$20 LCR ESR Transistor checker project

[madires]

It might also be useful to tell him the settings for the fuse bits.

With avrdude and make installed: make fuses

Manually:
- l fuse: f7
- h fuse: d9
- e fuse: fc

[snapper]

What is the measuring range for coils and with what accuracy?

[madires]

The standard inductance measurement starts about 10 µH and goes theoretically beyond 10 H. An inductor with a large resistance can't be measured and high value inductors with an iron core cause issues. Accuracy should be about <= 5 % (please see Karl-Heinz's excellent documentation for the graph). The k-firmware's sampling ADC mesurement method can go down into the nH range, and the LC meter hardware option for the m-firmware starts about 1 nH.

[GarySlaghammer]

I blew up my ESR02 Pro some time ago with a charged capacitor, and have finally got around to trying to replace the MCU (with a 328PB). This was successful, though I went round in circles for a whole evening since the USBasp wouldn't power up the Atmega and recognise it for programming. After changing the MCU again with no result, I tried through AVRDUDESS with a PICKit 2 and it worked fine. I guess the USBasp was a dud.

I have zero experience of programming with these things. I have previously followed a kit to programme a PIC which is basically painting by numbers so I didn't have to understand anything.

I am pleased I've managed to try three different firmwares and they all work. I have set the fuses to what I hope are correct values for an 8MHz external crystal (L:F7, H:D9, E:FC).

However, I can't get the menu to work. Long presses don't do anything, and single or double presses don't do anything either. If I connect a component and then press a button it boots up (takes a while) and seems to identify the component OK, but subsequent button presses do nothing. I have to wait for it to turn off. There is no self-calibration.

Have other components been damaged on the board, or have I messed up the firmware part?

[snapper]

Hi, I programmed one of those things years ago. Unfortunately, I can't say whether a Pickit with an Atmel microcontroller does it correctly. I've attached my firmware collection. I assume you didn't make any mistakes while soldering. Otherwise, just get a new USBasp; they don't cost much! Best regards to the island ;-)

[madires]

However, I can't get the menu to work. Long presses don't do anything, and single or double presses don't do anything either. If I connect a component and then press a button it boots up (takes a while) and seems to identify the component OK, but subsequent button presses do nothing. I have to wait for it to turn off. There is no self-calibration.

Sounds like an issue with the test button, e.g. wrong port pin (should be PD7).

[GarySlaghammer]

All I've done is remove and replace the MCU. An opportunity to break out my hot air rework station (I do very little SMD). I even protected the surrounding area with kapton tape. The new MCUs were drag-soldered and I checked with the microscope that there were no bridged pins. AVRDUDESS has PICKit 2 in its list of programmers. I used the pinout from here: https://reversatronics.blogspot.com/2018/07/batteries-pickit2-and-atmega328-pro-mini.html and AVRDUDESS reports success. Reading back the fuse settings is OK, too, as is recognising the MCU type.

So, I haven't touched the buttons. They're on the other side of the board, and I'd have to desolder all the component sockets on the front panel to get to that. By PD7, do you mean I ought to check the buttons are linked to that pin, or that I should monitor what happens on PD7 when the button is pressed? It's having some effect, as pressing it with the unit off makes it turn on and initiate the first probe cycle.

[indman]

It's having some effect, as pressing it with the unit off makes it turn on and initiate the first probe cycle.

I would recommend that you make a short video so that we can see how your tester works. This is almost always much better than describing the problem only in words.
I don't think the problem is related to the Test button, because you used firmware that worked fine for other people.

[madires]

So, I haven't touched the buttons. They're on the other side of the board, and I'd have to desolder all the component sockets on the front panel to get to that. By PD7, do you mean I ought to check the buttons are linked to that pin, or that I should monitor what happens on PD7 when the button is pressed? It's having some effect, as pressing it with the unit off makes it turn on and initiate the first probe cycle.

Yes, please check the input signal at PD7. When the test button is pressed the signal should go low. In case you don't have the schematics: https://disk.yandex.ru/d/yW8xa5NJgUo5z/BSideESR02(DTU-1701)/Schematic%20diagram/DTU-1701_V3.1.JPG

[GarySlaghammer]

Thank you for the assistance.  Here is a video showing the non-responsive buttons (51MB): https://we.tl/t-Fm9V1nS3Vs

I often think that there should be a course on technical description in schools. Despite several post-graduate degrees, I always find describing a mechanical or electronic problem with only text is a huge test of skill and one I often fail at! Academic writing doesn't really do the same thing, and rarely with the brevity suitable for a forum anyway.

[indman]

Here is a video showing the non-responsive buttons (51MB): https://we.tl/t-Fm9V1nS3Vs

Unfortunately, this resource is not available in my country.
Can you upload your video to googledrive, for example?

[snapper]

https://www.filemail.com/d/cfajcpjekiqjitp

[indman]

https://www.filemail.com/d/cfajcpjekiqjitp

Ok,thanks.
Now the problem becomes obvious.
1. The tester shows that you have a weak battery. Replace the battery or use an external 9-12V stabilized power supply.
2. The tester has detected a phantom capacitance between pins 1-2! This should not be the case, and it indicates that there may be a problem with the tester's hardware.
What kind of flux did you use to solder the new controller? Show us a photo of the PCB?

[mnaevis]

Over the last couple of hours I read/watched a lot about these LCR testers and after going through a lot of Aliexpress listings, videos and pictures, it seems that all of the TC1, T7, ... you can get nowadays are bad counterfeits without the 328 chip (one listing claimed 328, but I read that those are just re-labeled ones, so even worse).
My main focus is capacitor ESR (preferably while in circuit) and I think someone here (madires?) wrote that this especially is worse in the fake ones.

I came to the conclusion that the only device where I could get a real 328 are the GM328 devices (see link below) where you get the raw pcb without any casing - preferably with a DIP as the Chinese do not counterfeit them (though someone here also wrote that the newer version also has a real 328). I'd also prefer this anyway as I dislike fixed LiIon batteries as I would use it very rarely.

Questions:
1) Am I wrong? :)
2) Can I power the device with a 9V PSU or will this lead to problems because the DC might not be perfect as with a battery?
3) Anyone know what the differences between the "old" and the "new" versions (https://aliexpress.com/item/1005003678038543.html) are - except the processor packaging of course?

Thanks

[madires]

1) Am I wrong?

No. But in your case a proper LCR meter could be the better solution, albeit more expensive.

2) Can I power the device with a 9V PSU or will this lead to problems because the DC might not be perfect as with a battery?

An external PSU is fine too. Just stay away from cheap SMPSU wall-warts with poor filtering.

3) Anyone know what the differences between the "old" and the "new" versions (https://aliexpress.com/item/1005003678038543.html) are - except the processor packaging of course?

The GM328A you've linked looks to be the classic version which is still a good choice.

[Per Hansson]

3) Anyone know what the differences between the "old" and the "new" versions (https://aliexpress.com/item/1005003678038543.html) are - except the processor packaging of course?

The GM328A you've linked looks to be the classic version which is still a good choice.

To me it looks like the "new" vs "old" in that page might be if the Atmega is a DIP or SMD version, I'd go for the DIP version.

[dibro]

I have this one: https://eleshop.eu/gm328a-component-tester.html
Has an 'old' through hole gm328a chip.

[Seekonk]

I lost my open board transistor tester in my recent move.  These worked great for the price.  I just got the FINRST LCR-P1 to replace it because it had a case, zener capability and rechargeable battery.  Thought this is rather old technology and that even they could make it without issue.   But this still has the same software issues every other product they make has, ship it and we will figure it out later.  It took less than 5 minutes to find that it would identify a center pin as an emitter and other transistors would read differently if they were rotated 180 degrees.  This has been a couple years in production.  I have ver 1.0.8 and I hear that ver 9 has the same issue. Download procedure hasn't worked for me. No response from SUPPORT.  I'm sending this back.  Now I feel snakebit.  Anything cute out there that works?

[indman]

Anything cute out there that works?

There is, undoubtedly! The sweetest thing is a device that is assembled by your own hands and in which a piece of your soul and heart is embedded!

[GarySlaghammer]

https://www.filemail.com/d/cfajcpjekiqjitp

Ok,thanks.
Now the problem becomes obvious.
1. The tester shows that you have a weak battery. Replace the battery or use an external 9-12V stabilized power supply.
2. The tester has detected a phantom capacitance between pins 1-2! This should not be the case, and it indicates that there may be a problem with the tester's hardware.
What kind of flux did you use to solder the new controller? Show us a photo of the PCB?

Thank you - I'm sorry that family issues meant I hadn't replied sooner. The video was taken using a bench PSU, so the battery reporting was awry too. I took your advice of looking for flux residue and cleaned the board again. This seems to have fixed the problem, so now I know that something so simple can have such a large effect! The new firmware is more effective than the one it shipped with. I guess some protection of the input would be beneficial so I don't blow up another chip...

[GarySlaghammer]

Actually, I tell a lie. For a period this evening it was working perfectly with 1.13k - one momentary button press to turn on and test the connected component, with another momentary press to start testing again. Now, whether using a battery or a stabilised PSU a single momentary press results in the screen flashing for the short period the button is pressed, without the unit booting. Holding it very slightly longer doesn't start up into Test mode but brings up the selection menu, which previously only came up when booting the unit with a long press. Earlier this evening I could cycle this menu with a momentary press, but now a press of a second or more is required to change the selection. To select an option requires an even longer press. After testing a transistor, I have to hold a button down long enough to return to the menu, before selecting 'transistor' again, which makes testing very long-winded. What have I done wrong?

[indman]

What have I done wrong?

Interesting question? In this topic there are no wizards or psychics who know the answers to all questions in advance.
I can assume that you still have problems with the hardware of the tester and cleaning the board did not completely eliminate them.

[madires]

Measuring an electrolytic cap, e.g. 1000 µF, could give an indication if the fuse settings for the MCU clock are correct.

[GarySlaghammer]

I don't expect a psychic, but you were able to quickly pinpoint the issue before so I have great faith in your distance-diagnosis capabilities! 

And...it seems you've done it again. I scrubbed the board once more with flux cleaner and dried with compressed air. So far, the problem has been eradicated. Despite it looking pristine after the last clean (before my original post I hadn't cleaned it at all, wanting to test in case I needed to remove the chip again. I'm not experienced with SMD and the mid-century PCBs I'm used to dealing with have no problems with flux residue!) it was obviously still not quite clean enough.

I'll do something nerdy like gain-match my transistors and see if the button responsiveness continues. Thank you.

[indman]

If the tester is working now, then my advice is this. Go to the menu and perform the calibration(selftest) and let us know the results of the T1-T7 hardware tests in text form.