DUTs

After days of fighting computers, I needed to solder something, so I built a new resistor tester. Full credit to Paul Horowitz for the proper definition of DUT.

This puts 30 volts across a 50 ohm 1% 1206 resistor (18 watts) until it fails.

formatting link

formatting link

formatting link

--
John Larkin         Highland Technology, Inc 
picosecond timing   precision measurement  
 Click to see the full signature
Reply to
John Larkin
Loading thread data ...

What's Z438 mean?

--
 Thanks, 
    - Win
Reply to
Winfield Hill

We assign a part number to practically every prototype or experiment, and document the gadget and any interesting results. Some are PCBs and some are hacked things and some are just test setups.

We have a folder on a server J:\Protos with all these in subfolders, and a log file.

--
John Larkin         Highland Technology, Inc 
picosecond timing   precision measurement  
 Click to see the full signature
Reply to
John Larkin

After days of fighting computers, I needed to solder something, so I built a new resistor tester. Full credit to Paul Horowitz for the proper definition of DUT.

This puts 30 volts across a 50 ohm 1% 1206 resistor (18 watts) until it fails.

formatting link

formatting link

formatting link

--
John Larkin         Highland Technology, Inc 
picosecond timing   precision measurement 
 Click to see the full signature
Reply to
Carl Ijames

I think I can detect the 30 VRMS fault condition in way under a millisecond, so I should be OK.

I have a Conservation of Energy problem; there's a bunch of stored energy and if it doesn't go where I want it, it has to go somewhere else. COE is damned unforgiving.

formatting link

The 50 ohm resistors are basically a flyback snubber. The average voltage at CLAMP is about -1.5 normally, but hits -60V peak, -16V average, 30V RMS when things go wrong.

The MAX809 is doing power supply UVLO and powerup reset. We add several shutdowns onto that.

I designed a circuit that turns out to have an extensive list of failure modes, several of which I missed on the rev A board. I discovered one of them during a customer demo.

--
John Larkin         Highland Technology, Inc 

lunatic fringe electronics
 Click to see the full signature
Reply to
John Larkin

John Larkin wrote in news: snipped-for-privacy@4ax.com:

DUT stands for "Device Under Test".

Reply to
DecadentLinuxUserNumeroUno

Anything special about the resistor? Aluminum oxide or silicon nitride?

On your torture fixture, the 50 ohm resistor is soldered to two large copper heat sinks on the PCB. Apparently, the heat will transmit through the solder blobs and copper much quicker than anticipated resulting in a 300 msec burnout. If you monitor the heat spreading on the copper heat sinks with your FLIR camera, I think you'll find a rapidly expanding and large heat affected zone. (You might need to paint the copper black in order to see it).

If you have time, try elevating the 50 ohm resistor on two thin wires above the PCB, or cut small islands under the resitor ends, and see how quickly the resitor will blow up. Since my 78 msec number was based on only the mass of the 1206 resistor, with no consideration for the solder blobs or copper heat sinks, methinks the torture test, with the now minimal additional heat sinking, will be fairly close to my guess(tm).

--
Jeff Liebermann     jeffl@cruzio.com 
150 Felker St #D    http://www.LearnByDestroying.com 
 Click to see the full signature
Reply to
Jeff Liebermann

It might also be useful to run the test at an elevated ambient temperature (what ever your spec calls for the equipment).

Reply to
tom

RthJA time curves for SMTs level out in the seconds. It'll begin to be relevant here -- but not of primary concern.

What is of primary concern, is the element already being at its melting point by then! :)

You estimated 155C though -- the resistor is just fine _operating_ at 155C. I've desoldered resistors before, they can take much more than that (if not for too long, presumably :) ). The enamel probably starts melting in the red-hot area, and the element melting and oxidizing in the same range, with hot spots beyond yellow hot.

Indeed, I think you do have your answer, very nearly, in the plot! The initial ramp down, is probably the element heating up, showing a slight +TC. That seems to be going well for the first 50 or 100us. But then it curves up, perhaps a hotspot is igniting and melting and the gathered material has a lower resistance? Then it moves, the gap opens, resistance increases further, and so on until pssst--it arcs out?

A high speed closeup video would be neat. In dark lighting, I bet the hotspot becomes reddish and visible by 50us or so.

IR too, but those are notoriously hard to come by.

Tim

--
Seven Transistor Labs, LLC 
Electrical Engineering Consultation and Contract Design 
 Click to see the full signature
Reply to
Tim Williams

For these resistor exploders, it's Device Under Torture.

--
 Thanks, 
    - Win
Reply to
Winfield Hill

Winfield Hill wrote in news: snipped-for-privacy@drn.newsguy.com:

"torture" implies that the "victim" will survive.

This would be device under destructive test. DUDT.

Reply to
DecadentLinuxUserNumeroUno

Device Under Termination.

--
John Larkin         Highland Technology, Inc 

lunatic fringe electronics
 Click to see the full signature
Reply to
John Larkin

Ordinary 1% 1206 thickfilm. I tried a 47 ohm 5% and got similar results.

I only need tens of microseconds of endurance, and the heat won't get far in that time scale.

That would be interesting. We run 1206s at 1 watt with heat-sunk end caps. There will certainly be some heat sinking in the hundreds of milliseconds time frame.

A resistor is thermally complex, so any calculations short of 3D FEM will be approximate.

Now I have a fixture to torture future generations of resistors.

--
John Larkin         Highland Technology, Inc 

lunatic fringe electronics
 Click to see the full signature
Reply to
John Larkin

Here it is:

formatting link

formatting link

62 milliseconds.
--
John Larkin         Highland Technology, Inc 
picosecond timing   precision measurement  
 Click to see the full signature
Reply to
John Larkin

Fun. Not the failure mode you'd expect with it attached to a board, perhaps.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
 Click to see the full signature
Reply to
Phil Hobbs

John Larkin wrote in news: snipped-for-privacy@4ax.com:

If this application is as an 'HV snubber' the best device is a 'bulk medium' carbon composition resistor.

Reply to
DecadentLinuxUserNumeroUno

I need the paralleled transzorb to keep the fault peak voltage down, to avoid blowing up other stuff. The transzorb limits the voltage across the resistors but it dissipates about 70 watts in the fault case. So I need the fast shutdown circuit. So the 1206 resistors are OK.

--
John Larkin         Highland Technology, Inc 
picosecond timing   precision measurement  
 Click to see the full signature
Reply to
John Larkin

In the early days of teleconfusion, DUT meant "Dial Up Terminal".

When I was playing editor, DUT meant "Don't Use This" which is a diplomatic hint to throw the whole thing out and start over.

There are other meanings: (5 pages)

Duts all folks...

--
Jeff Liebermann     jeffl@cruzio.com 
150 Felker St #D    http://www.LearnByDestroying.com 
 Click to see the full signature
Reply to
Jeff Liebermann

I made a bad assumption in my original guess(tm). I assumed that it was desirable for the resistor to maintain it's 50 ohm value to within

1% tolerance, which is my definition of "survive". I also assumed that the resistor would use an aluminum nitride substrate, not alumina.

Here are the original calcs (as corrected for my screwup):

1206 weighs about 10 grams / 1000 pcs Specific heat of alumina = 800 Joules/Kg*Temp The Al Nitride chip is rated to survive to 155C. Ambient temp at 25C

Specific_Heat = (Watts * Seconds) / (kg * dTemp) J/Kg.K

800 = (18watts * sec) / (0.00001Kg * (155-25))

Seconds = 800 * 0.00001 * 130 / 18 = 57 milliseconds

What would need to change to match the test figure is the increase is specific heat and mass from the solder blobs and copper pours, change chip specific heat from aluminum nitride to aluminum oxide, and change maxi um survival temperature for the 1206 resistor (which I don't know if it's thick or thin film). In other words, change everything.

Agreed, except for the last part. The gap doesn't open and the resistor doesn't melt. In both thin and thick film, the resistor material is a form of glass. It takes too much energy (watt-seconds) to melt glass. Instead, the narrowest part of the resistor (usually where it was laser trimmed to 1%) produces a hot spot, which then expands the glass. The rest of the resistor isn't going to move, so glass cracks along the thermal gradient.

Yeah, that would be interesting. High Speed IR Footage with FLIR's SC6000 I couldn't find a price.

--
Jeff Liebermann     jeffl@cruzio.com 
150 Felker St #D    http://www.LearnByDestroying.com 
 Click to see the full signature
Reply to
Jeff Liebermann

Ah yeah, that makes sense. I was thinking in terms of metal (and thick?) films. And for a glass glaze, once that cracks, the crack will propagate quite effectively into a Al2O3 substrate, that stuff can't handle hardly any strain. (Or it cracks first, whatever.) And from there, you get... John's latest test result! ;-)

Tim

--
Seven Transistor Labs, LLC 
Electrical Engineering Consultation and Contract Design 
 Click to see the full signature
Reply to
Tim Williams

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.