5 Transistor ESR Meter Design

[E_Pepegamer]

*** Edited the message because I had a proper look to the schematic and many answers to my questions could be found there ***

Hi, thx for this version Anakiev. I am trying to make this one without having much electronics knowledge, however I am struggling a bit to find the correct parts.
I have exported the BOM list from kicard board. This generates the attached Excel.

1- From the resistor sizes I think these resistors should be 0.25W ones. But not quite sure.
2- Condensers do not show the volts parameter, only for the electrolitic ones.
3- R22 value is set to 'X', in the photo it seems a 18K resistor but again not totally sure.

Anyone can help me a bit with this?

[sdouze]

Thank you for your time to reply.

I have one more question, probably my meter is not very good.

With μΑ scale out of the box I can get the needle to go at 0μΑ (with the screw on meter) some times show 10μα sometimes less ore more (sometimes if move the meter can change) and  when I short the leads goes far right 50μΑ so I turn the pot to have exactly 50μΑ (0 esr) sometimes can not keep the zero adjustments.

I think that you are right, I think you have a bad meter movement. You can test the meter with a power supply, or battery and a few resistors. Use a  9V battery and a 330 kOhm resistor, that should give you around 27uA or 55% of full-scale. A 9V battery and 1M Ohm should give 9uA or 18% of the scale.

Regards,
Jay_Diddy_B

Hello,
I repaired the meter, and now it is working properly. I used a 15K resistor.

[sdouze]

*** Edited the message because I had a proper look to the schematic and many answers to my questions could be found there ***

Hi, thx for this version Anakiev. I am trying to make this one without having much electronics knowledge, however I am struggling a bit to find the correct parts.
I have exported the BOM list from kicard board. This generates the attached Excel.

1- From the resistor sizes I think these resistors should be 0.25W ones. But not quite sure.
2- Condensers do not show the volts parameter, only for the electrolitic ones.
3- R22 value is set to 'X', in the photo it seems a 18K resistor but again not totally sure.

Anyone can help me a bit with this?

Hello,
1.Yes 0.25W resistors are enough.
2.Most of the capacitors are ceramic and the minimum(I think) ceramic capacitor voltage is 50V also your power supply is 9V battery so you are ok with common ceramic caps.
3.The R22 value it depends on the meter you have.Check reply 422 in this thread where OP answered my question on how to choose R22 value.

[E_Pepegamer]

Hi sdouze, Jay_Diddy_B

Thx for your reply. I like it putting only 100 mV on the capacitor under test. This is really good compared to other simpler DIY solutions around. So I will try with the "anniversary edition" with 7 transistors.

I am in the middle of understanding everything. I found this great video explaining the Jay_Diddy_B 5 transistor meter:



I also read about other DIY solutions and there I found comments like "an open capacitor or with a high ESR will allow pulses (non-sinus vawes) to reach the zenner diodes and may burn them", which may not be the case for this design. Is the red LED detecting > 25 Ohms ESR related to this as a protection or could I skip that circuit part and just look at the meter?

I am going to start it right now in a breadboard, then make it in the perfboard 5x7 form factor, I will put all the feedback I can get with a 30€ oscilloscope /signal generator here. XD

[E_Pepegamer]

For those not getting 100Khz on the square signal generator the frequency depends on C3 ~1500p capacity, so test several ceramics until you find the one that gets closer to the 100KHz. The higher capacity the lower frequency you get.

I had to use two capacitors (with the most similar capacity i found) in parallel to get it right. None of the 1500p ones I had matched the proper value.

Some live action:

[Anakiev]

Sorry about the delay. I will add this information in my first post.

The BOM was auto generated and it's probably wrong. I've never looked at it.
I've used 1/4W 1% resistors and 25v/50v ceramic capacitors but 16v should be fine.
R22 is different for different uA meters. There is quite a lot information in the thread.

About the frequency:
I've noticed it's not 100kHz after I've made the scale and I didn't want to redo it. I haven't tried capacitor in parallel/series but I've tried with different standard values and the 1500pF was the closest.
I've also made the oscillator part with smd resistors to see if the wirewound resistors were the problem but they weren't.

@E_Pepegamer
What values have you used (in parallel) for 100kHz?
If there is something I've missed you can send a PM or ask again.

[BeetleJuice]

any chance to share the file you made with the flux capacitor?
thank you

[BeetleJuice]

Hi group,

I promised an update after I have tested the prototype.

I received the partially assembled board assemblies from JLCPCB. This is the first time that I have ordered assembled boards. The process was very smooth. I chose only to have them install parts in their basic library. You can have them install other parts, but there is a $3.00 per line item charge.



Here is a picture of the assembled board.

Schematic

This is the revised schematic:



The changes are:

1) The regulator was changed from the 7805 to AZ1117-5. This has a lower dropout voltage, but requires larger capacitors.

2) C4 was added. This is a d.c. blocking capacitor to allow shorted capacitors to be detected.

3) The circuitry Q4 and Q5 was added. The LED will illuminate if the capacitor being tested is less than 25 .

The rest of the circuit is the same as the original version.

Notes: This was redrawn in KICAD version 5.00 there is no connection between the component references in this version and the previous version or the LTspice model.

R22 must be chosen depending on the resistance of the 50uA meter movement. In my case the resistor is around 18k

If you search 5 transistor ESR meter on the internet you get quite a few hits from all around the world. A lot of people have done the right thing and given credit back to the original design presented in this thread. Other people have not. It would be nice if you make a copy or a derivative to credit the original work.

Files

I have attached a zipfile that should contain enough information to open the project in KICAD. Let me know if anything is missing.

I have also attached the LTspice model.

Regards,
Jay_Diddy_B
(Attachment Link)
(Attachment Link)

i made this version and it works fine, the voltages are correct. I bought a Japan-made 50uA ammeter, not the Chinese one. this meter when measured in ohms with a good multimeter (even with a milliohm meter) gives 1152ohm resistance. I think most ammeters have higher resistances.

i think its impossible to determine the value or r22, i even changed it with a 50k trimmer but no luck.i tested it with r22 shorted (!) and i think it works better!
Does anyone have any idea what's wrong with r22 or my meter?
thank you

[E_Pepegamer]

I have no Idea about R22, for my meter it was 27K Ohms. The 10k potentiometer bolt must be in middle way position.
I changed C3 with two 680p condensers in parallel (But these measured 720p in the multimeter). I measured 105KHz at the collector of Q2. However that was not the right place to measure. Frequency should be measured after the collector of Q3, because frequency slows down to 93Khz in my case.
So C3 has to be less than 1500p, but I do not know the exact value, maybe 1300p. The smaller capacity the faster frequency, and the opposite, the higher capacity the slower frequency you get.

Regarding the "short circuit detect" circuit I would also like to disable it at will, Is it a good idea to put a button that closes the circuit at R14 on press (instead of a jumper), or this will create "spikes"?

How can I make and print a good scale?

And about boxing it what should I do? Cut a plastic box myself ot there is anything a bit better?

PD: Mine is also working, I will put the "power supply" part in a different board, to make it a bit modular.

[kripton2035]

And about boxing it what should I do? Cut a plastic box myself ot there is anything a bit better?

using or investing in a 3d printer is the best thing to make on demand boxes for electronic projects nowadays...

[BeetleJuice]

guys i really am confused, my meter works the opposite way!
measured all voltages and are correct.
TP7 is connected to a 10K trimmer, the middle point of the trimmer is connected to the + of my 50uA meter, - of meter to TP8.
R22 is now 1500, still not working. replaced it with a 10/20/50/100K trimmer, i got strange results.

the most confusing issue is : when powering the meter it goes to 30uA, no matter what i do with the trimmer is impossible to bring it to zero uA !
moreover, when i measure a resistor, if it is 10ohms the needle goes to 10uA , but when i measure 20ohm resistor it goes to 20uA, meaning that with higher resistances the needle goes right not left!

how is this even possible! can't find any mistake!
 

[Anakiev]

@BeetleJuice By "flux capacitor" you mean the scale, right? I don't keep mine but there are examples in some of the first posts.

@E_Pepegamer I've made my scale with Inkscape but for a windows user there might be an easier tool. With this resistor chain I've measured all values I want:

It's made of 0.1 1   and 10 1% resistors. In theory it's not a good idea to use wirewound resistors because with the high frequency it "adds resistance". The cheapest option was to use smd resistors but I couldn't find all values with 1% locally. There were 5% but the error should be bigger. Although in my comparisons I've never seen any difference between smd and wirewound resistors. Definitely don't chain the resistors in breadboard - solder them. Shortening the leads should help but I don't think you will be able to see the difference.
Look here:
https://www.eevblog.com/forum/projects/5-transistor-esr-meter-design/msg1069445/#msg1069445
https://www.eevblog.com/forum/projects/esr-meter-adapter-design-and-construction/msg343855/#msg343855
After that with the original scale I've noted the shown uA. For example if 1 resistor shows 36.5 uA and the scale is 90 degrees with 50 divisions.then 90/50=1.8 and 36.5*1.8=65.7. So the deflection for 1 is 65.7 degrees.
The hardest part is to measure the dimensions of the original scale (and the mounting holes) to make the new one the exact size. If you have a datasheet take the data from there.
Don't forget to zero the meter. Print the scale in "actual size".

Edit: looking at your photo again: Use as less connections as possible. I am talking about all of the dupont wires you have.

My box is from plastic junction box. Jay_Diddy_B's one looks metallized from the inside. I guess this is to keep the high frequency inside the box rather then from interference.

[Jay_Diddy_B]

guys i really am confused, my meter works the opposite way!
measured all voltages and are correct.
TP7 is connected to a 10K trimmer, the middle point of the trimmer is connected to the + of my 50uA meter, - of meter to TP8.
R22 is now 1500, still not working. replaced it with a 10/20/50/100K trimmer, i got strange results.

the most confusing issue is : when powering the meter it goes to 30uA, no matter what i do with the trimmer is impossible to bring it to zero uA !
moreover, when i measure a resistor, if it is 10ohms the needle goes to 10uA , but when i measure 20ohm resistor it goes to 20uA, meaning that with higher resistances the needle goes right not left!

how is this even possible! can't find any mistake!
 

If all the dc voltages are correct, check the capacitors and any parts of the circuit that have capacitors in them.
The problem that you are seeing is probably in the Q7 end of the schematic.

Jay_Diddy_B

[Jay_Diddy_B]

Quote:

My box is from plastic junction box. Jay_Diddy_B's one looks metallized from the inside. I guess this is to keep the high frequency inside the box rather then from interferences.


My box is not metallized, it is metal. It is not necessary to build this circuit in a metal box or a shielded enclosure. The circuit will work fine in a plastic box.


Jay_Diddy_B

[Anakiev]

I've added a jumper to the new board to be able to disconnect R14 which completely removes the voltage offset and the current through TP5 and TP6. Now I have both 5 and 7 transistor versions in one.

Regarding the "short circuit detect" circuit I would also like to disable it at will, Is it a good idea to put a button that closes the circuit at R14 on press (instead of a ), or this will create "spikes"?

There will be bouncing and the led may flicker. But too fast to be seen.
The important thing is to make the scale with the "short circuit detect" connected because when I remove/put the jumper there is a small difference in the deflection BUT only while testing resistors. It doesn't affect the deflection when a capacitor is being tested (with and without the jumper).

I measured 105KHz at the collector of Q2. However that was not the right place to measure. Frequency should be measure after the collector of Q3, because it slows down to 93Khz in my case.

You have different frequencies at those two points? Is that right? I have to open mine to see.

[BeetleJuice]

well, believe it or not, i just stopped powering the unit from my bench PSU, i left the LDO out cause i burned it accidentally, so I gave the unit 5 volts from my PSU.
now the ldo is back and everything works fine!
also, i refreshed the solder to the caps near q7 but didn't find anything wrong.
either way, it works perfectly now, did the calibration and gonna do the scale also. i measured every 1uA how many ohms it sees.

any instructions for the scale?
do you think a switch to short the output would be useful or its better to short the leads at the end? maybe we should have a very short cable length?

thank you for the advice , the schematic and the pcb! 

[Anakiev]

did the calibration and gonna do the scale also. i measured every 1uA how many ohms it sees.

This is kind of the opposite way. How did you do that? If you've used a decade box this proven to be wrong:
https://www.eevblog.com/forum/projects/esr-meter-adapter-design-and-construction/msg343855/#msg343855

maybe we should have a very short cable length

Does it really matter if you zero the meter every time you change the test leads?

[BeetleJuice]

for me its easier to see a close-up 90 degrees of the needle to the marks of the original uA scale, then use a milliohm meter to measure the resistance.
either way you use your eye to see the needle and the marks, you can't have more accuracy than your eyes provide. i will make the new scale over the marks of the original scale of the panel meter.

sometimes after powering up the needle doesn't go to zero, that's the purpose of the trimmer, to make it zero. so yes i use the short a lot to check if the meter is always to zero. if you do it with crocodile clips it's a little boring to do it all the time. no, you don't have to zero it so many times, most of the time it keeps the zero.

the first test i did was with precision resistors and noting the uA. After that i completed the gaps i had by watching the needle and then measuring the resistance. it all fits very well if you see the table i posted. there is no error, the values agreed between the two methods.

[E_Pepegamer]

I measured 105KHz at the collector of Q2. However that was not the right place to measure. Frequency should be measure after the collector of Q3, because it slows down to 93Khz in my case.

You have different frequencies at those two points? Is that right? I have to open mine to see.
[ Attachment Invalid Or Does Not Exist ]
[/quote]

Indeed, measure frequency after Q3 to get the proper adjustment. I also noticed that you have to wait some minutes for capacitors to cool down after soldering in order to get definitive reads.

Thanks for the detailed scale adjustment explanation, that fine resistor necklace is a great piece of jewelcraft.

PD: I found a gameboy advance SP power adapter in the trash, it provides stable and clean 5.2 V. I bought the connector, I think I will skip the 9V converter / filter with that.

[Anakiev]

@BeetleJuice I've asked how do you do that: with potentiometer or with decade box or by calculating the resistance with the Ohm's law. And to warn you that the decade box may give you wrong results.

@E_Pepegamer My opinion is not to use switched-mode power supply for test equipment.