$20 LCR ESR Transistor checker project

[Fuzzy Star]

Sorry, honestly I have never edited a Makefile, I would need your help, to activate
SamplingADC for quartz test what should I put in the Makefile?

Looks like you have a modified makefile. In the original makefile that section looks like this:

Code: [Select]

# Option WITH_XTAL enables additional Xtal / Ceramic resonator tests.
# Option can be set only together with SamplingADC=1 and OP_MHZ=16
#CFLAGS += -DWITH_XTAL
See reply #8784 of Madires above. Download and examine the pdf document and the makefile inside the mega328_GM328 folder. Also examine the readme file inside the m-firmware source. Probably you'll find all the answers and you can learn how to compile, it's not difficult. Then you can experiment with different configuratons and try them on your tester.

[Maniaxx]

Or is it better to buy an assembly kit, and then I can chose  resistors and caps with tighter specs   for more accurate measurements. Like this one: aliexpress.com/item/32706710335.html

I ordered this (12Eur). Shipped to Germany in 9 days. Has 16MHz quartz.

The sellers assembly instructions are on a chinese cloud storage service though that only allows low-quality preview for unregistered users. For assembling its ok but the schematics.pdf (kicad) resolution is too low. I couldn't manage to register there (they request mobile phone number).

https://www.aliexpress.com/item/32706710335.html
https://www.jianguoyun.com/p/Df9U95oQweDQCxjm0YcFIAA

Is someone already registered there or can hack it or whatever to get the high resolutions schematics.pdf?

Edit:
1) Is there an overview/gallery somewhere that shows all the alternative icons and font sizes?
2) How to re-flash the Atmega328P DIP without in-circuit ISP? Or in other words, any good advice how to remove the chip easily out of the socket again (without an official pull tool)? I'm planning to put some strings under the chip to carefully lift it up. I once done that on PLCC Bios chips (hotswap flashing).

[madires]

That's an AY-AT/GM328A. You'll find the schematic and more at https://disk.yandex.ru/d/yW8xa5NJgUo5z/M328Kit%2BTFT. We don't have a gallery of all the fonts and symbols. If you like to program the ATmega out of circuit you can use an universal programmer or an ISP programmer plus adapter board.

[vitiv]

Is someone already registered there or can hack it or whatever to get the high resolutions schematics.pdf?
(hotswap flashing).

You have a PM

[Maniaxx]

Please post it as attachment.
Schematic.pdf and pcb.pdf should be sufficient.

[Obelix2007]

Maybe the resolution will help you

[Maniaxx]

Thanks. Its not only about resolution though. Its a clone and who knows if they changed something. Having the "clone specific" source wouldn't be wrong. You never know.

Edit:
When i measure larger caps (100uF) the tester shuts off everytime during detection. It shuts off sometimes with 2.2uF and it works properly on very small caps. What's wrong here? I have enabled colors, symbols, esr and e-values.
Edit2:
Maybe Brown-Out @4.3V? Fuses were set by 'make fuses'.
$ avrdude -c usbasp -P usb -p m328p -B 20 -U lfuse:w:0xf7:m -U hfuse:w:0xd9:m -U efuse:w:0xfc:m

I will try a lower bod.
Edit3:
confirmed, needs lower bod.

m-firm 1.50
gm328a/ay-at 16mhz

[vitiv]

Schematic.pdf and pcb.pdf

[Maniaxx]

Thanks vitiv!

The fixed cap for self-adjustment needs an ADC pin and an I/O pin for the resistor. Therefore it's an option for an ATmega324/644/1284 based tester.

1) No TP_CAP in 'config_328.h'? Isn't the 324 the same as 328?

2) When i just write the flash rom ($ make prog_fw) to keep my 'self-adjustment data' (in EEPROM) i get a 'checksum error' on 'boot screen' and profiles are gone. I'm on #define DATA_FLASH. How can i keep the data after flashing?

[madires]

The ATmega324 has more I/O pins than the ATmega328. Keeping the old profile(s) isn't supported when flashing a changed/new firmware.

[siealex]

A small bug with the E24 resistor series in 1.50m. This series contains "43", in the firmware it's "42".

Open variables.h and find this declaration:
const uint16_t E24_table[NUM_E24] MEM_TYPE = {100, 110, 120, 130, 150, 160, 180, 200, 220, 240, 270, 300, 330, 360, 390, 420, 470, 510, 560, 620, 680, 750, 820, 910};

Change 420 to 430 in this line, save and rebuild.

[majortom]

To everyone who uses my firmware. You can correctly identify JFET pins by looking at smaller Cg= readings ... Sample measurement in the picture...

Hi Yuriy_K,
I have been having fun playing around compiling and testing the different k-firmware, m-firmware.
I have seen you posted some firmware attachments, with .hex, .eep, and a Makefile in a a few posts.
Where would I find your source repository? I'd like to compile it and try it out as well, on this GM328A that I have.
See if I can get your modifications going.

Thanks,
majortom

[Yuriy_K]

Hi majortom,

I do not publish my sources due to their strong difference from the last version of the Karl-Heinz sources. In the sources, the font and formulas for calculating inductance and capacitance have been changed, and many color control commands have been added. All the changes are too long to list. Everything that can be published, you can try suitable for MG328A.

[Obelix2007]

Hi majortom,

the open source idea is interpreted one-sidedly by some people!

Greeting Horst

[indman]

the open source idea is interpreted one-sidedly by some people!

Address these reproaches to Chinese friends who churn out these devices in thousands of pieces and sell them without paying any attention to open source idea!

[Maniaxx]

Anything wrong with this self-adjustment setup? It always shows 'Error!' after adjustment (still shows value screen though).
I just have 2 clamps so i've put some shorter solid core cable in TP2. Maybe TP2 cable needs to be same length/type as well?
Cable 1+3 are stranded wire.

Short cables (2x solid core cable) for self-adjustment do work.

[madires]

Please post the values displayed during the self-adjustment. My guess would be that the probe resistance is too high.

[indman]

Short cables (2x solid core cable) for self-adjustment do work.

Such short cables are desirable. Long and thin wires with high internal resistance as well as wires that introduce significant additional capacitance and inductance must not be used with this device!

[Maniaxx]

Long and thin wires with high internal resistance as well as wires that introduce significant additional capacitance and inductance must not be used with this device!

So, thick wires are better? I will try 1.5mm² and somewhat shorter then.

Please post the values displayed during the self-adjustment.

I've attached the self-adjustment procedure with the setup shown above.

Is there any documentation for the self-adjustment values beside the source code annotations in 'adjust.c'?

[madires]

At step A6 the capacitance between probes #2 and #3 is 113pF for the first run. The limits are noted in the README:

Code: [Select]

Limits:
- probe resistance   < 1.50 Ohms for two probes in series
- probe capacitance  < 100 pF
- IO pin's internal resistance in low mode (RiL)   < 25 Ohms
- IO pin's internal resistance in high mode (RiH)  < 29 Ohms

Have you measured a film cap (100nF - 3.3µF) three times before running the self-adjustment?

[Maniaxx]

Looks like its the PSU (H+H SN1000). I suspect the multi plug interface (its somewhat loose). The device itself shouldn't be that bad. It's a heavy trafo (not a small switching one). With battery everything's fine.

1) When i measure 680Ohm (0.1%) i get (quite stable) 677.5Ohm no matter how its adjusted or with/without cables. 470k (0.1%) shows 469.5k that should be ok i guess.
Can i shift the RL value? There is 'RH_OFFSET' in config.h but no 'RL_OFFSET'.

2) I have a 470uF 63V cap. When testing in normal mode it shows 471uF 0.38Ohm. In 'C-Monitor' it shows 453uF with a slightly higher ESR. Do they have different testing methods?

[horo]

Hi,

If you want to create your own firmware version, choosing the right compiler version can noticeably reduce the code size.
I recently did a comparison of different avr-gcc versions. I took all the versions available for Linux (Debian stable) and compiled a 1.50m (with an identical setup for my AY-AT with 20 MHz xtal).
These are the results:

Code: [Select]

avr-gcc v. 12.1.0
Program:   35988 bytes (109.8% Full)
Data:        269 bytes (13.1% Full)
EEPROM:       22 bytes (2.1% Full)
avr-gcc v. 11.1.0:
Program:   35252 bytes (107.6% Full)
Data:        269 bytes (13.1% Full)
EEPROM:       22 bytes (2.1% Full)
avr-gcc v. 10.1.0
Program:   35196 bytes (107.4% Full)
Data:        251 bytes (12.3% Full)
EEPROM:       22 bytes (2.1% Full)
avr-gcc v. 9.2.0
Program:   35190 bytes (107.4% Full)
Data:        251 bytes (12.3% Full)
EEPROM:       22 bytes (2.1% Full)
avr-gcc v. 9.1.0
Program:   35190 bytes (107.4% Full)
Data:        251 bytes (12.3% Full)
EEPROM:       22 bytes (2.1% Full)
avr-gcc v. 8.3.0
Program:   32546 bytes (99.3% Full)
Data:        251 bytes (12.3% Full)
EEPROM:       22 bytes (2.1% Full)
avr-gcc v. 8.2.0
Program:   32536 bytes (99.3% Full)
Data:        251 bytes (12.3% Full)
EEPROM:       22 bytes (2.1% Full)
avr-gcc v. 8.1.0
Program:   32536 bytes (99.3% Full)
Data:        251 bytes (12.3% Full)
EEPROM:       22 bytes (2.1% Full)
avr-gcc v. 7.3.0 (Arduino-Version)
Program:   32612 bytes (99.5% Full)
Data:        251 bytes (12.3% Full)
EEPROM:       22 bytes (2.1% Full)
avr-gcc v. 5.4.0 (Debian-Stable-Version)
Program:   32742 bytes (99.9% Full)
Data:        251 bytes (12.3% Full)
EEPROM:       22 bytes (2.1% Full)

Based on these results, I set 8.3.0 (the last of the major version 8 ) as default.
For later review, I enabled "link-time optimisation" with the compiler switch "-flto", which reduced the code size even further:

Code: [Select]

avr-gcc v. 8.3.0 (my default)
Program:   31770 bytes (97.0% Full)
Data:        244 bytes (11.9% Full)
EEPROM:       22 bytes (2.1% Full)

avr-gcc v. 7.3.0 (Arduino-Version)
Program:   31810 bytes (97.1% Full)
Data:        244 bytes (11.9% Full)
EEPROM:       22 bytes (2.1% Full)

avr-gcc v. 5.4.0 (Debian-Stable-Version)
Program:   32096 bytes (97.9% Full)
Data:        244 bytes (11.9% Full)
EEPROM:       22 bytes (2.1% Full)


Martin

[madires]

1) When i measure 680Ohm (0.1%) i get (quite stable) 677.5Ohm no matter how its adjusted or with/without cables. 470k (0.1%) shows 469.5k that should be ok i guess.
Can i shift the RL value? There is 'RH_OFFSET' in config.h but no 'RL_OFFSET'.

You could try to adjust R_LOW.

2) I have a 470uF 63V cap. When testing in normal mode it shows 471uF 0.38Ohm. In 'C-Monitor' it shows 453uF with a slightly higher ESR. Do they have different testing methods?

No, both use the same measurement functions. Different probes?

[Maniaxx]

Same probes. ZIF socket shows same behavior. Tried different power supplies and different caps. Testing resistors is ok.

'Default Test' vs. 'C-Monitor':
1063uF <> 1016uF
463uF <> 444uF
103uF <> 100uF
222nF <> 222nF

The first two measurements of 'C-Monitor' are higher then the following. I have 'hold' values enabled. Maybe the test cycles are too fast?
Not sure if relevant, i have MCP1702 installed with Vref and 2.2k still in place (but disabled in config.h).

Can someone confirm or falsify this behavior?

[indman]

Same probes. ZIF socket shows same behavior. Tried different power supplies and different caps.

Madires, I have the same effect on my clones, but I didn't pay much attention to it.
I made 2 short videos so you can better understand what I'm talking about. I measure the same capacitance 1000uFx10V on ZIF pins 1-3. In the RLC monitor mode, the tester first shows the same capacitance value as in normal testing, but then the result decreases jump. I don't know what is the reason for this behaviour?
https://drive.google.com/file/d/1DSnLMSiL0VmlU29ebRCczFxiez6BJ08C/view?usp=sharing
https://drive.google.com/file/d/1lwazwLodpxfv2d3wuUisQfvdH8jAHm17/view?usp=sharing