$20 LCR ESR Transistor checker project

[madires]

Will these corrections only affect the display or will they also affect other measurements?
---------------------------
Sorted it out like. I put 150 and the voltage more or less began to coincide. The measurements did not seem to have changed either.

The BAT_* values only affect the battery monitoring, i.e. the battery voltage displayed and the low-battery shutdown.

[Yuriy_K]

In the HilandM644 an extended Frequency and crystal measurements carried out.

Hiland has two menu items for frequency measurements - up to 2 MHz and over 2 MHz. Both measurements are shown in the photo. So there are no problems with the software, you need to look at your diagram ...

[Obelix2007]

Hiland has two menu items for frequency measurements - up to 2 MHz and over 2 MHz. Both measurements are shown in the photo. So there are no problems with the software, you need to look at your diagram ...

Hello Yuriy_K,

the HilandM644 does not cause any problems, it was only given as an example, since it uses an additional prescaler by 16 for the higher frequencies. This is not the case with the T7, but the k software always assumes an additional prescaler by 16 for the Atmega644 MCU and higher frequencies, which then leads to incorrect (multiplied by 16) results after the first measurement.

Due to my limited knowledge of English, I occasionally use the Google translator. This may have led to problems of understanding.

Greetings Horst

[ilcaccillo]

Do you think it's possible or do you have the patience to explain that in other words so an ignorant (not knowledgeable) person like me could understand?

There's a chapter about the inductance measurement (also for the SamplingADC method) in Karl-Heinz' excellent documentation. And for the m-firmware you can find an explanation in inductance.c. It's boring physics. Or see any suitable physics book/text/webpage about inductance.

Still I cant get around to understand how Inductance is measure without an AC signal being use.
I'm having problems understanding the concept

[Obelix2007]

Still I cant get around to understand how Inductance is measure without an AC signal being use. I'm having problems understanding the concept

Hello ilcaccillo

Extract from Karl-Heinz Kübbeler's documentation:

Transistor tester with AVR microcontroller Version 1.13k

Page 44

3.10 Measurement of coils

The normal measurement of the inductance is based on the measurement of the time constant of the
current grow. The detection limit is about 0.01mH, if the resistance of the coil is below 24Ω. For
bigger resistance values the resolution is only 0.1mH. If the resistance is above 2.1kΩ, this technique
can never be used to detect coils. The measurement results of this normal measurement is shown in
the second line (resistance and inductance). With the samplingADC method a resonant frequency
of coils can be detected with greater inductance values. If this effect is noticeable, the frequency and
the quality factor Q of the coil is shown additionally in line 3.
The method of resonant frequency measurement can also be used for the determination of the in-
ductance value, if a sufficient big capacitor mith know capacity value is connected parallel to a little
inductance (<2mH). With a parallel connected capacitor the normal measurement of inductance
can no more operate well. If the resonant frequency let assume a parallel connected capacitor, the
inductance of the normal measurement is not shown and the resistance value is shown in line 1. For
this resonant circuit the quality factor Q is also computed and shown behind the frequency value in
line 3. You can identify this type of measurement with the inductance value at the first position of
line 2, followed by the text “ if ” and the value of the assumed parallel capacity. The value of this
parallel capacitor can currently only be set with the calibration function (1-||- 3 10-30nF(L)).
For displays with only two lines, the content for the third line is shown time-delayed in line 2.

Greetings Horst

[bffargo]

Excellent thread, I learned a lot and tried a lot. I upgraded a GM328A (1.8 color TFT) from 8MHz to 20MHz crystal and the 1.41m mod.  The menus are so much zippier with the faster clock.

I tried to do a similar upgrade to the really old near first M328 that uses a 1602 display. I swapped the 8MHz to 16MHz crystal and installed the "Rev804En(16MHz)" on a fresh 328PU chip. Everything seems to work -- except testing electrolytic caps.  If I have just turned it on after having it off a while, it usually will make it through testing a 10uf 50V 'lytic just fine.  A second press of the button and it makes it 2 to 6 "dots" scanning and then it immediately shuts down. If I power it back on it will keep crashing repeatedly until I put something other than a 'lytic to test.  All ceramic caps test fine.

I have tried it with Alkaline (9.2v), NiMH (9.6v) and LiIon (8.4v) 9V batteries all with the same "crash" results.   Before I put in the faster crystal I had tried the "Rev804En(8MHz)" file at 8MHz clock and it did not crash.  I also tried programming a different chip, same results. The same chips at 8MHz clock and 8Mhz program works fine.   (In fact, I programmed the chip for the 20MHz version for the GM328A mentioned at the top and it had no problems, so I doubt it's the ATMEL itself being able to run at 16 or 20).

What's causing the crash at 16?  Anyone see this on the older units?   (Pic below is from the archive site with all of the precompiled firmware zip's, but my tester looks the same)

[indman]

bffargo,What type of +5V regulator is installed on your M328 clone? Show T1-T7 hardware test results at 16MHz?

[anwfeb]

In the HilandM644 an extended Frequency and crystal measurements carried out.

Hiland has two menu items for frequency measurements - up to 2 MHz and over 2 MHz. Both measurements are shown in the photo. So there are no problems with the software, you need to look at your diagram ...

Can you try for a frequency above 50KHz, for example 55KHz, I started to find this problem when the generator freq number shows a number above 50KHz, and I have tried with 900Khz and it works no error, so to make sure please try again at 55KHz frequency - 100KHz if on hiland644 it works, then this bug is specifically for T7 upgrade atmega644 only, please provide check result information between 55-100KHz.

[madires]

Still I cant get around to understand how Inductance is measure without an AC signal being use.
I'm having problems understanding the concept

This might help:

[Yuriy_K]

Can you try for a frequency above 50KHz, for example 55KHz,... please provide check result information between 55-100KHz.

See the results ...

[anwfeb]

Can you try for a frequency above 50KHz, for example 55KHz,... please provide check result information between 55-100KHz.

See the results ...

Okay thanks Yuriy_K if it's like this then firmware "k" is not suitable for TC7 for Frequency Counter function, because TC7 does not have additional components that exist in hiland644 and the pdf schematic is written in version 1.13k

[Yuriy_K]

pdf schematic is written in version 1.13k

Show this pdf schematic or provide a link ...

[anwfeb]

pdf schematic is written in version 1.13k

Show this pdf schematic or provide a link ...

Page 25, Figure 2.16. Extended Transistor Tester circuit with ATmega644.

[Yuriy_K]

My 3 homemade Ttesters are assembled according to this scheme ...

[madires]

Okay thanks Yuriy_K if it's like this then firmware "k" is not suitable for TC7 for Frequency Counter function, because TC7 does not have additional components that exist in hiland644 and the pdf schematic is written in version 1.13k

I had a quick look at the issue. The precompiler directives for ATmega644 in GetFrequency.c are causing the 16:1 prescaler problem. When using an ATmega644 the "#if"s add the code for switching the external prescaler including auto-ranging. If the frequency is higher than a specific threshold the auto-ranging will enable the 16:1 prescaler. That's what you've noticed for higher frequencies, i.e. the first measurement is fine and after that the frequency is multiplied by 16. Adding another constant (e.g. USE_HFREQ) to the "#if" logic should fix the problem. I'll inform Karl-Heinz.

[anwfeb]

Okay thanks Yuriy_K if it's like this then firmware "k" is not suitable for TC7 for Frequency Counter function, because TC7 does not have additional components that exist in hiland644 and the pdf schematic is written in version 1.13k

I had a quick look at the issue. The precompiler directives for ATmega644 in GetFrequency.c are causing the 16:1 prescaler problem. When using an ATmega644 the "#if"s add the code for switching the external prescaler including auto-ranging. If the frequency is higher than a specific threshold the auto-ranging will enable the 16:1 prescaler. That's what you've noticed for higher frequencies, i.e. the first measurement is fine and after that the frequency is multiplied by 16. Adding another constant (e.g. USE_HFREQ) to the "#if" logic should fix the problem. I'll inform Karl-Heinz.

Thanks madries, I had previously reported to Karl-Heinz via email, but I don't know where the problem is coming from. Thanks for investigating this case and hopefully Karl-Heinz can fix the "k" version soon for the Frequency Counter fix especially for the T7 upgrade 644

[bffargo]

bffargo,What type of +5V regulator is installed on your M328 clone? Show T1-T7 hardware test results at 16MHz?

The test screens aren't numbered. I stitched them all together in order of display.  The regulator is a AMS1117.  I measured 4.97V out using a separate DMM.

Yes, I modified the version string before burning so I would remember it was the 16MHz one.

[indman]

The test screens aren't numbered. I stitched them all together in order of display. 

You showed the calibration results.I asked you to show the T1-T7 hardware test results. Read in the manual (chapter 5.5) how they are displayed. Also compare the test readings and performance of the latest firmware, which I have attached to the message.

[bffargo]

Thank you.  The new ZIP file you attached fixed the problem!  It no longer crashes.

Even so, here are the T1-T7 screens based on the new image in case this is of any use.

[indman]

It is highly desirable for you to pick up 3 accurate 470kOhm resistors. In the T3 test, a difference in readings of no more than 4-5 units is acceptable. You have this difference of 7 units. The T2 test looks more reasonable on your clone.Ideally,for the most reliable measurements, the readings in these tests should be close to 0 and the same.

[bffargo]

Thanks. I tested all of the resistors in circuit (ATMEL pulled) with a couple DMM's with the following results.   So I should just pull a trio of resistors as close to 470K as possible and between each other and replace R2/R4/R6 ?  Any of the others below look far enough off to be of a concern?

Note the issue with R8 that comes off the 9V input to the transistor. It's an order of magnitude off; accd to the schematic posted here; or the schematic is wrong: https://disk.yandex.ru/d/yW8xa5NJgUo5z/M328/Schematic%20diagram/M328.JPG

Code: [Select]

        DMM1 DMM2
R1 680 s/b 680 706.8
R2 474K s/b 470K 474.6K
R3 680 s/b 680 706.2
R4 476K s/b 470K 475.8K
R5 680 s/b 680 706.2
R6 473K s/b 470K 473.8K

R7 8.7K s/b 10K 9.99K
R8 3.29K s/b 33K ??? 3.317K
R9 3.31K s/b 3.3K 3.331K
R10 27.1K s/b 27K 27.16K
R11 99K s/b 100K 99.2K
R12 9.98K s/b 10K 9.999K
R13 3.28K s/b 3.3K 3.324K
R14 558 s/b 560 584.0
R15 26.9K s/b 27K 26.95K

[indman]

  So I should just pull a trio of resistors as close to 470K as possible and between each other and replace R2/R4/R6 ?

Yes!

Note the issue with R8 that comes off the 9V input to the transistor. It's an order of magnitude off; accd to the schematic posted here; or the schematic is wrong:

The schematic is correct,probably a mistake on the circuit board. Your problems with the tester malfunctioning and disconnecting it may be related to this.

[bffargo]

All 3 resistors are now swapped. Here are the results (only showing the screens that were different than the last batch I uploaded).  Some of them flipped between values during the test, I only included the last values shown per test before the next test started.

I think the 3.3K vs. 33K resistor most impacted the raw voltage reading on the battery, it reported closer to actual once that was swapped out. The regulator was still outputting around 5V either way.

Thanks again. Let me know if there are any other concerns I should look into to make it better.  Otherwise the new firmware definitely helped with the crashing first.

Also, where did the new Rev897 zip come from? I didn't see it on the yandex site that was last updated with Rev804 back in May.  Was that just a one off build off the current branch you did for me? If so, thank you!

[anwfeb]

Hi.. how do I adjust the value of the measurement result? I'm using firmware 1.13k and the capacitor measurement results have to increase 0.5pf while the resistor measurement results have to reduce 0.05 ohm

Sample case,
I measure Capacitor 1pf, then the measurement result is 0.51pf

Sample case,
I measured a 0.33 ohm resistor, then the measurement result is 0.38 ohm.

How do I adjust the measure value in code? Calibration has been performed but does not solve the problem.

[indman]

Was that just a one off build off the current branch you did for me?

Yes, this is a special version for your clone!