$20 LCR ESR Transistor checker project

[madires]

When I turn the device on, I get nothing but seemingly random numbers scrolling quickly across the top of the screen, and the device is unresponsive. The display is very high-contrast but you should be able to make out the numbers in the attached pic to see what I'm talking about.

Have you programmed the EEPROM too?

[madires]

I tried to program the m software too but there are some issues.
It doesnt want a capacitor to calibrate ony want to short probes.
The SELFTEST and ADJUSTMENT in the menu works the same - they run the calibration.
Capacitor measurement doesn,t work as it should, sometimes doesn,t detect, sometimes shows only capacitance and very rarely ESR.

Please read the 'Self Adjustment' section in the README file. The self-test is different from the self-adjustment.

[fudmuffin]

When I turn the device on, I get nothing but seemingly random numbers scrolling quickly across the top of the screen, and the device is unresponsive.

That's new to me. Please post photos of your board without LCD.

M328 is inserted correctly?

Perhaps something to do with the rotary encoder. Try disconnecting it.

The M328 (ATMEGA328P-PU) is connected correctly. I have re-programmed 6-7 times now between the r523 firmware mentioned earlier (which works) and the latest one from the repository (which I have the trouble with).

I have now tried compiling the latest firmware myself both with and without the "DWITH_ROTARY_SWITCH" and get the same result as the precompiled from the repository (scrolling numbers along top of screen).

I tried the 1.12k_r523 firmware again to check: The rotary encoder works with the 1.12k_r523 firmware (you can see the r523 firmware running in the attached picture).

Photos are attached of front and back of board, as well as another one where the scrolling numbers (that I am getting with the latest firmware) are easier to see.

Have you programmed the EEPROM too?

If you mean the .eep file to 'data memory', then yes. I am also setting the fuses as per the attached picture (which I sourced from elsewhere in the thread).

I guess unless there are other suggestions then tomorrow I will try removing the rotary encoder completely, and programming using avrdude instead of the TL866A - but I'm not confident either will make a difference...

Also - how do I check which revision/version number is the latest in the repository?

Thanks for your ideas!

[doktori]

Hi this recently threw me for a curve.  Using the TL866 you must load the .epp file as Intel HEX format to the data memory.  Loaded as a BIN format the display will only show a scrolling line at the top of the display for the M12864 board.

[eemes]

I bought a kit for AX (based on AtMega328P with color display 160x128 rotating encoder):
https://www.aliexpress.com/item/2016-DIY-KITS-ATMEAG328P-M328-Transistor-Tester-LCR-Diode-Capacitance-ESR-meter-PWM-Square-wave-Signal/32785578867.html

I replaced these  parts.
Actaully 16Mhz quarz instead 20MHz.

I managed to run both FWs (for 16MHz I have to recompile both with the right options). But I have some questions:
- what does 16MHz quartz bring? Faster execution? Is there any other (side) effect, like precision?
- since I replaced 2.5 voltage reference (and also 5V Vcc). I wonder if I have to compile the FWs with some other options.
 In particular M-firmware. Should I enable/uncomment: HW_REF25 (and also set UREF_25 to the proper value (which one)?

Are there some other options that I should enable/disable after replacing the above mentioned parts?

BTW: big thx to both developer of the tester, and really really good update (!) documentation!!! 

PS: I tried the idea that I mentioned  here (to build and run both FWs together) I've managed to get the flash, which is, unfortunately, ~64kB big so I could not try on my AtMega328P  .  It has sens as each FW has already ~32kB.

[madires]

With 16MHz the display output is faster, especially for color displays, and time related measurements (capacitance, inductance) are a little bit more precise while supporting lower limits. If you want to use an external voltage reference with the m-firmware please enable HW_REF25 and set UREF_25 to the typical value stated in the reference's datasheet (or measure the voltage with a 6.5 digit or better DMM). No other changes are required besides changing the MCU clock rate in the Makefile.

[Seekonk]

I had an interesting transistor the other day that was shorted.  It indicated as two diodes.

[mebel]

I tried to program the m software too but there are some issues.
It doesnt want a capacitor to calibrate ony want to short probes.
The SELFTEST and ADJUSTMENT in the menu works the same - they run the calibration.
Capacitor measurement doesn,t work as it should, sometimes doesn,t detect, sometimes shows only capacitance and very rarely ESR.

Please read the 'Self Adjustment' section in the README file. The self-test is different from the self-adjustment.

I have read that and now I understant the difference but still the tester does not show esr for capacitors bigger than 470 micro.
Usuaally 470 is  (not always) ok but 1000 micro and higher not.
Sometimes shows  0,00 ohm but usualy shows only capacitance and leakage.

Ri -20,4ohm
Ri+22.8ohm
C0 41pF
R0 0,14ohm
vref 1060mV
Vcc 5001mV
Acomp -47mv

[eemes]

With 16MHz the display output is faster, especially for color displays, and time related measurements (capacitance, inductance) are a little bit more precise while supporting lower limits. If you want to use an external voltage reference with the m-firmware please enable HW_REF25 and set UREF_25 to the typical value stated in the reference's datasheet (or measure the voltage with a 6.5 digit or better DMM). No other changes are required besides changing the MCU clock rate in the Makefile.

So 20MHz would be even better in terms of performance and precision?
Is it better (=more precise measurements) to use external voltage reference (or to stay with Vcc in case if both are quite accurate)?
For which component is the value of UREF_25 in the config.h
is 2500mv typical voltage for LM4040AIZ-2.5?

[madires]

When the tester doesn't display the ESR for electrolytic caps then the ESR measurement failed for some reason. For example one reason would be if the measured ESR (including R0) is lower than R0 (the resistance of the probe leads/pins), usually caused by contact problems.

[madires]

So 20MHz would be even better in terms of performance and precision?
Is it better (=more precise measurements) to use external voltage reference (or to stay with Vcc in case if both are quite accurate)?
For which component is the value of UREF_25 in the config.h
is 2500mv typical voltage for LM4040AIZ-2.5?

Not so much for precision but for display speed in case of a high resolution color display like ILI9341. An external voltage reference helps if its tolerance is about 10 times better than the tolerance of the voltage regulator. IIRC, the LM4040AIZ-2.5's typical voltage is 2.495V.

[mebel]

When the tester doesn't display the ESR for electrolytic caps then the ESR measurement failed for some reason. For example one reason would be if the measured ESR (including R0) is lower than R0 (the resistance of the probe leads/pins), usually caused by contact problems.

Maybe, but why it works on k firmware?

[madires]

I don't know. I can't test my firmware with all the different versions and variants of the testers available. It's simply impossible.

[eemes]

An external voltage reference helps if its tolerance is about 10 times better than the tolerance of the voltage regulator. IIRC, the MCP1702-5002E's typical voltage is 2.495V.

So in case of using MCP1702-5002E regulator for Vcc (tolerance 0.4%), and heaving M4040AIZ-2.5  (0.1%) as 2.5V external reference it is not big sense to use the external reference? Or it still has sense if the external reference is available?

Edited: replaced M4040AIZ-2.5 with MCP1702-5002E (copy-paste probelm)

[madires]

The MCP1702-5002 is a 5V LDO and not a 2.5V voltage reference. If you're using the MCP1702-5002 an additional 2.5V reference isn't recommended.

[dave356]

Hi guys,
I have a T4 that will not run the self-test
I want to try to reprogram it,
Is the 6pin header area to the left of the chip able to be used for programming if I install a pin header?
Thought I noticed somewhere that it was used for programming after installing pins.

Also, what is a T4 with "strip-grid"
I noticed T4 with different traces on back called this.

[fudmuffin]

Hi this recently threw me for a curve.  Using the TL866 you must load the .epp file as Intel HEX format to the data memory.  Loaded as a BIN format the display will only show a scrolling line at the top of the display for the M12864 board.

Thank you very much. This was exactly what I was doing wrong with the TL866. Loading the .eep as INTEL HEX format did the trick, and the latest firmware is now working. I guess my mistake was not apparent when loading the older firmware as there was no vital .eep data in it. Thanks to all and I look forward to continue learning from this thread

[mebel]

I don't know. I can't test my firmware with all the different versions and variants of the testers available. It's simply impossible.

I wrote that on K software my tester works but thats not true, after some time it displays message "uncalibrated"

I found this information in documentation of k software:
At this place I will give you an additional important hint. Never do a measurement with connected ISP plug!
The ISP interface influences the measurement.

Could it be that the programming cables cause the issues with your software too? By the cables I mean about 4cm long from microcontroller to 6 pin isp plug. The cables are lying on the pcb.

[madires]

I have a T4 that will not run the self-test
I want to try to reprogram it,
Is the 6pin header area to the left of the chip able to be used for programming if I install a pin header?
Thought I noticed somewhere that it was used for programming after installing pins.

Also, what is a T4 with "strip-grid"
I noticed T4 with different traces on back called this.

Yep, it's the ISP. I have no idea about a 'T4 strip-grid' but Karl-Heinz uses the term 'strip grid' for a tester built on a veroboard. It has a different pin assignment to keep the wiring simple.

[madires]

I wrote that on K software my tester works but thats not true, after some time it displays message "uncalibrated"

I found this information in documentation of k software:
At this place I will give you an additional important hint. Never do a measurement with connected ISP plug!
The ISP interface influences the measurement.

Could it be that the programming cables cause the issues with your software too? By the cables I mean about 4cm long from microcontroller to 6 pin isp plug. The cables are lying on the pcb.

The reason is that three pins of the MCU are shared between ISP and probe/test resistors. The length of the traces (and wires in your case) should be kept as short as possible, and the layout should keep the traces away from noisy stuff. I don't think a few centimeters are a problem. More likely the wires are picking up some signal from another trace/wire. A few mV can already screw up the ESR measurement.

[mebel]

I wrote that on K software my tester works but thats not true, after some time it displays message "uncalibrated"

I found this information in documentation of k software:
At this place I will give you an additional important hint. Never do a measurement with connected ISP plug!
The ISP interface influences the measurement.

Could it be that the programming cables cause the issues with your software too? By the cables I mean about 4cm long from microcontroller to 6 pin isp plug. The cables are lying on the pcb.

The reason is that three pins of the MCU are shared between ISP and probe/test resistors. The length of the traces (and wires in your case) should be kept as short as possible, and the layout should keep the traces away from noisy stuff. I don't think a few centimeters are a problem. More likely the wires are picking up some signal from another trace/wire. A few mV can already screw up the ESR measurement.

Exactly that was the problem. The wires were under the ZIF socket and cause all the problems. I unsoldered them and problem disappeared.
Thank you very much  for help madires.

[eemes]

I want to power my Component/LCR tester (see here) with a (permanent) 9V battery. So I wonder what's power  consumption (=current).
I could measure ~30mA when running, and less than 0.2mA at stand-by (waiting for the button to be pressed). So I believe that I do not need a switch to completely disconnect the battery when I would not need the tester. Anyway, the best test will be seeing how long the battery will last.

Has someone on you already measured the consumed current?
From the current consumption in  stand-by I would say that AtMega328P is in deep sleep.

[madires]

The current is about 20nA for the power-off state (standard power circuitry, MCU switched off). Sleep modes are used while the tester is running, e.g. for delays like waiting for user feedback.

[texaspyro]

The "Fish..." boards had a design defect.  The battery test voltage divider was across the power input and not after the power switch transistor.  It would drain the battery after a few weeks.

[JonnieCache]

Is this all the stuff I need to start programming the chip???

Could I ask you for the links to these, please?