$20 LCR ESR Transistor checker project

[indman]

Could you please try "ENCODER_PULSES 4"?

Hi, madires! I checked this mode, but I liked operation of an enkoder with digit 2 more.
 I reported about a problem with determination of resistance below 0.1 Ohms also earlier. It doesn't depend on crystal frequency. The great influence is exerted by quality of test contacts and a reliable clamp of the checked detail. But even, if to satisfy these conditions, the tester determines the low-impedance resistance not always.

[johwaa]

The Atmega pins drive a transistor which in turn drives the relays.

NPN with base resistor?

I think so, but I'll post the schematic tomorrow. It's on my work laptop, and I don't have it handy at the moment.

[madires]

I reported about a problem with determination of resistance below 0.1 Ohms also earlier. It doesn't depend on crystal frequency. The great influence is exerted by quality of test contacts and a reliable clamp of the checked detail. But even, if to satisfy these conditions, the tester determines the low-impedance resistance not always.

Of course the contact resistance might prevail, which causes some fun quite often. It's not a 4-wire measurement with gold plated kelvin probes. But I check all possibilities to be sure. Sometimes two (or even more) different problems add up. A few posts ago I've explained the forward and backward measurements for the resistor check. For values below 0.1 Ohms it allows a deviation of 500%. For example, when I put a 82mOhms resistor into a tester with a ZIF one direction measures 0.03 and the other one 0.08.

[Strada916]

It would be unrealistic to try and measure sub 1 ohm on a device such as this. Just my 2c worth.

Sent from my SM-G900I using Tapatalk

[indman]

Sometimes two (or even more) different problems add up.

Yes! The main problem with the  measurement low impedance is, that the resolution of the ADC
is unsufficient to get a 0.01 Ohm resolution.
The measurement current is only about 7mA, so 0.01 Ohm will only result to 0.07mV .
The ADC resolution with 1.1V voltage reference is about 1.07mV .
Other problem is that the tester sometimes doesn't define such low resistance at all!

[hapless]

Didn't realize there were so many naysayers around. Just sayin'. You're basically denying that what's already working can work, and that's after the person who made it work has told you how he's going to make it work even better.  Nobody's trying to defy the laws of physics here, but some amazing things can be done with a few tricks and a firm belief in success. Words of discouragement don't help.

[casinada]

We all know the limitations of the hardware. It is a nice little device They have done a great job with the software and greatly expanded its use. I couldn't ask for more. The only thing is to make it more reliable, of course, sometimes is due to the hardware problems like socket connection to the microcontroller, long testing wires, cheap ZIF socket, noisy power source, questionable sourced components, and it still manages to give decent results. There are lots of different hardware versions to add to the software complexity.
Probably we should define better the limits of the device so each time we go and measure something we can get similar results.  

[johwaa]

The Atmega pins drive a transistor which in turn drives the relays.

NPN with base resistor?

Yup, and I'm using a 2.2k resistor. BTW, I fixed the footprint, and re-laid out the board, zipped it all up and uploaded it to here. It hasn't been posted as of yet however. And I just realized that I forgot to fix my relay driver... Sigh...

[madires]

I've changed the measurement function for low value resistors (saving a few bytes and it should also help with inductors) and the tolerance logic in the resistor check function (replaced the 500% tolerance for <0.1Ohms with a better method applied to resistors <2Ohms) to cope with probe contact issues and low value resistors.

[13hm13]

Newbie to this thread, so some fundamental questions about the "$20 LCR ESR Transistor checker project".

What's the operating freq. of the project meter as of this writing?

Where can I find the schema and BOM?

Great project idea, BTW!

[jh15]

i agree, don't want to sift through 4000 posts

[madires]

Usually the last 10 pages will give you all the links and latest news. Every two pages someone asks for firmware and documentation while excusing for not reading all 144+ pages

[madires]

What's the operating freq. of the project meter as of this writing?

Where can I find the schema and BOM?

Great project idea, BTW!

MCU clock is 8 or 16 MHz, optionally 20 MHz with m-firmware. Documentatiion is avaliable at https://github.com/svn2github/transistortester/blob/master/Doku/trunk/pdftex/english/ttester.pdf for example. There's also a directory with PCBs for the tester and hardware options.

[amspire]

What's the operating freq. of the project meter as of this writing?

Where can I find the schema and BOM?

Great project idea, BTW!

MCU clock is 8 or 16 MHz, optionally 20 MHz with m-firmware. Documentatiion is avaliable at https://github.com/svn2github/transistortester/blob/master/Doku/trunk/pdftex/english/ttester.pdf for example. There's also a directory with PCBs for the tester and hardware options.

That repository (https://github.com/svn2github/transistortester) also contains the two principle open source versions of the software. The trunk contains the version from Karl-Heinz Kübbeler based on the work of Markus Frejek. The other branch by Markus Reschke (madires in this forum) is in the "Markus" folder in the repository.

They function slightly differently and have some different options, so it is worth testing both.

You do not need to install the full Atmel development software to compile.

For Windows, you will need WinAVR to program the processor, and this includes a complete GCC compiler.
WARNING*** For some people installing WinAVR, their PATH system variable gets overwritten. It is a very good idea getting a copy of your current path before installing WinAVR. You can do this with the DOS command
PATH > path.txt
https://sourceforge.net/projects/winavr/
This post then recommends you update this to one of two particular versions of the compiler.
https://www.eevblog.com/forum/testgear/$20-lcr-esr-transistor-checker-project/msg825078/#msg825078

I did originally try compiling with the latest GCC for the smaller Atmega168 processor and the code wouldn't fit in the processor's space. When I changed to the recommended versions, it just fitted.

Compiling the code yourself is worth it as you are able to turn on and off the features you want, and possible tweak some constants. For both versions, you usually only have to edit the config.h file. Markus's config.h file is very well self-documented, so it might be the easier version to try first.

I hope that is an accurate summary.

[Sredni]

Last week I assembled my first transistor tester (bought the kit on Amazon). It appears to be working  but there are a few strange things going on. Maybe you can help me in troubleshooting it? It's the kit in a red PCB board with

  k1036038s
  2578AY-AT

It's the same layout as shown in the attached picture (taken from the net, probably this very forum).
It came with firmware version 1.12k and a spare 7550 (for which I was unable to find a datasheet) and I have already done the calibration procedure twice - from the menu.

Here are the oddities, in no particular order:

0 - After hours of non use, battery voltage is reported lower than what it actually is: at startup it says "Battery 8.9V while the actual battery with no load is 9.2 V (IIRC). This could be just a loading effect, but it is worth mentioning it. See also point 4

1 - When I repeat the measure of a resistance I get increasingly lower values. A 12.5k resistor is given as 12100 ohm the first time, 11980 ohm the second time, 11648, 11245, 10956, 10437, ... (I am making up the values right now, but that's the trend - with the resistor untouched and only subsequent pressing of the knob I went down under 9k with no fixed point in sight). And this is bad.

2 - The same is going on for caps, but I could attribute that to the fact the the measuring procedure is charging them a bit 9316 pF, 9281pF, 8953pF, 8619pF, 8307pF,... and these are actual values just measured with the time between them due to typing the numbers in this post.

3 - The spare 7550 is not identified. Is this normal for this firmware version?

4 - I just noticed that battery value too is getting lower at each pushing of the knob, only to come back to 8.9V if I leave it off for a long time.

What do you think?
there mus be a cap somewhere that is not discharging?

[Strada916]

Last week I assembled my first transistor tester (bought the kit on Amazon). It appears to be working  but there are a few strange things going on. Maybe you can help me in troubleshooting it? It's the kit in a red PCB board with

  k1036038s
  2578AY-AT

It's the same layout as shown in the attached picture (taken from the net, probably this very forum).
It came with firmware version 1.12k and a spare 7550 (for which I was unable to find a datasheet) and I have already done the calibration procedure twice - from the menu.

Here are the oddities, in no particular order:

0 - After hours of non use, battery voltage is reported lower than what it actually is: at startup it says "Battery 8.9V while the actual battery with no load is 9.2 V (IIRC). This could be just a loading effect, but it is worth mentioning it. See also point 4

1 - When I repeat the measure of a resistance I get increasingly lower values. A 12.5k resistor is given as 12100 ohm the first time, 11980 ohm the second time, 11648, 11245, 10956, 10437, ... (I am making up the values right now, but that's the trend - with the resistor untouched and only subsequent pressing of the knob I went down under 9k with no fixed point in sight). And this is bad.

2 - The same is going on for caps, but I could attribute that to the fact the the measuring procedure is charging them a bit 9316 pF, 9281pF, 8953pF, 8619pF, 8307pF,... and these are actual values just measured with the time between them due to typing the numbers in this post.

3 - The spare 7550 is not identified. Is this normal for this firmware version?

4 - I just noticed that battery value too is getting lower at each pushing of the knob, only to come back to 8.9V if I leave it off for a long time.

What do you think?
there mus be a cap somewhere that is not discharging?

I'll be looking at the documentation for these testers(look at repository manual) and reading the section. "Chinese clones"
Check the power up circuit. eIt can be draining the battery?
I'd also check the battery voltage with a known meter whilst the unit is powered up.

Edited. Reread comment. Disregard y first sentence.


Sent from my SM-G900I using Tapatalk

[pepe10000]

It came with firmware version 1.12k and a spare 7550 (for which I was unable to find a datasheet).

www.e-ele.net/DataSheet/HT75XX-1.pdf

A greeting.

[madires]

4 - I just noticed that battery value too is getting lower at each pushing of the knob, only to come back to 8.9V if I leave it off for a long time.

What do you think?
there mus be a cap somewhere that is not discharging?

Please power the tester with a lab PSU and check again. Also observe the current draw.

[amspire]

Last week I assembled my first transistor tester (bought the kit on Amazon). It appears to be working  but there are a few strange things going on. Maybe you can help me in troubleshooting it? It's the kit in a red PCB board with

  k1036038s
  2578AY-AT

I have attached the documentation for the board. I think the circuit is pretty right, except I don't think your board has diode D1 from the battery.

0 - After hours of non use, battery voltage is reported lower than what it actually is: at startup it says "Battery 8.9V while the actual battery with no load is 9.2 V (IIRC). This could be just a loading effect, but it is worth mentioning it. See also point 4

The battery voltage is measured after going through transistor T1. There could be a small voltage drop across it. Also if you compile your own code, you can calibrate the battery voltage. I did some notes on this a few posts back. Since you have a socket, it is a really good idea to have a spare processor IC (or a few). It is probably a good idea anyway getting some 16MHz crystals of ebay, and modifying the code for 16MHz.

You will need something to program the chip in. If you have an arduino board, you can hook up an arduino as a programmer for the atmega processor.

But Version 1.12k, that should be fine - you shouldn't have to do anything.

...

4 - I just noticed that battery value too is getting lower at each pushing of the knob, only to come back to 8.9V if I leave it off for a long time.

Just check the circuit around T1 is working correctly, When the tester is off, there should be no volts at all across the 33K base resistor R14. This is the one on the left edge of the board next to the 10uF cap and the 9012 transistor.

[hapless]

[...]

What do you think?
there mus be a cap somewhere that is not discharging?

I think your regulator may not be regulating. Observe your V_CC over time during operation. Also, extreme negative temperatures of the climate in your region may have something to do with this. 

[upsss]

What determines whether the tester is using the Vcc (5V) as reference or the external 2.5V reference and how do you force either one to be the reference?

[hapless]

Why would you want to force that? It uses the 2.5V reference (if it is present) to see how far the V_CC is from the ideal 5.0V.

[upsss]

@hapless, you didn't answer my question.  Just because the 2.5v reference is present it should NOT be the default reference.  If the 2.5V reference is a crappy TL431 and the Vcc regulator is much more accurate then I may want to use the Vcc as a reference and not the 2.5V.  On the other hand, if I have a very precise 2.5v 0.1% reference (not a TL431) and the Vcc regulator is a crappy 7805 5% then I may want to use the 2.5V as a reference. 

Again, the question is what determines whether the tester is using the Vcc (5V) as reference or the external 2.5V reference and how do I select which one I want to use?

[madires]

k-firmware: If a voltage around 2.5V is detected at the pin for the external reference, it's used. You can disable the external reference by removing it or modifying the source.
m-firmware: If HW_REF25 is defined and a voltage around 2.5V is detected at the pin for the external reference, it's used. You can disable the external reference by removing it or undefining HW_REF25.

[hapless]

@hapless, you didn't answer my question.  Just because the 2.5v reference is present it should NOT be the default reference.  If the 2.5V reference is a crappy TL431 and the Vcc regulator is much more accurate then I may want to use the Vcc as a reference and not the 2.5V.  On the other hand, if I have a very precise 2.5v 0.1% reference (not a TL431) and the Vcc regulator is a crappy 7805 5% then I may want to use the 2.5V as a reference. 

Again, the question is what determines whether the tester is using the Vcc (5V) as reference or the external 2.5V reference and how do I select which one I want to use?

You "select" it by having or not having it connected, that is what determines its use, unless you want to mess around with the firmware like madires describes above.

Your premise makes no sense though. If the 2.5V reference is not an order of magnitude more accurate than your regulator, then what is it doing there anyway? You don't want your circuit to have parts that do nothing. Remove it. Then the firmware will use the 5V without comparing it to a known reference.