$20 LCR ESR Transistor checker project

[Feliciano]

No joy with the suggested modifications. I post here the relevant sections:

Code: [Select]

*
 *  SPI bus
 *  - might be required by some hardware
 *  - could be enabled already in display section (config_<MCU>.h)
 *  - for bit-bang SPI port and pins see SPI_PORT (config_<MCU>.h)
 *  - hardware SPI uses automatically the proper MCU pins
 *  - uncomment either SPI_BITBANG or SPI_HARDWARE to enable
 */

//#define SPI_BITBANG                /* bit-bang SPI */
//#define SPI_HARDWARE               /* hardware SPI */
//#define SPI_RW                     /* enable SPI read support */
//#define SPI_SLOWDOWN               /* slow down bit-bang SPI */


Code: [Select]

/*
 *  ST7735
 *  - 4 wire SPI interface using hardware SPI
 */

//#if 0
#define LCD_ST7735                      /* display controller ST7735 */
#define LCD_GRAPHIC                     /* graphic display */
#define LCD_COLOR                       /* color display */
#define LCD_SPI                         /* SPI interface */
/* control and data lines */
#define LCD_PORT         PORTB          /* port data register */
#define LCD_DDR          DDRB           /* port data direction register */
#define LCD_RES          PB4            /* port pin used for /RESX (optional) */
//#define LCD_CS           PB?            /* port pin used for /CSX (optional) */
#define LCD_DC           PB3            /* port pin used for D/CX */
#define LCD_SCL          PB7            /* port pin used for SCL */
#define LCD_SDA          PB5            /* port pin used for SDA */
/* display settings */
#define LCD_DOTS_X       128            /* number of horizontal dots */
#define LCD_DOTS_Y       160            /* number of vertical dots */
#define LCD_OFFSET_X     2               /* enable x offset of 2 or 4 dots */
#define LCD_OFFSET_Y     1               /* enable y offset of 1 or 2 dots */
#define LCD_FLIP_X                      /* enable horizontal flip */
//#define LCD_FLIP_Y                      /* enable vertical flip */
#define LCD_ROTATE                      /* switch X and Y (rotate by 90°) */
//#define LCD_BGR                         /* reverse red and blue color channels */
#define LCD_LATE_ON                     /* turn on LCD after clearing it */
/* font and symbols: horizontally aligned & flipped */
#define FONT_10X16_HF                   /* 10x16 font */
//#define FONT_6X8_ISO8859_2_HF           /* 6x8 Central European font */
//#define FONT_8X8_ISO8859_2_HF           /* 8x8 Central European font */
//#define FONT_8X12T_ISO8859_2_HF         /* thin 8x12 Central European font */
//#define FONT_8X16_ISO8859_2_HF          /* 8x16 Central European font */
//#define FONT_10X16_ISO8859_2_HF         /* 10x16 Central European font */
//#define FONT_8X16_WIN1251_HF            /* 8x16 cyrillic font */
//#define FONT_8X16ALT_WIN1251_HF         /* 8x16 alternative cyrillic font */
#define SYMBOLS_30X32_HF                /* 30x32 symbols */
//#define SYMBOLS_30X32_ALT1_HF           /* 30x32 alternative symbols #1 */
//#define SYMBOLS_30X32_ALT2_HF           /* 30x32 alternative symbols #2 */
//#define SYMBOLS_32X32_HF                /* 32x32 symbols */
//#define SYMBOLS_32X32_ALT1_HF           /* 32x32 alternative symbols #1 */
//#define SYMBOLS_32X32_ALT2_HF           /* 32x32 alternative symbols #2 */
/* SPI bus */
#define SPI_HARDWARE                    /* hardware SPI */
//#endif
And as I said, it makes no difference to me to include or not the

Code: [Select]

#include <avr/io.h>
on SPI.c

I'm using WinAVR2010 with the replaced dll and toolchain over W10.

[indman]

On such a small display 1.8 we should not notice any difference in speed, or are there any other important factors that indicate the usefulness of hardware SPI?

A slightly smaller firmware size, as Feliciano already pointed out. And the most visible impact would be a faster screen clearing.

To my surprise, the size of the test firmware with SPI BitBung turned out to be smaller than SPI Hardware?!!! 
Feliciano, check both test firmware on your clone and compare their behavior. Unless, of course, I made a mistake somewhere with the settings?

[Feliciano]

indman, are you messing with me? 

• The bitbang_SPI version boots, in continous mode, in Russian, rotated (from top to bottom), and flipped horizontally
• The hardware_SPI version gives me a white screen. Regardless, did you modify something else compared to the configurations I shared earlier?
• It also amazes me that the hardware subroutine takes more memory than the software subrutine. I'll try to confirm/deny if I find how to fix the compiler error

[EDIT] I found the error, it was the toolchain of my WinAVR. I can confirm that hardware SPI should be faster, but takes a little more memory on the m-firmware (33358B vs 33396B in my case). (Interesting, I got less memory usage with the other toolchain, but errors when trying to enable hardware SPI).

[indman]

indman, are you messing with me? 

No, I didn’t even think about it! I have my own method for setting up configuration files, sorry.
I don’t like the 10x16 font at all, it looks awkward and it significantly increases the size of the firmware.
For ATMega328 and ATMega324 this is critical, so for clones with these controllers I only use the 8x16 font.
It’s strange that your firmware launched in Russian, because I chose English in the settings, this font supports the Latin alphabet.
Try 1 more firmware option from bitbung. Now I applied a 10x16 font and rotated the display orientation.

[Feliciano]

(I know you weren't messing with me, that was the wink for).
The new binaries work: correct orientation, and in continous mode. A callibration was needed out of the box, though. After that, it shows a 23V zener, because this unit only generates that much voltage. With DUT in place, it seems to work fine.

And about the bitmaps, we know bigger fonts means more memory, but for older eyes like mines is a trade-off. I haven't rendered the 8x16. By any chance do you have a side by side comparison against the 10x16 (small screen)? Is the 10x16 still the one I graphed on the attached?

I also noticed that more symbols were added to the bitmaps on recent versions of m-firmware, like battery icon, onewire, and maybe other stuff. I wonder whether those symbols are loaded into memory regardless whether we choose to display it or not? (I mean, whether the whole bitmap is loaded, or just the part of it we're declaring it's going to be used).

[indman]

Feliciano, this was a test firmware, the rest of the nuances and fine settings must be chosen by the user himself. This doesn’t interest me much anymore, since I don’t have such a clone and I hope there won’t be one. The main thing is that we managed to adapt m-firmware in a relatively simple way. Say a big thank you again to madires, who provided very flexible port settings in the configuration files.

[madires]

To my surprise, the size of the test firmware with SPI BitBung turned out to be smaller than SPI Hardware?!!! 

That's intriguing!

[Feliciano]

Indeed.

I publically express my gratitude again to madires, for keeping this project going for so long, and keeping improving it on his free time, and to the other people here that help, like indman of course, and others (like 2hry in this case). And because I appreciate that, I also try to collaborate a little bit every now and then.

There are some question marks remaining for this other variant, like the probable hardware SPI support on such, but that's not a priority, I will give it another try at another moment.

[madires]

I also noticed that more symbols were added to the bitmaps on recent versions of m-firmware, like battery icon, onewire, and maybe other stuff. I wonder whether those symbols are loaded into memory regardless whether we choose to display it or not? (I mean, whether the whole bitmap is loaded, or just the part of it we're declaring it's going to be used).

If none of the display features requiring additional characters/symbols are enabled the additional characters/symbols aren't included in the firmware. The additional characters in the font bitmaps are the probe numbers with reversed colors and the battery icons. Addtional symbols are question mark, Zener diode, quartz crystal and one-wire.

[carl1961]

New GM328A  running 1.51m   I had to use Semi-ST7735, took me a great while to figure how, I came across a post searching here.

#if 1
//#define LCD_ST7735                      /* display controller ST7735 */
#define LCD_SEMI_ST7735                /* display controller SEMI ST7735 */

Attached is files changed

I could no get 16 Hz crystal working on this board, it messes with the display driver.

[YouCanDoIt]

Sorry guys for disturbing the discussion, but I need your help to build a Mega 4gsl.
I found these two versions of the pcb (2022, 2019):


I only saw version 2015 developed from Nick Lanchak, the original author. Did he also released the above two versions?
I would like to use a color LCD, but the seller said that he only partially succeded. The output shows up, but very slowly (he said it is due to too high resolution). Plus there was a problem with the encoder, most likely the encoder needs to be transferred to other ports.
Someone can recommend a compatibile color LCD without such problems?

UPDATE
The tested screens with the above problems are ST7789v and ILI9341 on 1.48m. The board used is rev2023:

Works fine with with LG12864E:



I am wondering if the board revision plays a role.

[rddube]

rddube,The lower range of inductance measurements on m-firmware starts at about 10 microHenry. If you want to measure inductance values ​​lower than 10 microHenry, you can build an additional LC meter attachment or use the resonant method in k-firmware.

Thank you Indman for your reply and patience. I am relatively new at this and didn't know that the K firmware could support lower values of inductance. I have a TT with version 1.12K, is that good enough or is there a newer version and if so, where can i find it? I'd also like to read about the resonant method if you point me in the right direction. And again, thanks!

[Feliciano]

The last k-firmware version at the moment is 1.13k. You can find it, as well as the extense documentation (including measuring methods) at madires' github.

[carl1961]

Just received this one, with original ATmega328P. They are still out there. 

https://www.aliexpress.com/item/1005005497252414.html?spm=a2g0o.order_list.order_list_main.5.52321802KA7hcS

That's the same one I been posting about.  here I show pictures how I connected and programed. both 1.3k and 1.51m  https://www.eevblog.com/forum/testgear/lastest-gm328a/msg5327048/#msg5327048

[madires]

The last k-firmware version at the moment is 1.13k. You can find it, as well as the extense documentation (including measuring methods) at madires' github.

Just some snapshots. You'll find the latest k-firmware at https://github.com/kubi48/TransistorTester-source and its documentation at https://github.com/kubi48/TransistorTester-documentation.

[Feliciano]

Got it.

On the other hand, I was checking the 8x16-iso8859-2_hf font, and I corrected the uppercase O umlaut, and added some missing commas. Attached.

[indman]

Feliciano,here is the schematic diagram for your clone! Check if there are any gross errors in the arrangement of parts?

[madires]

On the other hand, I was checking the 8x16-iso8859-2_hf font, and I corrected the uppercase O umlaut, and added some missing commas. Attached.

Thanks! The commas aren't missing because the last element of an array doesn't need one. A comma after the ç (c with cedilla) is added when FONT_EXTRA is defined (two lines below ç).

[Feliciano]

indman, I modified the LCD and ISP connections. There are other minor errors, like the nunmbering of resistors (R20-pin19, R21-pin20, R22-pin21, R23-pin22, R17-pin23, R16-pin24, R14-pin4, and so on), although I understand that's not critical to fix (except for a newbie-intended documentation).

[indman]

Feliciano,Initially in the clone the LCD ports are assigned like this
#define LCD_DC           PB3            /* port pin used for D/CX */
#define LCD_SDA          PB5            /* port pin used for SDA */
Now what did you show in the last diagram?
#define LCD_SDA          PB3            /* port pin used for SDA */
#define LCD_DC           PB5            /* port pin used for D/CX */
Where is right?

And one more question - PB1(41pin) is connected to +5V through resistor R7. It is also directly connected to somewhere else through a via hole. Where?

I updated the circuit diagram for the FNIRSI clone in my previous post, taking into account corrections from Feliciano.

[Feliciano]

Thanks for your concern. For sure it could be more errors lying around (and I could have made a mistake) as we are adapting one diagram to another. To help, I attach a high res. photo of the area of interest.

[indman]

Feliciano, I just updated the circuit diagram. I think that if there are errors or something is missing, then this is not too important. The main thing is that the thoughts of FNIRSI engineers remained in the fog! 

[Feliciano]

Yep. I'm glad and thankful we could sort this out  . Now I can pass my old and trusty and improved WEI-M8 to a friend.

Another question: are we now using an small right-pointing-arrow instead of the ~ (the top part of the ñ) char.? I ask because I saw that on the 8x16-ISO8859-2_hf bitmap, and I might create a similar font for being able to keep with the m-firmware increasing code size.

BTW: I would prefer the show font feature on the m-firmware would do something like the attached picture, instead of the actual block and pages based format.

[BILLPOD]

I bought one of the units on Amazon, (USA),...this one:  https://www.amazon.com/dp/B07WT9VVZB?psc=1&ref=ppx_pop_dt_b_product_details, which describes the item as:
LCR-T4 Mega328 Digital Transistor Tester.    However, the MCU is an APT32F172K8T6.   It looks at an IGBT transistor and says it is a resistor.
I also hooked it to a UJT, (Unijunction Transistor), and it described it as a double diode.  And some of the FETS I connected it to were identified as plain old resistors.  I bought a T1 Transistor tester back in 2019 and it
identified all of these items accurately.   This new one is a piece of crap and I do not recommend that anyone waste their money on it.  If I find my old one, this one is going in the bin.

[Feliciano]

Some remarks about the measurements with the new CTester running 1.51m, after measuring some components and callibrated profile:

• The Capacitance is showing an ESR tenfold the expected (around 24Ω for 27nF and 7Ω for 100nF film Capacitors, or 1Ω for 100µF, regardless of ESR method and/or ESR tool enabled/disabled)
• A short on the ZIF detects around 8µH
• Apparently no problem with semiconductors

EDIT: Similar high ESR measurements with indman's 1.51m, or with 1.46m.