$20 LCR ESR Transistor checker project

[grossd]

Hi, I would like to clarify the situation regarding the discussed 2.5V reference. I have my AY-AT equipped with a 5V MCP1702-5002 regulator and a 2.5V reference LM4040AIZ-2.5. If I disable the use of the 2.5V reference in the config.h file (//#define HW_REF25), is it also necessary to remove the 2.5V reference from the tester?  Will the tester work properly if the LM4040AIZ-2.5 remains physically connected in the tester?

[madires]

You can leave the LM4040AIZ-2.5 in place. Disabling HW_REF25 simply removes all related code, i.e. the external reference will be ignored. When running the k-firmware you would have to remove the external reference, IIRC.

[CaptainBucko]

The tester turns off, "you (are) cut off", the display goes off during the measurement of capacitors, mostly of large capacity. When measuring through the menu, there is usually no such effect.

This problem is typical for software from Karl-Heinz k-firmware (1.12-1.13k).The software from Markus Reschke (m-firmware) also has this effect.
What do you mean it turns off? That is, for some reason the MK resets or the control program fails. The most likely reasons:
1. The power supply voltage of the MK and/or display disappears for a short time;
2. Power supply of the tester (MK) from a pulsed source with large ripples, interference and noise;
3. Triggering of WDT (watchdog timer) - enable/disable is set by the appropriate fuse before compilation or in the software;
4. BOD activation (supply voltage level monitoring system) - enable/disable, the threshold is set by the corresponding fuses before compilation.
5. MK defective (quite possible, given the strong rise in price of original chips and the slag that the Chinese industry is now trying to “sniff”).

I've recently been on my own journey with the known problem of shutdown during >100uF and <1000uF capacitor measurement. My issue has been on the AY-AT GM-328 clone with genuine Atmel MCU328 (DIP28). I narrowed it down to:
   - Capacitors => 100uF and <= 1000uF. Oddly, >1000uF are fine.
   - Measured using ZIF pins (1-3) or (2-3). Oddly, I could never cause the issue to occur using ZIF pins 1-2.
   - External power supply 12v DC. Varying DC supply voltage made no difference.
   - Using Siglent DSO, I could see no significant voltage drop on Vcc during the measurement.
The MCU that came with this kit (MCU #1) had 1.12k. Upgrading this MCU to 1.53m made no difference. Problem occurred in both firmware's.
Upgrading the supplied 7550 +5v regulator for a MCP1702-5002E, checking and replacing decoupling capacitors and add 10uF Tantalum direct on AVR supply pins also made no difference.
I purchased a 2nd Atmel MEGA328 MCU from Digikey (MCU #2) and it worked flawlessly on the same GM-328 board. BOD = 4.3v
I purchased a 2nd AY-AT GM328 kit, took the Atmel MEGA328 MCU from it (MCU #3) it worked flawlessly on the same GM-328 board. BOD=4.3v
I checked MCU#1 ADC 0,1,2 (ZIF 1,2,3) inputs against MCUs #2 and #3 (in the same PCB) - Diode drops to GND and VCC (A-K and K-A) and resistance to GND and cross compared all MCUs, they were all about the same.
Eventually (and thanks to much information posted in this thread), I realized I set the Extended Fuse to = 0xFC (4.3volts) in MCU #1. Looking at my notes, the kits were delivered with the extended fuse = 0xF9 with the kit (which is invalid but BOD is probably disabled in this state?) and later changed it to 0xFC upon erasing MCU #1 moving from 1.12k to 1.53m.

So now, I can get MCU #1 to work if the BOD (Extended Fuse) BOD is set to FD (2.7 volts).

Cal values for MCU #1, 2 and 3 are attached. MCU #1 definitely looks different - the negative Acomp is a concern. Is MCU #1 is a dud?

The 28-pin MCU IC socket is at risk of loose morals so I will stop here and leave MCU #2 installed for now.

Does it appear MCU #1 is damaged or just

EDIT: Photos of MCU ID for each MCU attached

[Cosmic_Starlight]

The tester turns off, "you (are) cut off", the display goes off during the measurement of capacitors, mostly of large capacity. When measuring through the menu, there is usually no such effect.

This problem is typical for software from Karl-Heinz k-firmware (1.12-1.13k).The software from Markus Reschke (m-firmware) also has this effect.
What do you mean it turns off? That is, for some reason the MK resets or the control program fails. The most likely reasons:
1. The power supply voltage of the MK and/or display disappears for a short time;
2. Power supply of the tester (MK) from a pulsed source with large ripples, interference and noise;
3. Triggering of WDT (watchdog timer) - enable/disable is set by the appropriate fuse before compilation or in the software;
4. BOD activation (supply voltage level monitoring system) - enable/disable, the threshold is set by the corresponding fuses before compilation.
5. MK defective (quite possible, given the strong rise in price of original chips and the slag that the Chinese industry is now trying to “sniff”).

I've recently been on my own journey with the known problem of shutdown during >100uF and <1000uF capacitor measurement. My issue has been on the AY-AT GM-328 clone with genuine Atmel MCU328 (DIP28). I narrowed it down to:
   - Capacitors => 100uF and <= 1000uF. Oddly, >1000uF are fine.
   - Measured using ZIF pins (1-3) or (2-3). Oddly, I could never cause the issue to occur using ZIF pins 1-2.
   - External power supply 12v DC. Varying DC supply voltage made no difference.
   - Using Siglent DSO, I could see no significant voltage drop on Vcc during the measurement.
The MCU that came with this kit (MCU #1) had 1.12k. Upgrading this MCU to 1.53m made no difference. Problem occurred in both firmware's.
Upgrading the supplied 7550 +5v regulator for a MCP1702-5002E, checking and replacing decoupling capacitors and add 10uF Tantalum direct on AVR supply pins also made no difference.
I purchased a 2nd Atmel MEGA328 MCU from Digikey (MCU #2) and it worked flawlessly on the same GM-328 board. BOD = 4.3v
I purchased a 2nd AY-AT GM328 kit, took the Atmel MEGA328 MCU from it (MCU #3) it worked flawlessly on the same GM-328 board. BOD=4.3v
I checked MCU#1 ADC 0,1,2 (ZIF 1,2,3) inputs against MCUs #2 and #3 (in the same PCB) - Diode drops to GND and VCC (A-K and K-A) and resistance to GND and cross compared all MCUs, they were all about the same.
Eventually (and thanks to much information posted in this thread), I realized I set the Extended Fuse to = 0xFC (4.3volts) in MCU #1. Looking at my notes, the kits were delivered with the extended fuse = 0xF9 with the kit (which is invalid but BOD is probably disabled in this state?) and later changed it to 0xFC upon erasing MCU #1 moving from 1.12k to 1.53m.

So now, I can get MCU #1 to work if the BOD (Extended Fuse) BOD is set to FD (2.7 volts).

Cal values for MCU #1, 2 and 3 are attached. MCU #1 definitely looks different - the negative Acomp is a concern. Is MCU #1 is a dud?

The 28-pin MCU IC socket is at risk of loose morals so I will stop here and leave MCU #2 installed for now.

Does it appear MCU #1 is damaged or just

to add to this issue

i also have well more had the same issue if not pretty similar on another AY-AT unit mine being a factory 16Mhz unit running 1.13K on its supplied MCU

i have this 2200uf electrolytic capacitor that upon trying to meassure it the unit will shut down 100% of the time no matter if its ZIF 1-3, 1-2, 2-3, (ran the test 60 times 20 on each pin combination discharching the cap fully in-between measurements) other lower capacitor values were fine

i tried a higher psu voltage and behaviour was the exact same, 7.0v 9.0v, 12.3v nothing worked

swapping that mcu to another unit the fault followed it with the exact same behaviour no change at all

i also suspect the Brownout detection, however i cant change the EX fuses due to the lockbits preventing any sort of change

replacing it for a new mcu fixed the issue, its currently running 1.54m without a problem using the supplied parts HT7550 and the others and no hardware modification whatsoever and i can test that same 2200uf cap without any problem

[indman]

Does it appear MCU #1 is damaged or just

Note the resistance readings of the Ri- Ri+ I/O ports!!!
It differs significantly from the readings of MCU2 and MCU3. I and many other users have experienced this problem many times.
Your MCU1 is a newer generation controller, it does not mean that it is faulty, but the zone of stable operation of such controllers without failures and BOD reset is much narrower than that of older generation controllers. Therefore, it is highly desirable to set BOD=2.7V.
Perhaps madires will explain this effect in more detail.

[madires]

There isn't much I could add. The first things to try are checking if the LDO is able to provide sufficient current and to add an additional buffer cap (electrolytic) to the 5V rail. If the power supply is fine then try a lower BOD voltage level. Some ATmegas are a bit finicky, i.e. you need to set a lower BOD level or shouldn't clock them beyond 16 MHz. I had fun with that too.

[genix]

I have GM328A kit with such MCU + bought separately 2 spare MCUs as well.
I found that it crashes in DischargeProbes() because this newer MCUs (all 3 of mine) has lower internal resistance so current "optimistic" 800mV pull-to-ground threshold is simply too high and causes overcurrent protection / shutdown, it is also more likely to reproduce with hi-value low-esr capacitors, this threshold must be lowered to 400mV or less - then it works fine without any changes in BOD fuses etc.
It has nothing with VR/power delivery.
All my samples works flawlessy at 20MHz as well

[madires]

That's interesting! What are the values of RiL and RiH for your ATmegas?

[indman]

I found that it crashes in DischargeProbes() because this newer MCUs (all 3 of mine) has lower internal resistance so current "optimistic" 800mV pull-to-ground threshold is simply too high and causes overcurrent protection / shutdown, it is also more likely to reproduce with hi-value low-esr capacitors, this threshold must be lowered to 400mV or less - then it works fine without any changes in BOD fuses etc.

Thank you, this information will be very useful for many new generation MCU owners. I correctly understood that you changed the value of 800 to 400 in the Probes.c file?
else if (U_c < 800)                 /* extra pull-down */

[genix]

I'm separated Ri for each probe plus measured resistance in both directions with average, also binned and re-soldered R1-R6 (it was 1%+ originally) for <0.05% spread.
For current MCU sample:
Ri- 12.2 12.8 12.8 (87 91 91 mV) - very consistent
Ri+ 16.6 16.0 16.0 (119 114 114 mV) - very consistent
R0: ~0.18 - +/- 0.03, 2-3 tends to be lower

else if (U_c < 800) /* extra pull-down */

this should be changed to < 400

[indman]

this should be changed to < 400

Yes, this is understandable. And in K-Firmware the same threshold of 800?

[madires]

The threshold for the k-firmware is currently 1000mV (EntladePins.c).

[indman]

I found that it crashes in DischargeProbes() because this newer MCUs (all 3 of mine) has lower internal resistance so current "optimistic" 800mV pull-to-ground threshold is simply too high and causes overcurrent protection / shutdown, it is also more likely to reproduce with hi-value low-esr capacitors, this threshold must be lowered to 400mV or less - then it works fine without any changes in BOD fuses etc.

I still don't quite understand one thing. If this is a overcurrent protection trip, then the same protection trip should occur when we measure the capacitor parameters through the menu. After all, the discharge procedure is similar. And this effect is not there. Why?

[genix]

I still don't quite understand one thing. If this is a overcurrent protection trip, then the same protection trip should occur when we measure the capacitor parameters through the menu. After all, the discharge procedure is similar. And this effect is not there. Why?

I didn't make screenshots from my DSO but in dedicated menu there is zero voltage + [discharge + test with predictable polarity + discharge] + long wait to get rid of any transition effects, then repeat.
In auto-mode there is multiple rapid charge-discharge cycles with opposite polarity then [discharge + test with predictable polarity + discharge].
If MCU crashes and reboots - then it may crash again even on first DischargeProbes() before doing anything - trying to discharge cap from its previous "life".
For me it happens with solid-polymer caps, I couldn't remember any issues with regular liquid electrolytes so I guees that low-ESR + chemistry (high dielectric absorption?) matters somehow

[Yuriy_K]

I will add my measurements to the topic discussed on this page.
I have no problems with measuring large capacitors on ATMega328, purchased earlier than last year. But with a batch of 10 pieces purchased in the middle of last year, there were problems. Low readings of R0, Ri_Hi, Ri_Lo do not allow Testing to pass.. - after this message, the Tester automatically switches off. This is for the range > 10 μF and < 50 μF. If you measure after any capacitor that has passed testing, then capacitors of the specified range are measured normally. The problem is in CheckPins.c, but it is problematic to search there. The solution is to reduce the BOD to 1.8v (FE) for now, after that I have not found any problematic capacitors in my stocks.

[phreich]

Hello,

I did a stupid and accidentally connected lead1 and lead3 to a 12v battery and the tester is now dead.

I looked over all the components and didn't see any visibly "fried", but obviously something was damaged.

When something like this happens to these, is there a specific component that usually is damaged that can be replaced?

I have the following version,

Here's the back:

[indman]

When something like this happens to these, is there a specific component that usually is damaged that can be replaced?

Remove parts Q5 and P6KE6V8 from the board and try to turn the tester back on. If it doesn't work, you are out of luck and the ATmega328 microcontroller is burned out.

[Cosmic_Starlight]

Hello, so ive been running 1.54m @16Mhz on my AY-AT without a problem everything works as expected i just came looking mostly for some clarification of some specific parameters on the config files

so on makefile i changed settings accordingly to its user manual, however reading on here about compiler optimizations i enabled 3 of them on the make file namely

MakeFile

Code: [Select]

OPTIMIZE_VECTORS = yes
CFLAGS += -flto
LDFLAGS += -Wl,-relax

and from what i remember madires saying, is that you should be wary playing with other optimizations not defined in the makefile because they could alter test resutls,
but would these 3 settings together change the results at all ?

also on config.h i changed some setting as suggested to optimize on the user manual and i could get it to read the wall adapter pretty accurately and lowered the shutdown voltage enough to be able to power it from a smartphone charger should i ever need to (highest quality psu i can get right now)

anyway on this config file i had doubts about these

Config.h

Code: [Select]

#define UREF_25
#define HW_REF25
#define UREF_VCC

so i was taking measurements with the DMM as suggested and i noticed that the VCC voltage on the HT7550 is much more stable than the VREF of the TL431A
(i know and have seen scope captures that it dips during measurementss and my dmm cant read that)

because even the device being idle on the main menu the vref voltage is fluctuating while the vcc voltage is pretty much rock solid (no more than a 0.001 variance) while vref was varying 0.02+-
so idk if i did more harm that good by disabling vref on fw without a scope measurements

even then with vref disabled measuring caps is very very stable, measuring a 1uf polyester cap doesnt vary more than 0.001 and same for esr 0.01 at worst (i think with vref enabled it was almost if not the same results)

also two other questions

the manual mentions a contrast setting being available for SOME lcd modules being available to change with the rotary encoder,
i assume this isnt available for the ST7735 the AY-AT uses ?

and also last question is there a slightly thicker font available ? i mean closer to whatever font K firm used, maybe quite not so thick but it seems like the default one is like 1 pixel width at best

also thank you all for your continued development of this firmware am very happy with it

[madires]

and from what i remember madires saying, is that you should be wary playing with other optimizations not defined in the makefile because they could alter test resutls,
but would these 3 settings together change the results at all ?

So far I haven't noticed or heard about any problems with those. The problem with some optimizations is that they optimize specific code critical for a measurement, e.g. changing the timing.

even then with vref disabled measuring caps is very very stable, measuring a 1uf polyester cap doesnt vary more than 0.001 and same for esr 0.01 at worst (i think with vref enabled it was almost if not the same results)

The external 2.5V reference makes only sense when it is at least 10 times more precise than the LDO. With a stable LDO you can measure its output voltage and update UREF_VCC.

the manual mentions a contrast setting being available for SOME lcd modules being available to change with the rotary encoder,
i assume this isnt available for the ST7735 the AY-AT uses ?

Yep. Colour displays usually don't support a direct contrast setting. But you could play with the charge pump voltages (power control registers).

and also last question is there a slightly thicker font available ? i mean closer to whatever font K firm used, maybe quite not so thick but it seems like the default one is like 1 pixel width at best

Have you tried FONT_10X16_HF and FONT_10X16_ISO8859_2_HF?