$20 LCR ESR Transistor checker project

[Maniaxx]

Is there anything one should pay attention to when getting a GM328A from China? I read something from "old batches" or "round edges". I want to flash custom firmware and replace crystal (16MHz). Would this be a good pick? They call it 'new version'.

[jwperry]

indman,
We are starting to get somewhere. Both versions boot and allow for testing of components. I just using a 4700uf 35v cap for testing purposes. However there is still some interesting behavior from the tester. I can test components with the Chinese NoStripe firmware but the display font is very dim.  On the third firmware the font takes a second to "warm up" but is eventually bold and legible.  Also the third firmware keeps requesting calibration which I attempted. I can test a component with the third firmware and I receive values but it does ask for calibration each time. I may need to read the documentation on how to properely calibrate. I would start wondering about components on the board or the display itself, however, when flashing the first firmware you provided (with the odd characters), the font is is completely fine.

Breakdown:
First Firmware: Bold, legible odd characters
ChineseNoStripe: Faded hard to read, tests components successfully
Third Firmware: Screen looks mostly fine. Tests components successfully, but might have some calibration issues

Thank you for your efforts so far sir!

-John

[Yuriy_K]

I want to flash custom firmware and replace crystal (16MHz). Would this be a good pick? They call it 'new version'.

If this clone has a BGR gamma display, it cannot run at 16 MHz. See page 277 for details.

For "k" firmware corrected the measurement of Darlington transistors. Source and pictures are attached. The source requires more memory, check for exceeding the limit.

[indman]

For "k" firmware corrected the measurement of Darlington transistors.

What are the markings on the transistors you showed in the screenshots?

[Yuriy_K]

What are the markings on the transistors you showed in the screenshots?

These are NPN transistors - KT972B, PNP - 2SB1093 - datasheet on the previous page.

New firmware, taking into account previously published changes.

[vever001]

Hello everyone, and thank you for this great project!

I'm attempting to update my "Daniu" TC-1 clone with an atmega324pa.
I did not make any hardware modifications yet (no 10uF cap, no mod).
The v1.34m hex from the Yandex repo works fine https://disk.yandex.ru/d/yW8xa5NJgUo5z/LCR-%D0%A2%D0%A11/Firmware/Markus%20Reschke/English/16MHz

I'm now trying to compile the latest v1.47m but all I get is a white screen and I need to unplug the battery.
Here are the changes I made from the original v1.47m: https://github.com/vever001/Transistortester-Warehouse/compare/39cf8f2adc5921e460891bf5707916381a2d36cf...vever001:Transistortester-Warehouse:master
You can also check my commits here: https://github.com/vever001/Transistortester-Warehouse/commits/master

I don't know what's wrong... I tried to replicate the exact config from the yandex repo.
Can anyone spot the issue? or give any hints?

Thank you very much  ,
Hervé

[madires]

Have you tried the settings listed in the Clones file?

[vever001]

Ok I started again from scratch, applied the clones instructions and it works perfectly now... Sorry for that
Here's my config if anyone wants it (for a Daniu TC-1 with atmega324pa): https://github.com/vever001/Transistortester-Warehouse/commits/clones-settings

I made some more changes:
- flip the LCD X and Y (screen was upside down)
- remove the LCD offsets (pixel lines were showing up at the edge)
- use 8x16 cyrillic font (some chars were not showing properly)
- enabled UI_AUTOHOLD (for convenience)

Now I can start digging in the options and hardware mods

Thank you for the amazing work,
Hervé

[indman]

New!
Updated the comparison table of clones  - added 2 devices on fake microcontroller and a clone on LGT8F328.
Be careful when choosing a device to purchase from Chinese friends! 

[MaxFrister]

I bought a T7 ("LCR-T7(2)"?) recently to try. It appears to mostly work except for zener testing.

For low voltage zeners, it seems to test but test high.

For high voltage (>20V) zeners, it seems to power them as soon as connected even before test button is pressed.  Reports results but high.  However the zeners are hotter than the sun after removed.

Before I open the case and void my warranty ;-) I wanted to check that this is not  the expected behavior.

[madires]

IIRC, the boost converter of the TC-1, T7 and the other models with the same case is running as long as the tester is powered on.

[indman]

MaxFrister,In all TC clones, the current in the tested zener diode is limited by a resistor of a certain value in units of milliamps(no more than 5mA). If the zener diode being tested is very hot, this means that a more significant current flows through it, which can damage it.This is a malfunction!Open a dispute with the seller, unless of course you somehow shorted the K-A contacts between each other or with one of the test contacts 1-2-3.Then you are the culprit of the malfunction.

 I also have a suspicion that you have a clone inside with a fake processor. You are no longer the 1st who complains about the heating of the tested zener diode.I think this malfunction is caused by a serious error in the circuitry of Chinese engineers.

[MaxFrister]

I'm pretty sure it is defective although I was unable to quantify how defective.  I tried to measure the current using a series resistor and managed to melt the probe grabber on my scope.  I don't think I'll try that again.

As for shorting K and A on the tester, I didn't do that but isn't that a normal diode failure mode?

Thanks.

[indman]

As for shorting K and A on the tester, I didn't do that but isn't that a normal diode failure mode?

This is really not dangerous if a current stabilizer or its limiter is provided in the circuit. If it is not present or it is incorrectly calculated, this will lead to the failure of the measuring port of the controller.

[Feliciano]

To characterize the boost converter you have in front, with the aid of your DMM:
a)you could measure the open-circuit voltage, it should be around 30V
b)then you could put a resistor, like 1K (around 1W to be on the safe side), and measure the voltage again, calculate the current, and check whether the internal current limitiation is working as it should. For testing a Zener, a few mA should be enough

[indman]

Feliciano,this measurement operation is true for the classical tester scheme. But in the clone version with a fake processor, the LM358 operational amplifier is present in the zener diode measurement circuit.What function does it perform in the scheme? A clear and understandable schematic diagram of the clone is needed to definitively determine the cause of this error.

[Feliciano]

For sure an schematic (official or reverse-drawn) would be better, but as you know a boost converter should follow an inductive (classical) or capacitive (voltage multiplier) principle, and we could try to guess a rough Thevenin or Norton equivalents of it around a given operating point. If the zener gets too hot, the Voc is too high, and/or the internal resistance is too low. Knowing which is the case, a modification could be devised to the circuit, trying to make it safe.

Regarding the op-amp in this case, it could be part of a not-so-regulated (cheapo) boost converter, or it could be part of the V sensing for protecting the microcontroller. That would deserve analysis after the initial findings. On the other hand, an LM358 can only source a few tens of mA, bellow it's Vcc. That's why if I were MaxFrister, I would start by trying to understand how the circuit behaves and then think whether a mod can be made.

[indman]

Look closely at the PCB - pins A are shorted to ground through a 40 ohm resistor(59x)! The output voltage from the DC-DC switch (28V) is directly fed through the diode to pin K. It turns out that the current through the zener is limited only by this resistor and open diode (and it's ~ 0.7A!!!) That's why the diode D2 damage and the zener diode is very hot. If I saw the right track on the board.
For the experiment I suggest you solder a 5-10 kΩ resistor in series with the diode.It's not hard mod.

[madires]

Look closely at the PCB - pins A are shorted to ground through a 40 ohm resistor(59x)! The output voltage from the DC-DC switch (28V) is directly fed through the diode to pin K. It turns out that the current through the zener is limited only by this resistor and open diode (and it's ~ 0.7A!!!)

Ouch! A constant current source with a BJT and a few diodes and resistors is simple and inexpensive. Just a series resistor causes a too large variation of the test current, which in turn also impacts the measurement (Zener voltage). AFAIK, the TC-1 and T7 Plus have a 10k series resistor. We would have less trouble if those cloners would stick to the recommended design instead of trying to cut even more corners.

[indman]

A constant current source with a BJT and a few diodes and resistors is simple and inexpensive.

You're right of course, a simple current constant regulator is called for in this design. But this is only one obvious problem that was detected. And I'm sure there will be others.

[MaxFrister]

Look closely at the PCB - pins A are shorted to ground through a 40 ohm resistor(59x)! The output voltage from the DC-DC switch (28V) is directly fed through the diode to pin K. It turns out that the current through the zener is limited only by this resistor and open diode (and it's ~ 0.7A!!!) That's why the diode D2 damage and the zener diode is very hot. If I saw the right track on the board.
For the experiment I suggest you solder a 5-10 kΩ resistor in series with the diode.It's not hard mod.

I was able to install a 5.1k resistor for current limiting.  I chose to install it after the cap rather than before.  I initially tried a 8.2k resistor but the tester reported a zener of 29.5V with nothing in the test socket.

On my unit, "A" is connect to ground via a 5-ohm resistor.  I'm not really sure why. 

It works now and best of all you can no longer use the DUT as a soldering iron.

[indman]

I chose to install it after the cap rather than before.  I initially tried a 8.2k resistor but the tester reported a zener of 29.5V with nothing in the test socket. On my unit, "A" is connect to ground via a 5-ohm resistor.  I'm not really sure why. 

I'm not quite sure how you installed the extra resistor? Can you show a high quality photo of this section of your PCB with the resistor installed?

[vklimk]

I was able to install a 5.1k resistor for current limiting.  I chose to install it after the cap rather than before.  I initially tried a 8.2k resistor but the tester reported a zener of 29.5V with nothing in the test socket.

On my unit, "A" is connect to ground via a 5-ohm resistor.  I'm not really sure why. 

It works now and best of all you can no longer use the DUT as a soldering iron.

Try to play with that 5 ohm resistor instead of 5,1k. It seems that 5 ohm resistor is a shunt (current sensor) in current supply unit (similar to power supply for led). Put variable resistor in place of 5 ohm and try to change current using it.
It is just a theory, I don't have such a device to check. But if sdb628 is in use here then 120 ohm shunt should limit current to 5mA. 5ohm shunt should limit current to 120mA.

[mebel]

Sorry for obvious question.

The link doesn't work anymore. Where from can I download this file?

1. To download from the author's site at the link https://www.mikrocontroller.net/svnbrowser/transistortester/Software/trunk/  the current distribution kit of firmwares.
For this purpose to click in the bottom of the page on a line of ""Download GNU tarball"".

[Fuzzy Star]

Sorry for obvious question.

The link doesn't work anymore. Where from can I download this file?

1. To download from the author's site at the link https://www.mikrocontroller.net/svnbrowser/transistortester/Software/trunk/  the current distribution kit of firmwares.
For this purpose to click in the bottom of the page on a line of ""Download GNU tarball"".

https://github.com/kubi48/TransistorTester-source/

https://github.com/madires/Transistortester-Warehouse/tree/master/Firmware/k-firmware