$20 LCR ESR Transistor checker project

[madires]

Some more background info on the symbol sizes in the m-firmware At the moment there are two bitmap sets, 24x24 (horizontal & vertical bit orders) and 32x32 (horizontal only). As larger the symbols as larger the firmware. Based on the display controller you have to select a bitmap with a specific bit order. ILI9341 and ST7735 need horizontal order, and ST7565 is page-wise vertical. For ILI9341 and ST7735 you can change SYMBOL_RESIZE in the driver source file to resize the symbols by that factor. I've choosen the defaults to match the display size. For the ST7735 you can choose between SYMBOLS_24X24_H (default) and SYMBOLS_32X32_H in the config_328.h and change the sizing factor (default: 2) via SYMBOL_RESIZE in ST7735.c.

In case someone likes to draw some nicer symbols, I've written a tool to convert the bitmaps from and to different bit orders, including page-wise grouping.

[madires]

There are problems in case of measurement of the low-impedance resistors and inductance. I will try to show them in the following post.

Unusual result of measurement of inductivity 240 µH at a frequency of quartz of 20 MHz. Also uncertainly the low-impedance resistors are defined. An example on a photo.

Currently I'm looking into the inductance measurement issue and found some error in the control logic, causing the delayed check to be run instead of the high current one for a <1mH inductor. The resulting mismatch worsens with the MCU frequency. This problem was introduced with the improvement of the measurement resolution in 1.16m. I'll also fine-tune the offsets.

[indman]

Hi, madires! I well understand that measurement of inductivity is not the simple task. The result depends on the frequency of measurement, material of the core, diameter of a wire which wound the coil. I put some more photo, I hope that it will help you with the decision.

[madires]

A few posts ago flywheelz reported some funny ESR business with low value film caps and different MCU clock rates. I'm able to reproduce that and even have some extra fun. flywheelz measured a 220nF cap with 8 and 16 MHz clock rates and the 16MHz tester shows an ESR about twice the value of the 8MHz tester. For a complete desaster I've thrown in additionally a 20MHz tester which shows an ESR about 50% higher than the 8MHz tester. What I've also found is that the 16MHz tester can measure the ESR of low value film caps the 8MHz tester can't. For comparison I've checked all caps with a LCR meter and the ESR values are higher (3.2 @ 1kHz instead of 0.95 for example).

The measurement uses charge/discharge pulses while synchronizing them with the ADC. The 8 and 16MHz testers use the same ADC clock, while the 20MHz got a higher ADC clock, caused by the limited fixed prescalers. So the pulse length stays the same for 8 and 16 MHz clocks (besides the time for switching), but the 16MHz clock will create more harmonics when switching the pulse (a wink from Mr. Fourier). For the 20MHz clock the higher ADC clock seems to mitigate the problem. My guess is, that the switching harmonics of the MCU frequency in relation to the ADC clock creates the different ESR values for low value caps.

Edit: The 20MHz tester can also measure the ESR of low value caps the 8MHz tester can't.

[belzrebuth]

I have the EZM build with madires' firmware.
Since the unit is able to measure frequency,I'd like to try this.
I should connect the output of ,say, a function gen to the atmega pin pd4 (pin6) as stated in the manual or I could do that with the test leads?
Also what could be the maximum voltage I should apply ?


edit:I think I have to build this additional breakout circuit to measure frequency with my build..Just browsed thru the documentation and it's right there..

[madires]

Yep, that's the circuit for the simple frequency counter. I wouldn't recommend to go above 5V peak.

[madires]

A quick update on the m-firmware. I'm still investigating the inductance measurement. The inductance values from the low current measurement are too high, and the values from the high current measurement are too low. And there are variations across the different MCU clocks. Besides this I've done following:
- Changed FrequencyCounter() to support also ATmega 324/644/1284.
- Fixed error in voltage reference handling for ATmega 324/644/1284.
- Improved detection of turning velocity of rotary encoder to cope with different values of pulses per step or detent.
- Added hardware SPI to all drivers for SPI based displays.

When I'll have solved the inductance puzzle I'll release the new version.

[pepe10000]

Hi, I would like to know if you can use the IR-Sensor and IR-Decode function in an LCR-T4 like this:

https://www.eevblog.com/forum/testgear/$20-lcr-esr-transistor-checker-project/msg659867/#msg659867

Thanks and best regards.

[madires]

Yes, simply select the IR detector in the menu and connect the IR receiver module to the probe pins as explained in the README file (m-firmware). Of course, the firmware should be compiled with SW_IR_RECEIVER enabled in config.h. If your IR receiver module requires 5V also enable SW_IR_DISABLE_RESISTOR.

[pepe10000]

Thanks, that's why I could not find it since I use the trunk firmware.

It is possible to activate it for this trunk firmware ?.

A greeting.

[madires]

Sorry, the k-firmware doesn't support IR receiver modules (yet?).

[pepe10000]

Thanks, I have since yesterday investigating how to compile the Markus version to work correctly with this Chinese LCR-T4.

These are the things I have had to change to work decently with my LCT-T4 with 16MHz xtal and installed Pull-UP resistors:

In the Makefile:

FREQ = 16

In the config.h file:

//#define HW_ENCODER
//#define ENCODER_PULSES     2
//#define SW_ENCODER
#define CYCLE_DELAY            12000
#define CYCLE_MAX               1

In the config_328.h file

#define LCD_RESET            PD4            /* port pin used for /RES */
#define LCD_A0                 PD3            /* port pin used for A0 */
#define LCD_SI                  PD1            /* port pin used for SI (LCD's data input) */
//#define LCD_OFFSET_X                       /* enable x offset of 4 dots */
//#define LCD_FLIP_Y                            /* enable vertical flip */
#define LCD_CONTRAST     15              /* default contrast (0-63) */

But I have found more inconveniences than advantages:

The measurement of quartz crystals is lost.
The square wave generator is lost by not having the rotary encoder.
The Spanish language is not available.
To enter the menu, the test leads must be short-circuited.
When you press to check a component, make two measurements.

I would like to know how to modify the last two things, that you can enter the menu with the test button and only take one measurement at a time.

Thank you for everything and greetings.

[madires]

You can enter the menu also by pressing the button two times briefly. Please see the README file for the operation modes (continously/auto-hold). I think you're running the tester in continious mode, which doesn't require any button press for cycling. If you press the button, the tester skips the pause between two measurements. You could support the project by translating the text in variables.h into Spanish The squarewave generator of the m-firmware allows you to use the full frequency range, not just a few fixed frequencies from a list. That's the reason for the rotary encoder. Doing that with just the push button would be a nightmare. The "Sampling ADC" is already on my to-do list, but I can't give you any time frame. Since the tester clones are cheap, I'd suggest to get a second one, maybe that nice kit with the rotary encoder. So you can have one for each firmware

[pepe10000]

Actually I have two identical units of the tester, one operative and one damaged in the process of repair (soon I will receive the components to repair it).

Once I have the second unit repaired I will proceed to install the firmware m, so I will have the best of both firmwares.

And I will try to add a rotary encoder even if I have to modify the box where I have it installed.



Can you recommend a rotary encoder as small as possible to ask for it on Ebay?

Also I will look at the file variables.h to try to translate it into Spanish in case I can support the project.

I attach a Spanish translation of the file langSPANISH.h of the version k with some modifications in the texts in the measurement of crystals of quartz and in the option meter of frequency for that appears like "Not available".

A greeting.

[madires]

I'd recommend an ALPS EC11 series rotary encoder.

[stj]

PM'd

[pepe10000]

Hi, thanks to your advice I already bought the rotary encoder, but it will take me a month and a half to come from China.

Meanwhile I will continue to learn and prepare the other components that are needed for installation.

According to the manual I only need two resistors from 10K to Vcc and two other resistors from 1K to PD1 and PD3.

But in some schemes I have seen an additional 100nF capacitors.

For this Chinese Yellow LCR-T4 I need exactly, just the four resistors and the rotary encoder?

In case you are supporting the project I have done the Spanish translation of the firmware m but replacing it with the English in the file variables.h.

So if this file is used to compile with the English language by default the meter will be in Spanish.

Attached file for interested parties.

A greeting.

[madires]

Thanks for the Spanish translation! I'll add it to the next release. Basically you only need the four resistors, since the debouncing is done in software. The caps could cause trouble, because the rotary encoder is in parallel with the LCD display (ATmega 328 based testers).

[stj]

here is me fitting a rotary encoder inplace of the original push-button on a different tester.
it may give you an idea how to do it nicely. 

http://www.badcaps.net/forum/showthread.php?t=50720

[pepe10000]

They are very good ideas, I had more or less something in mind, although in my case I will have to make some modifications since the circuit tracks are somewhat different, and it is possible that I also have to make some hole, but the idea is the same .

Thanks and best regards.

[hapless]

here is me fitting a rotary encoder inplace of the original push-button on a different tester.
it may give you an idea how to do it nicely. 

http://www.badcaps.net/forum/showthread.php?t=50720

Is there a way to see the images for people who have not registered there? I, for one, would not want to go through the registration process just to look at pictures.

[stj]

no, sorry.
i dont have them all on the pc now to re-upload here.

[Seekonk]

Just found something interesting.  I have a unmodified tester, forget the taobao version no encoder.  Just tried testing some vintage 1N55 diodes.  It returns as a small pf capacitor on one of the unconnected leads. All three of the diodes I tested read OK .417 on a DVM.  Schottky diodes test fine at .206V as well as silicon.  Can't imagine why it can't test germanium.

[madires]

The old firmware versions have a detection problem of some Germanium diodes with very low Vf or a high leakage current. A freezer spray might help. The current versions got some improvements (1.12k trunk, >= 1.24m), especially the m-firmware.

[siri_]

Hi allI 

have one of those, and when I power it from a adjsutable power supply, it says "Bat. 10V OK". But i'm supllying it 9V, at least that's what's my multimeter (uni-t ut61E) says...

Is it something that could be a problem or it doesn't matter?

Thanks!