LCR T7 tester problem

[Crystaleyes]

Hi all.

Several months ago, whilst testing a box of old used caps which I had 'assumed' were discharged, the T7 LCR tester I was using was damaged.
When the button was pressed, the display would briefly light up and then go dead.

So, I stuck the T7 in the freezer for ten minutes and 'Bingo!' it came back to life.

Only thing was that after it 'defrosted' at room temperature, the problem returned.

After freeze spraying components one at a time, I managed to pin the problem down to a SOT-8 chip. An STC 15L104W. Which until now, I had just been spraying each time I used the device.

Anyway, I decided to buy a new chip, to fix the problem, however this only makes the problem worse. The device screen merely flashes briefly and that is that!

Replacing the original, bad chip, just takes me back to square one.


My question is, "do these STC 15L104W's need to be programmed or flashed?"


Adding some photos to help see what is inside the tester.


Any suggestions will be greatly appreciated

[stj]

they need to be programmed - but we dont have the code.
i have the tools but not the code.

the fix is to replace it with a small 2 transistor circuit and then re-flash the atmel.
it will have much better battery life afterwards.

as a bonus, if your going to do this then swap the atmel for an atmega644 or 1284 and a 20MHz crystal

[stj]

hmm - looking at your pictures they changed the board.
you cant upgrade to a 644 or 1284

you can still do the crystal upgrade though.

is it still a colour display or did they fuck that too?

[Crystaleyes]

Easy fella. Still has a colour display.

What difference will the crystal upgrade make?

[Crystaleyes]

Originally Posted by stj

they need to be programmed -

That would make sense.
Does both the Atmel and the STC get programmed at the same time?

[Crystaleyes]

Originally Posted by stj

the fix is to replace it with a small 2 transistor circuit and then re-flash the atmel.
it will have much better battery life afterwards.

And this circuit is then squeezed into the casing?

[stj]

the atmel and stc talk to each other to control the power.

the circuit can be built onto the pcb pads - i'v done it, or on stripboard.
look here:
https://github.com/madires/Transistortester-Warehouse

[Crystaleyes]

So do you think the STC 15L is damaged and should be replaced?

Or just that it only works when cold due to the firmware being damaged?

As mentioned, if it is freeze sprayed then it keeps taking readings one after the other, until Ileave it without testing for half a minute or so.
Therefore it seems that it only needs to be chilled when 'beginning' the test readings

[Crystaleyes]

Hang on...

I might have been talking shit...


..Right next to the STC 15L is a small diode, D2. With 'S4' written on it. TBH I hadn't paid that any attention seeing as when I freeze sprayed the chip the device worked.
That said, I just did a few runs spraying 'only' the diode...

... And that made it work!

(where is the Face-Palm smiley?)

Let me rewind and focus on the diode before I think about flashing firmware and inserting circuitry

[stj]

that link should have the T1 schematics - it's very similar

[stj]

here

[lotas]

Schematic diagram

[Crystaleyes]

Originally Posted by stj

that link should have the T1 schematics - it's very similar

Excellent.

Very similar indeed.
My concern is that the schematic is for the design which has the transformer whereas my tester is the cheapo version with no trafo.

In the schematic, D2 is an SS18 whereas my suspect D2 is an S4 (which I believe equates to an SD103 amongst others, inc. the SS54? and 1N5819 in through hole form).

I'll grab a diode tomorrow from the shops and confirm one way or the other

[stj]

the diodes are just 1amp schottky,
the circuits are similar but yours uses inductors to boost the voltages instead of the transformer - a better way of doing it i think.

[Crystaleyes]

Right..
Even though the D2 diode tested fine after pulling it, I installed a 1N5819 1A Schottky to see if that made any difference, but it didn't.

I also pulled the Atmega238 and froze that separately, but that too made no difference.

Then I carefully froze the STC15L, being sure to avoid all other components, and the tester worked again.

So, the issue 100% lies with the STC15L104W.

Any suggestions as to how to resolve this?

Or just keep it for spares?

[stj]

replace the stc with the circuit i posted and re-flash the atmel with the software from madires at that link
if you cant build the atmel software then i can create it for you.
do you have tools to program AVR chips?

btw, some people say to re-flash the stc chip,
i tried that for a friend but the stc chips are shit and dont always go into standby - result is it drains the battery overnight!

these testers (T1 / T7) have a couple of known problems.
one is the stc chip fucking up, another is a couple of ceramic caps shorting in the boost circuit for the zener testing.

[lotas]

1. Getting ready to flash the U4 power controller. We solder the USB-UART converter with a stabilizer to the board.
(3.3V to 3.3V, Gnd to Gnd, Tx to P1, Rx to P2).
2. We connect the converter, wait for it to be detected by the system, launch the software, select the com-port, select the firmware from the archive (p. 1), set the frequency to 12 MHz, apply power, the processor starts, does not go into boot mode, is not detected by the programmer .
3. Experimenting with the sequence of power supply to the processor and pressing the "Check MCU" & "Download/Program" buttons.
4. We study the datasheet on U4, we find that the Test button of the tester is connected to the Reset output. Click "Check MCU", click the Test button, the microcontroller resets and is detected. In the same way, we launch the firmware, and finally, we flash U4.

[stj]

yes, but we are flashing firmware for the new atmel code,
i dont think it works with the original chinese code - not on T1 anyway.

[Crystaleyes]

Originally Posted by stj

do you have tools to program AVR chips?

Can that be done with an Arduino Uno?

[stj]

yes, atleast i'v seen people doing it.

personally i just use a USBasp programmer - why not when they cost about $2!!
saves a lot of messing around with patch wires and INO files.