$20 LCR ESR Transistor checker project

[indman]

Feliciano,Now your clone shows adequate results on both firmware versions.
You should not expect from a simple tester that the measured capacity will exactly match what is written on the marking (I think these are truisms and there is no need to further explain the reasons).

[indman]

I think these results are acceptable, and maybe it can be fine-tuned with compensation for each of the 3 ranges. Still, there's some difference between normal mode and monitoring mode. Being the same algorithm, I don't see the reason.

Yes, they can be easily adjusted if necessary!
As for the difference when measuring capacitance in AUTO mode and monitor mode, madires has already explained this behavior earlier, and also made a note in the manual(hint). I also think that the slight difference in results is most likely due to the fact that AUTO mode involves a lot more preliminary checks and steps before measuring capacitance. And in monitor mode, the algorithm is immediately configured to measure a certain parameter. But it is a little surprising that the first measurement in monitor mode is always very close to the value obtained in AUTO mode, but then decreases.

[Feliciano]

Yes, I mostly agree with you. I'm only elaborating because this is a public discussion, and different readers are at different levels.

Anyways, I backported my 1.51m to 1.50m, and it works, so something on my configuration or my build environment breaks 1.51m monitoring. At least now we know is something like that, and it will require further reviewing.

Regarding my measurements, Normally the first value is higher than expected, the second lower, and then the others are intermediate.

Just for reference, my 1000uF on 1.50m, with my same settings, measures:

• on normal mode: 1033uF, then 1023uF, then 1026uf or so
• on RCL monitoring: 1029uF, then 991uF, then 998uF or so

I also noticed that if I short the leads in LCR monitor mode, then it reads 0.05R 8.36uH.

And a short is still measured as 8uH on both modes.

[indman]

Yes, I mostly agree with you. I'm only elaborating because this is a public discussion, and different readers are at different levels.

This is true, but nevertheless, if a person is involved in amateur radio and he needs this device not only as a beautiful toy, then he must have basic knowledge and concepts in electronics. And it's never too late to learn

[Feliciano]

We agree again.

To rule-out my W10 build environment, I took indman's config and made 1.51m, with the result of LCR monitoring broken too. So it's not my config files, it's my setup. Hence I ask:
indman, what toolchain are you using?
With the default toolchain of WinAVR-20100110, I can build a hardware SPI (which we don't use for this), but I cannot make your config files fit (105% vs your 99.6%), and the LCR monitoring gets broken on 1.51m -but not on olders-.
And with the avr8-gnu-toolchain-installer-3.5.1.87 I cannot make a hardware SPI (the compilation errors I reported before).

It may be related with:

v1.51m 2023-12
- Changed GetThirdProbe() into convenience function UpdateProbes2() to reduce firmware size and adapted calls in various other functions (suggested by Viktor Klimkovich).
- Optimized DataStorage() and ManageAdjustmentStorage() to reduce firmware size (thanks to Viktor Klimkovich).
- Added option to slow down bit-bang SPI in case of a high MCU clock rate (SPI_SLOWDOWN, suggested by Viktor Klimkovich).
- Changed calculation of delay loop in IR_Detector() to support MCU overclocking (thanks to Viktor Klimkovich).
- Improved voltage drop algorithm in function LargeCap().
- Function LargeCap() takes now also negative zero offset into account.

[indman]

Feliciano, I am using WinAVR-20100110 + AVR-GCC 8.3.0.
Use the search, not so long ago in this thread we discussed the most favorable combination of software for compilation to get a compact firmware size.

[Feliciano]

Thanks, I remember that discussion, but the conclusion was not clear to me.

I'm trying now that toolchain, downloaded from ZakKemble's github, but it gives me compile errors. I'm troubleshooting what it is.

Code: [Select]

avr-size: invalid option -- C


[indman]

While you are figuring out how to install the software, you can give me your customized configuration files and font, I will compile the 1.51m firmware for you.

[madires]

To rule-out my W10 build environment, I took indman's config and made 1.51m, with the result of LCR monitoring broken too. So it's not my config files, it's my system. Hence I ask:

Have you replaced the 1.51m's cap.c file with the fixed one (repo: cap.c-1.51m.tgz)?

[Feliciano]

As a little give-back to the project, here I present an alternate 8x16 hf western font. I composed it to be heavier and less serif than the existing one, and shifted a bit to the right, so I hope it gets better readability overall.

It's my first font from scratch, so please check it before adding to upcoming releases.

[Feliciano]

Have you replaced the 1.51m's cap.c file with the fixed one (repo: cap.c-1.51m.tgz)?

No, but I can fetch it an try again, once I fix the actual 8.3.0 toolchain incompatibility with the WinAVR2010.

In the meantime, I posted my config files on a previous message, and the custom font is on the previous post. So if somebody can build the binaries we can pinpoint the abnormal C monitoring differences bug. But me must note whether is with the december's cap.c or the current one.

EDIT:
I fixed the toolchain incompatibility, and made 1.51m with the old cap.c and with the new cap.c. And I can confirm now the monitoring bug can be specifically related to 1.51m and the old cap.c.

Special thanks to indman, and also Kim Christensen, 2hry and madires for the support.

[indman]

Feliciano, I'm glad you're over the problem! 
 I compiled the firmware with your font, it's in the archive.
There were a few errors in your config.h file, I fixed them. I also believe that for your clone, a correction of 0(-1%) for large capacitances is optimal. If you set +1.0, the tester will overestimate the capacitance on large capacitors, for example 4700.

[Feliciano]

Tried last indman's build.

And thinking about the compensation, I went to the other extreme, and tried a very large capacitor. I couldn't get a reading in this case: maybe the timeout for LCR monitoring is a tad short?

Note: I will check config's differences later.

[indman]

And thinking about the compensation, I went to the other extreme, and tried a very large capacitor. I couldn't get a reading in this case: maybe the timeout for LCR monitoring is a tad short?

This is already interesting! I have not tested such large capacities in monitor mode. Tomorrow I'll check it on my clones.

[Kim Christensen]

Tried last indman's build.
And thinking about the compensation, I went to the other extreme, and tried a very large capacitor. I couldn't get a reading in this case: maybe the timeout for LCR monitoring is a tad short?
Note: I will check config's differences later.

I tried a 10,000uF and it worked in both modes with the "1.51mEn(8x16)Feliciano firmware". (Read as 11.78mF)
Then I paralleled another 8,200uF and it continued to work in both modes.  (Read as 19.6mF)
Crammed another 3900uF.   (Read as 26.19mF)
Note: My jumper wires did add some extra "ESR" and these, except for the 3900uF, were used parts...

Not sure what the point of your "ESR" test mode is, since it only works with capacitors... ie: RCL monitor does that already. Speed?

Feliciano, I am using WinAVR-20100110 + AVR-GCC 8.3.0.
Use the search, not so long ago in this thread we discussed the most favorable combination of software for compilation to get a compact firmware size.

Thanks. I'll try this instead of continuing with the "whack a mole" compiler error game that I'm currently playing... (Bit of a C compiler noob here, so the binaries and config files are a godsend)
FYI, the "FNIRSI 1.51mEnTestMonitor" binaries also worked fine with my unit in all modes.

EDIT on Mar 5th, 2024:
Here are the changes I made to my ComponentTester-1.51m code's makefile so it compiles using WinAVR-20100110 and AVR-GCC 8.3.0 and works on the TC-1 (FNIRSI-4A10 version) tester.
The LDFLAGS makes the code smaller so it'll fit.
"avr-size" is no longer supported and gave errors, so that line was commented out and "avr-objdump" was added instead.

MCU = atmega324p
FREQ = 16
LDFLAGS += -Wl,-relax
#   @avr-size -C --mcu=${MCU} $<
   @avr-objdump -Pmem-usage ${NAME}

See attached makefile which includes the above changes:

[Feliciano]

Not sure what the point of your "ESR" test mode is, since it only works with capacitors... ie: RCL monitor does that already. Speed?

From the manual:

The ESR tool measures capacitors in-circuit and displays the capacity and ESR if the measurement detects a valid capacitor. Make sure that the capacitor is discharged before connecting the tester! Values could differ from the standard measurement (out-of-circuit) because any component in parallel with the capacitor will affect the measurement.

Note: My jumper wires did add some extra "ESR"

That's why we have the "Adjustement" and "Save" for. In my case, I use the two connection methods by turns, so that's why I select the saved profile at the begining, but other people can prefer to adjust and not to save.

(Bit of a C compiler noob here, so the binaries and config files are a godsend)

From my part, a pleasure to help, and contribute a bit to the developement. And I think we also shall thank indman and the other people for helping support this clone variant, and ironing the code (either with the "say thanks" button, or in written).

BTW: I increased the cycle time, and now I can detect a 22mF. I suppose if we increase it more we could detect larger capacitors (like 44mF), but it would become a trade off on waiting when measuring "normal" values on regular basis.

[madires]

As a little give-back to the project, here I present an alternate 8x16 hf western font. I composed it to be heavier and less serif than the existing one, and shifted a bit to the right, so I hope it gets better readability overall.

It's my first font from scratch, so please check it before adding to upcoming releases.

Thanks! I've corrected the index addresses for the additional characters and removed a comma.

[Feliciano]

Thanks. I updated my original post for further readers.

[carl1961]

For who may want this  a alternative for WinAVR     Transistor-Tester-AVR-Compiled-files-On-Windows
Mainly did this so later when i forget how.

also the github link where anyone can fork and improve.

https://github.com/carl1961/Transistor-Tester-AVR-Compiled-files-On-Windows

[carl1961]

Got the 644 Adapter board in today.. Now the fun begins. Remember anyone who wants one, there free for what a stamp and envelope cost.

[carl1961]

Got the Board soldered and Programmed, having trouble with the 2.8 TFT
The firmware right now is von Boleslaw J. (boleslaw_43) pre compiled Adapter_m328_m644_AY-AT_mod_hard_SPI_firmware 1.49m
@indman can you have share how you connected the 2.8 TFT to AH-AT?

[Yuriy_K]

can you have share how you connected the 2.8 TFT to AH-AT?

Look at the message and at the fragment of the diagram...
https://www.eevblog.com/forum/testgear/$20-lcr-esr-transistor-checker-project/msg4882043/#msg4882043

The link works

[Boleslaw_J]

Got the Board soldered and Programmed, having trouble ...

In order for the microcontroller to work with an external crystal
resonator, the microcontroller's fuses must be properly positioned.
When programming these keys, an external resonator (e.g. 8 or 16 MHz) must be
connected to the microcontroller!

Valid key values are:
Fuses:
L: 0xF7
H: 0xD9
E: 0xFC

Lock bits:
LB: 0xFF  (do not touch!)

[carl1961]

Got the Board soldered and Programmed, having trouble ...

In order for the microcontroller to work with an external crystal
resonator, the microcontroller's fuses must be properly positioned.
When programming these keys, an external resonator (e.g. 8 or 16 MHz) must be
connected to the microcontroller!

Valid key values are:
Fuses:
L: 0xF7
H: 0xD9
E: 0xFC

Lock bits:
LB: 0xFF  (do not touch!)

Thanks Boleslaw_J ,

I already have a 16 MHZ Crystal on the main board, that's why the crystal is missing from your  644 Adapter. Do i need to remove it and install on adapter? I under stood to leave it off the crystal from Adapter if already have one on the board.

[Boleslaw_J]

Thanks Boleslaw_J ,

I already have a 16 MHZ Crystal on the main board, that's why the crystal is missing from your  644 Adapter. Do i need to remove it and install on adapter?

Of course not. The location of the external resonator is not important.