$20 LCR ESR Transistor checker project

[Spongey]

I haven't yet checked for any output from the USB port.

For that I just connected a terminal program to the virtual COM port the CH340N creates on a PC. I don't know what the serial port speed is though. A scope connected to the header near U1 could work that out.

[elecdonia]

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.

I have a recently manufactured LCR-TC1 clone containing two LM358 opamps. I will endeavor to post a schematic of my unit. This unit employs APT32F172K8T6 MCU. I have the MCU pinout so I can show which of its I/O pins are used for zener diode test when I draw the schematic.

[elecdonia]

I also ordered replacement U7 devices at the same time. I will be testing the revised circuit you recommend as soon as these arrive.

For me, your efforts to restore this clone from FNIRSI look ridiculous - it's like trying to make a chocolate candy out of shit!? FNIRSI has released a crude, low-quality product to the market, which does not correspond to the level of this project in any way. It is desirable to return such products to the seller, to demand a refund of the money spent for wasting time.

If it was from Amazon I would have returned it. But this was eBay purchase shipped direct from China. I have bought from same large Chinese eBay vendor many times with no problem. So I think I will keep it and use it for experimentation.

The build quality of its PC board is actually better than some of my older transistor tester units. The LCD display is sharp & bright. The plastic case is well designed and it looks great. Aside from failure of U7 in the 30V step-up circuit for zener diode test, this LCR-TC1 is reasonably functional. So far I observed it gives accurate results for resistors and for the uF of capacitors. I still need to make detailed tests for all other items which are identified and measured accurately by a traditional transistor tester.

RE: Zener test feature, I temporarily inserted a 10K series resistor at the output from the 30V step-up converter. This works OK while waiting for arrival of replacement U7 chips. Note: This 10K series resistor was in the very first schematics published for implementing zener diode test feature.

I think these are the main faults of transistor testers containing non-Atmel MCU:
   1) Firmware cannot be changed or upgraded
   2) Several features of the traditional transistor tester aren’t implemented
   3) 30V power source for zener testing does not have functional current limiting

[indman]

I have said before that we will get more unpleasant surprises if we test the clones now on the new CPUs in all modes.

I carefully study the customer reviews of clone models with APT32F172K8T6 and LGT8F328 chips and can add a few more differences that are characteristic of a tester based on these chips:
1.Incorrect measurement of ESR capacitors.For most serviceable electrolytic capacitors, the ESR is not defined at all or is designated as 0.00. For some types of ceramic SMD capacitors, the ESR, on the contrary, is overestimated by 10-100 times.
2.Some faulty capacitors are detected by the tester as 2 counter diodes.
3.Some BJT transistors (for example,2N5089,MPSA12,MPSA13, etc.) are defined as thyristors or resistors.
4.Some JFET transistors (e.g. J112,J113,2N5457, etc.) are not defined at all or are defined as BJT.
5. Diode assemblies with a common cathode are incorrectly determined.
I think that this is not a complete list of differences from the original project and it will be supplemented over time.

[elecdonia]

These new MCU chips APT32F172K8T6 and LGT8F328 are likely to have totally different characteristics for their I/O pins.

As I understand it (please correct me if I’m wrong) measurements of very small resistance require direct connection of the tested component between one MCU I/O pin outputting 5V and another MCU I/O pin outputting 0V. The residual internal output resistance of these I/O pins are the reference resistors for this measurement. Not the 680R or 470k probe resistors. Evidently all flavors of Atmel AVR MCU have similar and relatively stable internal output resistance for their I/O pins. This provides capability to share similar software among all Atmel AVR MCU varieties.

I discovered a very old post which states the output resistance of an Atmel AVR I/O pin is about 20R:
     $20-lcr-esr-transistor-checker-project/msg889515

However, other MCU are likely to have considerably different internal resistance for I/O pins set in output mode. For example some of these new MCU are said to have 80mA source or sink capability on their output pins. Their residual output resistance would be much lower than Atmel AVR.

One thing I find of interest is that both the APT32F172K8T6 and the LGT8F328 can operate with 5V power supply. Because of this their I/O pins can directly generate 5V output levels for “logic high.” Compare this to nearly every other modern MCU: Their I/O in output mode can produce only 3V3 maximum for the logic high output state. This could explain why there aren’t any transistor testers using STM32 or other similar MCU families. The capability to directly generate a 5V output from an I/O pin is essential to the traditional transistor tester architecture. Otherwise an additional driver circuit separate from the MCU itself is necessary.

I’m beginning to think about “output driver” circuits which are independent from the MCU itself. Would it be possible to design such a circuit to be both simple and precise?

[Feliciano]

As you say, the output current of these MCUs is limited, but also keep in mind we are often using small-battery-powered units, and most of the times we are talking about "chinese" regulators (and other components), so asking for pulses of 50mA or so can generate different kinds of troubles/errors (and quickly exhaust the battery).

Fortunately, the developers took that into consideration, and that's why the calibration procedure was introduced, which allows us to obtain a reasonable accuracy on the measurements, if we remember to calibrate the instrument every now and then, specially if you are alternating leads and ZIF measurements (for which the m-firmware offer the option to have different profiles).

[elecdonia]

As you say, the output current of these MCUs is limited, but also keep in mind we are often using small-battery-powered units.

The Atmel AVR MCU chips (ATmega328, ATmega324, ATmega644) share similar output current ratings from their digital I/O pins. But we know very little about the characteristics of these alternate MCU chips from the Chinese manufacturers. Some are said to offer 80mA output current. These new Chinese MCU chips are targeted at low-cost applications where BOM cost and PC board real estate is saved when the output pins can drive loads directly without needing additional components.

…and most of the times we are talking about "chinese" regulators (and other components), so asking for pulses of 50mA or so can generate different kinds of troubles/errors (and quickly exhaust the battery).

Yes, the stability of the 5.0V power supply rail is critically important for the transistor tester. This is why it is often recommended to install a higher grade LDO voltage regulator IC in place of the commonly used 78L05. Larger input and output capacitors near the voltage regulator are also recommended. I intend to make detailed measurements of the 5V rail with my oscilloscope to learn how large the variation is when the transistor tester is measuring small resistance, such as while testing capacitor ESR, and also during self-calibration.

Fortunately, the developers took that into consideration, and that's why the calibration procedure was introduced, which allows us to obtain a reasonable accuracy on the measurements, if we remember to calibrate the instrument every now and then, specially if you are alternating leads and ZIF measurements (for which the m-firmware offer the option to have different profiles).

Unfortunately the Chinese engineers who ported the traditional transistor tester software to these other (non-Atmel) MCU chips were not fully aware of these issues. They were on a tight budget for expense and development time. Therefore they released firmware which was only partially functional.

At this time my experience with non-ATmega328 transistor testers extends only to one LCR-TC1 unit containing the APT32F172K8T6 MCU. So far the only improved feature I’ve observed is that measurements finish much faster in comparison to traditional transistor testers with ATmega328 clocked at 8MHz. I estimate the speed of the LCR-TC1 to be 4 or 5 times faster.

[vklimk]

Small resistors measurement needs not more than 5v/720R=6,9mA additional current. If you want to load TTto investigate power supply stability then you should measure inductance - it takes up to 5v/38R=130mA additional current (for a short period of time of course). lgt8f328p takes even more since it has lower internal resistance of io-ports than atmega MCU has. And I know nothing about APT32F172K8T6, but inductance measurement will load it hard as well I believe.

[elecdonia]

Small resistors measurement needs not more than 5v/720R=6,9mA additional current.

If I understand this correctly: Small resistor measurement includes one 680R probe resistor in series with internal resistance of two MCU output ports (one I/O port at 5V, other I/O port at 0V). Total resistance of 720R is 680R + (2 * 20R).

If you want to load TT to investigate power supply stability then you should measure inductance - it takes up to 5v/38R=130mA additional current (for a short period of time of course).

For this case there is only the internal resistance of two MCU output ports in series. Again, one I/O port is at 5V and other at 0V. The unknown inductor is connected between these 2 I/O ports.

lgt8f328p takes even more since it has lower internal resistance of io-ports than atmega MCU has. And I know nothing about APT32F172K8T6, but inductance measurement will load it hard as well I believe.

At this time I will re-read all of the TT manuals and documentation. Then I will measure the parameters of the output ports of the APT32F172K8T6 in my LCR-TC1.

[vklimk]

I would suggest you to have a look into source code of TT, it contains very nice comments where measuring methods are described. And there is a document from Karl where he describes the methods as well.
And I am not sure how can you find internal resistance of mcu's ports if you do not have developer toolchain for APT32F172K8T6 to compile and run simple tests on it (you need pull io-port up or down, connect current source to the port, measure voltage drop on the port and calculate the resistance). We do not even know how that fnirsi bootloader works...

[indman]

If I understand this correctly: Small resistor measurement includes one 680R probe resistor in series with internal resistance of two MCU output ports (one I/O port at 5V, other I/O port at 0V). Total resistance of 720R is 680R + (2 * 20R).

Here you will find a very brief and practical explanation of how the tester works.
http://www.pa3fwm.nl/technotes/tn11b.html
All very clear and accessible.Very much I advise to acquaint all. This certainly does not replace the grand work done by the authors of the firmware to create detailed documentation.

[eti]

If I understand this correctly: Small resistor measurement includes one 680R probe resistor in series with internal resistance of two MCU output ports (one I/O port at 5V, other I/O port at 0V). Total resistance of 720R is 680R + (2 * 20R).

Here you will find a very brief and practical explanation of how the tester works.
http://www.pa3fwm.nl/technotes/tn11b.html
All very clear and accessible.Very much I advise to acquaint all. This certainly does not replace the grand work done by the authors of the project to create detailed documentation.

Also here, for those who prefer visual learning:

[indman]

Also here, for those who prefer visual learning:

Yes, this video is also very useful for those who want to understand in detail the principle of the tester! 

[eti]

Amazon UK sourced LCR-T4 product info update:

https://www.amazon.co.uk/gp/product/B07J2Q4VY9/ref=ppx_yo_dt_b_asin_title_o01_s00?ie=UTF8&psc=1

Note: In the event this listing is edited or removed, I have archived the Amazon page in the current (incorrect listing) state, for your records: https://web.archive.org/web/20221231203216/https://www.amazon.co.uk/gp/product/B07J2Q4VY9/ref=ppx_yo_dt_b_asin_title_o01_s00?ie=UTF8&psc=1

That archived URL seems to fail ONLY when clicked here, so I added it to a "tinyurl.com" URL:
https://tinyurl.com/amaz-B07J2Q4VY9-listing

This item arrived with the telltale signs of it being a Chinese counterfeited IC:

# No crystal
# USB-C socket
# CH-340N IC

Suffice to say I immediately went on CS chat, sent them these photos (attached and on the post immediately following this one) and they refunded without question, and allowed me to keep this product.

[eti]

Amazon UK sourced LCR-T4 product info update:

https://www.amazon.co.uk/gp/product/B07J2Q4VY9/ref=ppx_yo_dt_b_asin_title_o01_s00?ie=UTF8&psc=1

This item arrived with the telltale signs of it being a Chinese counterfeited IC:

# No crystal
# USB-C socket
# CH-340N IC

Suffice to say I immediately went on CS chat, sent them these photos (attached and on the post immediately following this one) and they refunded without question, and allowed me to keep this product.

Second photo (front) attached:

[indman]

This item arrived with the telltale signs of it being a Chinese counterfeited IC:

It's a lottery. Even when the images show a genuine ATmega it doesn't mean that you'll get one with your new tester.

Madires spelled it right.Unfortunately, you could not win this lottery, like many others who want to get an authentic product from Chinese comrades.

[eti]

This item arrived with the telltale signs of it being a Chinese counterfeited IC:

It's a lottery. Even when the images show a genuine ATmega it doesn't mean that you'll get one with your new tester.

Madires spelled it right.Unfortunately, you could not win this lottery, like many others who want to get an authentic product from Chinese comrades.

Due to the sheer size of this thread and the variety of opinions, sketchy sometimes, elsewhere online and here, would you please clarify:

# What is “wrong” with the LCR-T4 units with no crystal, a counterfeit “mega328p” and a USB-C socket

# What firmware they’re running

# Whether there is a way to mod them and flash in a special version of the original project  firmware

?

Thank you. As much detail as possible please. Happy 2023! Blessings.

[indman]

"# What is “wrong” with the LCR-T4 units with no crystal, a counterfeit “mega328p” and a USB-C socket"

eti,the answers to your questions are on this page in a message from elecdonia
https://www.eevblog.com/forum/testgear/$20-lcr-esr-transistor-checker-project/msg4610050/#msg4610050
and my message below, in which I listed the detected flaws of these clones with fake chips
https://www.eevblog.com/forum/testgear/$20-lcr-esr-transistor-checker-project/msg4610314/#msg4610314

"# What firmware they’re running"
This is a proprietary firmware from the clone manufacturer.

"# Whether there is a way to mod them and flash in a special version of the original project  firmware"
There is no such possibility yet!

[eti]

"# What is “wrong” with the LCR-T4 units with no crystal, a counterfeit “mega328p” and a USB-C socket"

eti,the answers to your questions are on this page in a message from elecdonia
https://www.eevblog.com/forum/testgear/$20-lcr-esr-transistor-checker-project/msg4610050/#msg4610050
and my message below, in which I listed the detected flaws of these clones with fake chips
https://www.eevblog.com/forum/testgear/$20-lcr-esr-transistor-checker-project/msg4610314/#msg4610314

"# What firmware they’re running"
This is a proprietary firmware from the clone manufacturer.

"# Whether there is a way to mod them and flash in a special version of the original project  firmware"
There is no such possibility yet!

I appreciate your rapid help. Okay so it’s as good as useless, got it. I’ll keep it in my “might use one day” pile. 😁

[madires]

# What firmware they’re running

# Whether there is a way to mod them and flash in a special version of the original project  firmware

They use a modified k-firmware with a more fancy UI, but had to remove features to free up flash memory for the fancy UI. At the moment Karl-Heinz and I have no plans to support any other MCU. We have to wait and see what will happen after chipageddon is over, i.e. if ATmegas will be inexpensive and easy to source again, or not. However, a user is working on a port of the m-firmware for the Logic Green MCU.

[py-bb]

I haven't yet checked for any output from the USB port.

For that I just connected a terminal program to the virtual COM port the CH340N creates on a PC. I don't know what the serial port speed is though. A scope connected to the header near U1 could work that out.

When I opened my T7 up it had like 1 IC. if there is a USB->UART in there it is integrated into the main IC - in which case all you'll see is USB traffic. You wont see the baud rate from this.

That said 9600 is a good guess - do you know anything it's supposed to respond to?

"screen /dev/ttyUSB0" tends to set the baud rate to 9600 but it will do lots of non-trivial stuff for you (TTYs are a big topic) and set it up for you, this tends to abstract things like \r expected for newlines, but if you are stuck I find "echo -n -e 'WHATEVER\r' > /dev/ttyUSB0" works quite well (This is a Bashism)

 

[eti]

# What firmware they’re running

# Whether there is a way to mod them and flash in a special version of the original project  firmware

They use a modified k-firmware with a more fancy UI, but had to remove features to free up flash memory for the fancy UI. At the moment Karl-Heinz and I have no plans to support any other MCU. We have to wait and see what will happen after chipageddon is over, i.e. if ATmegas will be inexpensive and easy to source again, or not. However, a user is working on a port of the m-firmware for the Logic Green MCU.

Thank you. Who is "a user" btw?

[Spongey]

When I opened my T7 up it had like 1 IC. if there is a USB->UART in there it is integrated into the main IC - in which case all you'll see is USB traffic. You wont see the baud rate from this.

That said 9600 is a good guess - do you know anything it's supposed to respond to?

Some of the fakes have a CH340N chip connected to the MCU which look like a UART from their pin description. The same pins go to a 6 pin header nearby: presumably for flashing the device. A scope on there should reveal the baud rate.

I've tried a range of speeds including 9600baud and got no consistent output and nothing like readable ASCII. Maybe it's just binary data in a proprietary format?

I've no idea what data should be sent on there: presumably only FNIRSI know. I was hoping it would output something like a version number or the value of whatever it was measuring and maybe some usable commands, nope.

[madires]

Thank you. Who is "a user" btw?

A member of this forum.

[elecdonia]

Amazon UK sourced LCR-T4 product info update:
This item arrived with the telltale signs of it being a Chinese counterfeited IC:
     # No crystal
     # USB-C socket
     # CH-340N IC
Suffice to say I immediately went on CS chat, sent them these photos (attached and on the post immediately following this one) and they refunded without question, and allowed me to keep this product.

The MCU (U1) in the photo of this LCR-T4 PC board is NOT the Atmel ATmega328P.
I received an LCR-TC1 a few weeks ago which contains an identically (mis)labeled MCU. Evidently someone in China created a totally fake Atmel label and is applying it to this other MCU chip: APT32F172K8T6. The only thing in common between the ATmega328P and the APT32F172K8T6 is they are both housed in 32-pin TQFP packages.
(Personally I find deliberate mislabeling of electronic components to be extremely annoying and insulting. It is bad enough when the original label has been removed with sandpaper )