$20 LCR ESR Transistor checker project

[MrSqueaky]

Very disappointing, the mega328 was almost 1/4 the price of the tc1 tester and now not getting it to work any time soon!
I get it there is no firmware available for it yet, or the chip not available to buy!
Clone tester with a clone chip from a clone! 
Hope the mega328pb is still ok, to keep for the t4, just in case another charged cap comes around!

A 5V Arduino Pro Mini glued on to the nice big ground plane and lots of enamelled jumper wires and it'll be good as new 

There seems to be a ~US$5 price difference between the LCR-TC1 and LCR-TC2 on Aliexpress. I'm guessing that covers the price of a real Atmel chip

[elecdonia]

There seems to be a ~US$5 price difference between the LCR-TC1 and LCR-TC2 on Aliexpress. I'm guessing that covers the price of a real Atmel chip

BINGO! Spot-on!  

[Thelmos]

I just received a LCR-TC2 from China...

I didn't like the first test, short-circuiting the three pins didn't launch the calibration, it directly measures the short-circuit resistance, from what I've read in this thread this may be because the resistors are out of spec (https://www.eevblog.com/forum/testgear/$20-lcr-esr-transistor-checker-project/msg4726172/#msg4726172).

The manual doesn't even mention how to do the calibration.

I have also done the test that @elecdonia mentioned in https://www.eevblog.com/forum/testgear/$20-lcr-esr-transistor-checker-project/msg4692413/#msg4692413, results: 12.68V 

Seeing the board, my fears have come true...

[MrSqueaky]

I just received a LCR-TC2 from China...

Looks like cost cutting has finally caught up to the LCR-TC2 (PCB labelled T7-PLUS V2.0) or it a knock off. What markings are on the MCU? It is not a ATMega324 (pin count wrong) and not a ATMega328P (power pins don't match). Some sort of clone chip.

The circuit for the zener diode voltage appears OK (U7, R18, R19, L2, D4, C15). R20 should limit the current. The voltage between D4 and R20 should be around 27V

The 6V8 TVS diode has been omitted. Smaller crystal. Configurable power-off time is back. There appears to be a little push button (?) in the bottom left labelled "OFF" for manual shutoff? Strange

[elecdonia]

I just received a LCR-TC2 from China...  I didn't like the first test, short-circuiting the three pins didn't launch the calibration, it directly measures the short-circuit resistance, from what I've read in this thread this may be because the resistors are out of spec Seeing the board, my fears have come true...

Your LCR-TC2 is likely to contain the LGT8F328P MCU chip. This is a Chinese “?clone?” of the Atmel ATmega328P MCU. However the pinout is different. Also the architecture is sufficiently different enough to require a rewrite of the existing TransistorTester firmware. Check out these other posts for more info:

https://www.eevblog.com/forum/testgear/$20-lcr-esr-transistor-checker-project/msg4316005/?topicseen#msg4316005

https://www.eevblog.com/forum/testgear/$20-lcr-esr-transistor-checker-project/msg4606612/?topicseen#msg4606612

As for the failure of your tester to enter the “self test” you should measure the resistance of the three 680 ohm resistors which connect from the MCU to test terminals 1, 2, and 3. These 3 resistors must be matched to each other within 0.1%. I fixed a similar issue by adding a parallel resistor to one of the 680 ohm resistors which was above its proper value by +1.8%

https://www.eevblog.com/forum/testgear/$20-lcr-esr-transistor-checker-project/msg4726172/?topicseen#msg4726172

[elecdonia]

Delivered my T7 with a clone Mega328, pcb similar to LCR-TC1(2) from the file TableClonesEN.pdf by indman@EEVblog
https://yadi.sk/d/yW8xa5NJgUo5z

The MCU chip in your LCR-T7 is the LGT8F328P made by a Chinese company known as “Logic Green.” It is somewhat similar to the Atmel ATmega328P, but it certainly isn’t a direct replacement. First of all the pinout is different. More important, its architecture and software are considerably different. Furthermore it doesn’t support the Atmel “ICSP” (aka “ISP”) programming interface.

Here are some related EEVBlog topics discussing the LGT8F328P:

https://www.eevblog.com/forum/microcontrollers/arduino-pro-mini-two-copies-have-different-sleep-currents/msg3184622/#msg3184622

https://www.eevblog.com/forum/projects/anyone-here-interested-in-the-logic-green-avrs-lgt8f328p/msg2717650/#msg2717650

https://www.eevblog.com/forum/microcontrollers/lgt8f328p-clone-of-atmega328-with-lots-of-extras-anybody-use-them/

[elecdonia]

Very disappointing, the mega328 was almost 1/4 the price of the tc1 tester and now not getting it to work any time soon!  I get it there is no firmware available for it yet, or the chip not available to buy!
Clone tester with a clone chip from a clone! 

Hope the mega328pb is still ok, to keep for the t4, just in case another charged cap comes around!

At least the manufacturer of this particular Transistor Tester was ethical enough to keep the original APT32F172K8T6 label on the MCU chip.

A couple of months ago I received a Transistor Tester where the APT32F172K8T6 device had been deliberately re-labeled as Atmel ATmega328P.   

[elecdonia]

Let's say we received a new Chinese T7 clone, a TQFP32 MCU without any marks @16MHz. Is there any way to make an educated guess on what knock-off it can be, LGT8F328P or other?

If there is a crystal mounted on the PC board next to the MCU chip:   Then MCU is Lucky Green LGT8F328P

If there is no crystal on the PC board:   Then MCU is APT32F172K8T6. Evidently this MCU has internal clock oscillator

Note: During the past several months I have seen and tested several recently manufactured  “LCR-x” Transistor testers.
None of these units with the MCU chip in a 32-pin TQFP package contain the Atmel ATmega328P MCU.
However, some of these 32-pin MCU chips did have realistic looking (counterfeit) ATMega328P  labels laser engraved on them.

I recently bought one LCR-x unit with a real ATmega324 in a 44-pin TQFP package. It cost about $9 USD more than the units containing imitation 32-pin MCU chips. To the best of my knowledge none of the low-cost Chinese MCU chips come in 44-pin packages (yet).  So if the MCU has 44-pins it is likely to be a real Atmel ATmega324.

[davidut5]

Yes, i guess 328p does not support infrared and maybe zenner test. That's why they use the 44 pin 324, but now they made their own ic. This was the cheapest lcr-1 tester i can get in my country, probably the other ones had the 324 atmel. Did not think it would be different.

[madires]

You can have the IR receiver/decoder (via the three probes) and the Zener check (usually PC3) with an ATmega328 without any problem.

[alex_D]

If  I had a custom PCB made using AY-AT design but altered the board substituting the ATMega328 with a ATmega1284p package, it should just work right?
assuming I compile the the firmware for the ATmega1284p and I would have 4 times the flash?

[madires]

Yep! But you could make some small changes to improve the display performance:
- move the rotary encoder to dedicated pins
- go for hardware SPI for the display

[elecdonia]

If  I had a custom PCB made using AY-AT design but altered the board substituting the ATMega328 with a ATmega1284p package, it should just work right?
assuming I compile the the firmware for the ATmega1284p and I would have 4 times the flash?

I've been thinking about the same thing. In fact there are ATmega644 and ATmega1284 devices sitting on my workbench right now. They are housed in 40-pin DIP packages.

I plan to scratch-build a Transistor Tester on a "perma-proto" board with a 40-pin MCU socket.

Another of my ideas is to design a separate analog section to interface with the 3 test pins. 
It will have a high-resolution A-D converter and dedicated low-resistance drivers for the test pins.
All signals to/from the MCU will then be logic level. This way it won't matter whether the MCU operates at 5V or 3.3V. 
Eventually I might make the test voltages/currents for the 3 test pins programmable: Ideally 0-100V and 0-2A.

[Aldo22]

To conclude, the fix is:
1) Cut the PC board foil leading to pin K and then connect 2K2 resistor from junction of D2 and C14 to pin K,

Thank you!
There isn't much space for a 1206 SMD, but it works.
This is my first attempt at SMD soldering. 

[Per Hansson]

A question for Madires:
I had a device with a seemingly working Rohm 1SR159-200 diode rectifying the output from a switch mode transformer but it provided unexpected low output voltage.
In the end I found that the diode had failed in such a way that it became a regular diode, not a ultra fast diode!
I first replaced it with a UF4007 and it got me much closer to the expected output voltage.
So I bought a even faster diode (Diotec ES1D) that should exceed the originals specifications on paper and I got the expected output voltage!
My question is if this type of testing (speed of the diode) could be implemented in the transistor tester?

[indman]

A question for Madires:
I had a device with a seemingly working Rohm 1SR159-200 diode rectifying the output from a switch mode transformer but it provided

And what test results do you see now with this diode on a transistor tester? What is the CpF? What are the results compared to a serviceable ES1D? 

[Per Hansson]

And what test results do you see now with this diode on a transistor tester? What is the CpF? What are the results compared to a serviceable ES1D? 

The bad Rohm 1SR159-200 with trr value from datasheet: <50ns (23ns typical)
Uf=570mV
9.3pF - 3.6pF @ 0 - 5V
Ir=0.18µA

A new Diotec ES1D with trr value from datasheet: <15ns
Uf=541mV
24pF - 15pF @ 0 - 5V

But to do a proper test I should desolder one of the other Rohm diodes from the unit and measure those, since the Diotec has better specs in the datasheet.
But first I want to hear what madires thinks too

[gipetto]

If  I had a custom PCB made using AY-AT design but altered the board substituting the ATMega328 with a ATmega1284p package, it should just work right?
assuming I compile the the firmware for the ATmega1284p and I would have 4 times the flash?

I recall years ago seeing one of the first avr transistor tester prototyped on a dip atmega32a. i think the page is only accessible on the web archive. however the support for that mcu seems to not be listed any more.
If you want any help in designing the pcb and keeping it open source i can offer it. I'd like it to have an isp header, something that is missing from most designs. If nothing else i can check for mistakes.

[madires]

My question is if this type of testing (speed of the diode) could be implemented in the transistor tester?

The ATmega's slow ADC plus some processing time would limit that measurement to about 30µs, I'd guess. The reverse recovery time of a classic 1N400x can be somewhere between 30 and 2µs (UF400x: 50/75ns). So it wouldn't help much.

[Yuriy_K]

The reverse recovery time of a classic 1N400x can be somewhere between 30 and 2µs (UF400x: 50/75ns). So it wouldn't help much.

There is no need to add anything, both authors clearly show the dependence of the diode capacitance on its speed.

[indman]

The reverse recovery time of a classic 1N400x can be somewhere between 30 and 2µs (UF400x: 50/75ns). So it wouldn't help much.

There is no need to add anything, both authors clearly show the dependence of the diode capacitance on its speed.

Then why is the lower capacitance 1SR159-200 diode not working properly?

[Yuriy_K]

Then why is the lower capacitance 1SR159-200 diode not working properly?

Because, Ir=0.18µA

[indman]

Then why is the lower capacitance 1SR159-200 diode not working properly?

Because, Ir=0.18µA

According to the datasheet 1SR159-200, the allowable Reverse Ir can reach 10µA

[Per Hansson]

My question is if this type of testing (speed of the diode) could be implemented in the transistor tester?

The ATmega's slow ADC plus some processing time would limit that measurement to about 30µs, I'd guess. The reverse recovery time of a classic 1N400x can be somewhere between 30 and 2µs (UF400x: 50/75ns). So it wouldn't help much.

Thank you for the explanation madires!

Yuriy_K: I desoldered a working Rohm 1SR159-200 diode now but unfortunately it measures almost identical to the bad one:
Uf=545mV
10pF - 3.5pF @ 0 - 5V
Ir=0.20µA

I thought it could be a nice addition to measure the reverse recovery speed (trr) since this was a pretty tricky failure mode, but if the ADC is too slow then that is not possible of course...

[indman]

Per Hansson,What operating voltage is present on this diode in the circuit?