to reduce firmware size (Makefile: OPTIMIZE_VECTORS, thanks to Viktor Klimkovich).
- Improved UJT detection to exclude potentiometers with a specifc resistance (3-15 k) and the wiper turned to one end (reported by wandows@EEVblog).
- Fixed issue with return value in DataStorage() (thanks to Viktor Klimkovich).
- Fixed detection issue of small capacitances which was caused by the newly added negative zero offset in LargeCap() (reported by Yuriy_K@EEVblog and Obelix2007@EEVblog).

Please download at:
- https://github.com/madires/Transistortester-Warehouse/tree/master/Firmware/m-firmware
- https://github.com/kubi48/TransistorTester-source/tree/master/Markus (soon)

[Adrian_Arg.]

Hello people, would someone be so kind as to make a video of the configuration and compilation steps, since I have a gm328, and I would like to update it

[indman]

madires ,thank you for adding very important clarifications to points T1-T6 selftest in the README! I think it would also be useful to add one more important point to this HINTS point in README:
"- Long test leads have a negative impact and worsen the values in T2, T3, T5 and T6."
- Disconnect all ICSP interface conductors - they significantly affect the selftest(adjusting) procedure and also interfere with the normal operation of the tester!

[carl1961]

Hello people, would someone be so kind as to make a video of the configuration and compilation steps, since I have a gm328, and I would like to update it

There is different ways, depending on what computer your using.

I learned a lot for this video   

this is how I did it on my Window 11, but should be same as for windows 10

https://github.com/carl1961/Transistor-Tester-AVR-Compiled-files-On-Windows

[Adrian_Arg.]

Thanks, I'll see if I can do it, with the last few attempts it's stuck.

[Kim Christensen]

Thanks, I'll see if I can do it, with the last few attempts it's stuck.

Don't forget that you must upload both the EEP and the HEX file. Uploading just the HEX file can cause it to malfunction.

[Adrian_Arg.]

Thanks, I already installed virtualbox and Ubuntu, I followed the steps at https://github.com/carl1961/Transistor-Tester-AVR-Compiled-files-On-Windows
  Now I have to see what files are modified for my GM328, like the one in the video

[Yuriy_K]

After surviving the April's Fool Day here's the real deal. Get your programmer ready and enjoy!

Why did you block diode mapping in JFET transistors? "Protective diodes are limited, but they do occur. I had to restore the existing solution myself.

[madires]

Didn't know that there are some JFETs with an integrated flyback diode. I'll add that to the todo list. And I also need to investigate the impact on symmetrical JFETs.

[Feliciano]

Same with some BJT, as you know. So I wonder, what if we add a thin diode symbol bitmap, and concatenate it to the normal symbol when required? In fact, from the capture of Yuriy_K's firmware, I guess he's doing that and the diode is in different color?

[indman]

Feliciano, The color of the diode does not play a special role here. For ATMega644-1284 it is easier to draw a few more additional symbols with a built-in diode - this will not significantly increase the size of the firmware. For ATMega328 with a limited memory size, the diode symbol option like Karl-Heinz is more preferable, but this forces the remaining symbols to be shifted to the left to make room for a possible diode, which I don’t really like.

[madires]

Didn't know that there are some JFETs with an integrated flyback diode. I'll add that to the todo list. And I also need to investigate the impact on symmetrical JFETs.

This is becoming more complex. I've tested a few JFETs with an additional diode between source and drain. For a n-ch JFET an optional flyback diode would work out of the box. And it would also determine source and drain. Reversing the diode also reverses source and drain. The detection of a p-ch JFET is totally messed up by a flyback diode, i.e. drain and source are the opposite of what they should be, and I_DSS and V_GS_off are much too high (because of the source/drain reversal plus the diode).

[Feliciano]

indman, the color of the diode itself is not a deal-breaker, but being in different color makes me think it was generated by another method and printed (to the screen) as needed. I think that would be a nice feature to have, that's my real suggestion. The symbol itself would require a few extra Bytes, my major concern would it be the code-size for inserting it to the right of the other symbol. And if slim, it won't affect much. Or, I don't know whether it would possible to overlap the other bitmap, or between the numbers, to avoid the shift.

And regarding to having the body diode or not by default, I and noticed it's present in some of the m-firmware bitmaps. For the majority of MOSFET I know of, they has it, so I'm not sure on which one would it be the best default option (but indeed in m-firmware we have both options to choose from). I haven't looked at kh's for that part yet, though.

[indman]

And regarding to having the body diode or not by default, I and noticed it's present in some of the m-firmware bitmaps.

Yes, it is. Look carefully at the title of the description: who created these character sets?

[Feliciano]

Of course madires and you, dear.

And now I found the explanation of what I'm seing, on an old Yuriy_K's post, which if the code size confirms it, it could be a reasonable alternative, as long as the symbol's letters are kept to the left of the wires (and of course we choose a diode-less symbols bitmap).

[Feliciano]

And as a matter of fact, a couple of posts down, you can notice inductor, crystal, and batteries symbols in the font set, instead of the symbol set. I guess that saves a little memory too. Maybe madires would like to check those alternative approachs again?

[madires]

Each new symbol display variant increases the complexity and we already have a few variants (-> maintainability).

[madires]

Back to JFETs with an integrated flyback diode. Yuriy_K has such an n-ch JFET. But are there also p-ch JFETs with a diode?

[Feliciano]

From what I see, for transistors he uses the standard symbols, and optionally takes the upward or downward diode from custom extra characters of the font, and prints it under the upper-right terminal number. For other components, takes a more traditional approach, but with a different character code. Attached some examples for differents scenarios.

Whether easy or not to implement, and/or the code-size versus tradicional solution, and/or the feasibility for 32KB, you're the expert.

[indman]

From what I see, for transistors he uses the standard symbols, and optionally takes the upward or downward diode from custom extra characters of the font, and prints it under the upper-right terminal number.

No, on the first photo in post #9362 we see that Yuriy_K has found a nicer way to display this additional diode.Whether he will share this way or not we do not know. All the previous examples you have shown do not make me want to apply them.