$20 LCR ESR Transistor checker project

[edgard22]

Hello everyone.
I am updating and modifying my GM328, with more functions and cosmetic changes.
And I would like to change the question mark in case of a failed test run to another position on the screen, for example above or below the message.
Could you tell me where I can make the change.

Regards

Hello, I have that same AY-AT and I use the font FONT_8X16ALT_WIN1251_HF which best suits this model.

[edgard22]

I want to thank madires for their selfless contribution by reading suggestions and updating the code, and also to all those who respond, especially those who, like me, do not have programming knowledge, therefore it is very difficult for us to do anything we want to change. Thanks to all.

[edgard22]

With some work I achieved what I had set out to do, wich was something to help me identify fake transistors based on the capacity of the base-emitter junction. I also added some lines to display.c to limit it to two digits after the decimal dot because it was off-screen. (I hope it doesn't cause conflicts in other visualizations).
In the next pictures you see comparisions between some real, and really bad 2SC5200 transistors.

[madires]

C_ebo is still on my to-do list. If you like to share your modifcation please send me a copy.

[Yuriy_K]

Here is my solution and examples of measurements on different firmware...

[madires]

You can simply use

Code: [Select]

Display_Value(Semi.C_value, Semi.C_scale, 'F');
as C_be is already measured anyway.

BTW, the next m-firmware version will have the new option SW_C_BE to display the base-emitter capacitance.

[ledtester]

Just today youtube user Defpom posted this video comparing ESR measurements between a FNIRSI TT clone and a DE-5000 and Shannon Tweezers:

🔴 A Surprising Result ? - FNIRSI DSO-TC3 ESR Measurement Testing - No.1273
https://youtu.be/1BhfNQuzeLs

In short, the DSO-TC3 always measured a very low ESR whereas the DE-5000 and Shannon Tweezers seems to give more reasonable results which also matched each other.

I realize the FNIRSI device is a clone which probably doesn't even use AVR hardware. My questions are:

Are there known limitations to the ESR measurements made by the TransistorTester methodology? Are there known circumstances where the measured ESR is likely to to be way off? Some TT clones come with warnings that the reported ESR tends not to be very accurate, e.g.:

https://www.eevblog.com/forum/testgear/%2420-lcr-esr-transistor-checker-project/?action=dlattach;attach=236653


FWIW, here are the measurements made in the video -- blanks indicate that particular measurement was not reported.

Code: [Select]

        CAP     VLOSS   ESR     FREQ
TC3     36.4u   0.2%    0.48R
DE      32.62u  0.24%   11.8R   100Hz
ST      32.93u          11R     100Hz

TC3     217u    0.4%    0.46R
DE      202.7u          3.9R    100Hz
ST      202.5u          3.915R  100Hz

TC3     7.266u  0.2%    0.28R
DE      7.388u          18.8R   100Hz
ST      7.348u          18.93R  100Hz


[madires]

The Transistortester's ESR measurement method delivers a suprisingly good approximation, but it will never be as good as a proper LCR meter. And it's optimized for the ATmega MCU family. Newer clones of the Transistortester often come with an alternative MCU and the manufacturers do a poor job of porting the firmware (while also hiding the OSHW nature of the project). Please ask FNIRSI why they screwed up the OSHW firmware.

[Kim Christensen]

Just today youtube user Defpom posted this video comparing ESR measurements between a FNIRSI TT clone and a DE-5000 and Shannon Tweezers:

I remember when a co-worker showed me one of the original units back when they came with genuine AVRs in DIP packages.
I didn't believe that such a device would give a reasonable capacitor ESR reading, so I borrowed it for a week and compared various caps, both good and faulty, with our HP4262 LCR meter and was duly impressed. It wasn't perfect, but seemed to be close enough for repair purposes when verifying bad capacitors after removal.

[indman]

The Transistortester's ESR measurement method delivers a suprisingly good approximation, but it will never be as good as a proper LCR meter. And it's optimized for the ATmega MCU family.

I will add one important addition to madires post - this is true for POLAR ELECTROLYTIC capacitors!
If you want to check the ESR of non-polar capacitors, especially those with ratings less than 1µf, be prepared that the results may differ from those measured on an LCR bridge meter by a factor of 10-100. The serviceability of such capacitors on the transistor tester can be checked, as it is correctly stated in the author's manual, only by comparison with a previously known serviceable capacitor of the same type!

For example, comparing the TT DTU-1701 with the LCR DE-5000 which I made relatively recently at the request of a forum member.
DTU-1701 is equipped with the author's firmware 1.13k. By hardware update - LDO MCP1702-5.0 and ION LM4040-2.5 are installed.
The test capacitors are not used or soldered (except the last bad one for clarity) in the photo below.
In order from left to right:
1. Hinano 4.7µFx50V 105°;
2. Hitano 47µFx160V 105°;
3. Hitano 220µFx50V 105°;
4. Hitano 470µFx100V 105°;
5. Hitano 1000µFx50V 105°;
6. Jamicon 2200µFx10V 105°;
7. Hitano 3300µFx25V 105°;
8. KZG 3300μFx6.3V 105° (defective, removed from computer motherboard).

Measurement results and comparison in the photo below:
From left to right - DE-5000 capacitance and ESR measurements at 100Hz,at 10kHz(or Rs if capacitance measurement is not available),at 100kHz(or Rs if capacitance measurement is not available),DTU-1701 test.
I'll explain the results a little bit:
Low ESR values of 0.00 ohms, especially near 0 on TT "float" plus or minus 1-2 units of the lowest digit, which is understandable and acceptable for this circuitry.
The capacitance of the faulty capacitor is overestimated compared to the DE-5000, but the high losses and ESR already show that it is unusable.
In principle, as I wrote before, you can trust the transistor  tester readings if the AUTHOR'S OSHW firmware is loaded, the device is in good working order and properly calibrated!

[Yuriy_K]

I will add to the previous message a comparison of readings with a much more accurate ESR meter. I show the difference in readings on short and standard connecting wires. All capacitors are new, I cut off the wires to demonstrate the difference...
I took measurements for a frequency of 20 kHz, at this frequency the readings of the k-firmware for such capacitors coincide.

[edgard22]

I think that in general the FNIRSI articles are of dubious quality, I have the 1014D oscilloscope and it is quite bad.

[Obelix2007]

... what do you expect from an oscilloscope for 75€ ?

Greeting Horst

[edgard22]

Yes, okay, but I never thought it was that bad. It is very imprecise in voltage measurements, for example. Now that I saw several reviews I should have bought an OWON SDS1104 or the Hantek DSO2D10 for almost the same price.   

[Feliciano]

If we start to talk about FNIRSI, we could spend pages discussing its bad build quality, low-grade components, and poorly ported firmware. I think the point has been already made several times on this thread. For those interested, a quick search would retrieve many examples.

[Strateg58]

Hi everybody. is it possible to increase the contrast of the ILI9341 display in firmware 1.53M?

[madires]

Color displays usually have some sort of auto-contrast. But you could play with the VCOM control 1 in file ILI9341.c, function LCD_Init():

Code: [Select]

  /* VCOM control 1 */
  LCD_Cmd(CMD_VCOM_CTRL_1);
  LCD_Data(FLAG_VMH_5000);         /* 5.000V, could be used to adjust contrast */
  LCD_Data(FLAG_VML_0600);         /* -0.600V */

For supported values please see file ILI9341.h (search for CMD_VCOM_CTRL_1).

[Strateg58]

Dear Magirus. Help me figure out my problem. I added the frequency meter function to my transistor tester LCR TC-1 v2.0. It measures the frequency, but the line with the frequency value on the screen does not stand steadily but blinks constantly. The higher the measured frequency, the more frequent the blinking. Firmware 1.53. On another tester with firmware 1.13, the line with the frequency value does not blink during measurements. I used two different schemes (in the attachment). Nothing changes. I am sending a signal from another transistor tester. The frequency meter input is made on Pin.40 PBO Atmega644PA.

[madires]

Which display does the LCR TC-1 tester have? ST7735? ILI9341/9342? And which display does the other tester have?

[Strateg58]

Both testers have ILI9341 displays. With firmware 1.51 2.8 inches. With 1.13K 2.4inch firmware.

[madires]

And you're using bit-bang SPI based on the schematics a few posts earlier? Bit-bang SPI with high resolution color displays is a bit slow as it takes more time to write the color data. If you would use PB5 for SDA and PB3 for D/C you could run hardware SPI.

[Strateg58]

Here are my settings in Config. 644H and Config.H PB3 is connected to the pin DC ILI9341. PB5 is connected to the pin SPI MOSI ILI9341.

[madires]

SPI_HARDWARE is already enabled in config_644.h. So you don't have to enable it also in config.h.

Regarding the frequency output, after each measurement the frequency is displayed as "f <frequency>Hz". Or in case of a bad measurement it's going to be "f -". For higher frequencies the gate time decreases and therefore also the measurement time, i.e. the frequency display is updated faster.

[Strateg58]

Madires thank you for giving me your time. If I understood you correctly, is the line with the frequency value supposed to flash like this? Then I don't understand why it doesn't flash on 1.13K firmware.

[madires]

I'm not sure about what you mean by blinking or flashing. The frequency measurement should just update the displayed frequency.