$20 LCR ESR Transistor checker project

[tom666]

The voltage reference 2.5V is not used in the original version. It can be additionally fitted (parts U3 and R19)

[klaus11]

Thank 

[indman]

In this new scheme Fish8840. Already the reference TL431 is not used?

No, it is the old diagram. After numerous experiments I came to a conclusion that application of TL431 doesn't lead to noticeable improving of parameters of a tester. It is more important to deliver the high-quality power supply voltage of the processor (Vcc=5.000V).

[indman]

I changed a firmware for Fish8840BlueDisplay a little - added a homepage with the version and revision of audit. It decelerates start a little, but is more evident and it is convenient!
Attention! This firmware only for a board, in which divider of a supply of R15/R8 = 10k/3.3k!

[madires]

In this new scheme Fish8840. Already the reference TL431 is not used?

No, it is the old diagram. After numerous experiments I came to a conclusion that application of TL431 doesn't lead to noticeable improving of parameters of a tester. It is more important to deliver the high-quality power supply conditioner of the processor (Vcc=5.000V).

Yup, a MCP1702 is a much better investment than a TL431. A dedicated voltage reference makes sense when you're using something like a 78L05 or if the voltage reference is more than 10 times as accurate as the voltage regulator (MCP1702 is 0.4%).

[klaus11]

Right now I'm trying to program Elnec ISP  smartprog2 and gives me VCC failure

[gojimmypi]

This looks awesome! I'm a newbie to this thread, wondering if someone could point me to the most recent hardware to buy. I found this:

http://www.ebay.com/itm/M328-Transistor-Tester-Capacitor-ESR-Inductance-Resistor-LCR-Meter-NPN-PNP-MOS/281538451480

Is it the right one?

I tried to register at aliexpress but depending on the page, either it shows only google play and app store buttons, or when I actually can fill in a registration form, it does not like any of the fields; they all give validation errors.

thanks

[tom666]

I recommend this kit:
http://www.ebay.com/itm/M328-LCD-12864-Transistor-Tester-DIY-Kit-Diode-Triode-Capacitance-LCR-ESR-Meter-/381351024208
or
http://www.banggood.com/DIY-M12864-Graphics-Version-Transistor-Tester-Kit-LCR-ESR-PWM-p-986954.html

or set with the cabinet:
http://www.banggood.com/DIY-M12864-Graphics-Version-Transistor-Tester-Kit-LCR-ESR-PWM-With-Case-p-997023.html

[CustomEngineerer]

I recommend this kit:
http://www.ebay.com/itm/M328-LCD-12864-Transistor-Tester-DIY-Kit-Diode-Triode-Capacitance-LCR-ESR-Meter-/381351024208
or
http://www.banggood.com/DIY-M12864-Graphics-Version-Transistor-Tester-Kit-LCR-ESR-PWM-p-986954.html

or set with the cabinet:
http://www.banggood.com/DIY-M12864-Graphics-Version-Transistor-Tester-Kit-LCR-ESR-PWM-With-Case-p-997023.html

Seconded. Recently got this one and so far seems to be great. Its already go the encoder instead of push button switches, the 0.1% resistors where it matters and a voltage reference (instead of zener diode), although the general consensus seems to be it would be better to use a ldo voltage regulator like a MCP1702 instead of the 7805 that this kit comes with. If using the MCP1702-5002 you don't need the voltage reference. Of course if you have a MCP1702 you can go ahead and use that if you want, or order one online. They are only $0.52 on mouser.com.

Edit: Also it has the nice display and seems to support all the newer features that have been added to the firmware recently.

[gojimmypi]

I recommend this kit:
http://www.ebay.com/itm/M328-LCD-12864-Transistor-Tester-DIY-Kit-Diode-Triode-Capacitance-LCR-ESR-Meter-/381351024208
or
http://www.banggood.com/DIY-M12864-Graphics-Version-Transistor-Tester-Kit-LCR-ESR-PWM-p-986954.html

or set with the cabinet:
http://www.banggood.com/DIY-M12864-Graphics-Version-Transistor-Tester-Kit-LCR-ESR-PWM-With-Case-p-997023.html

awesome, thanks!  Can you tell me where / if the source code is available? I tried searching on github & didn't find it.

[hapless]

Got one of those with the case more out of curiosity than anything else since I keep hearing about it. It's actually smaller than it looks in the pictures, especially the display. The socket leads needed to be carefully bent and the two pairs at the ends of it connected together. Used heat-shrink tubing to insulate them just in case they decided to touch each other when I wasn't looking  . I simply used some spare resistor leads to connect it to the main board, but the angle at which they connect is not pretty. The two banana plugs became frequency counter inputs. When soldering in 10K resistors, I measured them and made sure that the combination at the frequency input provided roughly 2.4V, as it says in the manual. Another reason to measure is that I wanted to make the divider for battery voltage testing to be as close to 10K/3.3K as possible.

The firmware that this thing came with was horrible: It would crash randomly at the end of measurement and the tester would sit there sucking about 60mA from the battery until it was disconnected, but no such symptoms were seen after reflashing.

I wonder what I can do about the voltage reference though. It is not nearly as precise as I would like. Would measuring it and changing the expected voltage in software be equivalent to trimming or does it absolutely have to be precise?

[hapless]

awesome, thanks!  Can you tell me where / if the source code is available? I tried searching on github & didn't find it.

http://www.mikrocontroller.net/svnbrowser/transistortester/

[tom666]

awesome, thanks!  Can you tell me where / if the source code is available? I tried searching on github & didn't find it.

In this thread you'll find plenty of information (I put them there as "tom_pin" ):
http://forum.banggood.com/forum-topic-92191.html
and
http://forum.banggood.com/forum-topic-92191.html?page=2

Videos (not mine):

[tom666]

I wonder what I can do about the voltage reference though. It is not nearly as precise as I would like. Would measuring it and changing the expected voltage in software be equivalent to trimming or does it absolutely have to be precise?

The 2.5V reference is used only to check and calibrate the VCC voltage, but is not required. The merits of using the 2.5V ref. source (for example LM4040 AIZ2.5 [0.1%], TL431, LT1004 CZ-2.5 [0.8%], LM385-2.5, LM336Z-2.5 [0.8%]) only makes sense if it is 10 times more accurate than LDO 5V regulator, as otherwise this may lead to inferior results of the measurement. If is the 78L05 replaced with low-drop LDO MCP1702T (with typical deviation of 0.4%) is the 2.5V reference unnecessary! Another benefit of using low-drop voltage regulator (for example MCP1702-5002, LP2950CZ-5.0, LE50CZ) is to reduce power consumption and lower voltage drop with the ability to work with a little lower input voltage than standard 78L05 stabilizer.

[tom666]

... a voltage reference (instead of zener diode) ...

The zener diode is used only in one type of tester (D1 on the attached picture), which is also sold in the variant without 8MHz crystal. Other types of testers (those that use the 2.5V reference source) use typically the TL431 in designs SMD or THT.

[hapless]

I wonder what I can do about the voltage reference though. It is not nearly as precise as I would like. Would measuring it and changing the expected voltage in software be equivalent to trimming or does it absolutely have to be precise?

The 2.5V reference is used only to check and calibrate the VCC voltage, but is not required. The merits of using the 2.5V ref. source (for example LM4040 AIZ2.5 [0.1%], TL431, LT1004 CZ-2.5 [0.8%], LM385-2.5, LM336Z-2.5 [0.8%]) only makes sense if it is 10 times more accurate than LDO 5V regulator, as otherwise this may lead to inferior results of the measurement. If is the 78L05 replaced with low-drop LDO MCP1702T (with typical deviation of 0.4%) is the 2.5V reference unnecessary! Another benefit of using low-drop voltage regulator (for example MCP1702-5002, LP2950CZ-5.0, LE50CZ) is to reduce power consumption and lower voltage drop with the ability to work with a little lower input voltage than standard 78L05 stabilizer.

True, but doesn't answer my question, I'm afraid. Even an MCP1702T can be off by enough that calibration against a known reference would be beneficial, even if it's just a lowly TL431 that we're checking against. TL431 is adjustable, so it can be trimmed to be significantly more precise than any voltage regulator (easily by an order of magnitude). But I'm asking whether the same result can be achieved by a simple change in the software. Something tells me that it should be the case.

[madires]

True, but doesn't answer my question, I'm afraid. Even an MCP1702T can be off by enough that calibration against a known reference would be beneficial, even if it's just a lowly TL431 that we're checking against. TL431 is adjustable, so it can be trimmed to be significantly more precise than any voltage regulator (easily by an order of magnitude). But I'm asking whether the same result can be achieved by a simple change in the software. Something tells me that it should be the case.

For the m-firmware you can set the voltage of the 2.5V reference via UREF_25 in config.h and for the k-firmware you have to change a value in Calibrate_UR.c, IIRC.

[torch]

I've planned to add the fancy pinout in this release too but had/have to deal with a bereavement, a broken heart and a flu. Why do bad things happen all at once? So I ask you all if I should release v1.19m soon without the fancy pinout or if you are patient to wait another month or so?

I am sorry to hear of your misfortunes. Unfortunately, sometimes bad things happen to good people, please accept my condolences.

As for waiting: take all the time you want. I for one am having trouble following all the improvements and changes as they seem to come so fast. I could use the pause to catch up!!! 

[hapless]

For the m-firmware you can set the voltage of the 2.5V reference via UREF_25 in config.h and for the k-firmware you have to change a value in Calibrate_UR.c, IIRC.

Thank you. What's funny is that I was actually looking at it and then got thrown off by the comment talking about "typical" and "datasheet". Could you elaborate on why you both have 2495mV as a default value there? I know that some (older?) vrefs are set to that instead of 2500mV, but we are all talking about precision references @2.5V here, and yet the software expects 2.495V by default?

[gojimmypi]

awesome, thanks!  Can you tell me where / if the source code is available? I tried searching on github & didn't find it.

http://www.mikrocontroller.net/svnbrowser/transistortester/

awesome! thanks for the software link. I ordered my hardware last night... expected delivery a few weeks from China.

in the meantime while reviewing the source code, is it supposed to be open source for community contributions or not? Every file has this text:

(c) 2012-2015 by Markus Reschke (madires@theca-tabellaria [dot] de)

that certainly does not sound very "open".

See:  http://opensource.org/licenses

it would be a shame to have something so cool be protected by copyrights.

I would willing to contribute if indeed it is open source.....

[tom666]

in the meantime while reviewing the source code, is it supposed to be open source for community contributions or not? Every file has this text:

(c) 2012-2015 by Markus Reschke (madires@theca-tabellaria [dot] de)

that certainly does not sound very "open".
See:  http://opensource.org/licenses
it would be a shame to have something so cool be protected by copyrights.
I would willing to contribute if indeed it is open source.....

I think that if you use it only for personal use, it's all okay

For this kit I recommend using software by Karl-Heinz Kübbeler (I have successfully tested):
https://www.mikrocontroller.net/svnbrowser/transistortester/Software/trunk/?view=tar

You can directly use the contents from this folder:
https://www.mikrocontroller.net/svnbrowser/transistortester/Software/trunk/mega328_st7565_kit/?view=tar

[madires]

Thank you. What's funny is that I was actually looking at it and then got thrown off by the comment talking about "typical" and "datasheet". Could you elaborate on why you both have 2495mV as a default value there? I know that some (older?) vrefs are set to that instead of 2500mV, but we are all talking about precision references @2.5V here, and yet the software expects 2.495V by default?

That value is based on the typical voltage of some voltage reference we're using. You could check out the datasheet of the reference you got or measure the voltage with a accurate DMM (5.5 digits or better).

[indman]

Thank you. What's funny is that I was actually looking at it and then got thrown off by the comment talking about "typical" and "datasheet". Could you elaborate on why you both have 2495mV as a default value there? I know that some (older?) vrefs are set to that instead of 2500mV, but we are all talking about precision references @2.5V here, and yet the software expects 2.495V by default?

Excuse for volume presentation, I hope that the theory will be clear!
In the program by default  reference voltage for ADC 5v (external voltage on AVCC).
In case of detection on  input analog-to-digitial converter(ADC) less 1v the reference switches to internal reference supply source 1.1v.
Affect on internal reference 1.1v we can't (only if to select the processor with its minimum low value).
External reference 5v on AVCC (accuracy) depends on us!
  The source of sample voltage as reference supply source for ADC in this program doesn't work at PC4, and serves for correction of value of the reference voltage 5v the result of operation of ADC accepted in case of calculation. In firmwares of both authors correction is carried out on a formula:
  V5_calc=v5_avcc*v25Uref_pasport/V_izm_ADC PC4, where:
- V5_calc - calculated value of reference voltage 5v the result of operation of ADC  accepted in case of calculation.
- V5_avcc - value of voltage 5v on AVCC (the constant changed by the user)
- V25Uref_pasport - passport value of voltage of the REF (the constant changed by the user at Markus Reschke, at Karl-Heinz Kübbeler - in a code)
- V_izm_ADC PC4 - (real) voltage measured by ADC on PC4
  where the constants  V25Uref_pasport and V5_avcc are (in mV)?

At Markus Reschke in the config.h file:
#define UREF_VCC 5001
#define UREF_25 2495

At Karl-Heinz Kübbeler in the config.h files
#define U_VCC 5000
and  Calibrate_UR.c (value 2495)
ADCconfig.U_AVCC = (unsigned long) (ADCconfig.U_AVCC (unsigned long) * 2495) / mv2500;

Now, taking into account constants the formula assumes an air: V5_calc=5000*2495/V_izm_ADC PC4.
From a formula it is visible that 2495/V_izm_ADC PC4 coefficient of correction of an error of operation of ADC and in case of =2495 it is equal  1.0, according to V5_calc=5000*1=5000!
  If it is very simple, that, then when the REF on 2495mV and real voltage will be connected to PC4 on AVCC will be 5000mV!

PRACTICAL RECOMMENDATIONS!
- to select AVCC and REF for parameters the close to AVCC=5000mV and URef=2495mv or in case of assembly of a firmware to enter the real
values of voltage of AVCC and URef.
- in case of less exact AVCC to use more exact REF, otherwise it is no need also an ear REF, otherwise it is no need and will worsen result!
Good results with MCP1702-5002 with specification of V5_avcc without REF.

[madires]

that certainly does not sound very "open".

See:  http://opensource.org/licenses

it would be a shame to have something so cool be protected by copyrights.

I would willing to contribute if indeed it is open source.....

Haven't you read the README explaining the license issue? The whole project is OSHW but without a proper license so far. Also a copyright doesn't contradict an open source license. The problem will be fixed in about 2.5 months

[hapless]

@indman
Although I already had my answer by the time you posted, I appreciate your effort. For the benefit of other readers, I'd recommend changing "tension" to "voltage" everywhere in your post.