Anyone know of any soak testing software?

Hello,
Normally I'd leave my things running at full pelt over night and check its still working properly in the morning but it would be good if there was some software that could monitor a test signal and record/make notes if it drops below a threshold temporarily or if THD momentarily increases etc... Things that otherwise I'd be missing currently. I use Arta for measurements but AFAIK it can't do this. Anyone have any ideas?
Cheers!

Looking for infant failures in manufacturing, generally we use simple barrier testing (works, measures good). 

Back in the 80s as a much smaller company I would burn in products over night and retest the next day before shipping.  To stress the products more than just running continuous, I used a simple timer to power up and power down the burn-in rack several times over night, so they would thermal cycle and cool down (I don't recall details but think they were off more than on to fully cycle through warm/cool temperatures.

JR

tc electronics used to have soak benches for their digital products (M2000/3000, G-Force...). The units were constantly on, but the programs kept changing under MIDI control and a PC-controlled signal monitor actually checked "something" depending on the actual program. It was a proprietary program.
For most products, soak test actually detects infant mortality; most of the times you don't even need to run signal through...
For the rest, no computer program can challenge the ingenuity of a candid user that's determined not to consider any caveats or RTFM, on account he's the owner.  ;D

abbey road d enfer said:

For the rest, no computer program can challenge the ingenuity of a candid user that's determined not to consider any caveats or RTFM, on account he's the owner.  ;D

Click to expand...

"A common mistake that people make when trying to design something completely foolproof is to underestimate the ingenuity of complete fools." - Douglas Adams

mjrippe said:

"A common mistake that people make when trying to design something completely foolproof is to underestimate the ingenuity of complete fools." - Douglas Adams

Click to expand...

Trust me I worked at peavey for 15 years, :  it was a constantly escalating arms race, but an interesting engineering challenge.  8)

JR 

And let me take this opportunity to say "Thank You, JR" for sharing those years of knowledge here 

mjrippe said:

And let me take this opportunity to say "Thank You, JR" for sharing those years of knowledge here 

Click to expand...

+1

Thanks chaps, guess I'll continue doing as I was.

> over night and check its still working properly in the morning

You call that a test? Traynor threw stuff off the roof. Mesa Boogie beat their amps with a hammer. I ran power amps on FULL for days. Hard-drive samples spend months on a heat-rack with constant seeking. Plymouth ran cars on the badlands behind the railroad yards. Another.

> drops below a threshold temporarily or if THD momentarily increases

Does that even happen? Electronic parts shouldn't have "drop-outs" like cheap tape. Either it works, or it doesn't, or there's a loose joint.

Your stuff should probably be good for 10,000 hours. Easily 10% of failures happen early in the first 1% of life. It is quite reasonable to "burn" 100 hours, alternate idle and hard-work. And beat it with a hammer while listening if loose-joints have been a problem.

hmm, if you have enough control on your manufacturing process, then you should be able to do a month punish test of 25+ units (as PRR suggests), then burn on regular models.

That's more or less how 95% of the chips you buy are done.

PRR said:

Traynor threw stuff off the roof. Mesa Boogie beat their amps with a hammer. I ran power amps on FULL for days. Hard-drive samples spend months on a heat-rack with constant seeking. Plymouth ran cars[/url] on the badlands behind the railroad yards.

Click to expand...

There's a big difference between pre-production tests, many of them being destructive (or at least trying) and QC tests.
I don't believe one second Traynor did throw every amp from the roof before shipping to the customer. 
Random testing is a complement to pre-prod and QC tests. It is well documented as in here https://qualityinspection.org/quality-control-basics/
And there must be some consideration for the expected use of the product: I would not test digital reverb, that's gonna sit in a rack, the way I would a guitar amp that's traditinally submitted to the worst treatments.

abbey road d enfer said:

There's a big difference between pre-production tests, many of them being destructive (or at least trying) and QC tests.
I don't believe one second Traynor did throw every amp from the roof before shipping to the customer. 
Random testing is a complement to pre-prod and QC tests. It is well documented as in here https://qualityinspection.org/quality-control-basics/
And there must be some consideration for the expected use of the product: I would not test digital reverb, that's gonna sit in a rack, the way I would a guitar amp that's traditinally submitted to the worst treatments.

Click to expand...

+1  at Peavey we had a reliability testing group and i would dare them to blow up new prototypes to see if the design is robust. They were kind of a blunt instrument and once shook capacitors right off a PCB on the shaker table. This was inside a rack mount cassette deck so unlikely to be stressed that hard in real life, but I added hot melt and extra securing for the massive film caps, rather than dispute the test. 

In production even 100% barrier testing has fallen a little out of favor in some industries. With semiconductors they found a small percentage of failures were actually caused by the handling during testing (but i suspect this has changed even more in the decades since, i.e. less hands touching stuff). That said back in the 70's I did 100% testing of all ICs and routinely screened out low single digit percentage of rejects. By the 80s IC quality was much better and i stopped 100% testing before use.

JR