Author Topic: $20 LCR ESR Transistor checker project (Read 4383146 times)

neslekkim · « **Reply #250 on:** January 13, 2014, 01:14:16 pm »

this?
svn://mikrocontroller.net/transistortester/Software/trunk/mega8

paulie · « **Reply #251 on:** January 13, 2014, 02:01:58 pm »

http://www.mikrocontroller.net/svnbrowser/transistortester/Software/trunk/mega8/

thanks. sorry but im not good with these code suppositories. with your hint and a little link fixup im there.

this is just the build i need because others im involved with already have m8 breakout boards and only need the 6 resistors and a button. when im back to my big pc i will play with the makefile to enable pullups so i dont have to put one on the sw or tie bat and ref hi. also disable timeout because these are powered off serial/usb dongle. maybe investigate strings in flash instead of ee option too if theres room.

anyway none of these are important atm. i put the button on reset instead of pd7 which works great. heres a couple screen shots. within a fraction of a percent of precision meter readings. and as seen some of my "mystery" to-92 are pnp and others nfet.

Gandalf_Sr · « **Reply #252 on:** January 13, 2014, 07:32:44 pm »

I've ordered one of these which is a little expensive but I like the zif connector approach.

I think this is going to use the atmega 328 but I may want to change resistors to better tolerances and/or improve the voltage reference.

Anyone already done this?

attila123 · « **Reply #253 on:** January 13, 2014, 09:03:51 pm »

Hi guys!

What are the opinions about this ESR meter? I would need it to repairing LCD monitors, power supplies, etc, so in-circuit testing would be useful.

http://www.ebay.com/itm/Inductor-capacitor-Tester-Milliohm-Meter-ESR-Meter-LC-Meter-/130888138780?pt=LH_DefaultDomain_0&hash=item1e798a801c

Thank you!

cellularmitosis · « **Reply #254 on:** January 13, 2014, 10:59:14 pm »

Gandalf, I just got that exact unit in, and then found this thread. I'd be happy to join in your efforts towards improving this little unit!

Gandalf_Sr · « **Reply #255 on:** January 14, 2014, 12:10:33 pm »

Atilla
These are all variations on a theme but I would beware of cheap Chinese copies.

Cellular
I'm mainly a hardware designer but can also do software - 60/40 is the split. I like the utility of the meter just for having around my workshop evne if I do nothing to modify it.

I am interested in coming up with a revised set of hardware and letting someone else tweak the software - is that someone you I wonder? Do you have any AVR studio skills?

Did you get schematics for yours or will we have to reverse engineer it - or more likely find out how close a clone it is of the original design?

I've read the whole thread and here's a summary (from memory) of things that can be improved:

1. The 2.5 volt reference voltage
2. Use 0.1% resistors where they matter
3. Put some thought into how best to lay out the component connection pads
4. Add an auto-discharge system (relay?) to discharge capacitors before testing

Stretch ideas:
a. Redesign the board to fit in a standard case
b. Add a frequency counter or does it already do this? - I have several iterations of a counter project that I did a while back using a PIC
c. Add a signal generator of some kind

Feel free anyone to chip in on this.

madires · « **Reply #256 on:** January 14, 2014, 01:26:02 pm »

Quote from: paulie on January 13, 2014, 12:28:57 pm

aha... i didnt know kubi was karl. what a team. thanks to both of you for bringing me up to speed on importance of ee in that build. im guessing putting those strings in ee is to save flash for the 8k version?

Exactly! Every byte counts ;-)

BTW, the 8k version has only a limited set of features and the 16k version has some limitations too.

madires · « **Reply #257 on:** January 14, 2014, 02:04:22 pm »

Quote from: Gandalf_Sr on January 14, 2014, 12:10:33 pm

I've read the whole thread and here's a summary (from memory) of things that can be improved:

1. The 2.5 volt reference voltage
2. Use 0.1% resistors where they matter
3. Put some thought into how best to lay out the component connection pads

Stretch ideas:
a. Redesign the board to fit in a standard case
b. Add a frequency counter - I have several iterations of a project that I did a while back using a PIC
c. Add a signal generator of some kind

The voltage reference helps if you got a poor voltage regulator. With a MCP1702 there's no need for it. Just the probe resistors should be 0.1% types. There are lots of PCB layouts in the support forum. Haven't looked into the possibility of a simple frequency counter yet (if it might be feasable with the standard pin assignment). The m-firmware includes a PWM tool (frequency is selectable from a menu and the ratio is adjustable in steps of 5%). For more information please see Karl-Heinz' excellent documentation (http://www.mikrocontroller.net/svnbrowser/transistortester/Doku/tags/english/).

Gandalf_Sr · « **Reply #258 on:** January 14, 2014, 02:24:09 pm »

Madires
Thanks for the advice. your link takes me to a page that appear to have pdf documents available but I can't seem to download or view them, do I need some special software?

When I eventually get the files, is there a source code file that I can compile in AVR studio or do I need another tool?

Is the source in C or assembly or mixed?

Thanks in advance.

cowana · « **Reply #259 on:** January 14, 2014, 02:30:19 pm »

To download those files, hit the "Download GNU tarball" link below the file names.

Andy

madires · « **Reply #260 on:** January 14, 2014, 02:48:38 pm »

Quote from: Gandalf_Sr on January 14, 2014, 02:24:09 pm

When I eventually get the files, is there a source code file that I can compile in AVR studio or do I need another tool?

The current version is always available at http://www.mikrocontroller.net/svnbrowser/transistortester/Software/trunk/ and should compile in AVR studio.

Quote

Is the source in C or assembly or mixed?

Mostly C, but a few pieces are in asm.

Gandalf_Sr · « **Reply #261 on:** January 14, 2014, 07:19:22 pm »

Thanks, got the files downloaded and into AVR Studio 6. It seems to compile OK for the 328 version.

Don't have the kit from China, probably won't for 10 days or more.

netdiver · « **Reply #262 on:** January 14, 2014, 10:12:34 pm »

Dear All,
first of all I would like to thank Markus and Karl-Heinz for being here and supporting. This project is really great! Thanks also to Markus Frejek who started this incredible little swisstool.

I'm totally new to the world of microcontrollers and I'm here after googling a bit about the Transistor Tester I bought on the bay.
After reading some times this thread, I decided that I should upgrade my new tester. Here it is:

I removed the zener D1 and soldered an ISP header (10 pin) to program it.
Being a complete newbie, I downloaded WinAVR and followed the direction given in the english documentation, and tried right away to put the 1.10k version on it, as it has only a Atmega168. I compiled the Makefile in the "default" directory, as it was designed for the 168.
But programming gave an error. After a little investigation, I found out that maybe the latest version does not fit anymore on a atmega168, as it is a bit larger than internal flash.
I succeded in installing 1.09k (i tried only the k versions, as the m version is only suitable for an atmega 328).
Is it true that from 1.10k on, Atmega168 will be insufficient to hold the newest versions of the firmware?

Another question: which flags on the Makefile of the default directory are suitable for this chinese clone? For this first flash, I left everything untouched, but I can't understand completely the meaning of the flags and put them in relation to the hardware.

I will try to solder the 100nF bypass caps where Karl-Heinz did, but... hey I need something very small to solder over these small atmega pins!

Thanks in advance.
Best Regards from Lorenzo

paulie · « **Reply #263 on:** January 14, 2014, 10:31:43 pm »

size is 17669 (0x4505) which is 1285 too much for m168. del a few features in the makefile and it will fit m168 or even m8.

i have a small problem myself. how do you run calibrate? whats really frustrating is it has been run twice before but im not sure how. shorting probes and hitting the button causes "not calibrated" message and instructions but then stops at the site address. hitting the button during or after just starts that again.

the maunal and message says short probes and hit button and confirm within 2s but this dont work. what am i doing wrong?

netdiver · « **Reply #264 on:** January 14, 2014, 10:55:10 pm »

Thanks very much paulie for your quick reply!
I commented out in the makefile the CFLAGS += -DWITH_UART an that did the trick! Now the code is at 99.3% and fits into the atmega 168!

As for your problem, I don't have this behaviour. Shorting the three test points together and pressing the test button I can see on the display "Selftest mode ?" If I press the test button again, the selftest / calibration starts. At a point it says "Isolate probe!": pulling out the short from the test points the calibration goes on, then the software asks for a capacitor greater than 100 nF and the selftest ends shortly after.

Thanks again for solving my issue!
Best regards
Lorenzo a.k.a. netdiver

netdiver · « **Reply #265 on:** January 14, 2014, 11:25:48 pm »

A little step further... It's all in the documentation! So sorry for asking without having read thoroughly the docs.
I just noticed that to enable the extended test T1 to T7 it is necessary to disable one of the hFE measurement circuits.
I remove comment from NO_COMMON_COLLECTOR_HFE and now I have the T1 to T7 tests in the selftest mode, but I had to downgrade to 1.09k, as with 1.10k I'm not able to enable those extended tests without going over capacity of ATmega168.

Best regards.
netdiver

paulie · « **Reply #266 on:** January 14, 2014, 11:46:36 pm »

hmmmm... strange thing is it worked at one time. i cant compile atm but use the 1.10 hex file below that markus sent me.

when i reset or hit button i get "testing 1-rr-2-rr-3 00ohm 00ohm" and that damn warning. its very accurate and stable in test mode but cant get it to calibrate again. the only thing different is tying the zener pin high but that shouldnt matter.

paulie · « **Reply #267 on:** January 15, 2014, 12:45:33 am »

problem solved.... reflashing all 3 chip made no difference but new eeprom upload fixed it. idk why that would make a difference but calibration is working ok now. looks like something was corrupted in the eeprom. maybe time to look at that bod fuse setting. i was at least able to suppress the warning in all of them by copying the ee data from the original but the defect got copied too. sorry for wasting everybodies time, all working great again.

i gotta say its astonishing how accurate this is. even without calibration they had less than 1% deviation but after less than 0.1% compared against thousand dollar precision meter. amazing, and all with less than a dollars worth of hardware. thanks again.

Gandalf_Sr · « **Reply #268 on:** January 15, 2014, 01:08:19 pm »

Netdiver
To put caps on the power pins, you can scrape at the copper ground fill to expose bare copper, prep it with flux, and then put solder on it (you'll need a strong iron to heat the ground plane). Then you can put one end of the cap on the Atmega pin and the other to the newly-created ground pad.

Can anyone point me to the absolute latest schematic for this project?

madires · « **Reply #269 on:** January 15, 2014, 01:58:04 pm »

Hi Lorenzo!

Quote from: netdiver on January 14, 2014, 10:12:34 pm

I removed the zener D1 and soldered an ISP header (10 pin) to program it.

That's the recommended procedure for the Chinese clones. Either remove the zener or replace it with a proper voltage reference.

Quote

I succeded in installing 1.09k (i tried only the k versions, as the m version is only suitable for an atmega 328).
Is it true that from 1.10k on, Atmega168 will be insufficient to hold the newest versions of the firmware?

The m-firmware supports the ATmega168 also. By setting the MCU to ATmega168 in the Makefile some features are disabled and the compiled firmware image will be small enough for 16kB flash and the smaller EEPROM. The k-firmware supports ATmegas down to 8kB flash with a limited set of features (read the documentation and edit the Makefile). If you like to have all available features you'll need an ATmega328.

madires · « **Reply #270 on:** January 15, 2014, 02:05:13 pm »

Quote from: Gandalf_Sr on January 15, 2014, 01:08:19 pm

Can anyone point me to the absolute latest schematic for this project?

The latest one is available at http://www.mikrocontroller.net/svnbrowser/transistortester/Doku/trunk/pdftex/english/ (click "Download GNU tarball" at the bottom).

netdiver · « **Reply #271 on:** January 16, 2014, 02:29:08 pm »

Quote from: madires on January 15, 2014, 01:58:04 pm

The m-firmware supports the ATmega168 also. By setting the MCU to ATmega168 in the Makefile some features are disabled and the compiled firmware image will be small enough for 16kB flash and the smaller EEPROM. The k-firmware supports ATmegas down to 8kB flash with a limited set of features (read the documentation and edit the Makefile). If you like to have all available features you'll need an ATmega328.

Thanks a lot Markus, so I can upload your firmware too!
Unfortunately my version has the SMD ATmega and my equipment lacks of SMD tools, so I'll stick with this version, until I find someone to sell it to!!

Thank you again for all your effort given in developing this project.

Lorenzo

SteveM · « **Reply #272 on:** January 21, 2014, 05:27:00 am »

Hi, I am new to this forum but I did read this entire thread and I have learned a lot about this little device. I noticed on ebay that the Chinese have put this tester into a new package with the brand name YIWANJIA. There are three buttons to the right of the center off/on/test buttons, but they are labeled in Chinese and there is no English manual for this tester. Does anyone know what those three new buttons do?

SteveM · « **Reply #273 on:** January 21, 2014, 08:59:03 am »

Here are a couple of pictures to my post above to save the trouble of looking it up on ebay. The specs seem to indicate it is the same circuit as the $20 LCR ESR transistor checker, but with updated software.

neslekkim · « **Reply #274 on:** January 21, 2014, 09:01:13 am »

according to other pictures, one button is power, and other is test, as seen here: http://www.ebay.com/itm/181164691633


EEVblog Main Site	EEVblog on Youtube	EEVblog on Twitter	EEVblog on Facebook	EEVblog on Odysee

Author Topic: $20 LCR ESR Transistor checker project (Read 4383146 times)

Share me