$20 LCR ESR Transistor checker project

[madires]

Is there any way for the component checker be adapted to identify the "foil side" of an un-polarized capacitor?

It seems important to know the orientation of the foil side so it can be attached to the low impedance side to help diminish noise pickup.

This is mentioned in Mr Carlson's Lab video:

Could be possible, but would require to earth the circuit's ground. The tester can measure down to 1mV and swap polarity on a DUT. And the ADC should be fast enough to sample 50/60Hz hum.

[madires]

i finally got some time to assembly this AY-AT tester. it was fun enough for me as soldering-newbie. 
but when powered it up, it didn't function normally.
after displays "turn-off, transistor & frequency" menu at first start (rotary encoder works for turn it on, but not functioning at this stage), it goes to "V-ext" measurement automatically and stays there until the power jack is unplugged.

does anyone has experienced this problem or have any clue about it?

If the ATmega came pre-programmed, I'd suggest to check the circuitry for the rotary encoder, e.g. the resistors, any shorts or bad encoder.

[hapless]

Is there any way for the component checker be adapted to identify the "foil side" of an un-polarized capacitor?

It seems important to know the orientation of the foil side so it can be attached to the low impedance side to help diminish noise pickup.

This is mentioned in Mr Carlson's Lab video:

Could be possible, but would require to earth the circuit's ground. The tester can measure down to 1mV and swap polarity on a DUT. And the ADC should be fast enough to sample 50/60Hz hum.

How about just using one of the pins to generate a signal and using that pin's connector as a sleeve into which a capacitor would be inserted? No need for grounding then and no need to use one's body as an antenna, but I wonder if it's going to be strong enough to be measured while weak enough not to penetrate the foil too strongly so that the difference on the pins can be measured reliably.

[madcat]

i finally got some time to assembly this AY-AT tester. it was fun enough for me as soldering-newbie. 
but when powered it up, it didn't function normally.
after displays "turn-off, transistor & frequency" menu at first start (rotary encoder works for turn it on, but not functioning at this stage), it goes to "V-ext" measurement automatically and stays there until the power jack is unplugged.

does anyone has experienced this problem or have any clue about it?

If the ATmega came pre-programmed, I'd suggest to check the circuitry for the rotary encoder, e.g. the resistors, any shorts or bad encoder.

please excuse my stupid question as i'm a newbie, but how to check for bad encoder?
does it by in-circuit check or i have to desolder it first?

[rastro]

Is there any way for the component checker be adapted to identify the "foil side" of an un-polarized capacitor?

It seems important to know the orientation of the foil side so it can be attached to the low impedance side to help diminish noise pickup.

This is mentioned in Mr Carlson's Lab video:

Could be possible, but would require to earth the circuit's ground. The tester can measure down to 1mV and swap polarity on a DUT. And the ADC should be fast enough to sample 50/60Hz hum.

How about just using one of the pins to generate a signal and using that pin's connector as a sleeve into which a capacitor would be inserted? No need for grounding then and no need to use one's body as an antenna, but I wonder if it's going to be strong enough to be measured while weak enough not to penetrate the foil too strongly so that the difference on the pins can be measured reliably.

Couldn't the test look for noise level (predominately 50/60Hz) differences between the two pins connected to the component even without a ground reference?   

[hapless]

please excuse my stupid question as i'm a newbie, but how to check for bad encoder?
does it by in-circuit check or i have to desolder it first?

Check the encoder's button for short circuit first. Desolder if there is continuity across its pins when it's not pressed.

Couldn't the test look for noise level (predominately 50/60Hz) differences between the two pins connected to the component even without a ground reference?   

The original design (not the one I proposed) relies on grounding to create shielding effect and attenuate the noise you pick up. Without proper grounding, your device's ground may make an even better antenna than your body, so the results would be wrong or inconclusive.

[madcat]

please excuse my stupid question as i'm a newbie, but how to check for bad encoder?
does it by in-circuit check or i have to desolder it first?

Check the encoder's button for short circuit first. Desolder if there is continuity across its pins when it's not pressed.

continuity only occurs between two big pins at side of the encoder (holder pins?). no continuity at 5 other pins.
if the encoder pressed, the two pins at top, which facing toward the LCD, has continuity.

is that looks ok?

[hapless]

Seems fine. Someone else had a similar problem and solved it (in post 2546) by changing transistors. Make sure that your transistors are all soldered with the correct pins in the right places. If they are, maybe one is faulty.

[madcat]

thank you. i'll check the transistors. 

[1413]

The thermal protection of 7550 kicks in and makes noise on the Vcc. You will find it hot when use 12v source but quite warm with 9v source. I use 2x 3.7 Li-ion battery connect in series, but found it troublesome with recharge circuit.
I am on process of modifying firmware based on 1.12k to work with single 3.7 li-ion without need of DC-DC step up circuit. Already got some progresses. But now I want to change xtal to 16MHz for the TFT draws faster.

PS: Picture below is my firmware *development kit*  LOL, which 8 ic sockets stack on each other for easy removing the atmega328p

[madires]

The thermal protection of 7550 kicks in and makes noise on the Vcc. You will find it hot when use 12v source but quite warm with 9v source. I use 2x 3.7 Li-ion battery connect in series, but found it troublesome with recharge circuit.
I am on process of modifying firmware based on 1.12k to work with single 3.7 li-ion without need of DC-DC step up circuit. Already got some progresses. But now I want to change xtal to 16MHz for the TFT draws faster.

With 3.7V the tester won't be able to check as much components as with 5V. Just saying

[hapless]

The thermal protection of 7550 kicks in and makes noise on the Vcc. You will find it hot when use 12v source but quite warm with 9v source. [...]

Sounds like your LDO is oscillating. Try adding capacitors or change the regulator to something like LM7805 for stable operation. Thermal protection would shut off the circuit, not add noise to it. It wouldn't be much of a protection if it just sat there making noise. This is one of those cases where it makes much more sense to change the hardware than the software.

[1413]

The thermal protection of 7550 kicks in and makes noise on the Vcc. You will find it hot when use 12v source but quite warm with 9v source. I use 2x 3.7 Li-ion battery connect in series, but found it troublesome with recharge circuit.
I am on process of modifying firmware based on 1.12k to work with single 3.7 li-ion without need of DC-DC step up circuit. Already got some progresses. But now I want to change xtal to 16MHz for the TFT draws faster.

With 3.7V the tester won't be able to check as much components as with 5V. Just saying

Make sense

Anyway, I recompiled the firmware with F_CPU = 16Mhz.

Turn out 7550 sucks as V-Out to be 6v!!! That's why m328p acted weird when I changed 8Mhz to 16Mhz crystal.

I still have no idea why just simple replacement of xtal can make change in voltage.

So, I bend middle leg of 78L05 to the side (into 2-3-1) and replace the trouble maker 7550. Now it gets 4.93v VCC on the TFT screen and similar value on my voltmeter.

Finally it works at 16Mhz, and noticeably faster in rendering text on the screen.

Edit: with extra cable to the TFT, it sometimes misses / misplaces some dots, thus the data sent to TFT is not integrity.

Sounds like your LDO is oscillating. Try adding capacitors or change the regulator to something like LM7805 for stable operation. Thermal protection would shut off the circuit, not add noise to it. It wouldn't be much of a protection if it just sat there making noise. This is one of those cases where it makes much more sense to change the hardware than the software.

Yeah, you're right. It also heat up fast because of differentiate in input voltage and output voltage, as it has to take out 7v out of 12v and transfer it into heat. Maybe 7v different is too much for 7550.

[upsss]

I have been using an 18650 Li-Ion battery (which lasts for ever) with exactly the same charging/protection board and step up converter as you are. 

The only thing I set the step up converter to 5V output and eliminated completely the on board original 5V linear regulator so everything is powered from the step up 5V output without any problems.  Se my original post, #3069.

[belzrebuth]

I find this tool highly useful for measuring ESR even in circuit but one thing that really bothers me is those grabber leads.

While they're okay for out of circuit component testing they're a pain for in circuit quick measurements.

Mine does not have a case so banana terminals is not an option.

I'm thinking to  buy a cheap set of multimeter probes cut off the shrouded banana plugs and solder the leads directly to the 1 and 3 pins of the tester.
Re-calibrating the tester to zero out any resistance/capacitance of the longer wires should do the trick ,right?
Another problem is that the pin holes are too small to fit a 12-14awg wire so I'll have to improvise a bit there..


Has anybody tried something similar?
I think that some kind smd tweezers or needle type leads would really make things easier for repair jobs..

[pepe10000]

I find this tool highly useful for measuring ESR even in circuit but one thing that really bothers me is those grabber leads.

While they're okay for out of circuit component testing they're a pain for in circuit quick measurements.

Mine does not have a case so banana terminals is not an option.

I'm thinking to  buy a cheap set of multimeter probes cut off the shrouded banana plugs and solder the leads directly to the 1 and 3 pins of the tester.
Re-calibrating the tester to zero out any resistance/capacitance of the longer wires should do the trick ,right?
Another problem is that the pin holes are too small to fit a 12-14awg wire so I'll have to improvise a bit there..


Has anybody tried something similar?
I think that some kind smd tweezers or needle type leads would really make things easier for repair jobs..

Something like that ?:





A greeting.

[Stecor]

Hello All, thanks for the great wealth of info thus far.

I also have a T4 Chinese clone, which I "upgraded" with a rotary encoder and upgraded the firmware to 1.12k (using a custom ISP cable connected to an USBASP adapter).

The clone has SMT resistors with 1% accuracy, the PCB is okay-ish, including a 2.5% Voltage ref and it's compliant to the newer version of the schematic.

I have the following questions:
1. After downloading the trunk, I could use the default MEGA328-T3-T4 .hex and .eep to upload the firmware, and it sort of works:
- the tester is now functional including the rotary encoder I added.
- calibration can be performed successfully, but:
Issues/Concerns:
Issue 1: Capacitor ESR is off - e.g. capacitors I used to read with a 0.1 - 0.2 ohm ESR now show as 0.00 Ohm ESR
Adding a .7 Ohm resistor in series with a 0.00 ESR capacitor yields a .35 Ohm ESR (that is off by .35 Ohm!!!).

Calibration was performed multiple times with both a "known good" very low ESR 2500 uF electrolytic as well as a 2uF ceramic (unknown quality). Really crappy caps (which used to read 2-3 Ohm, still show up with appreciable ESR, albeit I believe smaller)

Issue 2: Sometimes (especially when turning the rotary encoder during a measurement) I get the "Not Calibrated" message. If I restart the meter, the calibration is OK. Going in the menu, I can see the calibration data

2. I looked at the software, and in the Makefile, I could see the following config: CFLAGS += -DNO_AREF_CAP which is basically saying there is no 1nF cap at Aref. My board has one, so I would like to comment out the above. I could take the cap out, but it seems like a good thing to have, so would rather keep it, even if it slows down measurement some.

Issue is that, building the project with WinAvr yields 107% !!!! Flash utilization, thus I cannot flash the FW. I used the -O3 and -Os options, but no dice.
I have also downloaded the updated Atmel AVR toolchain and updated my WinAvr to the latest toolset, however in that case, I get an error related to EEPROM defines (__eerd_m328 or similar). I believe this is due to the EEPROM access functions.

Thus, please can you help me with the following:

1. Help troubleshoot the cause of the ESR discrepancy I see with the meter
2. Please point me in the right direction in order to compile the FW successfully (within the 32k of Flash available )


Thank you!!!

If I can provide any additional info, please let me know, thank you!

[madires]

If the 2.5v reference is just a TL431, please replace it with a LM4040 or LM285. If the 5V LDO is a MCP1702-5002 or another one with similar specs remove the TL431 without replacing it. The Transistor Tester is no precision instrument, but a resistor with 0.7 Ohms is measured correctly in most cases if the tester is adjusted. The best cap for the adjustment procedure is a film cap with 220nF up to a few µF. Are there any contact issues with the ZIF or probe leads? Those could cause low measurement values to be off a little bit. The issue with the rotary encoder is strange, since the check for an adjusted tester is based on a value stored in the EEPROM. The external cap at the AREF pin is recommended and its capacitance has an inpact on some timing of the ADC when switching voltage references. For a 1nF cap the k-firmware doesn't require to be told about that in the Makefile. But it's also fine to enable it. Older versions of avr-gcc are known to create large binaries. There are several posts in this thread about this and you'll find recommended versions in those posts. Could you please post the exact compiler error regarding the EEPROM functions?

[madcat]

Seems fine. Someone else had a similar problem and solved it (in post 2546) by changing transistors. Make sure that your transistors are all soldered with the correct pins in the right places. If they are, maybe one is faulty.

so i've desoldered the transistors and checked them with DMM.
9012 & 9014 looks fine (refers to some how to check transistor is bad/good videos on youtube).
as 7550 and TL431 are not transistors, i don't know how to check them.

i didn't desolder TL431 yet, but with 7550 i have these readings with DMM (diode test function):
* Vin  --> Vout  = 1070 ohm
* Vout --> Vin   = 560 ohm
* Vout --> GND   = 0L
* GND  --> Vout  = 825 ohm
* Vin  --> GND   = 740 ohm
* GND  --> Vin   = 0L

could that readings be used to tell if 7550 is bad/good?

[Stecor]

If the 2.5v reference is just a TL431, please replace it with a LM4040 or LM285. If the 5V LDO is a MCP1702-5002 or another one with similar specs remove the TL431 without replacing it. The Transistor Tester is no precision instrument, but a resistor with 0.7 Ohms is measured correctly in most cases if the tester is adjusted. The best cap for the adjustment procedure is a film cap with 220nF up to a few µF. Are there any contact issues with the ZIF or probe leads? Those could cause low measurement values to be off a little bit. The issue with the rotary encoder is strange, since the check for an adjusted tester is based on a value stored in the EEPROM. The external cap at the AREF pin is recommended and its capacitance has an inpact on some timing of the ADC when switching voltage references. For a 1nF cap the k-firmware doesn't require to be told about that in the Makefile. But it's also fine to enable it. Older versions of avr-gcc are known to create large binaries. There are several posts in this thread about this and you'll find recommended versions in those posts. Could you please post the exact compiler error regarding the EEPROM functions?

Thank you! I appreciate your taking time taking to noobs like me

First of all, yep, you nailed it completely - the ZIF went partially bad... I did not notice it right away, but the "main" 1-2-3 set (the one that is used most frequently) has a bent contact, hence the higher resistance value. Re-doing the tests in the un-damaged pins cleared the issues. I need to source a proper ZIF socket now.

Thanks for the guidance on regulators - unfortunately, both are less than stellar - the 2.5 reference is a 431, and the 5V is an 78L05. I guess next step is replacing them both.

On the compiler issue, I did get an issue with Avast (I use the free version) and it did complain about potentially malicious content during the install, so I simply disabled the AV shield and re-did the install - and marvel of marvels - it works beautifully... compiles with no issues @95% or so.

On the precision matter you mentioned, I still think it's a pretty useful tool. Unless I am grossly mistaken, you are the Markus that forked the trunk, so thank You, Karl-Heinz and the other Markus too   The main use case for me is capacitance & ESR (mostly ESR) and for capacitance I find it accurate enough for my needs by comparing to the other meters I have. ESR is the Godsend, since this is what I bought it for. BTW - if it would be possible to measure ESR at different frequencies, that would again be pretty useful - 100KHz to 1MHz would be awesome!!!


Thanks again and have a great day!

[Stecor]

so i've desoldered the transistors and checked them with DMM.
9012 & 9014 looks fine (refers to some how to check transistor is bad/good videos on youtube).
as 7550 and TL431 are not transistors, i don't know how to check them.

i didn't desolder TL431 yet, but with 7550 i have these readings with DMM (diode test function):
* Vin  --> Vout  = 1070 ohm
* Vout --> Vin   = 560 ohm
* Vout --> GND   = 0L
* GND  --> Vout  = 825 ohm
* Vin  --> GND   = 740 ohm
* GND  --> Vin   = 0L

could that readings be used to tell if 7550 is bad/good?

TL431 is a ref voltage supply, so you can easily check it in circuit - provided you have a decent voltmeter - you simply measure the voltage between PD4 and ground; it should be pretty close to 2500mV - WITH the tester working (when shut down, the voltage is 0...). Obviously, a 1% DMM is a must there... I am not sure what the 7550 is, but if it's the supply regulator, again, you can check the Vcc rail voltage... 5000mV +/- 1% and you should be fine.

As Madires mentioned above, you may wish to look for a better spec supply regulator and/or reference supply.

[1413]

Issue is that, building the project with WinAvr yields 107% !!!! Flash utilization, thus I cannot flash the FW. I used the -O3 and -Os options, but no dice.
I have also downloaded the updated Atmel AVR toolchain and updated my WinAvr to the latest toolset, however in that case, I get an error related to EEPROM defines (__eerd_m328 or similar). I believe this is due to the EEPROM access functions.

I use the cygwin + avr compiler comes with arduino ide for compiling and Avrdudess for fuse burning and firmware uploading

[hapless]

so i've desoldered the transistors and checked them with DMM.
9012 & 9014 looks fine (refers to some how to check transistor is bad/good videos on youtube).
as 7550 and TL431 are not transistors, i don't know how to check them.

i didn't desolder TL431 yet, but with 7550 i have these readings with DMM (diode test function):
* Vin  --> Vout  = 1070 ohm
* Vout --> Vin   = 560 ohm
* Vout --> GND   = 0L
* GND  --> Vout  = 825 ohm
* Vin  --> GND   = 740 ohm
* GND  --> Vin   = 0L

could that readings be used to tell if 7550 is bad/good?

The voltage regulator is an IC, and the way to tell if it's working is to measure the voltage at its output. It should be around 5V. Now that you have all the transistors out of the circuit, put back the regulator and short out the holes of where the emitter and collector of your 9012 would be. If I'm not mistaken, that's your T3. The tester should turn on and go through a normal test cycle. It should not go into menu or show V_ext. It will not react to any button presses in this state. Let me know how it goes.

[hapless]

2. I looked at the software, and in the Makefile, I could see the following config: CFLAGS += -DNO_AREF_CAP which is basically saying there is no 1nF cap at Aref. [...]

The name is misleading. This is how it should be for as long as you don't have the commonly used 100nF capacitor there. Could as well have been something like -DMUCH_SMALLER_THAN_100nF_OR_NO_AREF_CAP. If you get rid of that flag, it would assume 100nF (i.e., you'll be shooting yourself in the foot).

[madcat]

so i've desoldered the transistors and checked them with DMM.
9012 & 9014 looks fine (refers to some how to check transistor is bad/good videos on youtube).
as 7550 and TL431 are not transistors, i don't know how to check them.

i didn't desolder TL431 yet, but with 7550 i have these readings with DMM (diode test function):
* Vin  --> Vout  = 1070 ohm
* Vout --> Vin   = 560 ohm
* Vout --> GND   = 0L
* GND  --> Vout  = 825 ohm
* Vin  --> GND   = 740 ohm
* GND  --> Vin   = 0L

could that readings be used to tell if 7550 is bad/good?

The voltage regulator is an IC, and the way to tell if it's working is to measure the voltage at its output. It should be around 5V. Now that you have all the transistors out of the circuit, put back the regulator and short out the holes of where the emitter and collector of your 9012 would be. If I'm not mistaken, that's your T3. The tester should turn on and go through a normal test cycle. It should not go into menu or show V_ext. It will not react to any button presses in this state. Let me know how it goes.

ok, i've put back 7550 to its place and leave 9014 (2pcs) and 9012 out of circuit, then short out the holes of emitter and collector of 9012 (it should be the outer legs of 9012, right?) but nothing happens, it doesn't turn on. 

i just noticed that the board has top & bottom traces. does that mean the solder tin should goes through bottom-up on each component's pin?