$20 LCR ESR Transistor checker project

[JacquesBBB]

Some progress this will be useful for mojoe :

In the makefile, one  needs to comment the two line s

#CFLAGS += -DLCD_ST7565_H_FLIP=1
#CFLAGS += -DLCD_ST7565_V_FLIP=1

Also the font is better with

CFLAGS += -DFONT_6X8

I have still to correct the contrast

done with
LCD_ST7565_RESISTOR_RATIO = 5
instead of
LCD_ST7565_RESISTOR_RATIO = 4

 and to make it understand that the screen is much larger, or essentially to
add a new line at the end of each line.
Apart from that, it seems to work, with the proper graphics, although I have not fully tested it, and compared with the
original firmware.

[JacquesBBB]

Here is an update on my experience in using the
mega328_st7565   firmware  from Karl-Heinz Kubbeler v.1.11k on a
ebay cheap ESR meter.

After the modification of the Makefile as I mentioned before,
the  display is OK, but different from the original.

The nice thing of the board I have, is that I could just squeeze a zif socket on the original one.
This was extremely useful for testing various versions of the firmware.

In the attached  pictures, you have in the first three pics the New firmware that I   uploaded,
and in the next two the original firmware display, which is also supposed to be 1.11k,
according to the seller.

The last picture show the setting of the zif socket.

As you see,  the font is the same, but there is a blank line spacing in the original firmware.
I did not find any option to add these blank lines, which makes the output nicer.
Does anybody knows what to do to add these lines, and in general to
make better use of the large LCD display ? 

On the opposite, the "show data" item is missing in the original firmware,
but the version is visible in the selftest : also 1.11k

The nice thing is that I will now be able to upgrade the firmware, and to calibrate the  meter which is a
large improvement with respect to my previous one based on an atmega168 SMD  (with sanded brand).


 

[mojoe]

Hey, thanks for posting. I'll have to try what you did on mine and see if it works. Good work.

[mojoe]

Since you still have the original firmware, could you read it and post it somewhere so I have a copy? As I mentioned before, what I read from the SMT chip on mine is somehow corrupt and I can't flash it back with avrdude.

Thanks.

[mojoe]

Some progress this will be useful for mojoe :

Thanks to your posting, I now have my Fish8840 tester working again. Some of my makefile settings are different than yours.

makefile options:

FLAGS += -DLCD_ST7565_H_OFFSET=4
CFLAGS += -DLCD_ST7565_H_FLIP=1
# CFLAGS += -DLCD_ST7565_V_FLIP=1
CFLAGS += -DFONT_6X8
CFLAGS += -DBAT_NUMERATOR=66 (I measured the battery voltage and it matches with the displayed voltage with this)

To get the double-spaced lines, look in the file lcd-routines.c:

Change:
lcd_set_cursor(1,0)
lcd_set_cursor(2,0)
lcd_set_cursor(3,0)

To:
lcd_set_cursor(2,0);
lcd_set_cursor(4,0);
lcd_set_cursor(6,0)

The only anomaly I have is that although 3 is the best LCD contrast value that works, it really is a bit too light. It seems as though I need a 2.5 or so. The original firmware had the screen contrast set higher.

Joe

[mojoe]

One more anomaly. The transistor symbol in the lower right corner of the display is too low. Part of it is cut off. I tried setting the line spacing back to single and it makes no difference. There must be another setting to position the symbol. I just have to find it.

[mojoe]

Here is an update on my experience in using the
mega328_st7565   firmware  from Karl-Heinz Kubbeler v.1.11k on a
ebay cheap ESR meter.

When you test a diode or cap, do you get a graphic symbol in the lower right for it, or do you get the character symbol on the first line? For transistors, I get the graphic symbol. For two-terminal devices, I get the character symbol.

[JacquesBBB]

When you test a diode or cap, do you get a graphic symbol in the lower right for it, or do you get the character symbol on the first line? For transistors, I get the graphic symbol. For two-terminal devices, I get the character symbol.

For both the original and the  compiled firmware, I  have diodes, resistors, and capacitors graphics on the first line,
and transistors and mosfets with the graphic.

Thanks for  the lcd lines fix. It works, and now I have the same display as in the original firmware. I do not have to fix the battery voltage, as it is pretty close to the measure (8.6 instead of 8.7).

As you see in my pictures, the graphics is the same in the compiled firmware and the original one.

I believe that  the are some harware differences in the boards, and most probably mine is intended to be very
close to the original design. I did not check the values of the components, but it looks like.

[hapless]

That board looks exactly like the one with 1602. They just slapped a large display on an old design, I think. If it comes with 10-microfarad aluminum caps, it might benefit from replacing those, but otherwise it does follow the original design fairly closely.

I can't quite make it out on the pictures, but I think the offset may need to be set to 0 instead of 4. This type of display benefits from larger fonts, in my experience, but I'd like to make the icons fit in with the larger font... Maybe I'll learn to do that someday.


By the way, I've got another puzzle. I have a coil that's about 7.1 Ohm as measured with a DMM. My TT measures the resistance as 7.1 Ohm correctly. Then I insert a resistor in series to the remaining test point. It measures the resistor correctly, but suddenly thinks that the coil is about 19 Ohm! 

[madires]

By the way, I've got another puzzle. I have a coil that's about 7.1 Ohm as measured with a DMM. My TT measures the resistance as 7.1 Ohm correctly. Then I insert a resistor in series to the remaining test point. It measures the resistor correctly, but suddenly thinks that the coil is about 19 Ohm! 

Is the second resistor about 19 - 7.1 = 11.9 Ohms? Does that happen also for any other two resistors?

[hapless]

No, that would be too easy! It's 100 kilohm. Also tried with 500. Same result. Tried with two different testers, one with 1602 and one with 7565 display. Sometimes the coil's resistance is shown first, sometimes the other resistor's resistance is shown first, seemingly at random.

Seems to happen only with coils, but I don't have small value resistors to check at the moment.

[RedOctobyr]

Whew, this is quite a thread! Read through the whole thing, though that doesn't mean I understood all of it. Tons of great info! Thank you to Karl-Heinz, madires, and everyone who has contributed to this amazing tool/project!

By the way, thank you to the people who mention something like a hardware difference, and then explain the ramifications of the difference Just saying that one has a crystal, or doesn't have a 2.5V reference, etc, doesn't tell some of us what that really means, or why it's good/bad. Those kinds of explanations are appreciated 

I just ordered mine, I went with this one:
http://www.ebay.com/itm/251334183807?_trksid=p2059210.m2749.l2649&ssPageName=STRK%3AMEBIDX%3AIT

It's supposed to have the M328 chip, on a socket. I don't know if it can take an ISP header. It has the 2-line display, which should help with firmware compatibility. It lists an 8MHz crystal.

Now for the embarrassingly-easy question (I'm new at all this). One thing I'd like to use it for is in-circuit ESR testing. So I'll need some kind of pointed probes which can also plug into the tester. Presumably into the ZIF socket. Is there anything that's recommended for that?

I'm sure this is incredibly basic, which is why it never even comes up. But I'd like to order other things I'll need, while this is being shipped. And normal multimeter probe bananas will be too fat to plug in, unless I cut their ends off. There is one version (the one with a case) which includes leads, but I don't see much discussion of them apart from that.

As I understand it, I should have an appropriate capacitor available for use during calibration. And I will need to re-read some things to better understand what would be needed to update the firmware (no idea what mine will have, but almost certainly an old version; the listing says 2013 software).

[madires]

Now for the embarrassingly-easy question (I'm new at all this). One thing I'd like to use it for is in-circuit ESR testing. So I'll need some kind of pointed probes which can also plug into the tester. Presumably into the ZIF socket. Is there anything that's recommended for that?

I'm sure this is incredibly basic, which is why it never even comes up. But I'd like to order other things I'll need, while this is being shipped. And normal multimeter probe bananas will be too fat to plug in, unless I cut their ends off. There is one version (the one with a case) which includes leads, but I don't see much discussion of them apart from that.

Maybe some multimeter probes soldered to a pin header matching the ZIF socket. Add some heat shrink and have fun.

[madires]

No, that would be too easy! It's 100 kilohm. Also tried with 500. Same result. Tried with two different testers, one with 1602 and one with 7565 display. Sometimes the coil's resistance is shown first, sometimes the other resistor's resistance is shown first, seemingly at random.

Seems to happen only with coils, but I don't have small value resistors to check at the moment.

Thanks! I'm able to reproduce that issue with the k-firmware. For a 1mH coil the difference is minimal but for a 15mH coil it's 142 Ohms (with a 1k and 100k resistor).

Update: The issue seems to occur if the resistance of the coil is < 10 Ohms.

[hapless]

Thank you for looking into this. BTW, unrelated, but might be something for Karl-Heinz to look at for the next update, there's a new (I think) icon for bipolar transistors with a protection diode. If I use a transistor without a diode and add a diode to it in the "wrong" direction, this is reflected inline, but the icon remains the same. Kind of strange to look at. Not that it's a normal condition, of course, but still...

It's supposed to have the M328 chip, on a socket. I don't know if it can take an ISP header. It has the 2-line display, which should help with firmware compatibility. It lists an 8MHz crystal.

Now for the embarrassingly-easy question (I'm new at all this). One thing I'd like to use it for is in-circuit ESR testing. So I'll need some kind of pointed probes which can also plug into the tester. Presumably into the ZIF socket. Is there anything that's recommended for that?

I'm sure this is incredibly basic, which is why it never even comes up. But I'd like to order other things I'll need, while this is being shipped. And normal multimeter probe bananas will be too fat to plug in, unless I cut their ends off. There is one version (the one with a case) which includes leads, but I don't see much discussion of them apart from that.

As I understand it, I should have an appropriate capacitor available for use during calibration. And I will need to re-read some things to better understand what would be needed to update the firmware (no idea what mine will have, but almost certainly an old version; the listing says 2013 software).

That one looks just like the one I have, seems pretty popular. It has no ISP header nor pads for it. But the chip is removable, so the header is not really needed all that much if you're just going to update the firmware once in a blue moon.

I don't see why someone would buy a "naked" version now that ones with a case are available. I bought it for the sole purpose of mercilessly experimenting with the hardware, but if I needed one for just testing components, I'd get one with a case for sure. The one you mentioned is the one with a blue case, right? I believe it's very similar internally to what you have ordered. A quick overview of what's available on the famous auction site has shown that these also might come with leads: Mega328 Transistor Tester Diode Triode inductor Capacitance ESR Meter LCR LED and ATmega328 version inductor-capacitor ESR Transistor tester meter Digital lcd. If I were to get one for everyday use, it would probably be this: M328 transistor tester can be measured LCR / transistor / MOSFET / SCR (spelling and grammatical errors are courtesy of the sellers). It can take leads and also has the frequency option. Doesn't seem to come with leads though.

This is not something I'd recommend someone to do, but for quick makeshift test leads that fit into a ZIF socket I use flexible breadboard jumpers.

I'd go with what madires advised for now. I would try to use leads that are as short as possible to minimize stray capacitance and lead resistance.

As for the capacitor, as I understand it, your only concern should be that it's above 100nF and has the least possible amount of leakage. Nothing too special is required for calibration, which is one of the reasons I like this project so much.

[RedOctobyr]

madires, hapless, thank you both for the feedback. hapless, that last tester you linked to (with frequency functions) looks pretty nice. But I'll start with what I ordered, and see how that works for me. I'm guessing it'll be fine for my current purposes.

If I may ask, since mine is apparently like yours, are there modifications that I should plan on doing to it? I don't know if this design is missing something important. ESR/capacitance checking is what I forsee for uses at the moment, for trying to help troubleshoot issues.

I'll probably try multimeter leads to start, maybe cutting down their length if needed. The mention of flexible breadboard jumpers made me realize maybe I could use some similar small cables I have sitting around, thanks 

[Halfdead]

Has anyone else had their tester die?
 

I think the ATMEGA328P in my unit died, the unit draws over 200mA when turned on and the display doesn't power on properly, the display goes to nearly full brightness when you power the unit on and then dims very low nearly instantly after.

[madires]

Update: The issue seems to occur if the resistance of the coil is < 10 Ohms.

I've discussed that with Karl-Heinz. The cause of the issue is the ESR measurement which is also used to measure resistances less than 10 Ohms in the k-firmware. The measurement method doesn't deliver valid results for inductors. If an inductor is detected the low resistance measurement is skipped. And since the check for inductors is performed only if one single component is tested, an inductor and a resistor connected to all three probes are detected as two resistors. In that case the low resistance measurement is run and you'll see the wrong resistance value for the inductor. The m-firmware uses a different method for low resistance measurements (dedicated function) which also runs fine for inductors, but I don't know if Karl-Heinz will add it to the k-firmware (flash size constraints).

[hapless]

Interesting. So, if I use a low-value wirewound potentiometer and insert all three leads, it will not show the correct results? This is definitely something to be aware of.

How does your function compare to the ESR one in terms of resolution and accuracy? If he decides not to include it, I might just have to make a franken-firmware. I don't compile with UART and I enable the use of EEPROM, so should have enough memory for the modification. I'm also trying to figure out why the code for the frequency generator (function switch_frequency) is so lengthy while looking somewhat repetitive. Maybe folding it into a smaller function will free up some flash, too...

Has anyone else had their tester die?
 

I think the ATMEGA328P in my unit died, the unit draws over 200mA when turned on and the display doesn't power on properly, the display goes to nearly full brightness when you power the unit on and then dims very low nearly instantly after.

Is anything getting hot on the board? If the chip is removable, remove it and power on the unit. If it behaves the same way without the chip, check the regulator and its output capacitor.

If I may ask, since mine is apparently like yours, are there modifications that I should plan on doing to it? I don't know if this design is missing something important. ESR/capacitance checking is what I forsee for uses at the moment, for trying to help troubleshoot issues.

It should work well out of the box, but I would advise replacing those 10-microfarad capacitors with something better. On the output side of the regulator you may want to put a tantalum capacitor of about twice as much capacitance, and the one blocking the battery can be about 10 times higher. Mine are 47 and 100 respectively, but 47 might be a little overkill. This will improve the stability of the regulator and make the results a little more repeatable. Other than that, all modifications are totally optional. I like having a 2.5V voltage reference and a 1nF cap on the aref pin. Neither seems to affect the numbers too much for me though.

[Halfdead]

I removed the ATMEGA328P and the current drawn goes down to about 7 mA which seems reasonable and the display is now at full brightness but displaying gibberish as to be expected.


I think it's just a dead ATMEGA328P, but I've no idea how it died or where to get replacement firmware that's compatible with this unit.

Has anyone else had their tester die?
 

I think the ATMEGA328P in my unit died, the unit draws over 200mA when turned on and the display doesn't power on properly, the display goes to nearly full brightness when you power the unit on and then dims very low nearly instantly after.

Is anything getting hot on the board? If the chip is removable, remove it and power on the unit. If it behaves the same way without the chip, check the regulator and its output capacitor.

[madires]

Interesting. So, if I use a low-value wirewound potentiometer and insert all three leads, it will not show the correct results? This is definitely something to be aware of.

The displayed values for the potentiometer will be fine. Only if you add an inductor of a few mH or more the tester will show the wrong resistance value for that inductor. But if you got just the inductor connected to the probes the resistance diplayed will fine again.

How does your function compare to the ESR one in terms of resolution and accuracy? If he decides not to include it, I might just have to make a franken-firmware. I don't compile with UART and I enable the use of EEPROM, so should have enough memory for the modification. I'm also trying to figure out why the code for the frequency generator (function switch_frequency) is so lengthy while looking somewhat repetitive. Maybe folding it into a smaller function will free up some flash, too...

The resolution is the same (0.01 Ohms) and accuracy seems to be about the same also. If you use the optional functions I'd recommend to add a rotary encoder for a more convenient UI. In case you got a tester with text display you can run the m-firmware anyway.

[hapless]

Thanks, I actually have a couple of encoders lying in front of me just asking to be connected. Looks like the test button's behavior is automatically modified when the resistor dividers are connected. 

TBH, I don't remember what happened. I tried the m-firmware, liked it, then found something that I couldn't live without in k-firmware and switched back.  That was before I even started reading the forums. The problem is probably long gone by now.

I think I'm going to try to lift the function from m-firmware and transplant it into k (unless K.-H. beats me to it, which he probably will if he decides to use it). I wonder if there's a way to contribute back without the use of a versioning system? Maybe by posting diff files somewhere?

Interesting. So, if I use a low-value wirewound potentiometer and insert all three leads, it will not show the correct results? This is definitely something to be aware of.

The displayed values for the potentiometer will be fine. Only if you add an inductor of a few mH or more the tester will show the wrong resistance value for that inductor. But if you got just the inductor connected to the probes the resistance diplayed will fine again.

I see. Thanks for clarifying.

I removed the ATMEGA328P and the current drawn goes down to about 7 mA which seems reasonable and the display is now at full brightness but displaying gibberish as to be expected.


I think it's just a dead ATMEGA328P, but I've no idea how it died or where to get replacement firmware that's compatible with this unit.

This would mostly depend on the display and layout of the board. How about attaching pictures of what you have?

[Mafiawarsfourlife4]

In case anyone is interested in the graphics versions: they are described and sold within China on a Web site in Chinese only:
http://fish8840.taobao.com/ I prefer to open it in Google Chrome, which offers to translate it automatically, or on a right-click. I don't know whether the many sources of these devices buy from fish8840 (in which case they'd be all be of equal quality) or make their own copies.

Presumably fish8840 is the originator of the Chinese hardware versions (although they themselves may be imitated!), and weiweitm adapted the software. The reason for production of a graphics version given is that a text-only display in Latin characters is difficult for Chinese people, which makes sense; there are traditional and simplified Chinese, and English-language versions. There are two versions, an older one with a green display, and a later one with a slightly larger blue display; both are sold, with the blue one a little more expensive. The pictures don't show a crystal visible on the board; it could be there but hidden. As far as I can tell the later board was released in March 2014; the only software appears to be weiweitm's version 2.1 (with no mention of providing source code). I don't know if there is any provision at all for software updates (socketed chip, header, any intention to produce downloadable updates). There is a picture of an elegantly cased version, branded Yiwanjia, but no indication (such as price) that it exists.

Foreigner Markus F is acknowledged. The translation is difficult to understand in detail; I think Translate's "shelves without shell" are uncased boards!

Web pages of the two hardware versions (Chinese):
Larger screen, blue, backlit 25/3/2014, software shows as 2.1 on pictures:
http://item.taobao.com/item.htm?spm=a1z10.5.w4002-2323427594.16.IhvBUf&id=36323329572
Smaller green graphics screen version:
http://item.taobao.com/item.htm?spm=a1z10.5.w4002-2323427594.19.IhvBUf&id=37659767669

i need to repair this transistor tester and i dont know where to start?any help would be great.

[RedOctobyr]

I just ordered mine, I went with this one:
http://www.ebay.com/itm/251334183807?_trksid=p2059210.m2749.l2649&ssPageName=STRK%3AMEBIDX%3AIT

...
And I will need to re-read some things to better understand what would be needed to update the firmware (no idea what mine will have, but almost certainly an old version; the listing says 2013 software).

Mine arrived today, ordered 11 days ago. I'm pleasantly surprised, I was expecting it to take longer.

Another nice surprise, it has FW 1.11k. That version appears to be dated January 5th, 2015, from what I can see, and my unit was ordered about a week after that. I'm glad they're at least loading current firmware, so I don't have to worry about doing that right away.

I tested some capacitors and resistors that I had sitting around, very cool! A 100 Ohm 2% resistor measured as 98.7 Ohms. I did the calibration routine, using a 10uF capacitor. The readings didn't seem to change meaningfully after the calibration; perhaps it was calibrated before shipping.

[BlueBill]

Is there anywhere FW 1.11k can be downloaded?

Never mind, found it. Currently 1.16
http://www.mikrocontroller.net/svnbrowser/transistortester/Software/Markus/