Pasternack attenuator blew up in seconds at rated power

[maxwell3e10]

I was testing a microwave amplifier using a Pasternack PE7048-30 10W attenuator. Literally within seconds of reaching 10W (10dB output of the attenuator) it blows up, the input impedance is now 13 Ohm. Luckily I didn't pay retail price for it, but I opened the plastic packaging myself, so it wasn't abused previously. I thought Pasternack is a respectable company, but now would question it.

[bdunham7]

Assuming you have the amplifier connected to the attentuator which is attached to a power meter of some sort (details?), perhaps the power meter was reading low for some reason and you actually provided more than 10W to the attenuator.  Is the amplifier capable of much more output?  Pasternack is pretty reputable and that wasn't a cheap part, as I'm sure you know.

[maxwell3e10]

No, I checked the power detector and the amplifier itself is rated at 10W too, so it couldn't have been off by more than 1-2dB. I guess I should have used an attenuator rated to much higher power than the amplifier, but for a quick test it should have been OK.

[bdunham7]

Then I'd contact Pasternack and explain exactly what happened.  If it was new-in-package they may be willing to listen.

[james_s]

I wonder if it's a counterfeit part?

[bdunham7]

Well....here's an eBayer that has them in bulk for $24, about 95% off MSRP....

https://www.ebay.com/itm/224535865962

And here's another...

https://www.ebay.com/itm/164812659676

[maxwell3e10]

You guys are quick to guess, yes it was the second of these listing. It did come sealed in the plastic packaging, which is exactly the reason I picked it. I already threw away the packaging, can't tell if it looks slightly off from authentic.

[james_s]

I bet it's counterfeit, any time a seller has a bunch of new items at an absurdly low price that should be highly suspect. Contact Pastermack and tell them you have one of their products that failed and that you suspect it might be counterfeit, I would expect they will offer tips on identifying whether it is genuine or not.

[bdunham7]

What frequency were you testing at?  There are cheap attenuators available that don't go to 12+ GHz, perhaps a counterfeit could be based on that.

[maxwell3e10]

It was at 6 GHz.  With MW components it's hard to tell what is an "absurdly low price" since MSRP is often absurdly high and many functional items can be had for a small fraction of it.

[Bud]

Since it is dead anyway, get it open to see what is inside. Won't be surprized if it has a few axial resistors instead of anything high frequency stuff.

[coppercone2]

More than 10 available 10 sold / See feedback

ITS A FAKEEEEEE!!!!!!!

[coppercone2]

how do you think the military and nasa end up with fake parts? There is some sophisticated scammers out there.

[joeqsmith]

I would suggest not trying to take it apart unless you are certain that the materials used do not pose any health risks (BeO used to be fairly common).

Looking at the Pasternack datasheet, they derate the part linearly down to 2 Watts @ 125C.    We were playing around with some very low cost Chinese parts recently.   Maybe something there would help.

https://www.eevblog.com/forum/rf-microwave/cb-and-ham-radio-techs-love-their-bird-wattmeters/msg4423153/#msg4423153

[maxwell3e10]

Here is what it looks like inside. Never opened something like this, so tell me what you think? Not obviously fake but it seems somewhat doubtful that these little surface resistors, even if on a fancy substrate, can hold 10 W.

[ejeffrey]

It was at 6 GHz.  With MW components it's hard to tell what is an "absurdly low price" since MSRP is often absurdly high and many functional items can be had for a small fraction of it.

If it is listed as "new" and a name brand for less than 10% of the MSRP it's a fake, pretty much by definition.  How would the seller get it?  It's not like pasternack sells via wholesalers at such a dramatically lower price.  It's possible to get unopened parts from liquidation sales, but rarely in high quantities, and if genuine will usually be priced higher.  The only question is what kind of fake is it?

[maxwell3e10]

Based on package labeling I would say this is NOS. It wouldn't surprise me if someone had a box of these that was surplused.
I tested the other side of the attenuator with DC voltage and at 0.5A (12.5W) it started to fail after a couple of minutes, the resistance started going up. Without having blown up enough of these, I don't know how closely such attenuators are typically rated.

[G0HZU]

If it helps, a general rule of thumb for high power attenuators is that you should only run them at full duty cycle at about a quarter to a fifth of the rated power if not cooled or regulated externally. The attenuator can be run at the full rated power if the case temperature is regulated or if it is not run at full duty cycle.

So, I would only rate that attenuator at 2.5W average power for continuous use.

I'd expect it to last more than a few seconds at 10W though. How many seconds did it last for? It might be fake, or it might not have been assembled properly at the factory. To achieve a low thermal resistance between the attenuator chip and the metal carriers and the black heatsink there has to be a solid thermal connection between each of them. I'd expect the inner parts to be tightly bonded together in some way because of this.

[G0HZU]

Here's an image showing the insides of my old Bird WA1000 40dB attenuator.

https://birdrf.com/Products/Test%20and%20Measurement/Attenuators/1kW/1000-WA-Series_RF-Attenuators.aspx

This huge attenuator is about 50cm long and is rated at 1kW although this isn't for full duty cycle. You can see that each white attenuator section is firmly bonded (using thermal expoxy or solder?) to a copper spreader and this is then screwed down to the main heatsink.

This attenuator is directional and the first couple of attenuator chips only attenuate by about 0.5dB. Each section after this has more and more attenuation. This distributes the heat evenly along the length of the attenuator. Even so this means there could be 100W of heat dissipated in some of the white attenuator chips. These chips are a couple of cm wide.

I'd expect your attenuator chips to be bonded to the round copper carriers in some way, but they appear to be a loose fit?

[bdunham7]

Based on package labeling I would say this is NOS. It wouldn't surprise me if someone had a box of these that was surplused.

You can just ask Pasternack.  Their reputation is top-notch and they supposedly have 24/7 global support.   I would be shocked to find these 'surplused', as they are a completely current product and quite marketable.  Stolen perhaps, but I still doubt we'd see them at these prices or that yours would have blown up. 

Also consider that these have obviously already been cloned as replica versions that don't claim to be Pasternack and they appear all over eBay.  So the shell and hardware are already available, all you have to do to make a fake is add some markings and a package.  The only mystery is why they are being so obvious and selling them as new.  They're worth more in used condition.

https://www.ebay.com/itm/274310407299

https://www.ebay.com/itm/174741192533

This could be a supply-chain security issue as well.  The Pasternack version is purportedly assembled and tested right here in California, but if the shell and connectors are subbed out to a Chinese contractor and the final product is assembled here using an expensive attenuator element, well you can guess what happens next.

Based on your tests and the fact that 12.5W will fry it, what you have may actually be an attenuator only good for 3GHz and thus you simply toasted it by actually exceeding 10W. 

[maxwell3e10]

If it helps, a general rule of thumb for high power attenuators is that you should only run them at full duty cycle at about a quarter to a fifth of the rated power if not cooled or regulated externally. The attenuator can be run at the full rated power if the case temperature is regulated or if it is not run at full duty cycle.

So, I would only rate that attenuator at 2.5W average power for continuous use.

Thanks, this is a good rule of thumb that I wasn't aware of.

I'd expect it to last more than a few seconds at 10W though. How many seconds did it last for? It might be fake, or it might not have been assembled properly at the factory. To achieve a low thermal resistance between the attenuator chip and the metal carriers and the black heatsink there has to be a solid thermal connection between each of them. I'd expect the inner parts to be tightly bonded together in some way because of this.

It only lasted 10-20 sec and wasn't even hot to the touch at the end.

I was a little surprised that there was no thermal paste on the outside of the brass rings. They fit reasonably snug into the hole of the heatsink, but can be easily shaken out. The resistor chips have little brass springs at the edges that fit snugly into the ends of the channel in the brass ring.

[G0HZU]

It only lasted 10-20 sec and wasn't even hot to the touch at the end.
I was a little surprised that there was no thermal paste on the outside of the brass rings. They fit reasonably snug into the hole of the heatsink, but can be easily shaken out. The resistor chips have little brass springs at the edges that fit snugly into the ends of the channel in the brass ring.

Yes, it doesn't appear to be assembled with a good thermal interface from the attenuator chips to the metalwork.

A good example of that attenuator would be able to withstand 10W for quite a while. However, I'd expect the thermal stress to eventually degrade something within the attenuator especially if the attenuator was run at full rating quite often. If the attenuator is run at reduced duty, then it will last a lot longer.

At work, I've seen a lot of high-power attenuators that eventually degrade if run too hot, too often. The internal resistances can change and sometimes the internal connections can become intermittent.

[maxwell3e10]

You can just ask Pasternack.  Their reputation is top-notch and they supposedly have 24/7 global support. 

What would I ask them? If I say their attenuator failed, they will say "sure, it is fake". If I ask them for pictures or drawings of internal construction, I am not sure if they would share them.  Not sure how one can check short of buying one from them directly and opening it up

[james_s]

You can just ask Pasternack.  Their reputation is top-notch and they supposedly have 24/7 global support. 

What would I ask them? If I say their attenuator failed, they will say "sure, it is fake". If I ask them for pictures or drawings of internal construction, I am not sure if they would share them.  Not sure how one can check short of buying one from them directly and opening it up

Tell them you have a product claiming to be theirs that failed and that you suspect it may be counterfeit, I would assume they'll ask you for some pictures and will be able to tell from that.

[ogden]

Stainless knurling is shallow. I would agree that yours is fake. For comparison this looks legit to me: https://www.ebay.com/itm/203214649459