TC1 lcr-meter transistor-tester fix

[mtcaus]

There is one more area you could check for your issue before you determine its ATmega.  Check your testing sockets for any possible "shorts". I once bought a ESR02 PRO and it's working "fine". However once in a while after testing a few components, I would get calibration screen or show pin 1(3) and pin 3(2), (random pins) shows a cell, resistor or capacitor just like yours, but there is none on the testing sockets. Upon close inspection, I saw a solder sticking out from 1 test socket to another.  They weren't touched each other, but enough to cause weird problems. I removed it and the problem solved. It's a manufacturing defect.

[masster]

I didn't find any short between input pins.
Instead, I took another thermal image (see attached) and there is a hotspot on filter capacitor C7. Fingers crossed. By the way, what value have the filter capacitors? 10 uF and 0.1 uF ?
The output voltage of first step up converter is 6.67V (correct according to resistor divider (0.6 x (1+100k/10k)) but lower than 2 Volts over stabilized voltage of 5V.
The output voltage of the second step up converter is 22.6V, again correct.

[TurboTom]

No, the hot spot is not C7. The optical overlay over the thermal image isn't accurate at these close distances, the hot component is the Atmel. Better disable optical overlay when working at close distaces and use common sense... 

[masster]

What optical overlay are you talking about? There is no such thing. The image on the left was taken with a regular phone camera. The image on the right came directly out of the FLIR device, and it is not in overlay mode.
I have cropped the left photo to look similar to the framing of the FLIR photo. Moreover, what common sense are you talking about? Condemning the ATMega chip by default? I'm looking for a short circuit on VCC bus. What is wrong with a capacitor being shorted? What do you know about the precision and resolution of the FLIR device I'm using? Nothing.

[TurboTom]

Sorry I didn't mean to sound rude or to offend you. But if you look at the colour grading of the heat distribution, you can clearly see the "X" shape of the chip carrier inside the Atmel. And with "common sense", I mean, that at a maximum temperature of sixtysomething degrees, just use your finger for a quick check. I'm absolutely sure that in this case, it's the Atmel that's getting hot. And the thermal image surely looks as if there is a visual overlay. The thermal image itself is quite out of focus, so the sharp structures visible in there cannot be part of the thermogram. I hope this makes sense...

P.S. FYI, I attached a true in-focus close-up thermogram of my (old style) TC1 PCB after maybe three consecutive testing cycles. You can clearly see the metal chip carrier inside the AtMega's enclosure. But you can also see the (reflective metal) terminals that don't emit heat. Obviously, temperature difference of this intact unit is much lower than you found on yours. If you look carefully, you see two of the tiny SMD resistors get slightly "warmish" -- the high side of the zener voltage divider and the shunt of the voltage reference --  and that the copper tracks below the silkscreen dissipate some of this heat. Thermal camera is a "pimped" Flir E4 with an additional close-range ZnSe lens with a focal length of 75mm IIRC. Without this optics, close-range thermographs aren't possible with this entry-level thermal imager.

[Per Hansson]

What optical overlay are you talking about? There is no such thing. The image on the left was taken with a regular phone camera. The image on the right came directly out of the FLIR device, and it is not in overlay mode.
I have cropped the left photo to look similar to the framing of the FLIR photo. Moreover, what common sense are you talking about? Condemning the ATMega chip by default? I'm looking for a short circuit on VCC bus. What is wrong with a capacitor being shorted? What do you know about the precision and resolution of the FLIR device I'm using? Nothing.

Please look at the thermal picture of the crystal, and where it is in the photograph from the Flir:

[masster]

After I removed capacitor C7, the current consumption dropped to 57 mA when testing (from 170 mA). No more overheating on the board.
The tester looks functional now. I have to replace the faulty 78L05 chip, the SRV05-4, and the filtering capacitors I removed.
Thank you for your replies.

[TurboTom]

Yes. Of course. Good luck with your repair!

[shved]

Been there, done that. I've also had excessive power draw. Double check the caps, esp. C6 & C7. Also check for corrosive flux. Give a quick read to the rest of the thread.

[masster]

Quote from: TurboTom on Today at 12:53:48 am>

Yes. Of course. Good luck with your repair!

Despite your useless sarcasm, the device works. Learn to accept you were wrong. Good bye.

[TurboTom]

@masster:
I'm not questioning your repair success. But I would really like to understand the undisputable findings that in your thermal image, the microcontroller heated up while this cannot be clearly told for the capacitor. The thermal part of your image is too far out-of-focus to tell.

I assume you took the images with a FLIR ONE or a similar device which has the thermal and the optical cameras installed side-by-side, so at close distance you've got a parallax error. The wide angle optics will result in this problem to vanish at higher distance. Moreover, there may be some active correction for this parallax error built into the thermal imager's software. FLIR's more recent cameras offer an optical overlay mode that they call "MSX" which is enabled by default. And that's contributing the apparent in-focus structures in your thermogram. Try disableing MSX display mode and experiment a little with your TIC.

If the replacement of the apparently shorted capacitor C7 (and maybe the 5V linear voltage regulator) brought your TC1 back to life as you reported, and the microcontroller no longer heats up, the only explanation for this is that the designer of this PCB made a bad error: He may not have connected all the supply lines of the ATMEGA to Vcc individually and have the node where C7 is attached floating. This means, C7 will buffer this supply line but power will flow over the metallization of the ATMEL's die. And if C7 shorts this supply, this metallization layer can well become a heating element. It would be interesting to check that on the current TC1 version's PCB... As I reported before, my TC1 is of the previous design with the single voltage converter and the transformer.

Depending on the ruggedness of the ATMEGA die and the current limit of the voltage regulator, this may or may not cause permanent damage to the microcontroller.

That's IMO the only reasonable explanation that a shorted capacitor C7 may result in the microcontroller to heat up.

[MarkMLl]

I'm just about to lob one of these into the spares bin, I've spent more time on it than it's worth. However, a couple of comments if I may:

The PCB is as shown in Kamlakkannan's #78 with silkscreened "LCR-TC" and "XR-112" markings, and an unpopulated 8-pin outline next to the ISP header position.

C11, 12, and 13 appear to be in roughly the same place and have the same function on all boards discussed (all appear good in my case).

C6 and in particular C7 appear to vary. On my (and most) boards they're associated with U4 (8-pin chip probably STC) while on the board discussed by Masster C7 is associated with the Atmega's pin 38 (Vcc).

I agree with the points made by various people about the parallax error between the optical and IR optics on IR cameras. I'm surprised that there isn't an offset table, either on the side of the camera or at least in the manual.

MarkMLl

[PcarSBA]

Hi,

I blew up my Atmega324 based TC-T7-H transistor tester by connecting a 470uF capacitor charged to 5V to pin 1 and 3    It took out the TVS diode array. With it removed the Atmega324 comes on again but I get the dreaded 4.1V cell voltage between 1 and 3. Pin 3 has 56Ohm to Vcc. And that resistance remains with it lifted off the PCB when measuring the pin directly to Vcc.

So I conclude that I need to replace the Atmega. Can someone help me decide what version of main firmware I should or can install onto the new Atmega324. I read through a lot but am still not clear as to whether I also need to change firmware on the STC15W104 when I pick one of the versions of Github?

Thanks for any help

[abdulbadii]

original C is non polarized C isn't it, how can be replaced by polarized one?

[madires]

So I conclude that I need to replace the Atmega. Can someone help me decide what version of main firmware I should or can install onto the new Atmega324. I read through a lot but am still not clear as to whether I also need to change firmware on the STC15W104 when I pick one of the versions of Github?

If you want to run one of the two OSHW firmwares you need also to reprogram the STC15W104 or replace it with a simple 2-transistor circuit.

[madires]

original C is non polarized C isn't it, how can be replaced by polarized one?

Which cap do you mean?

[Siriano53]

Salve a tutti. Grazie per le vostre discusion sui possibili guasti del tester tc1.Il mio tc1 si è risvegliato sostituendo il 78L05 e zener l'SRV05.

[gianb_diazm]

Hi everybody,

I'm also the proud owner of a malfuctioning T7 component tester , but it's not suffering from the two typical failures described here. My tester is stuck in the Zener testing mode and won't budge from there. I was testing some Zener diodes, and having forgotten that you're supposed to use the KAA section of the ZIF connector, I was using the 123 section . I realised my mistake and started moving the wires around when I saw it: the little spark of death. Long story short, I derped out and gave the 123 testing probes a good ZenerZap™, and now it thinks everything is a Zener.

The tester is this one exactly: https://a.co/d/3DXImdQ

The PCB looks exactly like this one: https://vrtp.ru/uploads/post-39-1674996347.jpg , it says C5.1.3_2022/5/25

I have tried desoldering different components from the board to try and break its Zênerie (Zener rêverie lol), to no avail. It turns on, charges the battery, decodes IR and even speaks serial to the Arduino IDE at 500000 baud.

Any ideas?

[madires]

Most likely the ATmega's probe pins are zapped, and the ATmega needs to be replaced.

[indman]

madires, if it has exactly the same circuit board as he indicated in the link, then this is a Chinese processor, for which there is no working firmware yet.
gianb_diazm, remove the U7 chip from the board and turn on the tester again. What will be on the display?

[gianb_diazm]

Hello again and thanks in advance,

Even though the boards are identical, mine has a real ATmega 328P, instead of that U1 chip on the picture I linked to. I didn't catch the difference until after indman pointed it out.
Also, I removed U7 and it's still showing Zener. I have removed before, out of desperation more than anything, L1, the R6 and R7 voltage divider, U6, D1, R16, D2, basically anything less than 2 traces away from the KAA probes. In all cases, it shows the Zener test and a random 3.4ish voltage.

I forgot to mention, whenever I short all the 123 probes, the tester goes into selftest mode no problem, so there might be some hope left for the device.
I wouldn't mind losing the Zener functionality as long as the rest of the tester works, as I mostly use it for MOSFETs and the like.

[indman]

Even though the boards are identical, mine has a real ATmega 328P, instead of that U1 chip on the picture I linked to. I didn't catch the difference until after indman pointed it out.

Unfortunately, you are mistaken. If you have the same board as I showed in the photo below, then this is not a real ATMEL, but a Chinese fake. In fact, the APT32F172 processor is installed, it’s just a fake marking on the case. This has already been proven 100% and many buyers of similar clones have already successfully won the dispute and returned the wasted money. This Chinese controller is damaged on your board with 99.9% certainty, so return to life this clone is unrealistic.

[gianb_diazm]

Yes, it's exactly that board, down to the markings on all the chips. A very convincing forgery!

Well, such is life. I guess I'll have to replace the microcontroller then. 

It's a real shame that these MCUs are not reprogrammable, because I could've reused the chip for a digital pins-only project.

Thanks a lot for the help and the patience! 

[elecdonia]

Even though the boards are identical, mine has a real ATmega 328P

Unfortunately, you are mistaken. If you have the same board as I showed in the photo below, then this is not a real ATMEL, but a Chinese fake. In fact, the APT32F172 processor is installed, it’s just a fake marking on it

Two dead give-aways for identifying LCR-Txx units with APT32F172 MCU:
     1)  MCU has 32 pins
     2)  There is no crystal anywhere on the PC board

Other details often found in LCR-Txx units with APT32F172MCU:
          -  A pair of LMV358 dual op-amps at locations U9, U10 (I'm still in the process of learning exactly why these opamps are used)
          -  D2 cathode connected directly to test socket "K" pin for zener diode test. This circuit is guaranteed to fail. Before using the zener test feature the trace between D2 and pin "K" must be cut and a series resistor installed. 2.2k ohms worked well for me.

This is important! DO NOT test zeners before making this series resistor modification!

[elecdonia]

Perhaps someone with PC board layout talent will design a small adaptor PC board for replacing the APT32F172 with either ATmega328P or ATmega324?
Although the pinouts are considerably different this still might be feasible.