$20 LCR ESR Transistor checker project

[timofonic]

No, it not a separate tester. The basic firmware from Karl-Heinz Kübbeler with function of measurement of voltage reference diodes is used. Changes are made only in the tester hardware 

Well then. I guess I was almost completely wrong. How extensive are the hardware changes? Is this fork being made available in the same way as the original Karl-Heinz transistor tester?

Hi indiman!

About your "fork" and  explain more detailed your modifications

I'm  very interested on your modification, thanks a lot!

You can use your native language if others could translate it. Please, it would be greatly
* Can you say what's your native language?

- Some suggestions:
* How it measures 78xx and other voltage/current ICs, procedure and limitations.
* Manufacturing considerations: Layout, interferences, recommended shielding, connectors, other advices.



About possible future additions and improvements

- Are there possibilities of having a better firmware integration?
- Can other features be added easily without going into feature creep? 74xx and 40xx testing like in minipro, for example.

My reasons to build this device and contribute by making it in KiCad


I have and use TONS of salvaged parts

- A device like this might be really useful to check salvaged parts.
- It would save me time at the end.

I'm a cursed idealist, unable to change that. I'm a unadaptable person by nature.

- I'm an extreme copyleft believer:
* FOSS, OSHW and Open Access are just parts of it. I want to promote it very actively in every place.


Suggestion: Making your improved design a collaborative OSHW project and using KiCad (FOSS)

- Do you use KiCad? What about sharing schematics over GitHub?
* I could try to recreate your design in KiCad, submit it at GitHub and ask others about bugs and improvements. I'm awful at EDA, so please be polite and constructive

- Would you consider switching to KiCad 4 RC1/RC2/nightly?
* Maybe someone could help you by translating it to your native language. I'm doing to translate it to Spanish by using Transifex.com (very easy).

Other reasons: I have ADHD, economical issues, etc.

-----
About language

Can you say us where are you from and what's your native language?

- The first part is optional, but you are free to say it it but depending on your situation.
- I'm from Spain, so Spanish.

Maybe some of us could assist you with communication and even help you at English:

- My English is very far from perfect:
* I consider to have very good English reading skills.
* I think to have enough good written English skills if not writing it too fast or very tired.
* My speaking English is awful, but slowly improving.

I emphasize with you a lot. I had the same problem when starting to use Linux in the early 2000:
- I did my first hard baby steps when Linux kernel and other software needed to be compiled by hand (configure && make && make install):
* There weren't nice source based package repositories and sometimes you needed to use bleeding edge software for certain functionalities. * Debian was a mess at packages, Red Hat okaish but had issues too.
- I had lots issues but that ironically made it a funny challenge.
- I asked my issues over IRC by using online translators and dictionaries, but I was lots jess skilled than you:
* Fortunately some Mexican people helped me to explain my issues to others and act as interpreters.
* Despite my lack of proper formal English learning because school was awfully bad at teaching it and I hated to assist, I improved by practice.

- These days I use Archlinux and Linux Mint. I'm considering to switch back to Gentoo or similar, i just need to realize a twisted plan and get new hardware:
* I liked the extreme configurability of ebuilds.
* I need to consider the viability to use remote hardware to accelerate compiling.


Kind regards.

[madires]

I've got one for you, too though.  When the tester shows that it detected multiple diodes (like 4), the message preceding that says that no component is found. That's rather contradictory: It knows that something is connected, yet says otherwise.

The original idea was to give a hint about possible diodes when no component was identified. Since the output of the pinout for more than 2 diodes would be a nightmare, the Show_Fail() function is simply used for that too, to save a few bytes flash memory. I'll look into it.

[madires]

When starting up,  it shows "MTester v2.07...by Efan &Haoqixin"

Sorry, that's a modified firmware. You have to ask the seller for a current one.

[madires]

You could use my asm SPI code.
http://nerdralph.blogspot.ca/2015/03/fastest-avr-software-spi-in-west.html

I'll have a look. Thanks for the hint!

[indman]

Do you use KiCad? If so, what about sharing schematics over GitHub?

What's your language? Maybe some of us could assist you with communication. I'm from Spain.

No, I use for drawing of electric circuits and design of printed circuit boards of the program from ABACOM (SPLan 7.0 and Sprint-Layout 6.0). Link to this software:
http://abacom-online.de/uk/html/demoversionen.html
What specifically interests you according to this diagram? I will try to answer questions. My native language Russian!

[Kalvin]

Would there be any idea to add another Atmega328p to be used as an LCD driver? That would leave more room for the actual measurement application in the main Atmega328p. The two processors would communicate over SPI bus, and the SPI bus could be unidirectional thus making things quite simple. The BOM would not suffer too much either. Not sure how much this would complicate the actual code, though.

[timofonic]

Do you use KiCad? If so, what about sharing schematics over GitHub?

What's your language? Maybe some of us could assist you with communication. I'm from Spain.

No, I use for drawing of electric circuits and design of printed circuit boards of the program from ABACOM (SPLan 7.0 and Sprint-Layout 6.0). Link to this software:
http://abacom-online.de/uk/html/demoversionen.html
What specifically interests you according to this diagram? I will try to answer questions. My native language Russian!

I modified the post after you replied, sorry.

Another smart Russian guy. Nice!

sPlan is proprietary and for Windows. I want to use KiCad. I'm a very long time Linux user.

Can you export the project in some compatible formats please? Netlist, IDF, DXF.

Is the PCB routing so hard to make? I might try to do the export with and ask for help out there.

Are you brave enough to try latest KiCad too?

Kind regards!

PS: To all the involved parties in this project: Would you accept some good
Spanish food as donation? Wine or others. Send PM!

[madires]

Would there be any idea to add another Atmega328p to be used as an LCD driver? That would leave more room for the actual measurement application in the main Atmega328p. The two processors would communicate over SPI bus, and the SPI bus could be unidirectional thus making things quite simple. The BOM would not suffer too much either. Not sure how much this would complicate the actual code, though.

That wouldn't solve the problem, because we still would have to use a bit-bang SPI to interface the display. With an ATmega 644 or 1284 there are more I/O pins which could be used for the test resistors to free up the hardware SPI, which in turn could be used to drive the graphic LCD. The 644 example in Karl-Heinz' documentation does exactly that.

[Kalvin]

Would there be any idea to add another Atmega328p to be used as an LCD driver? That would leave more room for the actual measurement application in the main Atmega328p. The two processors would communicate over SPI bus, and the SPI bus could be unidirectional thus making things quite simple. The BOM would not suffer too much either. Not sure how much this would complicate the actual code, though.

That wouldn't solve the problem, because we still would have to use a bit-bang SPI to interface the display. With an ATmega 644 or 1284 there are more I/O pins which could be used for the test resistors to free up the hardware SPI, which in turn could be used to drive the graphic LCD. The 644 example in Karl-Heinz' documentation does exactly that.

Maybe I expressed my intention a bit poorly. I will try to be more specific this time.

The would be two Atmega328's:

1. The main Atmega328 "Main MCU" running the measurement application without actual display code.
2. The display controller Atmega328 "Display MCU" driving the LCD (or whatever display).

- There is a simple, unidirectional serial bus from the Main MCU to the Display MCU.
- The Main CPU will be a master, and the Display MCU will be a slave.
- There are specific, simple messages passed from the Main MCU to the Dispaly MCU.

    For example, the message might contain following information:
        - Component test
        - Transistor
        - Pin 1 E
        - Pin 2 B
        - Pin 3 C
        - NPN
        - Gain=200
        -etc.

Thus, the Main MCU will just send simple messages to the Display MCU, and the Display MCU
application will decide how to display that information. This would free quite a lot of code memory
from the Main MCU for the measurement and extra algorithms. The Display MCU would contain all
code displaying the measurement information, fonts, graphical bitmaps and language data.

[timofonic]

Would there be any idea to add another Atmega328p to be used as an LCD driver? That would leave more room for the actual measurement application in the main Atmega328p. The two processors would communicate over SPI bus, and the SPI bus could be unidirectional thus making things quite simple. The BOM would not suffer too much either. Not sure how much this would complicate the actual code, though.

That wouldn't solve the problem, because we still would have to use a bit-bang SPI to interface the display. With an ATmega 644 or 1284 there are more I/O pins which could be used for the test resistors to free up the hardware SPI, which in turn could be used to drive the graphic LCD. The 644 example in Karl-Heinz' documentation does exactly that.

Why not go MIPS or ARM? Or it would overcomplicate, more expensive and require rewriting code?

[madires]

Why not go MIPS or ARM? Or it would overcomplicate, more expensive and require rewriting code?

The ATmegas run with 5V which is the ideal voltage for testing components. 3.3V is simply too low, i.e. we would have to add external switching for a higher voltage, external ADC and so on.

[hapless]

@Circuiteromalaguito, Kalvin, et al:

This, of course, is just my opinion, but...

Every now and then someone comes who wants to add lots of bells and whistles to the existing schematic. The charm of the original tester was in the simplicity of its hardware. Take that away, and you are left with something other than the transistor tester that we are all discussing in this thread. Sure, you can do everything that you suggest, but what reason is there not to do it as a separate project? Then you can use, say, ARM and a voltage converter, and do many other interesting things... but I believe you're trying to horseshoe this particular project into something that it's not.

[CustomEngineerer]

You could use my asm SPI code.
http://nerdralph.blogspot.ca/2015/03/fastest-avr-software-spi-in-west.html

Very nice, thanks for info. Assembly was one of my favorite classes when I was in school but I haven't hardly touched it since.

[CustomEngineerer]

@Circuiteromalaguito, Kalvin, et al:

This, of course, is just my opinion, but...

Every now and then someone comes who wants to add lots of bells and whistles to the existing schematic. The charm of the original tester was in the simplicity of its hardware. Take that away, and you are left with something other than the transistor tester that we are all discussing in this thread. Sure, you can do everything that you suggest, but what reason is there not to do it as a separate project? Then you can use, say, ARM and a voltage converter, and do many other interesting things... but I believe you're trying to horseshoe this particular project into something that it's not.

Fully agree with this.

[tom666]

@hapless:
I agree I have the same opinion.

[timofonic]

I partially disagree.

If the addons are simple enough and provide extra features... Why not?

If you like it simpler, just fork. If you want more features, fork. The most popular fork wins: Open Source evolutionary way

I have over 500x 74xx 40xx 45xx ICs tó test, some of them unlabeled.

I just found devices like this one but with IC TTL CMOS support too.

http://www.ebay.com/bhp/ic-tester

[madires]

I have over 500x 74xx 40xx 45xx ICs tó test, some of them unlabeled.

I just found devices like this one but with IC TTL CMOS support too.

http://www.ebay.com/bhp/ic-tester

A tester for logic gates is a complete different story. Some universal programmers can do that too but the Transistor Tester won't support that. My opinion about future developments is that we'll keep the simple 328 Transistor Tester as a basic version (inexpensive and easy to build) and add new options/features to a 644/1284 advanced version (more flash, RAM and I/O pins). And I'd like to keep maintaining the code feasable.

[con-f-use]

It shouldn't be the "eierlegende Wollmichsau", I like it for it's simplicity. When I was starting out in electronics, I could build a useful tool with the things I had in hand and learn something in the process. Cramming it with features would make it too complicated for beginners and make it actually less useful. The free hardware point was valid. If you want something more powerful based on the original software, nothing is stopping you from building it.

The only thing I think that is a bit overdue is a new reference design for the board that is well tested and maybe incorporates some of the more useful add-ons such as the rotary encoder.

[GnatGoSplat]

I have a Chinese tester from 2012 with the 2x16 display and an Atmega8L and no crystal.  Figured I'd goof around with it because I can't leave well enough alone.  Mine has no programming pins on the board, but the chip is socketed.  I downloaded the newest firmware (1.12k) that says it works with the Atmega8, and used a TL866 to flash TransistorTester.hex into CODE and TransistorTester.eep into DATA.  It boots up, displays "Version 1.12k", but doesn't detect devices and after awhile of "Testing..." it goes to "Timeout!"  I had backed up the original firmware, so I wrote those back in and all is well again on the original firmware, but what did I do wrong with flashing 1.12k?  I have a hunch I need to change some fuses to tell the software the configuration of my hardware, but I'm not familiar with microcontroller programming and I don't know how to translate the hex codes I see for fuses into the named checkboxes on the TL866 fuse setting screen.

[Macbeth]

The checkboxes you tick in the TL866 Miniprog settings are all representative of binary bits which all add up to a hex nibble (0-F), byte (00-FF), or word (0000-FFFF)... I don't have mine in front of me but I recall when you tick the checkboxes the combined value in bytes is also displayed. So if you have been told to set the fuses to a certain 0xFF or whatever take a look for that, or just add up the bits 1 2 4 8 16 32 ... etc

[timofonic]

I have over 500x 74xx 40xx 45xx ICs tó test, some of them unlabeled.

I just found devices like this one but with IC TTL CMOS support too.

http://www.ebay.com/bhp/ic-tester

A tester for logic gates is a complete different story. Some universal programmers can do that too but the Transistor Tester won't support that. My opinion about future developments is that we'll keep the simple 328 Transistor Tester as a basic version (inexpensive and easy to build) and add new options/features to a 644/1284 advanced version (more flash, RAM and I/O pins). And I'd like to keep maintaining the code feasable.

Are there plans for a 1284 version?

Are there some standalone 74xx 45xx 40xx checker using a similar hardware? I did see one for PIC published on EPE.

[indman]

I have a hunch I need to change some fuses to tell the software the configuration of my hardware, but I'm not familiar with microcontroller programming and I don't know how to translate the hex codes I see for fuses into the named checkboxes on the TL866 fuse setting screen.

Program the tester from fuses, as on the picture below. They for the internal generator of the chip on 8MHz. IntelHex need to load the TransistorTester.eep file in DataMemory as.

[tom666]

I have a Chinese tester from 2012 with the 2x16 display and an Atmega8L and no crystal. ... I have a hunch I need to change some fuses to tell the software the configuration of my hardware, but I'm not familiar with microcontroller programming and I don't know how to translate the hex codes I see for fuses into the named checkboxes on the TL866 fuse setting screen.

Hi GnatGoSplat,
the recommended setting of the fuses for ATmega8 (Int. RC 8MHz) by software author (KHK) is:
lfuse:0x24 hfuse:0xd9

Bye
Tomas

[madires]

The only thing I think that is a bit overdue is a new reference design for the board that is well tested and maybe incorporates some of the more useful add-ons such as the rotary encoder.

A forum member has planned to create a nice PCB with all hardware options included. So I haven't looked into a new reference PCB yet. Meanwhile there's a simple adapter in the SVN for the rotary encoder and the frequency counter which is inserted between the PCB and the display.

[madires]

Are there plans for a 1284 version?

Please see Karl-Heinz' current documentation There's a circuit for the 644/1284.

Are there some standalone 74xx 45xx 40xx checker using a similar hardware? I did see one for PIC published on EPE.

I've seen some on ebay, but they seem all to have the same basic design. I got no idea if those are PIC or ATmega.