$20 LCR ESR Transistor checker project

[elecdonia]

…I purchased a T7-Tester that is locally branded by german distributor Joy-It. I attached some pictures - as it seems, this still is "the real deal" on the inside?
This one was purchased from reichelt.de

This unit should be perfect for upgrading to current OSHW firmware, either k- or m- style. To the best of my knowledge all ATmega328P chips in 28-pin DIP packages are genuine. And because it is mounted in a socket it will be very easy to change if you wish to do so.

Another good feature about this PC board is that it has pads for installing the standard 6-pin Atmel ISP programming header. This permits uploading firmware through an AVR-ISP programmer without removing the MCU from the PC board. I use “Arduino as ISP” for this.

[elecdonia]

I can thoroughly recommend one of these units with the case. I've had mine now for a number of years and is the same board as this uses, there are other slight variations of the eBay for sale, but I can't vouch for those, but this one has given my very good service and is certainly reasonably accurate, checking it against bigger 4 lead devices.
https://www.ebay.co.uk/itm/155087140184

There is one major issue with these Transistor Testers which have their MCU in a 32-pin TQFP package (8 pins on each of the 4 sides of the package).

     These days there are 3 totally different 32-pin TQFP package MCU chips which may appear in Transistor Testers:

Atmel ATmega328P   — this is the MCU specified by the original OSHW Transistor Tester project. However its price has increased from $3 USD to $9-10 USD, so the low-cost Chinese vendors stopped using it.

Logic Green LGT8F328P — This is a Chinese MCU which costs <$1 USD. It is said to be “somewhat” similar to the original Atmel AT328P. But unfortunately several of its pins have different functionality. Therefore a Transistor Tester containing LGT8F328P cannot easily have its MCU replaced by the Atmel ATmega328P. Also the LGT8F328P has substantially different internal features, which means existing OSHW firmware won’t work without considerable modification.

Aptchip APT32F172K8T6.  — Another Chinese MCU which costs <$1 USD. It has nothing at all in common with the original Atmel ATmega328P. The pinout is totally different. Also the internal MCU architecture is totally different.

Warning: Some of these $1 USD Chinese MCU chips have been deliberately mislabeled as “Atmel ATmega328P.”

More discussions about this are found here:
   https://www.eevblog.com/forum/testgear/$20-lcr-esr-transistor-checker-project/msg4613737/?topicseen#msg4613737
   https://www.eevblog.com/forum/testgear/$20-lcr-esr-transistor-checker-project/msg4610314/?topicseen#msg4610314
   https://www.eevblog.com/forum/testgear/$20-lcr-esr-transistor-checker-project/msg4600771/#msg4600771

[gipetto]

Hello folks,

I just found out about this thread on eevblog forums about Markus' and Karl-Heinz' Testers recently.

After reading about the latest batches of LCR-TC1/TC2 containing counterfeit or clone MCUs, I tried not to head down the "Aliexpress Alley" - so I purchased a T7-Tester that is locally branded by german distributor Joy-It. I attached some pictures - as it seems, this still is "the real deal" on the inside?

This one was purchased from reichelt.de

After you posted this i decided to buy it too, since i was in need of one and they took bank transfer as payment. my research says there's a minor issue with power with the stc chip on that model that can cause a higher standby drain as well as crashes. i'm not worried so much about the power drain since it is rechargeable.
https://www.eevblog.com/forum/testgear/$20-lcr-esr-transistor-checker-project/msg3598639/#msg3598639

[Feliciano]

The high drain during standby was a major issue because the battery is small (and getting smaller). Nevertheless, it seems that part has already been solved by several of the manufacturers, if you want (or cannot migrate from) the chinese firmware.

[Specmaster]

I can thoroughly recommend one of these units with the case. I've had mine now for a number of years and is the same board as this uses, there are other slight variations of the eBay for sale, but I can't vouch for those, but this one has given my very good service and is certainly reasonably accurate, checking it against bigger 4 lead devices.
https://www.ebay.co.uk/itm/155087140184

There is one major issue with these Transistor Testers which have their MCU in a 32-pin TQFP package (8 pins on each of the 4 sides of the package).

     These days there are 3 totally different 32-pin TQFP package MCU chips which may appear in Transistor Testers:

Atmel ATmega328P   — this is the MCU specified by the original OSHW Transistor Tester project. However its price has increased from $3 USD to $9-10 USD, so the low-cost Chinese vendors stopped using it.

Logic Green LGT8F328P — This is a Chinese MCU which costs <$1 USD. It is said to be “somewhat” similar to the original Atmel AT328P. But unfortunately several of its pins have different functionality. Therefore a Transistor Tester containing LGT8F328P cannot easily have its MCU replaced by the Atmel ATmega328P. Also the LGT8F328P has substantially different internal features, which means existing OSHW firmware won’t work without considerable modification.

Aptchip APT32F172K8T6.  — Another Chinese MCU which costs <$1 USD. It has nothing at all in common with the original Atmel ATmega328P. The pinout is totally different. Also the internal MCU architecture is totally different.

Warning: Some of these $1 USD Chinese MCU chips have been deliberately mislabeled as “Atmel ATmega328P.”

More discussions about this are found here:
   https://www.eevblog.com/forum/testgear/$20-lcr-esr-transistor-checker-project/msg4613737/?topicseen#msg4613737
   https://www.eevblog.com/forum/testgear/$20-lcr-esr-transistor-checker-project/msg4610314/?topicseen#msg4610314
   https://www.eevblog.com/forum/testgear/$20-lcr-esr-transistor-checker-project/msg4600771/#msg4600771

I was not aware of this, however, if anything other than a capacitor or resistor failed, wouldn't it be better to replace the whole unit seeing as it is such a low cost unit, rather than run the gauntlet of facing a possible problem with fake chips? 

[elecdonia]

I was not aware of this, however, if anything other than a capacitor or resistor failed, wouldn't it be better to replace the whole unit seeing as it is such a low cost unit, rather than run the gauntlet of facing a possible problem with fake chips?

At the present time (March 2013) I can provide two recommendations worth considering. The choice depends on whether one wants a simple little device to toss into their toolbox for quick measurements (with limited accuracy), or whether they wish to learn more of “the details” of the Transistor Tester project with the intention of improving and customizing their device into an accurate and dependable bench instrument.

1)    For those who wish to get a functional Transistor Tester which won’t need any modification I specifically recommend the “LCR-TC2 V2.3E” model. Please expect to pay $30-40 USD for it. This  model contains a genuine Atmel ATmega324 MCU chip (44-pin TQFP square package).
Example:  https://www.eevblog.com/forum/testgear/$20-lcr-esr-transistor-checker-project/msg4753835/?topicseen#msg4753835

2) For those who wish to modify/customize/improve their Transistor Tester I recommend the “kit-form” AY-AT GM328 “color” units which come as a bare PC board and a bag of parts. These units use the Atmel ATmega328P (28-pin DIP package).
Example:   https://www.eevblog.com/forum/testgear/$20-lcr-esr-transistor-checker-project/msg4681222/?topicseen#msg4681222

3) Avoid purchasing “LCR style” units (plastic case with rechargeable battery inside) costing <$25 USD. Nearly all of these will contain cheap (<$1 USD) Chinese MCU chips. Most will be only partially functional. The OSHW Transistor Tester firmware does not currently support any of these $1 Chinese MCU chips.

[hapless]

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?

1) Some of us want a Transistor Tester which comes in an attractive “toolbox-friendly” finished plastic case with a self-contained USB rechargeable battery, rather than an unenclosed PC board with a loose 9V battery dangling by its cable.

2) High-quality LCR-style units contain the ATmega324 MCU in a 44-pin TQFP package. In contrast, “cheap” LCR-style units have 32-pin MCU chips which probably aren’t the Atmel ATmega328P. However there aren’t yet any alternative “cheap” MCU chips which come in 44-pin TQFP packages. Therefore if a Transistor Tester contains a 44-pin TQFP MCU chip this indicates the MCU is likely to be an authentic Atmel ATmega324, rather than some dodgy $0.25 USD MCU with unknown architecture running poorly designed Chinese firmware which cannot be upgraded.

3) Transistor Testers with ATmega324 can be upgraded to ATmega644 to get twice as much flash (64k vs. 32k). This enables activating all of the advanced features offered by some versions of the open-source Transistor Tester software. The ATmega324 and ATmega644 have the same physical dimensions and pinout.

None of this is what I was asking. Looking to build, not buy. Need info.

[indman]

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.

[madires]

None of this is what I was asking. Looking to build, not buy. Need info.

Please see https://github.com/kubi48/TransistorTester-documentation for Karl-Heinz' excellent documentation which includes schematics and a few hardware options. One benefit of an ATmega324 is that the additional pins allow you to assign things differently, making the use of hardware SPI possible (which is great for color displays). If you go for more features a 644 would be the better choice as the 328/324's flash is easily maxed out.

[hapless]

I see. Thanks.

A problem with a thread as huge as this is that even when you know what you're looking for, it's not always easy to find. It would be nice to have either the first post or a separate page (or web site?) that links to the current important stuff: Where the current repository is, where the most updated documentation is, where the different modifications are described (like how to wire up a display for hardware SPI for example).

A bit unrelated, I wanted to ask if you'd ever managed to solve the problem with LC add-on where it was counting too many pulses.

[Aldo22]

I know that many are unhappy with the tester that contains the APT32F172K8T6 (TC1 in my case).
But if you don't know anything about ATmega and firmware etc., it's not that bad for ~$10 including battery and case.
Despite the small battery, it has a really good endurance.
Once the problem with KAA range is solved (2K2 resistor) there is no reason to throw it away. Imo it's worth the money.

I randomly tested some components (cheap chinese stuff in a big bag). Here are the results.
Imho even the ESR values seem plausible.

Resistors:
680k -> 682k
220k -> 217k
51k -> 53k
3,3k -> 3.28k
100 -> 98.7

Capacitors:
6.8pF -> 6.2pF
470pF -> 520pF
3.3nF -> 3.5nF
220nF -> 240nF
470nF -> 504nF
1μF -> 0.974μF (ESR:4.1Ω)
2.5μF  -> 2.41μF (ESR:4.8Ω)
10μF -> 9.78 (ESR:0.98Ω)
22μF  -> 24.6μF (ESR:1.5Ω)
33μF  -> 32.1μF (ESR 0.8Ω)
47μF  -> 49μF (ESR:0.5Ω)
220μF  -> 228μF (ESR:0.2Ω)
470μF  -> 420μF (ESR:0.13Ω) (Aneng 8080 shows also 420μF).

Diodes:
1N4148 -> Uf=708mV C=2pf Ir=11na
1N4007 -> Uf=703mV C=9pf Ir=8na
FR107 -> Uf=666mV C=10pf Ir=10na

Zener:
3.6V -> 3.6V
5.1V -> 5.1V
12V -> 12.2V

Transistors:
A42:    NPN hFE=45  Ube=609mV Ic=5.9mA
BC547B: NPN hFE=344 Ube=774mV Ic=5.9mA
BC547C: NPN hFE=562 Ube=813mV Ic=5.9mA
C1815 : NPN hFE=382 Ube=773mV Ic=5.9mA
C8050:  NPN hFE=138 Ube=695mV Ic=5.8mA
C945:   NPN hFE=362 Ube=768mV Ic=5.9mA
S9013:  NPN hFE=151 Ube=675mV Ic=5.8mA
S9014:  NPN hFE=375 Ube=770mV Ic=5.9mA
S9018:  NPN hFE=71  Ube=735mV Ic=0.0mA
2N3904: NPN hFE=214 Ube=719mV Ic=5.9mA
2N2222: NPN hFE=289 Ube=717mV Ic=5.9mA
2N3904: NPN hFE=211 Ube=718mV Ic=5.9mA
2N5551: NPN hFE=154 Ube=681mV Ic=5.9mA

A93:    PNP hFE=116 Ube=729mV Ic=6.0mA
A733:   PNP hFE=385 Ube=890mV Ic=6.0mA
A1015:  PNP hFE=358 Ube=884mV Ic=6.0mA
2N5401: PNP hFE=162 Ube=761mV Ic=6.0mA
2N3906: PNP hFE=210 Ube=838mV Ic=5.9mA
S9012:  PNP hFE=184 Ube=776mV Ic=6.0mA
S9015:  PNP hFE=390 Ube=889mV Ic=6.0mA
S8550:  PNP hFE=345 Ube=868mV Ic=5.9mA

It definitely helps to put the components in the right drawer. 

[Specmaster]

I agree, those results are very similar to my findings and are certainly good enough to work with in the restoration / repair of other equipment.

[madires]

A bit unrelated, I wanted to ask if you'd ever managed to solve the problem with LC add-on where it was counting too many pulses.

The m-firmware expects the base frequency of the LC meter hardware option to be within a specific range. Do you mean this?

[elecdonia]

...A problem with a thread as huge as this is that even when you know what you're looking for, it's not always easy to find. It would be nice to have either the first post or a separate page (or web site?) that links to the current important stuff: Where the current repository is, where the most updated documentation is, where the different modifications are described (like how to wire up a display for hardware SPI for example).

For those who haven't noticed, this Transistor Tester topic is by far the largest on the eevblog "Test Equipment" board: 8,323 replies, 2,957,670  views. In fact this might be the largest topic on the entire eevblog forum.

The sheer size of this topic does suggest that adding a "table of contents" or "faq" would be a useful task.

[hapless]

A bit unrelated, I wanted to ask if you'd ever managed to solve the problem with LC add-on where it was counting too many pulses.

The m-firmware expects the base frequency of the LC meter hardware option to be within a specific range. Do you mean this?

Yes. I remember it counting too many pulses at one end of the range, probably above 500nF (but, of course, now I can't find where I saw that being stated  ).

[elecdonia]

I know that many are unhappy with the tester that contains the APT32F172K8T6 (TC1 in my case).
But if you don't know anything about ATmega and firmware etc., it's not that bad for ~$10 including battery and case.
Despite the small battery, it has a really good endurance.
Once the problem with KAA range is solved (2K2 resistor) there is no reason to throw it away. Imo it's worth the money.

I also have one of these LCR-style units with the APT32F172K8T6. My unit is reasonably accurate when testing R, C, L, or zener diodes.

However, I have noticed that it sometimes totally fails to correctly identify transistors by type. On a related topic it tends to display wildly inaccurate hfe values for some transistors which test consistently within the expected range by several other transistor testers which I have on hand. I will review the notes I made when I observed this discrepancy and post them.

[elecdonia]

None of this is what I was asking. Looking to build, not buy. Need info.

I'm starting construction on a scratch-built DIY Transistor Tester "development system:"

• I have ATmega644 and ATmega1284 MCU chips in 40-pin DIP packages. Enabling all optional features of current OSHW Transistor Tester software does require a 64k MCU (such as ATmega644). I read that a "dual boot" setup where the user can select between one k- version and one m- version might be feasible. That would require 128k flash. Alternately the board could support a micro SD card for storing several different firmware versions
• IL9341 or similar large color LCD display (I want a larger display than ready-made units)
• Constructed on “perma-proto” style board
• USB port supplying power and providing serial connection to host PC
• Bootloader for firmware updates through USB serial port
• Full protection for the 3 probe pins (SRV05-4, P6KE6V8A, plus a relay)
• High-quality rotary encoder with large knob for user input
• Zener test feature implemented (up to 50V, test current ~5-10mA)
  
  More info about DIY Transistor Testers:
       (Check this post periodically - I will add links as I find them)
  
  First take a look at this post by Madires (attached photo is Madires ATmega644/1284 development board):
       https://www.eevblog.com/forum/testgear/show-us-your-$20-transistor-tester-enclosures-and-mods/msg1201483/#msg1201483
  
  https://www.eevblog.com/forum/testgear/$20-lcr-esr-transistor-checker-project/msg3578117/#msg3578117
  https://www.eevblog.com/forum/testgear/$20-lcr-esr-transistor-checker-project/msg4637764/?topicseen#msg4637764
  https://www.eevblog.com/forum/testgear/$20-lcr-esr-transistor-checker-project/msg825537/#msg825537
  

[Messtechniker]

I have several ATmega644 and ATmega1284 MCU chips in 40-pin DIP packages.

Good idea. If socketed, makes for easy replacement since the inputs are unprotected (for a good reason).
Also dont forget to provide an USB data output. The firmware does provide for this.

[gipetto]

None of this is what I was asking. Looking to build, not buy. Need info.

I am starting construction on a DIY Transistor Tester myself:
I have several ATmega644 and ATmega1284 MCU chips in 40-pin DIP packages. Actually the mega644 is big enough in terms of flash size. Enabling all available features in the OSHW Transistor Tester software does require an MCU with more than 32k of flash.
I intend to use a larger color LCD display than what comes in the ready-made Chinese units.
I plan to build the first unit on a “perma-proto” style board.
My version will have a high-quality rotary encoder for user input.

It makes sense to build on this existing project. I have yet to find the original post in this thread about it.
https://github.com/madires/Transistortester-Warehouse/blob/master/Hardware/DevKit-644.kicad.tgz

[madires]

Yes. I remember it counting too many pulses at one end of the range, probably above 500nF (but, of course, now I can't find where I saw that being stated  ).

Larger capacitance means lower frequency (lower limit is 10 kHz). The base frequency is around 595 kHz (L_i 82µH, C_i 1nF), the theoretical measurement range is 3.3 fF - 3.5 µF, and what you get with the example circuit in the repo is 10 fF - 33 nF (120 nF with signal clean-up). A different LC oscillator circuit might extend that range.

[madires]

It makes sense to build on this existing project. I have yet to find the original post in this thread about it.
https://github.com/madires/Transistortester-Warehouse/blob/master/Hardware/DevKit-644.kicad.tgz

Here's a picture: https://www.eevblog.com/forum/testgear/show-us-your-$20-transistor-tester-enclosures-and-mods/msg1201483/#msg1201483

[Aldo22]

I also have one of these LCR-style units with the APT32F172K8T6. My unit is reasonably accurate when testing R, C, L, or zener diodes.

However, I have noticed that it sometimes totally fails to correctly identify transistors by type. On a related topic it tends to display wildly inaccurate hfe values for some transistors which test consistently within the expected range by several other transistor testers which I have on hand.

Hi!

I've updated my list with some transistor measurements.
Do you see a problem with these numbers?
I don't have enough knowledge to judge that.

[indman]

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

[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

Thank you!
No, I don't have any of these.
All I have is in the list above (21 types).

I found that calibration is really important with this device.
If it is not freshly calibrated, it sometimes even recognizes a transistor as a resistor.
After the calibration, however, it recognizes it correctly.
Maybe even the battery status makes a difference (just a guess).

I am sure that the tester with the APT32F172K8T6 has weaknesses (like other devices?).
I just think it's not completely useless and not that bad for $10.

[hapless]

10 fF - 33 nF

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.)