Does old test equipment really ever become truly obsolete?

[G0HZU]

My really old (and long obsolete according to Keysight) Infiniium 500MHz DSO can display magnitude and phase in FFT mode. It also reports the 'resolution' although I'm not sure this is the same as the resolution bandwidth. The 'resolution' doesn't change if I change the window type for example. Either way, the FFT function is fairly poor on this scope. Not as poor as some other scopes, but poor enough to only make me use it for really quick/casual measurements.

I also have a newer DSO here made by Tek (MSO4104) and the FFT function is not nice to use. I rarely use this scope, even for everyday  stuff because the UI is so awful. It does have a nice big display though. I think this Tek scope is only about 12-13 years old but it feels very dated. It has the latest firmware so this is about as good as it is ever going to be. The features on the newer scopes from manufacturers like Siglent really do show how restrictive and poor these older Tek scopes are today.

It's hard to imagine how this MSO4104 could ever be classed as 'obsolete' in the sense it could no longer be useful in a typical engineering lab. However, in 50 years' time who knows what will change. Maybe the 230V AC mains system will change by then or maybe there will be health and safety regulations that effectively make it obsolete.

I still have my first scope here (Tek 585) and I think it is over 60 years old and it still worked the last time I powered it up. I would class this scope as obsolete but I'm sure there are plenty of happy 5xx series scope users still out there.

[nctnico]

The fact that you can still "use" old equipment (if you have nothing better) is one thing.

What is repeatedly recommended to beginners in this forum is to buy a 30-year-old 20MHz CRO because it is "better" than, for example, a Hantek DSO2000 for a similar price.
Is that nonsense or can you say that with a clear conscience?

It is not nonsense, it is utter nonsense. The same goes for every bench DMM thread. Someone will come in and go harping on about some ancient 6.5 digit HP / Agilent DMM with little features and a crappy VFD or LCD display.

[coppercone2]

what the hell features are you looking for.

even the new stats feature that you get on the new DMM truevolts, try using that in a company. like I wanna see someone that actually made a decision with it. you end up writing down V or A or Ohm.

historgrams and graphs are open to interpretation and no one can make a decision with that.causes confusion

then you get some other BS thats not even optimized. The beeper is useful and the diode test, well you wanna get a fluke for that since it goes to a higher voltage anyway. Most people end up using the hand held 4.5 digit DMM because its calibrated and they can walk around with it, it has a clock display. then I end up with a high end DMM on my soldering workstation permanently in continuity mode


and TONS of stuff in modern electronics are too unstable to use with the high resolutions anyway

do i even use the true volt unless I have to? No. it takes too long to boot up. I have a 34401A under it that gets used all the time over the fancy 7.5 digit meter that has to spool up the OS for way too long and annoys me with its bright light

seriously it feels like a joke when you need to wait for equipment to setup in a work environment. and you know I will round it down to 10 mV because its going to just create confusion or be incompatible or cause too tight a spec for no reason or just drift way too much to be of any use

[DimitriP]

ancient 6.5 digit HP / Agilent DMM with little features and a crappy VFD or LCD display

HP /Agilent 6.5 digit display ARE  the "features".

The rest is eye candy, design, procurement, production and future service , calibaration and warranty "optimization".

Now, if I could get my hands on one of those "ancient, 6.5  digit HP /Agilent DMM with little features"  , "cheaply" I would be happy !

[tggzzz]

Don't be silly, DimitriP and coppercone2.

Newer is always better.

Directly something newer arrives, all older stuff becomes useless - and it is immoral to let newbies use anything except the newer stuff.

[David Hess]

I still have my first scope here (Tek 585) and I think it is over 60 years old and it still worked the last time I powered it up. I would class this scope as obsolete but I'm sure there are plenty of happy 5xx series scope users still out there.

The 547 held on for many more years after it was declared obsolete by Tektronix because it was their fastest oscilloscope that lacked a scan expansion mesh, which allowed it to display signal details which were not visible on later or faster instruments.

I have a Tektronix 545A but would only use it as a last resort, or if I needed a space heater.  My 547 does at least have the virtue of the best possible screen clarity, and it really shows.

historgrams and graphs are open to interpretation and no one can make a decision with that.causes confusion

I have been looking for a bench multimeter that can calculate standard deviation, but it has to be over a specific period of time or a specific number of samples to be useful.  Right now I do it manually on my calculator.

ancient 6.5 digit HP / Agilent DMM with little features and a crappy VFD or LCD display

HP /Agilent 6.5 digit display ARE  the "features".

I consider OLED displays to be an "anti-feature".

[DimitriP]

I consider OLED displays to be an "anti-feature".

OLED is not  "ancient" enough 

I have been looking for a bench multimeter that can calculate standard deviation, but it has to be over a specific period of time or a specific number of samples to be useful

I pushed the STD DEV button on my ancient 5335a with Voltmeter option installed and a number showed up , eventually

[shabaz]

I have been looking for a bench multimeter that can calculate standard deviation..

Not an exact match at all, but I thought I'd give it a bash with a Bluetooth handheld multimeter (screenshot attached).

The new code is just sitting on my local PC, I've not committed the standard deviation code to the web (mobile/desktop) app yet, but can do that if required.

I'm not very good at web apps, so I find it a bit of a struggle. SCPI + Python etc would have been far easier but that handheld device doesn't support that. Also, being a typical handheld device, it doesn't have a very fast update rate, so stats performance is not great; for instance 20 samples takes about 10 seconds.

The MP730624 multimeter I used is some Uni-T rebranded model I believe.

[nctnico]

I have been looking for a bench multimeter that can calculate standard deviation, but it has to be over a specific period of time or a specific number of samples to be useful.  Right now I do it manually on my calculator.

I'm looking at getting a UNI-T UT8805E bench DMM which can do that. It supports measuring a specific number of samples and calculating the standard deviation. But I'm sure modern day bench DMMs from Tektronix and Keysight have a similar feature as Uni-t must have copied the idea from somewhere.

[pqass]

I have been looking for a bench multimeter that can calculate standard deviation, but it has to be over a specific period of time or a specific number of samples to be useful.  Right now I do it manually on my calculator.

I'm looking at getting a UNI-T UT8805E bench DMM which can do that. It supports measuring a specific number of samples and calculating the standard deviation. But I'm sure modern day bench DMMs from Tektronix and Keysight have a similar feature as Uni-t must have copied the idea from somewhere.

The HP 3456A from 1981 can do this.  It can be had for about $250 on eBay and has a lovely LED display.
Enter the number of samples, delay between samples, and it will produce min, max, mean, and variance after a single trigger button press.
You'll have to √variance yourself for std dev though.

[2N3055]

I have been looking for a bench multimeter that can calculate standard deviation, but it has to be over a specific period of time or a specific number of samples to be useful.  Right now I do it manually on my calculator.

I'm looking at getting a UNI-T UT8805E bench DMM which can do that. It supports measuring a specific number of samples and calculating the standard deviation. But I'm sure modern day bench DMMs from Tektronix and Keysight have a similar feature as Uni-t must have copied the idea from somewhere.

The HP 3456A from 1981 can do this.  It can be had for about $250 on eBay and has a lovely LED display.
Enter the number of samples, delay between samples, and it will produce min, max, mean, and variance after a single trigger button press.
You'll have to √variance yourself for std dev though.

That function is also available on Rigol DM3068 and a slighty simpler version on DM3058. Also on Siglent SDM3055/3065.
You can also set various triggering options, etc.

[Alex Nikitin]

I have been looking for a bench multimeter that can calculate standard deviation, but it has to be over a specific period of time or a specific number of samples to be useful.  Right now I do it manually on my calculator.

I do have the HP3456A in my lab, however my new HIOKI DM7275 7.5 digit DC voltmeter can do it as well, I can set a manual/external trigger, a delay and a number of samples per a trigger event, and the statistics display gives max/min/av/p-p/stdev.

Cheers

Alex

[pdenisowski]

I have been looking for a bench multimeter that can calculate standard deviation, but it has to be over a specific period of time or a specific number of samples to be useful. 

Our HMC8012 bench multimeter can do this - calculates standard deviation based on a user-defined number of samples (up to 50,000)

https://www.rohde-schwarz.com/us/products/test-and-measurement/rs-essentials-meters-and-analyzers/rs-hmc8012-digital-multimeter_63493-44315.html

[S57UUU]

If you need to test modern stuff like 5G or something, using old instruments would make it very hard work if not impossible. New instruments have built in "personalities" that enable "one button press" measurements to test compliance with modern standards.

But I don't need that for my ham radio and radio astronomy hobbies. For that, my stack of old instruments from the 80's serves me very well. They are "just enough digital": they do calibration and error corrections, and can be computer controlled via HP-IB, but no windows virus gardens and similar.

For example, some five years ago I bought a HP8593E with the 26GHz option for 1200 euros. I don't think there is a modern alternative that would fit that budget.

I do have a nanoVNA and a SAA-2, also a tiny SA ultra, these are great when out in the field. But when at home, I prefer to use my HP8714ES and HP8720A, the user interface is just soooo much friendlier.

My HP8663 opt 103 is bulky and heavy, bought it for 600 euros. Even today, it is still among the cleanest signal generators. Also have an HP8673 up to 26GHz, don't know of any cheap modern alternative.

I cobbled together a HP3048 system with HP11848 and HP11729. It is run by a DOS/ISA computer, is slow and bulky, but it measures up to 18GHz and gives great results.

My HP8970 uses totally obsolete detector based technology, but measures noise figures very well, on par with modern equipment.

Admittedly, I am an old fart, I used these boxes as a young professional in the 80's, so they grew close to my heart. But I do believe they can still be very useful today. Service manuals, schematics and even EPROM firmware files for most of them are available on the web, also the through hole components make them easy to repair (for now, I mostly had to replace some electrolytic caps).

Some of my instruments were old stuff even in the 80's, like the HP3400, which I call a "true-true-RMS meter", and I still find it useful.

As far as scopes go, I just bought a rigol 814 and a set of "pc bite" self supporting probes, it is great for debugging serial stuff. But when just needing to see some signals, for example in the audio range, I still sometimes fire up my tek 2232 (and 2467 until the battery went bad and it forgot it's calibrations).

[coppercone2]

test equipment often evolves to
1) make tests run faster, because more tests of more devices is ALWAYS the easiest route for a test lab to take to show 'improvement' in a way that non technical people understand and will not question (don't want your budget to get interrogated, don't want to teach old dogs and some how gophers where there should be dogs new tricks). It is also a measuring contest between competitors, because volume means quality.
2) allow for less trained personnel to run tests (lower salary, allows for more workers doing tests, decreases cost of hated QC and R&D departments, easier to hire (they like to get irrelevant professions and pay little)
3) less chance of error (harder for sloppy work to screw up, ties into 1&2. i.e. irrelevant degrees & uninterested people slamming buttons)
4) decrease load on analysis (that is, have the company "approve" some calculated number so the buyer can say "well _____ says that this method of data analysis is the "standard" for this industry)..... (again, it might mean that a clueless TE company is making decisions about something it knows nothing about that govern standards of another unrelated industry... they often don't know what is best for the customer but the customer wants them to back them up as much as possible for liability reasons, and it sounds good. like yeah we are on the same page as ______)

These are IMO the driving factor for MANY equipment decisions. Its often not engineered to give you that much more capability because IMPLEMENTING new more precise and interesting stuff  for the companies buying the TE is HARD compared to doing more of the same.



I am sure there are engineers at TE manufacturers that wanna raise the bar of precision and deep thought, but they do end up making (alot of) stuff to be appealing towards customers. Business people hate buying equipment, these listed factors are a huge 'no argument' selling point that does not require visionaries to see the potential. Then of course sales will tell the TE designer that "this is the selling point now".

But if it actually helps improve things, instead of just being empty calories, I am not sure about that. Too much data is more and more common because they feel 'data rich' and 'confident'. It could also be seen as over weight with lots of ru(b)bles in the pocket

[nctnico]

You are leaving one very important item out: Adding more measurement functions. Like capacitance and temperature measurement on DMMs and protocol decoding, digital channels and bode-plotting on oscilloscopes. Back in the old days these measurements required expensive equipment. Digital channels and protocol decoding on oscilloscopes has helped making logic analysers becoming obsolete.