$20 LCR ESR Transistor checker project

[madires]

I've got a couple of request for features that might be useful - I did build them in to my own unit but it would be nice if they were officially supported.

Please send your modifications to Karl-Heinz (his email address is on the first page of his excellent documentation) for the k-firmware or to me for the m-firmware.

And a quick update on the m-firmware. I've tested several TRIAC type opto couplers and the detection works fine so far. Even a MOC3043 with the zero crossing circuit and a low inhibit voltage of about 5V is no problem. For BJT types the t_on/t-off delays are measured. Not very accurate, but it gives an indication. I think I'll release 1.22m at the weekend. Sorry, no hidden easter eggs

[Meth]

No problem, I'll try and sort out a patch in the next few days.
For reference I've just taken some screen shots of what you see.

Edit: going to back to the source it looks like the potentiometer shows like that in the standard code anyway - the change I'd made (which I don't have a picture for) was to add the total resistance on the next line below.
I'll submit it anyway and leave it to Karl-Heinz's discretion

Thanks!

[bitseeker]

Welcome to the forum, Meth. I like your enhancements. Looking forward to seeing them incorporated.

[MrJohnnny]

Hi, guys! It has been a long time since my last post since I was very happy with the tester  But, today I wanted to make it work from a single Li-Ion cell. I built a simple boost converter (3.7V-5V), and it works just fine; however the software I'm using complains about battery voltage being too low (3.8V) and won't even start the testing procedure. So, the question - is there any .hex available that will allow me to use the 3.7V cell?
 I'm not very good at programming, so I can't go and change the low battery threshold myself. I tried, but failed immediately 
P.S. I did hardware "hack" (sent a 2.5V to pin 28), works just fine. But that way I don't know what is battery voltage at a time. So, I ask for your help 

[madires]

For the k-firmware you can change the threshold voltage in the Makefile by changing BAT_POOR. For the m-firmware it's in config.h (also BAT_POOR).

[MrJohnnny]

I tried to do so. I'm using 'k' software with atmega328 and 16x2 LCD, so I went and opened Makefile in: trunk/mega328. I opened Makefile in Programmer's Notepad, changed BAT_POOR to 2900 (CFLAGS += -DBAT_POOR=2900). After I pressed "Make All" it did compile fine (exit code=0) so I flashed the uC, but it doesn't work, there's nothing on the screen. What did I do wrong? Maybe I need to include some files (.c or .h) before pressing "Make All"?

[madires]

Is the LED lit? Is there some brief output on the display or nothing at all? Does the tester stay powered on?

[MrJohnnny]

There's nothing appearing on the screen even for a split second. Maybe the project has been changed to accommodate large graphical displays so it won't work now with my boring 16x2 LCD? By the way in my build data bits 4-7 of the LCD are connected to PD0-PD3 on the Mega328, is this correct?  Yes, LED is lit and the device stays on. I can't even run a Proteus simulation, every .hex refuses to work

[MrJohnnny]

Please excuse me being a complete idiot  I used the wrong fuses, now it works excellent. I can't really be embarrassed any more, you guys (Karl and Markus) paid so much time&effort into making this and you did a GREAT job making it simple to modify, yet I still managed to fail, D'OH!

[kapuchy]

Changing hex code isn't necesssary, you can also change value of one resistor (1:2 divider instead of 1:4 for battery voltage measurement)

[MrJohnnny]

I'm aware of that, I did exactly the same to make it work (I mentioned about putting 2.5V (Vcc/2) onto PC5 of the Mega328 above). But that will give me bullshit readings, so I decided to do it the proper way 

[hendrick]

I think I'll release 1.22m at the weekend.

B-but I just managed to flash to 1.21m...

I have some questions regarding compiling the m-firmware for the DIY M12864 kit (the red one with the rotary encoder):

• Am I meant to leave the "MCU = atmega328" setting in the Makefile if my MCU is really atmega328p (I cannot compile if I set "MCU = atmega328p" in the Makefile)?
  
• My atmega328p came with original fuse settings [L=FF, H=D9, E=FF, Lock=3C], after reading through this thread I have put my fuses at [L=F7, H=D9, E=FC, Lock=3F] -- are these correct?

Thanks!

[madires]

Yes and yes.

[ttcole1254]

I just got one of these, believe it is a clone but seems to work nonetheless.

I found a problem though when attempting to measure capacitors of any size. The larger the capacitor, the larger off it seems to be. It reads extremely high.

I tested a 1000 microfarad cap, got 3254 microfarads with 0.02 ohm ESR and 1.4% voltage loss
An 0.47 mic cap reads 520.5 nano, 10 ohm ESR and 2% VLoss.

The lower you go, the more accurate it becomes. An 0.01 mic cap reads 10.64 nano with a VLoss of 1%.

I attempted the self test procedure, and the values are a little better, but still way off on the larger capacitors. These caps were sitting in a drawer and were fully dead, so that couldn't have been it. I tested all resistors on the back, and found a 100k resistor was actually doing 55k. Replaced it with a new 100k, and the readings didn't change. Resistors test correctly, showing 1011 ohms for a 1000 ohm resistor attached. Not sure what the problem could be now, but I can always return it for a new one. Figured I would try and fix it first

Edit: Another thing because I know the crystal can affect the readings. It says 8 MHz, but I have no way to test that. No version number appears when booting it up, just "MTester" at the top.

[all_repair]

Yes and yes.

Impressive !!! 

[madires]

I tested a 1000 microfarad cap, got 3254 microfarads with 0.02 ohm ESR and 1.4% voltage loss
An 0.47 mic cap reads 520.5 nano, 10 ohm ESR and 2% VLoss.

There's something wrong with your tester. The error normally increases to about 5% above 100µF.

Edit: Another thing because I know the crystal can affect the readings. It says 8 MHz, but I have no way to test that. No version number appears when booting it up, just "MTester" at the top.

It's a modified firmware version without the source code being available

[ttcole1254]

Yeah I was worried it was a clone when no version appeared anywhere, not even in the self test. Shame something's wrong with it because the circuit is relatively simple as far as the component count, and all the resistors check out on a multimeter. Guessing it's either something wrong with the atmega or possibly a software glitch due to it being a modded clone. Also reflowed all solder on the board thinking it could've just been a bad connection somewhere.

It does have the ISP port and I do have an arduino uno I could use to program it. But if I do change the firmware and it doesn't fix it, I have no chance at returning it. Do any of you know where I could buy a new one running a non-modded firmware?

[hapless]

The semi-DIY versions typically come with unmodified/somewhat unmodified firmware these days. Here, for instance, is a popular version that's not too bad:
http://www.banggood.com/DIY-M12864-Graphics-Version-Transistor-Tester-Kit-LCR-ESR-PWM-With-Case-p-997023.html
I still recommend reflashing as soon as you get it.

Reflowing these things is something I would do only as an act of desperation because, in my experience, resistors, while harder to destroy by heat than other components, tend to drift in value permanently when heated too much, and that's not good for the "precision" measurement resistors that are probably not so precise to begin with.

[madires]

Yeah I was worried it was a clone when no version appeared anywhere, not even in the self test. Shame something's wrong with it because the circuit is relatively simple as far as the component count, and all the resistors check out on a multimeter. Guessing it's either something wrong with the atmega or possibly a software glitch due to it being a modded clone. Also reflowed all solder on the board thinking it could've just been a bad connection somewhere.

Could be also wrong fuse bits, like using the internal RC oscillator instead of the external crystal. Component-wise there's not much to screw up, besides damaging the MCU by a charged cap.

It does have the ISP port and I do have an arduino uno I could use to program it. But if I do change the firmware and it doesn't fix it, I have no chance at returning it. Do any of you know where I could buy a new one running a non-modded firmware?

Actually any clone not running that MTester firmware. But it doesn't matter since you can flash one of both "official" firmwares as long as the display is supported.

[hendrick]

Some questions about the m-firmware (I have the DIY M12864 kit, red pcb):

• Is the calibrating capacitor included in the kit not needed for the self-test/adjustment with the m-firmware?
• It seems my rotary encoder isn't responsive, when I rotate it slowly, it cycles through the menu but changing to the next menu item doesn't take the same number of turns, e.g. it might change after 1 turn but some times as many as 4 or more turns are needed. If I rotate it faster in a clockwise direction (at a constant rate), it cycles downwards through the menu but also it may decide to go upwards or even up then down again. Is this a firmware or hardware issue?

[bitseeker]

Sounds like a bad or dirty encoder.

[madires]

Is the calibrating capacitor included in the kit not needed for the self-test/adjustment with the m-firmware?

For the m-firmware's self-adjustment you would check the cap three times at least (normal measurement) and then run the self adjustement. It's explained in the README file.

It seems my rotary encoder isn't responsive, when I rotate it slowly, it cycles through the menu but changing to the next menu item doesn't take the same number of turns, e.g. it might change after 1 turn but some times as many as 4 or more turns are needed. If I rotate it faster in a clockwise direction (at a constant rate), it cycles downwards through the menu but also it may decide to go upwards or even up then down again. Is this a firmware or hardware issue?

Have you set ENCODER_PULSES in config.h correctly? During the development of the rotary encoder support I've tested mostly ALPS encoders and I haven't got any feedback about issues with specific manufacturers or types yet. The algorithm detects Gray code violations and skips bad steps. A dodgy encoder maybe?

[stj]

Hello people.

for anybody using the EZM GM328A model, i have been running a thread for it here.
http://www.badcaps.net/forum/showthread.php?t=50720

    

[upsss]

Here is my final version.
Original LCR-T4 clone + Rotary Encoder + Protection Relay + Li-Ion battery + Charger with protection circuit + 5V Sep-Up Converter + Banana Sockets + Case.   Total cost <$20 (not including the battery) + at least a day of work, yes I am retired.

[stj]

is that the firmware i posted?
i asked for feedback on that so i could finetune it.