Author Topic: Modern day Curve Tracer?  (Read 3575 times)

0 Members and 1 Guest are viewing this topic.

Offline DaJMasta

  • Super Contributor
  • ***
  • Posts: 2150
  • Country: us
    • medpants.com
Modern day Curve Tracer?
« on: September 22, 2018, 06:39:40 am »
It seems like curve tracers stopped being manufactured quite some time ago, and while I'm sure the need for characterizing discrete semiconductors is way less than when they were being made, is there any modern equivalent piece of test equipment?

There's source measure units, which can be configured to do these kinds of graphs with a little extra software, or the fancy new graphical ones that can do it on screen directly, but in the time between the 80s CRT curve tracers and the modern graphical SMU, has there been any single device that gave you this visualization of DC performance?  Is there some other kind of device that's taken over the role as a secondary function?  Has everyone interested just moved over to using discrete power supplies and multimeters or SMUs in a mainframe kind of configuration with external visualization software?
 

Offline ivaylo

  • Frequent Contributor
  • **
  • Posts: 642
  • Country: us
Re: Modern day Curve Tracer?
« Reply #1 on: September 22, 2018, 06:49:55 am »
Yes, SMUs and these kinds of things- https://www.peakelec.co.uk/acatalog/dca75-dca-pro.html
You don’t have to write software for all SMUs, BTW. I have a Keithley 2400 series (clearly not graphical) which offers full plotting functionality over its web interface. So yes, old anchor weights got replaced by modern stuff.
 

Offline filssavi

  • Frequent Contributor
  • **
  • Posts: 431
Re: Modern day Curve Tracer?
« Reply #2 on: September 22, 2018, 06:52:15 am »
Curve tracers are still alive and kicking they just changed name to semiconductor device analysers, you probably are not aware of them since they are probably ludicrously expensive (like if you have to ask the price you can’t afford it expensive)

https://www.keysight.com/en/pd-582565-pn-B1500A/semiconductor-device-analyzer?cc=US&lc=eng
 
The main reason is that those machines are used only in large research institutes that do experimental device fabrication and fabs, both places where spending 2/300k on an instrument is not that bad (compared to things like clean room upkeep)


 

Online MasterTech

  • Super Contributor
  • ***
  • Posts: 1167
  • Country: 00
Re: Modern day Curve Tracer?
« Reply #3 on: September 22, 2018, 07:15:29 am »
Also curve tracers are still alive in the igbt area where a typical SMU won’t even turn on a little one of those bricks. I’m talking igbs with Vces >1 -2 kV and 100s of A in Ic
 

Offline TiN

  • Super Contributor
  • ***
  • Posts: 4399
  • Country: us
  • PPM Alert!
    • xDevs.com
Re: Modern day Curve Tracer?
« Reply #4 on: September 22, 2018, 11:14:57 am »
Except Keysight B1500, there are also older generation ones, like E5260/E5270B or even older Agilent 4142B mainframes. These targeted for remote operation in the ATE rack, but there are old variants of 4145 and similar which also have CRT display  :). 4142B with SMU modules (supported up to 8 modules mid-power or four high-power/high-voltage ones or a mix) can be had for under 1K$ if you patient. Writing python software to control these ain't that hard and can be done in few evenings and Raspberry Pi.

I have done curve tracing and JFET matching, using my 4142B some time ago, when was busy on fixing nanovoltmeter input stage.



Keithley have SCS 4200 system (old and facelifted one) and SMUs to do same thing. There are also few less known vendors that do high-current (hundreds amps) and high-voltage (kilovolts) SMUs for power applications. It is just most know less about this market, as semiconductor parametric analysis is usually business task, not that much of a daily hobbyist job.  :)
YouTube | Metrology IRC Chat room | Let's share T&M documentation? Upload! No upload limits for firmwares, photos, files.
 

Offline TheUnnamedNewbie

  • Super Contributor
  • ***
  • Posts: 1087
  • Country: 00
  • mmwave RFIC/antenna designer
Re: Modern day Curve Tracer?
« Reply #5 on: September 22, 2018, 11:57:29 am »
while I'm sure the need for characterizing discrete semiconductors is way less than when they were being made

I don't think this is true; There is still a massive market for discrete devices. In addition, there is a large need for very accurate measuring of the relationship between currents and voltages of devices in order to set up the simulation toolkits (PDK, process development kit) of semiconductor systems. These models have to be very precise and take a lot of work and precision to create. To do so, the measurements are done in massive automated systems - put in wafers, and using tiny needles they will automatically measure hundreds/thousands of devices. (You usually have to add on your own test equipment to this, they are just probing systems). A number of manufacturers exist for this - Formfactor, MPI, Cascade, Teradyne.

I think we don't see classical curve tracers as much because we now tend to test more in detail (this is hypothesis, I'm not a process engineer). Not just DC performance, but AC, RF, etc are tested in one go. As a result I think we tend to move to more complex systems built up of a number of pieces of equipment, perhaps in a rack/PXI/VXI test system. Instead of a true curve tracer, we could take a source and a number of multimeters and then program it to do a number of tests.
The best part about magic is when it stops being magic and becomes science instead

"There was no road, but the people walked on it, and the road came to be, and the people followed it, for the road took the path of least resistance"
 

Offline DaJMasta

  • Super Contributor
  • ***
  • Posts: 2150
  • Country: us
    • medpants.com
Re: Modern day Curve Tracer?
« Reply #6 on: September 22, 2018, 04:40:24 pm »
Seems at least to some degree, they've been hiding in plain sight!

My line of thought was more than with discrete devices being cheaper, being manufactured much more consistently, and not needing matching as often (since if you want matched you often just buy ICs), the demand had dropped - I figured fabs and R&D labs would still need something to help optimize their devices.


In any case, I wonder if I can justify one to have a play with.... I certainly don't need to match transistor gain very often.
 

Offline particleman

  • Regular Contributor
  • *
  • Posts: 115
Re: Modern day Curve Tracer?
« Reply #7 on: September 22, 2018, 05:08:34 pm »

In any case, I wonder if I can justify one to have a play with.... I certainly don't need to match transistor gain very often.



My 7CT1N is in use all the time. Gain matching is the reason I use it the least. Go no go,saturation,leakage, diode Vf,Vr,Vz are my most common uses. It is one of my prized most valuable tools.
 

Offline Old Printer

  • Frequent Contributor
  • **
  • Posts: 648
  • Country: us
Re: Modern day Curve Tracer?
« Reply #8 on: September 22, 2018, 09:59:23 pm »
This is a video demonstrating how the different tools included with the Analog Discovery can be configured as a curve tracer. I am not knowledgeable enough to understand what the limitations might be, but I think it would be worth your while to watch the video. Tom is a retired EE and quite knowledgeable.

 

Offline TheUnnamedNewbie

  • Super Contributor
  • ***
  • Posts: 1087
  • Country: 00
  • mmwave RFIC/antenna designer
Re: Modern day Curve Tracer?
« Reply #9 on: September 23, 2018, 11:09:35 am »


My line of thought was more than with discrete devices being cheaper, being manufactured much more consistently, and not needing matching as often (since if you want matched you often just buy ICs), the demand had dropped - I figured fabs and R&D labs would still need something to help optimize their devices.


Every device (including every single 0.0001 cent MLC capacitor or thick-film resistor) is usually tested. Sure, this might not be a true curve-tracer any more, but the curve tracer has just been replaced by more complex setups of testing systems. Curve tracers were popular because the nature of a CRT allowed easy analog measurements like that. Now, since you are going to digitize things anyways, might as well make use of more 'modular' systems and the more complicated processing you can do with it. Once you have the signals in the digital domain, all you need is more math to get more information out of it.
The best part about magic is when it stops being magic and becomes science instead

"There was no road, but the people walked on it, and the road came to be, and the people followed it, for the road took the path of least resistance"
 

Offline David Hess

  • Super Contributor
  • ***
  • Posts: 13567
  • Country: us
  • DavidH
Re: Modern day Curve Tracer?
« Reply #10 on: September 24, 2018, 12:22:57 pm »
My 7CT1N is in use all the time. Gain matching is the reason I use it the least. Go no go,saturation,leakage, diode Vf,Vr,Vz are my most common uses. It is one of my prized most valuable tools.

I rarely used my 7CT1N to characterize parts.  More often I use it to diagnose weird problems which would not be caught with simpler measurements like "soft" transistors which have developed things like poor saturation or high base spreading resistance.
 

Offline particleman

  • Regular Contributor
  • *
  • Posts: 115
Re: Modern day Curve Tracer?
« Reply #11 on: September 24, 2018, 01:43:27 pm »
I rarely used my 7CT1N to characterize parts.  More often I use it to diagnose weird problems which would not be caught with simpler measurements like "soft" transistors which have developed things like poor saturation or high base spreading resistance.

Exactly. It has saved me many hours of shotgunning and I credit it to several repairs that may never have been completed. If I was better a circuit analysis I probably wouldn't rely  on it as much as I do, but for a novice like me its a giant help.
 

Offline owiecc

  • Frequent Contributor
  • **
  • Posts: 297
  • Country: dk
    • Google scholar profile
Re: Modern day Curve Tracer?
« Reply #12 on: September 24, 2018, 01:49:49 pm »
We recently released an instrument that is capable of measuring both the switching losses and measuring the static characteristics at the same time. So basically a switching setup with built in curve tracer.
 

Offline David Hess

  • Super Contributor
  • ***
  • Posts: 13567
  • Country: us
  • DavidH
Re: Modern day Curve Tracer?
« Reply #13 on: September 25, 2018, 12:25:36 am »
I rarely used my 7CT1N to characterize parts.  More often I use it to diagnose weird problems which would not be caught with simpler measurements like "soft" transistors which have developed things like poor saturation or high base spreading resistance.

Exactly. It has saved me many hours of shotgunning and I credit it to several repairs that may never have been completed. If I was better a circuit analysis I probably wouldn't rely  on it as much as I do, but for a novice like me its a giant help.

When I repaired my Tektronix DC505, I initially found a pair of bad 2N3565 transistors.  This was suspicious so I put some other "good" 2N3565s from the DC505 on my 7CT1N and discovered that about half of them were "soft" as I described with a couple near failure so I replace them all.

We recently released an instrument that is capable of measuring both the switching losses and measuring the static characteristics at the same time. So basically a switching setup with built in curve tracer.

I would like something which measures the various small signal DC parameters like leakage, Early voltage, and base and emitter spreading resistance under different conditions.  Base-collector time constant would be nice as well but that is an AC measurement.

My own design for this is advancing but I doubt it will ever get beyond the test stage.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf