Author Topic: Help with ThaiKit Curve Tracer NOW W/ Calibrated Instrument OpAmp  (Read 39358 times)

0 Members and 1 Guest are viewing this topic.

Offline watchmakerTopic starter

  • Frequent Contributor
  • **
  • Posts: 383
  • Country: us
  • Self Study in EE
    • Precision Timepiece Restoration and Service
I changed the to account for the improvements.  See  end of the thread.



I assembled this kit and made a nice enclosure, etc.  I have twice blown up signal transistors.  I feel like a watchmaking student who keeps testing his watch hoping for an improved result with no changes.  Neophyte.  But I like the puzzle to solve.

I have the schematic and I also looked up the Elektor article for the waveform information.

I verified the 2 ICs are ok and the same with the two transistors via substitution.  Both chips give out clean crisp signals.

I am leaving both jumpers open; small signal, NPN.  Both relays click when jumpers are added.  Checked all diodes out of ckt.

So I am doing something incredibly stupid.

Below are my waveforms at the collector and base inputs.  Emitter is zero.  I looked for the step signal from pin 9 of the 4024 chip but it is a simple square wave.  No transistor being tested.  I get the sawtooth at the scope horizontal output but nothing at the horizontal where the elektor article sez I should get a step (again, I have no transistor being tested/smoked).

I get that the emitter feeds the last opamp before the power transistors.

Can someone who has built this confirm my signals?

Conditions are :

No transistor
NPN
Small Signal

THANKS!!!

2022677-0

2022683-1







« Last Edit: February 28, 2024, 04:39:19 am by watchmaker »
Regards,

Dewey
 

Offline Jwillis

  • Super Contributor
  • ***
  • Posts: 1716
  • Country: ca
Re: Help with ThaiKit Curve Tracer (Urghhhhh)
« Reply #1 on: February 19, 2024, 12:25:48 am »
Looks close to what I get with same settings. Sorry for the scope photo. Hantek USB doesn't work (piece of junk).  Pretty sure it's the same kit as yours. Added the signal from pin 9  of the 4024.
« Last Edit: February 19, 2024, 12:38:37 am by Jwillis »
 

Offline Jwillis

  • Super Contributor
  • ***
  • Posts: 1716
  • Country: ca
Re: Help with ThaiKit Curve Tracer (Urghhhhh)
« Reply #2 on: February 19, 2024, 12:40:18 am »
Which small signal transistor are you testing. If I have one I'll see if I can blow it up.

Now you have me wondering what's going on. Large transistors test fine but as you, small signal transistors quickly get hot and fail. I don't use many small signal transistor in special applications. Mostly power applications. So I didn't notice the problem until now.
So given the evidence, your probably not doing anything wrong and there is an issue with the kit.
« Last Edit: February 19, 2024, 01:06:34 am by Jwillis »
 

Offline Jwillis

  • Super Contributor
  • ***
  • Posts: 1716
  • Country: ca
Re: Help with ThaiKit Curve Tracer (Urghhhhh)
« Reply #3 on: February 19, 2024, 02:14:26 am »
Little update. The tracer works on most of my transistors. It won't test Darlingtons. BC517 Darlinton burns up and TIP142 (also a Darlinton) just shows a dot in the middle of oscilloscope and gets hot. Bost tested fine on a BSIDE ESR02 Pro.
 

Offline watchmakerTopic starter

  • Frequent Contributor
  • **
  • Posts: 383
  • Country: us
  • Self Study in EE
    • Precision Timepiece Restoration and Service
Re: Help with ThaiKit Curve Tracer (Urghhhhh)
« Reply #4 on: February 19, 2024, 03:26:53 am »
Thank you and I'll be damned.  I spent 3 DAYS on this following all the signals and marking the board with w/ the paths. |O

I just found this youtube.  He does an excellent job of explaining the design and its limitations.  He made me think I need to look at my 1 ohm resistor.


Regards,

Dewey
 

Offline Jwillis

  • Super Contributor
  • ***
  • Posts: 1716
  • Country: ca
Re: Help with ThaiKit Curve Tracer (Urghhhhh)
« Reply #5 on: February 19, 2024, 04:26:37 am »
Thank you and I'll be damned.  I spent 3 DAYS on this following all the signals and marking the board with w/ the paths. |O

I just found this youtube.  He does an excellent job of explaining the design and its limitations.  He made me think I need to look at my 1 ohm resistor.

Oh Wow. Great video. Now I can take a second look at mine and make those improvements. Thanks!
I was thinking about that resistor too and wondered if adding another rotory switch or switches to change the value the peak power. like on that other curve tracer he described.
 

Offline watchmakerTopic starter

  • Frequent Contributor
  • **
  • Posts: 383
  • Country: us
  • Self Study in EE
    • Precision Timepiece Restoration and Service
Re: Help with ThaiKit Curve Tracer (Urghhhhh)
« Reply #6 on: February 19, 2024, 09:23:59 pm »
2024651-0 |O |O |O

OK, I have no clue.  I have scope pics of every used IC pin and the Collector and Base signals.  Everything looks fine to me but the base signal, but what do I know and it eats transistors.

THANKS for any help!  I am not giving up, but I will put it away for a while.

According to the youtube above, the signal at the base should look like this:  https://youtu.be/7exRUNCQ-hI?t=827s

I even removed the relays to determine the contacts and operation.  These are marked on my PCB pic.

Conditions are:
NPN
Signal
No transistor in test

Mine looks like:

2024534-1

My Collector out looks like this:

2024540-2

My PCB looks like this:

2024651-3

Here is my marked up schematic:

2024552-4

The signals at the pins for the 4024 and the tl074 will be listed in the next posts



« Last Edit: February 19, 2024, 09:51:19 pm by watchmaker »
Regards,

Dewey
 

Offline watchmakerTopic starter

  • Frequent Contributor
  • **
  • Posts: 383
  • Country: us
  • Self Study in EE
    • Precision Timepiece Restoration and Service
Re: Help with ThaiKit Curve Tracer (Urghhhhh)
« Reply #7 on: February 19, 2024, 09:29:44 pm »
Here are the signals at the 4020 counter The caption is the location (A=4020 B=TL074)

Pin 1 4024
2024564-0

Pin 6
2024570-1

Pin 9
2024576-2

Pin 11
2024582-3

Pin 12
2024588-4
Regards,

Dewey
 

Offline watchmakerTopic starter

  • Frequent Contributor
  • **
  • Posts: 383
  • Country: us
  • Self Study in EE
    • Precision Timepiece Restoration and Service
Re: Help with ThaiKit Curve Tracer (Urghhhhh)
« Reply #8 on: February 19, 2024, 09:35:48 pm »
And the TL074 signals:

Pin1
2024594-0

Pin 2
2024600-1

Pin 8
2024612-2

Pin 9
2024618-3

Pin 12
2024606-4



Regards,

Dewey
 

Offline Jwillis

  • Super Contributor
  • ***
  • Posts: 1716
  • Country: ca
Re: Help with ThaiKit Curve Tracer (Urghhhhh)
« Reply #9 on: February 19, 2024, 09:54:03 pm »
Base and Collector look right. What Small signal transistor are you testing?
I'll check the other readings you got.
 

Offline watchmakerTopic starter

  • Frequent Contributor
  • **
  • Posts: 383
  • Country: us
  • Self Study in EE
    • Precision Timepiece Restoration and Service
Re: Help with ThaiKit Curve Tracer (Urghhhhh)
« Reply #10 on: February 19, 2024, 10:48:39 pm »
Thanks.  Be very careful about changing the voltage divider resistor (1 ohm) to 100 ohm 1 watt.  Put in a power transistor and the 100 ohm resistor burned right up.  Put the 1 ohm back in and it is happy.  I wonder if there is some kind of positive feedback going on under load????? (Superstitious thinking).

Ok.  I will check the morgue and see what has toe tags.  Thanks again!  I am also wondering (superstition again) if it is the diode cascade.  Substituted both op amps with same results on signals.  I keep looking at the bottom of the board but see nothing.  And I would think I would not get clean signals.

Also, tell me a transistor you have successfully (not barbecued) tested.  I may have it.

I KNOW it makes the most sense to build from the Elektor plans, but this is one helluva learning exercise.

My wife is getting concerned!
Regards,

Dewey
 

Offline Jwillis

  • Super Contributor
  • ***
  • Posts: 1716
  • Country: ca
Re: Help with ThaiKit Curve Tracer (Urghhhhh)
« Reply #11 on: February 19, 2024, 11:28:33 pm »
Most of the readings on 4020 and TL074 you got are the same as mine.
The picture below is pin 2  Px10 X10Mag 5mV/Div 500ms T/Div  from The V1050 because it produces better quality image than the Hantek at low levels.
I can get a better image if I used a probe grounding spring instead of the lead.
« Last Edit: February 19, 2024, 11:39:21 pm by Jwillis »
 

Offline Jwillis

  • Super Contributor
  • ***
  • Posts: 1716
  • Country: ca
Re: Help with ThaiKit Curve Tracer (Urghhhhh)
« Reply #12 on: February 19, 2024, 11:34:11 pm »
I need a ferrite beads on the transistor test leads. Any noisy oscillation from those and transistors get really hot. Which makes sense.

also need a better DSO. This Hantek is really P***ing me off.
 

Offline Jwillis

  • Super Contributor
  • ***
  • Posts: 1716
  • Country: ca
Re: Help with ThaiKit Curve Tracer (Urghhhhh)
« Reply #13 on: February 20, 2024, 06:46:33 am »

Also, tell me a transistor you have successfully (not barbecued) tested.  I may have it.


The transistors that work: 2n2222, 2n3904, 2n3906, A1015, C1815, BC547,BC548 up to BC558, S9013, S9014
Ones that won't work are BC517, BC337 and I'm not sure why.
Doing a little investigation it seems that transistors with an BEon voltage lower than 1V measure fine without heating up. The ones with an BEon voltage higher than 1V get hot and burn out after a while.
This occurs on both large and small settings so Darlinton power transistors like the TIP142, TIP122, TIP117 also won't work.


Edited The BC337 as being untestable. With correct connection it is testable.
« Last Edit: February 20, 2024, 09:55:49 pm by Jwillis »
 

Offline watchmakerTopic starter

  • Frequent Contributor
  • **
  • Posts: 383
  • Country: us
  • Self Study in EE
    • Precision Timepiece Restoration and Service
Re: Help with ThaiKit Curve Tracer (Urghhhhh)
« Reply #14 on: February 20, 2024, 10:27:08 am »
OK.  I will redo a 2222, this time I will solder it directly to the board.

BUT, what good is a test device that only works on some specific items and destroys the rest??  Are the reports of success from one time users?  I did ask the guy from that video to comment on this thread, but this may well be a dead end.

I looked at Microcap last nite and it seems both ICs are there and I placed components on the board to wire together.  Not sure what this will prove, but....
Regards,

Dewey
 

Offline watchmakerTopic starter

  • Frequent Contributor
  • **
  • Posts: 383
  • Country: us
  • Self Study in EE
    • Precision Timepiece Restoration and Service
Re: Help with ThaiKit Curve Tracer (Urghhhhh)
« Reply #15 on: February 20, 2024, 05:14:20 pm »
OK. To falsify hypothesis thatmy LED lighting is causing too much noise, I pulled at the old warhorse, the BK 501a.  Scrren shot of a 2N2222 is attached.

No difference with all lights on and all lights off.

As far as the ThaiKit, it smoked a 2222 soldered right to the board.

Interestingly, the creator of that youtube deleted my comment requesting he comment on this thread.

So, I am going to create MicroCap ckt.  Run it with a 2222 in the socket.  Measure the various currents, voltages and signals.  Then I am going to modify my board by adding test points for current (cut trace, drill for header pins).  I will compare the damned measurements against the sim.

I did notice the kit uses a tl074CN as opposed to a tl074 listed on the schematic.  I will look up the differences.

If I wasn't trying to learn, I would throw the thing away.  Now I am like a dog with an old bone.  The BK does not destroy what it does not like, and the Analog Discovery has a computer based curve tracer.  By the time you are done with a box, PSU, connectors etc, I think you are at 1/2 to 1/3 of the cost of either of those options with no headaches.

2026100-0
Regards,

Dewey
 

Offline jdragoset

  • Regular Contributor
  • *
  • Posts: 245
  • Country: us
Re: Help with ThaiKit Curve Tracer (Urghhhhh)
« Reply #16 on: February 20, 2024, 05:52:35 pm »
It seems to be missing a collector load resistor between the BD 139 and BD 140 emitters and the collector terminal.
Tektronix curve tracers had a switch so this could be set from low to high resistances.
The 1 ohm in the emitter circuit is the current shunt.
 

Offline watchmakerTopic starter

  • Frequent Contributor
  • **
  • Posts: 383
  • Country: us
  • Self Study in EE
    • Precision Timepiece Restoration and Service
Re: Help with ThaiKit Curve Tracer (Urghhhhh)
« Reply #17 on: February 20, 2024, 08:50:42 pm »
Thank you.  You prompted me to more closely compare the Eleckto ckt and the Thai Ckt.

OK.  I am new to semiconductors.

But first, when I look at the datasheets for the TL084CN and the TL074CN, especially the suggested apps in section 9 (and the block diagrams) these are two very different OpAmps.  I barely learned the 4 basic configurations, and maybe they are interchangeable, but it does not look like a direct interchange to me.

Among other things, the Elektor design very clearly specifies shielding the E,C,B and scope output lines.

No wonder it is so hard to get info from the seller.  He never even sent me the instructions and I populated the board based on the silkscreening although I upgraded the zeners.

Secondly, the Elektor ckt does not try to deal with signal vs power transistors.  The Elektor may be a candidate for the limiter resistor switching but it makes sense to start with the ckt as originally designed.

I think I will order some TL084CNs and play.  I have some prototype PCBs.
Regards,

Dewey
 

Offline Jwillis

  • Super Contributor
  • ***
  • Posts: 1716
  • Country: ca
Re: Help with ThaiKit Curve Tracer (Urghhhhh)
« Reply #18 on: February 20, 2024, 09:48:28 pm »
I think I'm getting close to figuring out the problem.
I redid the measurements on the transistors that didn't work. The BC517 Darlington won't work but the BC337 Does. I must have had the connection wrong before testing the BC337. So it seems Darlingtons are off the list of testable transistors.

At first I thought it may be internal bias resistors but 2 2N2222, in Darlinton configuration, is also untestable. Trailing transistor gets excessively hot. It's going into full saturation because of the gain increase!
Testing working  transistors base voltage is around 500 to 700mV and around 20mV across 1 ohm resistor or 20mA to the base , depending on transistor being tested. On Darlingtons the base voltage drops to 0V and 1 ohm resistor reads 655mV. That's 655mA to the base  It's drawing to much current for the tester to supply to the base. With a 1M resistor in series with base it becomes testable. This is because of  the high gain of Darilintons. Some extra rework here is needed because of the noise. A lot of noise due to the long test leads I have.
To much current to the base and the transistor goes into full saturation. Across a 1 Ohm resistor. the transistor exceeds it's maximum dissipation and burns out. So we need to limit the base current output on the tester for high gain transistors over 1000. The BC517 is a 30000 gain transistor.
 

Online Calambres

  • Frequent Contributor
  • **
  • Posts: 362
  • Country: es
    • Piso-Tones
Re: Help with ThaiKit Curve Tracer (Urghhhhh)
« Reply #19 on: February 21, 2024, 08:17:35 am »
Interestingly, the creator of that youtube deleted my comment requesting he comment on this thread.
I don't think he's deleting your commets. For some time now YouTube is not allowing links in the comments and a bot deletes all such comments in mere seconds.

Offline watchmakerTopic starter

  • Frequent Contributor
  • **
  • Posts: 383
  • Country: us
  • Self Study in EE
    • Precision Timepiece Restoration and Service
Re: Help with ThaiKit Curve Tracer (Urghhhhh)
« Reply #20 on: February 21, 2024, 02:18:03 pm »
Interestingly, the creator of that youtube deleted my comment requesting he comment on this thread.
I don't think he's deleting your commets. For some time now YouTube is not allowing links in the comments and a bot deletes all such comments in mere seconds.

Thanks!  My cynicism is showing.

I will leave a comment directing to this thread using verbal directions (and delete  my comment!.)
Regards,

Dewey
 

Offline watchmakerTopic starter

  • Frequent Contributor
  • **
  • Posts: 383
  • Country: us
  • Self Study in EE
    • Precision Timepiece Restoration and Service
Re: Help with ThaiKit Curve Tracer (Urghhhhh)
« Reply #21 on: February 21, 2024, 02:25:35 pm »
Attached is a marked schematic of the original ckt as described in Elektor.  There are some pretty interesting changes in the ThaiKit.

Two schottky diodes are changed to 4148s.  Now I have to learn about schottkys.  Another is that load resistor value.

Finally, he does not switch in IC1 pin 9 unless he wants to increase the traces.  His diode tree/waterfall for the step function is also smaller.

It will not be hard to revise the ThaiKit to the Elektor design while keeping everything on NPN so that I can see the difference in barbecue menus.

I am still befuddled that this ckt appears to work for some people.  Included components???

2027765-0
Regards,

Dewey
 

Offline Alex Nikitin

  • Super Contributor
  • ***
  • Posts: 1218
  • Country: gb
  • Femtoampnut and Tapehead.
    • A.N.T. Audio
Re: Help with ThaiKit Curve Tracer (Urghhhhh)
« Reply #22 on: February 21, 2024, 05:36:04 pm »
For starters you can increase the value of the current sensing resistor from 1 Ohm to 10 Ohm, that might considerably reduce the amount of unintended smoke (it will also change the vertical current scale to 100mA/V). The Schottky diodes would further lower the current limit and make things safer as well. As is, the current limit is close to 1A which would certainly fry any small device and might easily take out the output devices in the amplifier/integrator. A good idea would be at least a resettable fuse in series with the amplifier output, better still a proper current limit. For small devices you really want to limit collector current by about 50mA for a continuous loop. With modern digital scopes you might also investigate a "single shot" option, running only one curve tracing cycle and then stopping, keeping the dissipated energy at bay.

Cheers

Alex
 
The following users thanked this post: watchmaker

Offline jdragoset

  • Regular Contributor
  • *
  • Posts: 245
  • Country: us
Re: Help with ThaiKit Curve Tracer (Urghhhhh)
« Reply #23 on: February 21, 2024, 09:20:49 pm »
If R17 (10 ohm) is as shown in the diagram, it is only acting with C6 as an RC network rather than a load.
A load resistor would be between the emitter connection of BD 139/BD 140 and the "C" terminal of the DUT.
 

Offline Jwillis

  • Super Contributor
  • ***
  • Posts: 1716
  • Country: ca
Re: Help with ThaiKit Curve Tracer (Urghhhhh)
« Reply #24 on: February 21, 2024, 11:38:24 pm »
The tl071 and the tl081 are nearly identical. The most significant difference is TL071 has a lower noise.
You need to measure the current of the transistor under test. If the measurement across the 1 Ω resistor is to high the transistor will burn. Most transistors I tested measure 18mV to 23mV across that 1 Ω resistor. The ones that get hot and burn up over an extended time are Darlingtons only because of the Gain. The 1Ω  resistor has 400mV to 700mV depending on transistor being tested. By adding a voltage divider to reduce the Base voltage and current, I can see the curve trace of Darlingtons. I tested the BC517 this way and it doesn't get hot. With a 910K resistor between  base output and transistor base  and a 300K resistor from transistor Base to Emitter. This is because I'm essentially changing what the 100K and 7K5 current limiters do. This reduces the the base voltage and current there by reducing the current across  the collector emitter junction. 

You need to check the current across transistor under test by the voltage drop at the 1Ω resistor. That should be  around 20mV which is 20mA across transistor under test. 
Check the Base voltage of tracer with no load. That should be around 5V in NPN setting and -5V in PNP setting.
The Collector voltage  of tracer with no load should be around 5V in NPN setting  and -5V in PNP setting.
When a transistor is out under test, the base voltage should drop to about .7V
Re-check the current limiters. The 7K5 and 100K resistors. If those are not correct then to much current goes to the base of transistor under test. There should be only a few micro amps going to the base of the transistor under test when setting is at Small NPN.

On a 2n2222 when tested with the small NPN setting it does not get hot.Theirs only 23mV at the 1Ω resistor or 23mA across the 2n2222 CE junction. Base voltage should be around .63V. Emiiter voltage should be 23mV and Collector voltage is 4.4V. That works out to around 100mW that the transistor needs to dissipate.
If 2n2222 is tested with the Large NPN setting it gets really hot and fails after a minute or so.
If your transistors are blowing up theirs to much current going to the base.
« Last Edit: February 22, 2024, 12:17:51 am by Jwillis »
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf