Author Topic: EEVblog 1442 - DON'T DO THIS!  (Read 4448 times)

0 Members and 1 Guest are viewing this topic.

Offline EEVblogTopic starter

  • Administrator
  • *****
  • Posts: 37626
  • Country: au
    • EEVblog
EEVblog 1442 - DON'T DO THIS!
« on: December 03, 2021, 01:33:01 pm »
An innocent TO-220 regulator gets tortured using a mystery device from the bunker.
The art of environmental PCB testing.

 
The following users thanked this post: SeanB

Offline PartialDischarge

  • Super Contributor
  • ***
  • Posts: 1610
  • Country: 00
Re: EEVblog 1442 - DON'T DO THIS!
« Reply #1 on: December 03, 2021, 02:20:49 pm »
That's an interesting device. I've seen the larger ones on labs but not these tiny ones.
I guess that now I'd use motors with eccentric weights to do the same since they are widely available.
« Last Edit: December 03, 2021, 02:27:29 pm by MasterTech »
 

Offline Ice-Tea

  • Super Contributor
  • ***
  • Posts: 3062
  • Country: be
    • Freelance Hardware Engineer
Re: EEVblog 1442 - DON'T DO THIS!
« Reply #2 on: December 03, 2021, 04:32:03 pm »
The smallest ones on the market?

Well, no -> https://www.qsources.be/simcenter-qsources/

Offline Kleinstein

  • Super Contributor
  • ***
  • Posts: 14016
  • Country: de
Re: EEVblog 1442 - DON'T DO THIS!
« Reply #3 on: December 03, 2021, 06:03:56 pm »
The motor with excentric weight is used for the more brute force way, when you don't need to frequency to change over a large range. With just 1 axis the movement is more than just 1 axis.

The poor little 7812 did survive quite some time and quite some vibrations. Unless used in things like automotive I would not consider the free standing TO220 so bad.

The resonance was surprisingly wide (low Q), probably because of the plastics from the PCB involved. Mechanical resonances in metal structures can be quite narrow and thus high Q. So things can get a lot worse in resonance, though a bit tricky to find the resonances.
 

Offline Old Fart Analog Engineer

  • Contributor
  • Posts: 12
  • Country: us
Re: EEVblog 1442 - DON'T DO THIS!
« Reply #4 on: December 03, 2021, 07:47:06 pm »
I used to design electronics for car manufacturers, quads and motorcycles.  During R&D we needed to test for vibration, without the time and money for using an external lab, so I made some of these using sub-woofers.  The attached photo is the small one @ 8" with some relays attached to measure contact resistance variance due to vibration.  The rectangular plate matches a 200W x 2 power amplifier for a motorcycle.  The larger one is buried in my garage uses a 12" 1000 W rated subwoofer with a similar set up to test larger devices.

It was cheap and quick to build using aluminum (none of that "Al you min yum" rubbish!) plates bolted to the cone near the surround with a thick (~ 3" D) post bolted and epoxied in the middle to allow various mounting plates. The smaller one could take a few hundred Watts and move ~14mm p-p. 

The larger one could take 1200 Watts** and move ~ 20mm.  At 40 Hz the signal had to be in bursts of ~3 seconds on and ~7 seconds off to not burn up the voice coil.  This was ~ 10 g acceleration to the item being tested.  The 80-20 extrusion is a handle and a place to hang a counterbalance spring to cancel out the weight of the device being tested and the fixture (~6-8 lbs.).

How long will this last?  The larger unit ran over 2000 hours in a temperature chamber which varied between -20c and +70c  (none of that Fahrenheit rubbish).  Why? Well the Chinese vendor of the PCB cut a lot of corners and the 4 layer PCB's vias would randomly fail at between a few Ohms and a few kOhms in "field" use after 3-9 months.  They started at 600 hours and had > dozen failures after 800 hours.  The new vendors PCBs had zero issues > 1200 hours when we closed the test.

** a word about Watts in audio:  Amplifiers are rated with a continuous sine wave into a resistive load, speakers are rated with a bandwidth shaped pink noise.  If you power a speaker with a sine wave at its rated power , you will quickly smell and see the death of the voice coil.  Lesser quality speakers will not survive long, even with short pulses of a full power sine wave, so you have been warned.
« Last Edit: December 03, 2021, 07:50:22 pm by Old Fart Analog Engineer »
 
The following users thanked this post: thm_w, HighVoltage, Bud, DH7DN

Offline DH7DN

  • Regular Contributor
  • *
  • Posts: 128
  • Country: de
    • DH7DN Blog
Re: EEVblog 1442 - DON'T DO THIS!
« Reply #5 on: December 04, 2021, 12:05:07 am »
Very nice video Dave, I really enjoyed it. Everything you said was very accurate.

Amazingly the stuff from the 1970's still performs superb. Modern shakers are based on the same principles as the old ones (permanent magnet - moving part - coil). The membrane fixture doesn't perform very well at low frequencies (< 20 Hz) due to the hysteresis of the rubber. At higher frequencies (few kHz) they experience lateral resonances which cause unwanted deviations and disturbances of motions (e. g. during calibrations).

Modern (and very expensive) shakers can be bought with air bearings in order to reduce those unwanted transverse motions. The goal is to produce a rectilinear sinusoidal motion with low distortion. The shaker tables you mentioned in the video can really be huge/massive. They are used for all kinds of environmental testing: automotive parts, everyday stuff packed on a pallet, even high-tech parts like satellites or digital cameras are tested before shipping. Vibration testing is a very exciting field with very clear experiments. Demonstration of phenomenae such as vibration motion, resonance, dampening, mechanical impedance, transmission, structure-borne and air-borne noise, structural fatigue etc. are easily possible

Few ideas for future videos...
- Calibration of a transducer by comparison method
- Frequency response
- Chladni patterns
- Resonance phenomena (destructive experiments)  >:D
- Demonstration of Normal Modes on structures such as PCBs or cables/wires
- How to fix PCB and electronic parts properly in order to prevent failures by fatigue/vibration
vy 73 de DH7DN, My Blog
 

Offline Neilm

  • Super Contributor
  • ***
  • Posts: 1545
  • Country: gb
Re: EEVblog 1442 - DON'T DO THIS!
« Reply #6 on: December 04, 2021, 06:14:19 pm »
Watching this reminded me of an incident of testing on some equipment I was working on. It was for a tracked vehicle which has one of the worst vibration frequency requirements I know (not absolute shock loading).

A new filter assembly had been designed for the power supply. It included a large inductor that could handle about 60 Amps (from memory). We put this on the 3 axis vibration table we had and ran it through the profile gradually increasing the amplitude towards the specification amplitude. The boards hit a resonance and the inductor ripped itself clear of its mount and proceded to smash everything on the PCB. We had taken the front off to observe and we recorded the thick PCB (at least 3 mm) deform by about +/- 5 mm.

Needless to say that filter went back to the drawingboard
Two things are infinite: the universe and human stupidity; and I'm not sure about the the universe. - Albert Einstein
Tesla referral code https://ts.la/neil53539
 

Offline NiHaoMike

  • Super Contributor
  • ***
  • Posts: 8951
  • Country: us
  • "Don't turn it on - Take it apart!"
    • Facebook Page
Re: EEVblog 1442 - DON'T DO THIS!
« Reply #7 on: December 04, 2021, 10:01:53 pm »
The smallest ones on the market?

Well, no -> https://www.qsources.be/simcenter-qsources/
I'd imagine the smallest actuators of that kind would be the ones used in smartphones for tactile feedback.

On the other end of the scale, wouldn't compressed air cylinders acting as a mechanical amplifier of sorts (like a jackhammer) be the most economical way to really get a lot of vibration?
Cryptocurrency has taught me to love math and at the same time be baffled by it.

Cryptocurrency lesson 0: Altcoins and Bitcoin are not the same thing.
 

Offline joeqsmith

  • Super Contributor
  • ***
  • Posts: 11601
  • Country: us
Re: EEVblog 1442 - DON'T DO THIS!
« Reply #8 on: December 04, 2021, 10:25:24 pm »
I ran a few tests using an LVDT to close the loop on 4 small speakers.   PC is used to control it. 

About  26:30 in, you can see the whole setup.
« Last Edit: December 04, 2021, 10:51:56 pm by joeqsmith »
 

Offline Bud

  • Super Contributor
  • ***
  • Posts: 6860
  • Country: ca
Re: EEVblog 1442 - DON'T DO THIS!
« Reply #9 on: December 04, 2021, 10:41:34 pm »

Holy shit, a filled Coca-Cola bottle is not something i would like to use at home as a DUT for my shaker.... ???  ::)
Facebook-free life and Rigol-free shack.
 
The following users thanked this post: DH7DN

Offline Jay_Diddy_B

  • Super Contributor
  • ***
  • Posts: 2731
  • Country: ca
Re: EEVblog 1442 - DON'T DO THIS!
« Reply #10 on: December 05, 2021, 12:17:35 am »




Hi Dave,

When I watched the video, I kept thinking you are shaking the wrong plane.
I think it would fail a lot quicker if it was shaken so the PCB is vertical and the regulator is horizontal?

Regards,

Jay_Diddy_B
« Last Edit: December 05, 2021, 12:24:58 am by Jay_Diddy_B »
 

Offline Ice-Tea

  • Super Contributor
  • ***
  • Posts: 3062
  • Country: be
    • Freelance Hardware Engineer
Re: EEVblog 1442 - DON'T DO THIS!
« Reply #11 on: December 05, 2021, 08:32:51 am »
The smallest ones on the market?

Well, no -> https://www.qsources.be/simcenter-qsources/
I'd imagine the smallest actuators of that kind would be the ones used in smartphones for tactile feedback.

On the other end of the scale, wouldn't compressed air cylinders acting as a mechanical amplifier of sorts (like a jackhammer) be the most economical way to really get a lot of vibration?

Sure. But there's a difference between shakers you use in calibrated test setups and stuff that goes "brrrr" in your hand :D

Offline vk6zgo

  • Super Contributor
  • ***
  • Posts: 7547
  • Country: au
Re: EEVblog 1442 - DON'T DO THIS!
« Reply #12 on: December 05, 2021, 08:54:22 am »
I think I have referred t this before, but we had, as part of the ENG setup on a helicopter, a very small Sony Trinitron, (around 50mm) mounted in the instrument panel.
Although it was not "flight rated" we had a "special dispensation " to use it.
As happens, eventually the tube lost emission  & died.

With great difficulty, the tiny device was dismantled, a new tube fitted & adjusted for best picture.
Running on test, & with judicicious"bumps" it refused to fail, so it was refitted to the "chopper".

Sadly, it failed on its first outing, so back it came so we could investigate.
No faults could be found, so "just in case" several "possible spots" were "anchored" with neutral silicone sealant.

Failed again!

After a couple more iterations we apparently "hit the jackpot", as there were no more failures.
It was very unsatisfying, as we never knew which blob of sealant finally did the job!

The joke around work was that we had a "vibration desk that could fly!" ;D
« Last Edit: December 05, 2021, 08:56:08 am by vk6zgo »
 

Offline Kleinstein

  • Super Contributor
  • ***
  • Posts: 14016
  • Country: de
Re: EEVblog 1442 - DON'T DO THIS!
« Reply #13 on: December 05, 2021, 09:12:43 am »
The smallest ones on the market?

Well, no -> https://www.qsources.be/simcenter-qsources/
I'd imagine the smallest actuators of that kind would be the ones used in smartphones for tactile feedback.

On the other end of the scale, wouldn't compressed air cylinders acting as a mechanical amplifier of sorts (like a jackhammer) be the most economical way to really get a lot of vibration?

Compressed air is relatively low efficiency by nature.  There are more economic ways to create virations with drive from a motor, but these are usually not so flexible in the frequency and amplitude. AFAIK the very big ones to simulate earthquakes for bulding structures use hydraulic rams. However this is more high force for a short time, not so much moderate vibrations for days.
 

Offline DH7DN

  • Regular Contributor
  • *
  • Posts: 128
  • Country: de
    • DH7DN Blog
Re: EEVblog 1442 - DON'T DO THIS!
« Reply #14 on: December 05, 2021, 10:55:02 am »
(Attachment Link)



Hi Dave,

When I watched the video, I kept thinking you are shaking the wrong plane.
I think it would fail a lot quicker if it was shaken so the PCB is vertical and the regulator is horizontal?

Regards,

Jay_Diddy_B

I think he did it on purpose in order to demonstrate the effect of vibration fatigue in this particular arrangement. Every mechanical structure can be approximated as a beam (1D) or a membrane (2D) and due to its mechanical properties (usually its mass and spring constant/Young's Modulus) one can observe different types of resonances (longitudinal, transversal, rocking/bending/shearing). At even higher frequencies (few kHz to tens of kHz), according to the laws of mechanics, the "stiff" structures become "jelly-like" and lots of crazy stuff happens. Imagine a screw or a bolt turning into a spaghetti at high frequencies. That's how they behave :scared:

Electrodynamic shakers are mostly preferred because they can be very well controlled (electrically) within parameters such as force, displacement, velocity or acceleration (-amplitude) over a wide frequency range. There are always principal limitations (e. g. maximal displacement at low frequencies, shaker resonances) and therefore no "universal" shaker exists.

Compressed air is used for cooling or for reducing the friction (air bearings), it's not viable for generating the sinusoidal motion (actually it is viable in certain cases but very limited to few applications). Air has some undesireable properties such as compression heat for slow processes, nonlinear properties (compression modulus is a function of density, density changes when a gas is compressed), it's very expensive compared to electricity etc. Same with hydraulics (no compression problems) but they have niche applications (heavy structures, low frequencies <10 Hz)
vy 73 de DH7DN, My Blog
 

Offline EEVblogTopic starter

  • Administrator
  • *****
  • Posts: 37626
  • Country: au
    • EEVblog
Re: EEVblog 1442 - DON'T DO THIS!
« Reply #15 on: December 05, 2021, 10:55:34 am »
When I watched the video, I kept thinking you are shaking the wrong plane.
I think it would fail a lot quicker if it was shaken so the PCB is vertical and the regulator is horizontal?

Yes I know, this was just a quick and dirty test for fun using stuff to hand.
 

Offline Fungus

  • Super Contributor
  • ***
  • Posts: 16531
  • Country: 00
Re: EEVblog 1442 - DON'T DO THIS!
« Reply #16 on: December 05, 2021, 12:37:27 pm »
When I watched the video, I kept thinking you are shaking the wrong plane.
I think it would fail a lot quicker if it was shaken so the PCB is vertical and the regulator is horizontal?

Of course! But it's more interesting to see it fail in a plane where you might not expect it to.

Dave: You missed a chance to put the legs in a staggered pattern to show us the difference it makes.

(and will it still fail if it's shaken laterally as shown above...)
 

Offline ejeffrey

  • Super Contributor
  • ***
  • Posts: 3669
  • Country: us
Re: EEVblog 1442 - DON'T DO THIS!
« Reply #17 on: December 05, 2021, 07:31:53 pm »
The poor little 7812 did survive quite some time and quite some vibrations. Unless used in things like automotive I would not consider the free standing TO220 so bad.

It depends.  Any product can be transported long distances in a truck.  It's a low likelihood problem as evidenced by the many devices out there with free standing TO220s that survive just fine, but if you are building a product and 0.1% DOA is a problem you don't want this even if only for shipping purposes.

If I build something for personal use I wouldn't give a second thought to leaving a free-standing TO220, probably even with a small heat sink.  For a mass produced commercial product with a warranty it certainly isn't ideal.
 

Offline EEVblogTopic starter

  • Administrator
  • *****
  • Posts: 37626
  • Country: au
    • EEVblog
Re: EEVblog 1442 - DON'T DO THIS!
« Reply #18 on: December 05, 2021, 10:43:21 pm »
Of course! But it's more interesting to see it fail in a plane where you might not expect it to.
Dave: You missed a chance to put the legs in a staggered pattern to show us the difference it makes.

Yes I know, which is why I added that the video.
As I said, this was literally a quick and dirty test to see what would happen, absolutely no thought went into the video at all in terms of it being some sort of professional and instructive demonstration etc.
 
The following users thanked this post: DH7DN


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf