$20 LCR ESR Transistor checker project

[Yuriy_K]

I followed the instructions
Now it always gives me this.

Measure the input resistances A-1, A-2, A-3, COM of the multimeter should be connected to A.
For example, the readings of VC9808+ are 10.23 - 10.25 M\$\Omega\$, UT61E - 9.2 - 9.3 M\$\Omega\$.

[proditaki]

proditaki, I can see the jumper now. But I also see that you have a lot of flux residue and dirt on the board.Remove them and also very carefully clean the ZIF panel contacts.If the problem does not disappear, we will think further how to solve it.

Thanks. I cleaned it but that made no difference at all.

[proditaki]

I followed the instructions
Now it always gives me this.

Measure the input resistances A-1, A-2, A-3, COM of the multimeter should be connected to A.
For example, the readings of VC9808+ are 10.23 - 10.25 M\$\Omega\$, UT61E - 9.2 - 9.3 M\$\Omega\$.

1 and 2 is 83 K\$\Omega\$ 3 is 4.2\$\Omega\$

[Yuriy_K]

1 and 2 is 83 K\$\Omega\$ 3 is 4.2\$\Omega\$

You need to start by looking for a low value at input 3 using the diagram on the previous page.

[indman]

1 and 2 is 83 K\$\Omega\$ 3 is 4.2\$\Omega\$

Was the ATmega328 controller that you soldered new or had it already been installed somewhere?
I would remove this chip from the board again and check the resistance of measuring ports 1-3 outside the PCB.

[proditaki]

1 and 2 is 83 K\$\Omega\$ 3 is 4.2\$\Omega\$

Was the ATmega328 controller that you soldered new or had it already been installed somewhere?
I would remove this chip from the board again and check the resistance of measuring ports 1-3 outside the PCB.

the Atmega328p came from a arduino nano.

[Pygmalion]

​I have finally found the time to start the "Transistor Checker" project. I did some research and decided to go with the m-firmware as it is still under development. If you don't mind, I'd like to ask you a few questions now.

First, to use a new firmware, I believe you need to go through the config.h, enable individual options and then compile and upload the firmware via the SPI protocol using the Arduino IDE. Is that correct?

Second, I see in this PDF document https://github.com/madires/Transistortester-Warehouse/blob/master/Documentation/English/Clone-Comparison-Chart.pdf that some of the Chinese clones can be programmed. However, according to the document, the latest allowed firmware is 1.48m, while the latest available firmware is 1.55m. Why this discrepancy? Is it because at the time the document was written, the latest available firmware was 1.48m?

Third, I would prefer to build my own transistor checker because I want to have full control over my device. At first glance, this doesn't make sense from a financial perspective because building my own device will cost much, much more than Chinese clones. However, if I start from DevKit-644 https://github.com/madires/Transistortester-Warehouse/blob/master/Hardware/DevKit-644.kicad.tgz, simplify it and use displays and other electronic components I already own, it might work.  My idea is to

• use a widely available 5V USB charger as the only power source and
• use a cheap boost converter to power the Zener test.

Not only does this halfs the necessary electronic components, it also reduces the PCB dimensions under 10cm, which lowers its order price considerably. I estimate that such a transistor checker should be no more than three times as expensive as a Chinese clone (not counting the electronic components I already own). Would you agree with this estimate?

Fourth, I still need to understand a few things:

• Why is the circuit around TEST_BUTTON so complicated? Wouldn't a push button with a pull-up resistor be sufficient?
• What is the advantage of the optional automatic switch-off of POWER_CTRL? Is it so dangerous to simply switch off the power?
• Why is TEST_BUTTON connected to the power supply for the Zener test? Would it be enough to replace what's left of C19 with a 15V boost converter?
• What is the meaning of +5P, to me it seems to be the same as +5V?

I hope you can help me with these questions, and if I decide to do my own design, I will report the result in this forum.

[madires]

​First, to use a new firmware, I believe you need to go through the config.h, enable individual options and then compile and upload the firmware via the SPI protocol using the Arduino IDE. Is that correct?

Yep, config.h, config_<MCU>.h and Makefile.

Second, I see in this PDF document https://github.com/madires/Transistortester-Warehouse/blob/master/Documentation/English/Clone-Comparison-Chart.pdf that some of the Chinese clones can be programmed. However, according to the document, the latest allowed firmware is 1.48m, while the latest available firmware is 1.55m. Why this discrepancy? Is it because at the time the document was written, the latest available firmware was 1.48m?

Yes, the chart is older than the latest firmware. 1.55m will work fine too,

Fourth, I still need to understand a few things:

• Why is the circuit around TEST_BUTTON so complicated? Wouldn't a push button with a pull-up resistor be sufficient?
• What is the advantage of the optional automatic switch-off of POWER_CTRL? Is it so dangerous to simply switch off the power?
• Why is TEST_BUTTON connected to the power supply for the Zener test? Would it be enough to replace what's left of C19 with a 15V boost converter?

The test button is also a soft power switch and the automatic power-off is meant to prolong battery lifetime. If you don't need that you can go for a simple power switch and just a test button with pull-up resistor. There are several options for controlling the boost converter. It can be switched by the test button (original design), run all the time (ZENER_UNSWITCHED) or be switched by a dedicated MCU pin (ZENER_SWITCHED).

[indman]

My idea is to

• use a widely available 5V USB charger as the only power source and
• use a cheap boost converter to power the Zener test.

This is a bad idea. This project, despite the visible simplicity of the design, requires a very high -quality stabilized nutrition of MCU.
Any savings in this matter will only worsen the characteristics and stability of the device.

[Pygmalion]

My idea is to

• use a widely available 5V USB charger as the only power source and
• use a cheap boost converter to power the Zener test.

This is a bad idea. This project, despite the visible simplicity of the design, requires a very high -quality stabilized nutrition of MCU.
Any savings in this matter will only worsen the characteristics and stability of the device.

OK, then it's probably better if I just buy an 5V powered upgradeable Chinese clone. For my hobby use, I don't need high precision and it makes sense to keep the cost down. I just kind of miss building my own device.

[Per Hansson]

My idea is to

• use a widely available 5V USB charger as the only power source and
• use a cheap boost converter to power the Zener test.

This is a bad idea. This project, despite the visible simplicity of the design, requires a very high -quality stabilized nutrition of MCU.
Any savings in this matter will only worsen the characteristics and stability of the device.

OK, then it's probably better if I just buy an 5V powered upgradeable Chinese clone. For my hobby use, I don't need high precision and it makes sense to keep the cost down. I just kind of miss building my own device.

Well, to be honest here none of the Chinese clones have a even half-way decent 5v supply so there's that side of it too

[RoGeorge]

Well, it is not that it wouldn't work at all from a 5V wall adapter.  Just that it would be less precise.

If you enjoy making devices for your hobby, then go build your own. 
If you'll need more precision, you can always upgrade it later.

[Pygmalion]

OK, then it's probably better if I just buy an 5V powered upgradeable Chinese clone. For my hobby use, I don't need high precision and it makes sense to keep the cost down. I just kind of miss building my own device.

Well, to be honest here none of the Chinese clones have a even half-way decent 5v supply so there's that side of it too

Well, all those Chinese clones, even the ones with 5V voltage input, have a linear regulator, so it seems that my design would be even worse.

If you don't mind that I am not an electronics engineer, but an hobbyist, I will try to explain my intention. For someone like me, the DevKit-664 is really a complex design; it has three different PNP transistors, two different MOSFETS, and many other more exotic electronic components. Also, I don't want to use a battery at all and, if possible, only wildly available 5V power sources.

I thought I could replace BC640 with 2N3906 (with updated resistors), LM4040 with TL431B (still 0.5% tolerance) and a MEA1D0515SC with a MT3608 module (up to 28V), all of which are very common and all of which I already have on hand without compromising precision too much. And since I don't want to use a battery and don't need a soft turn-on, I could get rid of most of the power supply circuitry, including four more exotic transistors. Of course, I could keep a linear regulator (which if possible supplies both +5V and +5P lines) and two related capacitors, but that means I will always need a power source of more than 7V sticking around. I don't understand how removing linear regulator affects precision, considering that the 2.5V reference is already there (which voltage I am going to determine with lab-grade voltmeter).

So if my modifications result in the device being less accurate than the Chinese clone, then I seriously wonder what the point is.  Your thoughts, which of these simplifications do you think are reasonable and which are not?

[Pygmalion]

I'm sorry, my question was perhaps too specific. So I have investigated, delved into the code and can confirm that if the external 2.5V reference is present, the input on AVCC (reference voltage powered by either 5V USB power supply or MCP1702-500) is checked against this external 2.5V reference.

Now if you use TL431B with larger tolerance instead of LM4040 with smaller tolerance, but measure your TL431B with a lab-grade voltmeter with 0.001% accuracy and put measured value in the code, the practical advantage of LM4040 is lost.

Of course, the exact output of the typical 5V USB power supply is less predictable than that of the MCP1702-500, but since this voltage is checked by the code anyway, there should be no disadvantage.

Still, the question remains, how stable is the output of 5V USB power supply? I would assume that a professionally built 5V USB power supply rated at 2A should be more stable than the MCP1702-500 rated at 250mA, well at least against current draw changes.

As professionals, would you agree or disagree with the above reasoning?

[madires]

With an MCP1702-500 there's no need for an external 2.5V reference. The reference makes only sense if it's an order of magnitude better than the voltage regulator. The switching noise of a wall wart can cause measurement issues (-> add an additional LC filter).

[indman]

Still, the question remains, how stable is the output of 5V USB power supply? I would assume that a professionally built 5V USB power supply rated at 2A should be more stable than the MCP1702-500 rated at 250mA, well at least against current draw changes.

No, you are mistaken. USB Power will never be more stable nutrition than power from a constant power source, for example, 7-9V battery and a separate MCP1702 type stabilizer. The value of the current, which can be a source here is not a decisive factor. Once again, I repeat my experience gained on the example of a large number of tester clones - in this project it must be, pure, high-quality, stabilized meals +5V for MCU!
Any additional filters, power supply up schemes only complicate the project and are not mandatory.

[Pygmalion]

No, you are mistaken. USB Power will never be more stable nutrition than power from a constant power source, for example, 7-9V battery and a separate MCP1702 type stabilizer. The value of the current, which can be a source here is not a decisive factor. Once again, I repeat my experience gained on the example of a large number of tester clones - in this project it must be, pure, high-quality, stabilized meals +5V for MCU!
Any additional filters, power supply up schemes only complicate the project and are not mandatory.

Oh, I see you are the author of the comparison PDF file. I also looked at the circuits you provided and they are extremely helpful.

First, I see that all USB powered clones have boosters and then linear regulators - they are never powered directly from the 5V supply. So now I'm confused. Is there actually a clone with a direct external 5V supply that you have tested?

Second, I see that all the clones are quite sophisticated, (almost) as complicated as the DevKit-644.  This suggests that building a device myself is a really dumb idea, especially if I am not an expert. (In other words, these Chinese designers probably know better than I do.)

Third, when you write that the firmware can be upgraded, does that mean that all SPI pins are exposed somewhere?

[indman]

Is there actually a clone with a direct external 5V supply that you have tested?
Third, when you write that the firmware can be upgraded, does that mean that all SPI pins are exposed somewhere?

Here is a simple basic circuit that can be powered directly from a +5V source. This circuit can be assembled even by a novice amateur radio operator, it is not difficult to understand and manufacture. If you want to build this project for yourself, start learning from this simple circuit. Contacts for programming are also marked on the circuit diagram in the ISP connector.

[Pygmalion]

Is there actually a clone with a direct external 5V supply that you have tested?
Third, when you write that the firmware can be upgraded, does that mean that all SPI pins are exposed somewhere?

Here is a simple basic circuit that can be powered directly from a +5V source. This circuit can be assembled even by a novice amateur radio operator, it is not difficult to understand and manufacture. If you want to build this project for yourself, start learning from this simple circuit. Contacts for programming are also marked on the circuit diagram in the ISP connector.

Thank you for the circuit.

Even though I'm not an expert, I'm not so inexperienced that I wouldn't know how to send a low pulse with a push button and a pull-up resistor or how to add a linear regulator to the circuit! I could have drawn this circuit myself.  I actually completely understand all of Markus' design, including the cool current-limiting Zener part, except for the soft power-up and auto turn-off part.  But I want to avoid the latter part anyway.

To get back my question, let me clarify it further: Have you ever tested a device that was powered directly from a 5V USB power source and used a 2.495V reference (without a booster and a linear regulator in between)? What problems have you encountered exactly?

This information would be very helpful for me to know if it makes sense to continue with my project or abandon it and simply buy a Chinese clone.

[indman]

To get back my question, let me clarify it further: Have you ever tested a device that was powered directly from a 5V USB power source and used a 2.495V reference (without a booster and a linear regulator in between)? What problems have you encountered exactly?

Of course, I tested this type of food, otherwise I would not have informed you about my experience.
The +5V USB power supply is usually poorly stabilized (the switching power supply of the computer, which may include a complex set of circuitry PWM switching noise and interference). Using an external REF 2.5V smooths out these factors a bit, but not completely. Measurements that are more time-consuming, such as measuring the parameters of capacitors, will be particularly sensitive.
You will not be able to get reliable repeatable measurement results, they will float a lot. This applies not only to capacity, but also to secondary parameters. When measuring transistors, parasitic reverse currents may appear in the results, which will not be present with clean and stable power. I have given only a few examples from practice, but there may be many more.The Karl-Heinz manual also mentions this.

[Pygmalion]

Since my USB-powered DevKit 644 project is apparently worse than cheap Chinese clones that all use linear regulators, even the USB-powered ones, the only sensible way is to cancel my project and buy one of those clones.

I apologise to everyone for my distraction and thank you for your patience.

[indman]

As it has been reported many times before now the safest option is to buy either a GM328(DIP ATmega328) assembly kit or a similar ready to work variant. You can always upgrade it, it's easy to program. If you want, you can make an adapter to use the ATmega644. Other options for Chinese clones today involve playing the lottery, and your chances of winning are minimal.

[Pygmalion]

As it has been reported many times before now the safest option is to buy either a GM328(DIP ATmega328) assembly kit or a similar ready to work variant. You can always upgrade it, it's easy to program. If you want, you can make an adapter to use the ATmega644. Other options for Chinese clones today involve playing the lottery, and your chances of winning are minimal.

Does "similar ready to work variant" apply to this one? https://www.aliexpress.com/item/1005004914474349.html

[indman]

I don't have your link open. Just need to check with the seller to make sure the clone uses DIP MCU.
Usually the seller will provide several variations of the same tester,look closely at the pictures.

[Pygmalion]

I don't have your link open. Just need to check with the seller to make sure the clone uses DIP MCU.
Usually the seller will provide several variations of the same tester,look closely at the pictures.

No, it is not DIP MCU.  I don't understand why it should it be DIP MCU?