EEVblog Electronics Community Forum
Products => Test Equipment => Topic started by: mongobit on April 17, 2016, 12:51:40 am
-
Hello All. Long time lurker and video viewer and Amphour listener here. I am a Technician with 10 years experience in electronics manufacturing, testing and troubleshooting. I have recently started at a new company and have stepped up and taken over sole responsibility for the small electronics lab the company maintains at our site. Both predecessors in this position have left the company for various reasons leaving me with what amounts to an underfunded hobbyist level set up. My boss has asked me to look into expanding our capabilities. I have most of the basics (better solder rework equipment, board prep tools, and lab organization and component inventory) in the process of getting upgraded. One thing I know I want to add but haven't used or specked out before is a mid range semiconductor analyzer/curve tracer to the mix to do some preliminary failure analysis on various devices I am replacing. some requirements I have are:
Professional product, need something I can show my customers that they wont laugh at.
Computer connected for generating reports and possibly some data/trend analysis.
Auto component identification like the Atlas DCA75
$1-5KUSD price range.
-
2nd hand Keithley SMU ought to do it. See the videos by "The Signal Path" on curve tracing using the newer Keithley models.
Obviously, go new if you can, but they aren't cheap.
-
Ouch just looked at a few. Not sure my boss would go for that. Maybe sell it as a service we can market to other facilities the company has to defray the cost.
lol keep in mind exactly what expand our capabilities means has yet to be fully defined. I'm working up scenarios to keep it a 1-3 man shop with expanded test, troubleshoot, and rework capabilities to shooting for the moon going for 10-30 person LPKF PCB fab setup and a mixed tech assembly line in a separate building in a local industrial park.
-
Well the keithley SMU (or keysight, etc) are basically the go-to devices for curve tracing as far as I know. The next step up is called a "semiconductor parameter analyser", though I don't work in the semiconductor industry so I don't know if they are still used over SMUs. The things you are paying for are high voltage swings, accuracy and reliable PC interfaces. Also some brand name of course.
You might be able to get away with a DAQ card from national instruments which might be a little cheaper in the short run, but if you go with labview the licensing fees will cost you much more in the long run imo. That's not to say I recommend labview, but it is an option.
-
Yeah, a dual channel SMU is the easy way to do it. Keysight have some software that goes along with theirs that does curve tracing simply. But even second hand thats right at the top limit of your budget. What accuracy and ranges were you intending to use? A pair of USB controlled single quadrant power supplies might be suitable.
-
From what I'm told we are getting some labview licenses to build some custom test equipment for some major contracts that are partially fueling this expansion. I'm already looking into what I can do to get in on the training some of us would have to go through to fully utilize it.
This is all preliminary at this point. For all I know I could get called into a meeting next week and be told I have an official budget and an official operational requirement that I have to meet that could take this in a whole different direction. Or I could walk in Monday and oh hey we hired a consultant to set this up so your going to work with them to set up something that maximizes the consultants fees but minimizes resultant capability expansion.
-
If you've got a good programmable environment like Labview available then its just a question of what specifications you need the sources to achieve.
-
I'm thinking we would want to go for a more COTS solution for this rather than a roll your own labview solution. The field we serve presents the possibility of our work coming under scrutiny from several different investigative agency's if something went wrong. So from a liability standpoint I think we would want something with a proven engineering background backing us up when we say hey this is a problem and here is an industry recognized analysis saying this is the trend we see this is what needs done to correct the issue/protect the public.
-
There isn't much beyond the big 3 in curve tracing, Keithley, Keysight and Yokogawa all offer turn key systems but they're well above your price target. You can use their cheaper hardware to assemble a system with your own software, the older Agilent 66319 and 66309 could be a cheap basis for a reliable system. Either way you'll be on the hook for calibration cycles if you're working in such a strict environment.
-
Analog Discovery Curve Tracer maybe?
https://ez.analog.com/community/university-program/blog/2013/12/07/analog-discovery-bjt-curve-tracer-program
Admittedly if you need something that's, turnkey, calibrated to traceable standards and generally completes a box ticking exercise then you'll probably need to put your money where your mouth is I'm afraid.
-
Analog Discovery Curve Tracer maybe?
https://ez.analog.com/community/university-program/blog/2013/12/07/analog-discovery-bjt-curve-tracer-program
Admittedly if you need something that's, turnkey, calibrated to traceable standards and generally completes a box ticking exercise then you'll probably need to put your money where your mouth is I'm afraid.
I have one of them, and one of these too: https://ez.analog.com/community/university-program/blog/2015/05/16/adalm1000-analog-inputs-and-outputs
They are really neat, but are really more educational devices than commercial tools. They are curve tracers in the same way that a bicycle with training wheels is a Harley Davidson.
-
I might be able to bump up the budget to $5-10KUSD if we push things out till next fiscal year. But we will see like I've said things are really still up in the air in regards to how far down the rabbit hole we want to go. I'm just exploring options right now
-
Analog Discovery Curve Tracer maybe?
https://ez.analog.com/community/university-program/blog/2013/12/07/analog-discovery-bjt-curve-tracer-program
Admittedly if you need something that's, turnkey, calibrated to traceable standards and generally completes a box ticking exercise then you'll probably need to put your money where your mouth is I'm afraid.
I have one of them, and one of these too: https://ez.analog.com/community/university-program/blog/2015/05/16/adalm1000-analog-inputs-and-outputs
They are really neat, but are really more educational devices than commercial tools. They are curve tracers in the same way that a bicycle with training wheels is a Harley Davidson.
I have two! While I only use them on the bench occasionally (they are useful for simulating digital data for example), my primary use for them is when travelling. Sure I'd love to have a scope and signal generator with me all the time, but realistically that can't always happen. You're right though, they're hardly going to win in a performance shootout, but at that price point there's nothing else out there that comes close.
-
HP/Agilent 4142B if you feeling like some amount of programming code is not a big deal. It's 8-slot mainframe in which you can install plenty of SMUs, voltage source/voltage monitor modules.
I got one recently and made an article/teardown of it (https://xdevs.com/fix/a4142b/). There are number of frames w/o and with modules on ebay for few K USD, perhaps sellers will bulge for less if make an offer.
Only caveat - these beasts are long obsolete, and Keysight will not help you if something broke...
There are more new E5260A and E5270A but those are way more expensive.
-
The TEK 370, 370A or 370B curve tracer have a GPIB port that allow to configure and read measurements.
-
> up the budget to $5-10KUSD if we push things out till next fiscal year
Is leasing an option?
-
Leasing may be an option. I keep trying to get definitive information from higher up on requirements but that is not always easy.
-
I built this for a few hundred in pieces/parts.....and subsequently let all the smoke out of it :D
I call it POCASA (Power Off Component Analog Signature Analyzer).
It's my pet project...still...after 6+ years...and I'm still working on it.
https://www.youtube.com/watch?v=RkQoVHUZqmw (https://www.youtube.com/watch?v=RkQoVHUZqmw)
This version is a very early version, software is slow, parts aren't precision by any stretch, controlled by a 48Mhz PIC, no memory, parts hacked together, etc.
I'm up to a 7" RGB LCD w/capacitive touch, Li-Ion powered, WiFi/BlueTooth enabled, STM32F4 driven, variable everything, multi-channel goodness. Although the total cost of the pieces/parts has gone up a bit with all the new goodies.
-
If i am not mistaken this is basically a Huntron Tracker?
-
Ya, at this stage, pretty much.
I'm shooting for a bit of a mix between a 128 channel ProTrack 3200 (modelnumber???), a decent SMU, and a RF analyzer, crushed into the same box. Having a helluva time with RF as its not even close toy specialty, but I'm learning...slow but painful :)
-
Good to know. we have a Tracker 2000 that is awesome for troubleshooting. I am looking at some of their solutions for some of our other needs but we will see.
-
I was introduced to the Huntron while working in AFREP. If that rings a bell, you know where I'm coming from.
-
They look promising and its really made it easier to justify replacing some rather pricey components. I'd love to acquire a newer model but this one s still going strong.
-
That might work. i'll have to dig into it.
-
For these smaller solutions, doesn't Ic matter? Or even Beta and Ib? Let's face it, it will take a pretty solid power supply to test a 2N3055 and Beta will be pretty low at high Ic.
It seems to me that the devices requiring test will be a big factor in deciding what capabilities are needed. 30V @ 1A isn't all that impressive if we're dealing with power transistors.
-
For these smaller solutions, doesn't Ic matter? Or even Beta and Ib? Let's face it, it will take a pretty solid power supply to test a 2N3055 and Beta will be pretty low at high Ic.
It seems to me that the devices requiring test will be a big factor in deciding what capabilities are needed. 30V @ 1A isn't all that impressive if we're dealing with power transistors.
And at the same time having a minimum resolution (no accuracy specified for the CTR-101) of 2uA makes it hard to measure reverse characteristics or high gain darlington devices. Once I used a curve tracer for the first time (a Tek 571) I knew it was going to be indispensable forever more.
-
From what I'm told we are getting some labview licenses to build some custom test equipment for some major contracts that are partially fueling this expansion. I'm already looking into what I can do to get in on the training some of us would have to go through to fully utilize it.
If you're in the states, for situations like this, leasing is the best from a tax POV + you can line-item your customer the monthly, then also write off the item as a depreciation for the quarter. (That's why most companies end up leasing on 3 year cycles for all their gear.) Then when the lease term expires, buy it out. On paper you've paid a lot more, but ultimately your capital expenditure is minimized.
Or just build some octopus tracers from the 60s ;)