$20 LCR ESR Transistor checker project

[elecdonia]

Aldo22  Check these items from the list, if there are any in your household?

I'll test these transistors and JFETs with my Transistor Testers and post a spreadsheet of my results here. I have large quantities of the parts listed below:

3.Some BJT transistors (for example,2N5089,MPSA12,MPSA13, etc.) are defined as thyristors or resistors.
4.Some JFET transistors (e.g. J112,J113,2N5457, etc.) are not defined at all or are defined as BJT.

I have 3 LCR-style units at this time: One with ATmega324, another with LGT8F328P, and another with APT32F172K8T6.
All have the Chinese firmware they came with. Eventually I'll put different FW into my ATmega324 unit (probably upgrading to ATmega644 at same time) but I haven't done that yet.
I also have 2 AY-AT style units and one GM328A (monochrome graphic LCD). All have Atmel ATmega328P MCU with original Chinese firmware.

[madires]

That's... a disappointingly short range. What's involved in the cleaning up of the signal? (I thought it came as a simple square wave at the output of the comparator.)

The LM311 based LC oscillator circuit has a known issue with the signal starting around 30 nF (spurs at falling/rising edges). Another comparator or a 4093 will do the job. But starting around 120 nF the oscillator becomes unstable. It's no big deal as the LC oscillator is meant for measuring low value capacitance/inductance (below the normal range of the tester).

[elecdonia]

I like DIP because of the ease of swapping out the chip. What would be the advantage of an ATmega324 over ATmega328 to pay a premium for a tester based on it?

ATMega324 has more I/O ports than ATMega328 DIP. Therefore, you simply have more chanses and options for connecting peripheral sources for the tester. I don’t see any more advantages.

My thoughts about selecting MCU chips for scratch-built DIY Transistor Testers and also for upgrading/customizing "factory built" Transistor Testers:

The OSHW Transistor Tester requires the MCU to have GPIO ports which fully support 5V logic levels for both input and output.
There simply aren't many alternatives to the Atmel/Microchip AVR family.
The trend is for newer/faster/larger/cheaper MCU chips to operate only at 3.3V (although some can accept 5V logic signals as inputs on selected pins)
     To review: Since the Transistor Tester generates its test signals directly from MCU GPIO pins this means the MCU must be able to output a full 5V from its GPIO pins.

And with the OSHW Transistor Tester software gradually growing in features and size this makes it desirable to select an MCU with 64k (ATmega644) or 128k (ATmega1284) flash program memory. Fortunately I already have ATmega644 and ATmega1284 on hand in 40-pin DIP packages  I will use them for my scratch-built Transistor Tester.

ATmega324 TQFP-44 (square SMT package) has 44 pins vs. mega328P (TQFP-32) with only 32 pins
-   The additional pins are useful GPIO ports.
-   ATmega324 can be directly replaced by ATmega644. No alteration of PC board is needed.
-   So far none of the $1 USD Chinese MCU chips come in 44-pin TQFP.
-   Therefore a Transistor Tester which contains a 44-pin TQFP MCU is very likely to have a real ATmega324 (or ATmega644) MCU.
-   In contrast, there are at least 2 different $1 USD Chinese MCU which come in 32-pin TQFP packages.
-   This means Transistor Testers with 32-pin MCU chips could have any one of 3 different MCU chips (Atmel ATmega328P, Lucky Green LGT8F328P, or AptChip APT32F172K8T6.

[ayubando]

About to finish an exact build like this, with an atmega 1284, protoboard and an ili9488 3.5 inch screen

[Aldo22]

Aldo22, Check these items from the list, if there are any in your household?
https://www.eevblog.com/forum/testgear/$20-lcr-esr-transistor-checker-project/msg4610314/#msg4610314

Hi!
Could it be that the APT32F172K8T6 version has a problem with lower hFE values (~<50) on BJT transistors?
It also identifies some of my fishy A42 as resistors (4 of 10).
The same could be the case with your MPSA12 and MPSA13.

[elecdonia]

Could it be that the APT32F172K8T6 version has a problem with lower hFE values (~<50) on BJT transistors?
It also identifies some of my fishy A42 as resistors (4 of 10). The same could be the case with your MPSA12 and MPSA13.

I plan to measure the emitter and/or collector currents during the test. This will likely require a small value current sensing resistor (1 to 10 ohms) and my oscilloscope. To the best of my knowledge the duration of hFE (gain) measurement is rather brief. Therefore it can't be measured with a DVM.

When I look at my Transistor Tester results I observe widely differing collector (or emitter) currents reported by the Transistor Testers with Chinese MCU/firmware compared to units with Atmel AVR MCU running OSHW firmware. The gain of a BJT depends quite a bit on the collector current it is measured with. This is especially true for power transistors. Curiously NPN transistors have considerably lower gain at low currents vs. high currents compared to PNP transistors where the hFE is relatively stable over a wide range of collector currents.

I'm getting real close to measuring a whole bunch of different BJT and JFET with several different Transistor Testers.

[gipetto]

sounds like there's some kind of hardware bug with the mcu. the rp2040 for instance had an adc bug, but could be compensated for by notching out some spikes in software. link here although it's off topic.
https://www.hackster.io/news/raspberry-pi-confirms-it-is-investigating-a-flaw-in-the-raspberry-pi-pico-rp2040-adc-95c393b55dfb

[Aldo22]

When I look at my Transistor Tester results I observe widely differing collector (or emitter) currents reported by the Transistor Testers

I only have a TC1 with the APT32F172K8T6 and a Fnirsi DSO-TC2.
The DSO-TC2 reports a range of Ic from ~2.5mA (mostly for NPN) to  ~5.5mA (mostly for PNP) afaics, while the TC1 always shows ~6 mA (see my list above).
If you have a look a these curves, this might explain the big difference in some cases, e.g. for MPSA 42.
https://www.biophysicslab.com/2021/04/27/testing_transistor_hfe/
In most cases, however, 6mA seems to be a reasonable (average) value.

EDIT: Not quite true. The DSO-TC2 sometimes reports Ie.

A42, C1815

[elecdonia]

I only have a TC1 with the APT32F172K8T6 and a Fnirsi DSO-TC2.
The DSO-TC2 reports a range of Ic from ~2.5mA (mostly for NPN) to  ~5.5mA (mostly for PNP) afaics, while the TC1 always shows ~6 mA (see my list above).
If you have a look a these curves, this might explain the big difference in some cases, e.g. for MPSA 42.
   https://www.biophysicslab.com/2021/04/27/testing_transistor_hfe/
In most cases, however, 6mA seems to be a reasonable (average) value. EDIT: Not quite true. The DSO-TC2 sometimes reports Ie. A42, C1815

That URL contains an excellent article about testing BJT gain. The curves shown are typical of NPN devices.

Note: My APT32F172K8T6 TC1 does this:
- Pressing "Start" button 4 times in quick succession (from off) causes my TC1 to display: FNIRSI-TC1 BOOTLOADER.
- It displays FNIRSI-TC1  for ~2 seconds just after entering self-test mode by pressing "Start" with a short between 1-2-3.

I wonder if FNIRSI makes these APT32F172K8T6  TC1 units?
(It is also possible the actual maker "borrowed" the FNIRSI name to try to make their knock-off product look more legitimate.)

[Aldo22]

I wonder if FNIRSI makes these APT32F172K8T6  TC1 units?
I doubt it - I suspect the actual manufacturer "borrowed" the FNIRSI name to try to make their product look more legitimate.
Note: Pressing "Start" button 4 times in quick succession does cause my TC1 to display: FNIRSI-TC1 BOOTLOADER.

Yes, it also shows "Fnirsi-TC1" in self-test mode for me.
I'm not sure either. "Fnirsi" is not written anywhere else, neither on the case nor on the circuit board.
In any case, the DSO-TC2 has a different level of quality.

[.RC.]

I have ordered a through hole DIY kit.  Is it worth it doing the things in this mod just for general use?  https://github.com/blurpy/transistor-tester

[wandows]

is this a full version of the LCR ESR Transistor Tester Project?

[elecdonia]

I have ordered a through hole DIY kit.  Is it worth it doing the things in this mod just for general use?  https://github.com/blurpy/transistor-tester

That’s a very good set of recommendations. Note that it begins by explaining the importance of matching the probe resistors. Especially the three 680 ohm probe resistors. They don’t need to be precisely 680 ohms, but they need to match each other within +/- 0.5 ohms. The three 470k ohm probe resistors also need to match each other. Unfortunately the resistors which come with these kits usually aren’t matched. Replacing these six resistors with high-quality 0.1% precision resistors is a great place to start.

[madires]

is this a full version of the LCR ESR Transistor Tester Project?

No, it's my 644 developement board (https://github.com/madires/Transistortester-Warehouse/blob/master/Hardware/DevKit-644.kicad.tgz) and a display board with level shifters. The dev board features the Zener check with boost converter, rotary encoder, 2.5V reference, self-adjustment cap, SRV05-4 input protection and the SamplingADC cap.

[oitar]

Hi all, sorry- tried reading the whole forum, but it will take a few months..

I have a LCR-TC2 with a Mega324PA cpu, on a T7-plus V1.1 board, 6 pin connector to 160x128 display subboard.
Originally was showing v3.1E firmware, but had a very bad(although consistent) accuracy on certain resistor values only, and very low reading on zeners below 10V(showing less than 1V). Input ports were not blown- confirmed it by measuring equal resistance of few hundred kOhm to Vcc and Gnd.

Soldered headers for main cpu, and U3(in my case), tried a few pre-compiled firmware(.hex and .eep) files without success. Best I can get is startup, quick measurement, and then almost immediate off(with v1.34m).
Is there any alternative firmware for the U3 chip? It gets recognized by STC-ISP as STC15L104W, but the only firmware I have is tc1_u4_v0.3.1.hex(which flashes without error). Or is it the main program that's not right?

If I need to compile from a source file, why are there no mega324 subfolders inside the trunk folders of both developers?
Sorry about the noob questions, but have been battling for few days.

[madires]

Is there any alternative firmware for the U3 chip? It gets recognized by STC-ISP as STC15L104W, but the only firmware I have is tc1_u4_v0.3.1.hex(which flashes without error). Or is it the main program that's not right?

Nope, the STC15 firmware from https://github.com/atar-axis/tc1-u4 should also work for your tester. Alternatively you could replace U3 with a simple two-transistor circuit (https://github.com/madires/Transistortester-Warehouse/blob/master/Hardware/TC1-Mod.pdf). And yes, the m-firmware could be configured incorrectly. In the source archive is a file called 'Clones' which lists the recommended settings for various clones, including the TC family.

If I need to compile from a source file, why are there no mega324 subfolders inside the trunk folders of both developers?

The k-firmware needs a small modification for the ATmega324 (see https://www.eevblog.com/forum/testgear/$20-lcr-esr-transistor-checker-project/msg4637602/#msg4637602] and the next few posts). And the m-firmware doesn't provide any binaries as there are too many configuration possibilities.

[elecdonia]

The dev board features the Zener check with boost converter, rotary encoder, 2.5V reference, self-adjustment cap, SRV05-4 input protection and the SamplingADC cap.

I’ve read many posts claiming the “2.5V reference IC isn’t needed if the 5V regulator is very precise in voltage and free of drift.”

What is your current opinion about this?

[madires]

Still the same, i.e. the external 2.5V voltage reference should be at least 10 times more precise than the 5V voltage regulator to make sense. The dev board has an external reference for testing purposes (in this case an LM4040).

[Fuzzy Star]

...it's my 644 developement board... features the Zener check with boost converter, rotary encoder, 2.5V reference, self-adjustment cap, SRV05-4 input protection and the SamplingADC cap.

I'm curious about the MCP1702 at the boost converter input. Any reason to use that precision regulator there instead of an AMS1117-5.0 or 78L05?

Also curious about the 'logic probe' option. Do you think about including that circuit on the dev kit schematic?

The 'SamplingADC cap relay' option is on the schematic but firmware files say 'not suppoted yet', 'not implemented yet'. Is there any developement about it?

[madires]

Any 5V regulator is fine. I simply had a bunch of MCP1702s at hand. The logic probe option is a dedicated ADC pin with a voltage divider, quite easy to add. I don't think I'll add it to the dev kit to keep one ADC pin free for testing future options. The SamplingADC is currently supported only by the k-firmware, and I don't have any plans to add it to the m-firmware any time soon. Also, I don't know if Karl-Heinz already added support for the cap relay option.

[9voltbrain]

About the LCR-T7... The rechargeable battery is making me a little paranoid, plus I think it's annoying having to keep it charged, so I'm thinking of replacing it with AAA-batteries. How hard would this be? Would I need to regulate the voltage down from 4.5 volts (3xAAA), or can it handle 4.5? The rechargeable battery is 3.7v.

[gipetto]

i saw a post saying the stc chip in the t7 is only rated for 3v max, and so using it with a 3.7v li ion is overdriving. in that case, two aa cells in series would be sufficient.when (if) mine arrives i may do as others did and use a switch in series with the battery to prevent discharge if it is an issue. aa cells are expensive, i think you would be going backwards.

[Feliciano]

About the LCR-T7... The rechargeable battery is making me a little paranoid, plus I think it's annoying having to keep it charged, so I'm thinking of replacing it with AAA-batteries. How hard would this be? Would I need to regulate the voltage down from 4.5 volts (3xAAA), or can it handle 4.5? The rechargeable battery is 3.7v.

• The battery draining is annoying if you have one of the old versions in original state (with the STC power management MCU). That can be solved by replacing it with 2 transistors (plus some dioes and resistors) and flashing the m-firmware, with many other benefits discussed in this thread
• If you have one of the recent versions, the battery doesn't drain too fast (unless yours is damaged somehow)
• To feed the ctester with something less than 5V, and variable (as batteries goes flat) would give you all sort of troubles. For starter remember this is comparing voltages samples against references. So I would rather find a bigger Li-Ion battery and/or modify the circuit as I stated above. But if you really want to use AAA size, I would analize whether the existing charger/boost converter could handle that scenario

[thm_w]

Can't believe there is still no STM32 or other high end port of this tester. Some regular alix models are selling for $30.
Usually Chinese products are very quick to swap out the micro to a completely different model. But maybe the atmega clones are so cheap there is no incentive.

Anyone here working on it?

[9voltbrain]

About the LCR-T7... The rechargeable battery is making me a little paranoid, plus I think it's annoying having to keep it charged, so I'm thinking of replacing it with AAA-batteries. How hard would this be? Would I need to regulate the voltage down from 4.5 volts (3xAAA), or can it handle 4.5? The rechargeable battery is 3.7v.

To feed the ctester with something less than 5V, and variable (as batteries goes flat) would give you all sort of troubles. For starter remember this is comparing voltages samples against references. So I would rather find a bigger Li-Ion battery and/or modify the circuit as I stated above. But if you really want to use AAA size, I would analize whether the existing charger/boost converter could handle that scenario[/li][/list]

The battery that comes with it is already lower than 5V (3.7), so how would this be more of an issue with AAA? When the battery voltage drops too low for it to regulate to its desired operating voltage, it's going to tell me to "charge", then I'll just change the batteries. Am I missing something?