Stress testing a wireless router

[dkotschessaa]

This is not exactly engineering. Please move where appropriate if it's not.

I am testing some 4G netcomm wireless routers to see how much traffic they can handle in certain situations. We are trying to find ways to "impede" them. They have two antennae that can be unscrewed. Removing both pretty much messes them up completely - good to know. Removing one slows things down a bit, but doesn't grind it to a halt.

The best someone came up with at work to test was:

In our meeting I said that I believed this was the old "tin foil hat" thinking. The foil makes it EASIER for the aliens/government to track your thoughts, right? Anyway, in this case, is the can impeding or boosting? What would be a good way to impede the signal?

-Dave K

p.s. this is becoming a fun place to work. Let's hope I get hired full time.

 

[dkotschessaa]

Wait - aluminum cans stuck on wireless router antennas... this IS engineering, isn't it?

 

[dkotschessaa]

Noone?

 

[NTL2009]

Don't you want something more controlled and repeatable than that? Do those antennae screw on using SMA, or other standard connectors? If so, get a variety of attenuators. You will have a known loss.

 

[sophiecentaur]

You will have a known loss.

In principle; as long as the connector is connected in a pukka way to the circuit board. Even with an attenuator of known value, signal can be radiating any old way from the box. I have heard rumours of all sorts of dodgy practice with router design. Let's face it, some of them are a lot worse than others. They rely on the fact that very few people have the facilities to measure their performance.

 

[NTL2009]

True, if the antenna circuit isn't designed/implemented with a fairly close matching impedance, the loss may not be the same as the stated number on the attenuator.

But it should be repeatable on a particular unit, and maybe even across units of the same make/model if their 'dodginess' (is that a word?) is at least consistent. At least far more so than a coke can hung over the antenna!

 

[sophiecentaur]

The only way to test any receiver-plus-antenna is to present it with a signal of known field strength. That would involve quite a lot of trouble because you would need a calibrated source and a detailed knowledge of the whole system. Otoh, it would not be difficult to compare two different receivers.

 

[OCR]

... is that a word?

Yowza, it most definitely is .[COLOR=#black].[/COLOR]

 

[NTL2009]

Yowza, it most definitely is .[COLOR=#black].[/COLOR]

Thanks - I had to look up "pukka" earlier :) That one was new to me, I like it, now I need to remember to add it to my everyday speech!

 

[dkotschessaa]

Don't you want something more controlled and repeatable than that? Do those antennae screw on using SMA, or other standard connectors? If so, get a variety of attenuators. You will have a known loss.

I would love something more controlled and repeatable, but we are kind of in a position to make do with what we've got, as this testing is not a major QA requirement.

But to answer your question - well, can you tell me?

Edited to add: if I had actually looked up attenuators than I would have seen that this is probably the right thing, but I'd love if

a) Somebody can verify that from the above photo
b) where/what you recommend on getting.

(Remember, I'm a half-assed mathematician and even less an engineer/electronics guy).

-Dave K

 

[dkotschessaa]

True, if the antenna circuit isn't designed/implemented with a fairly close matching impedance, the loss may not be the same as the stated number on the attenuator.

But it should be repeatable on a particular unit, and maybe even across units of the same make/model if their 'dodginess' (is that a word?) is at least consistent. At least far more so than a coke can hung over the antenna!

To that - is the coke can impeding or actually helping in some way?

And just to repeat - this is a very low budget experiment - if that isn't obvious. :) Most of our real QA work is geared towards the devices the router is connected to.

-Dave K

 

[dkotschessaa]

Sorry I keep asking on the same question. So is this just like a coax/cable connector? In which case I could hook up a few of these and put the antenna right back into them? https://www.amazon.com/dp/B01HJNGWYS/?tag=pfamazon01-20

 

[Rive]

... is this just like a coax/cable connector?

Yes.
Actually sometimes it happens that the antenna of the router is changed to some directed or special one instead of that puny stick, and in such cases it is exactly used as a coax connector for cable.
However, I'm not sure that it'll be an 50Ohm one. You have to dig that up somewhere.

 

[NTL2009]

What is the purpose of this experiment? Is the "I am testing some 4G netcomm wireless routers to see how much traffic they can handle in certain situations. We are trying to find ways to "impede" them. " an attempt to mimic real world conditions? In that case, things like a Coke can may cause a variety of issues - reflecting excessive RF energy back into the router, for example. This may be very different from the typical scenario in a home/business, where the users are just far away from the router (or many walls in between), resulting in lower RF levels being seen by both the router and the user, and a lower S/N ratio.

As an analogy, if I were tasked to discover how adverse conditions might affect the fuel efficiency (mpg or km/L) of a vehicle, I might test it in cold weather, up hill, head winds, etc. But it wouldn't make sense to flatten all four tires, and expect to get anything meaningful from that. Your Coke can may be much like flattening all four tires. Too far from real-world conditions to translate in any meaningful way.

I suggest you step back and think about what it is you are trying to accomplish.

 

[dkotschessaa]

What is the purpose of this experiment? Is the "I am testing some 4G netcomm wireless routers to see how much traffic they can handle in certain situations. We are trying to find ways to "impede" them. " an attempt to mimic real world conditions? In that case, things like a Coke can may cause a variety of issues - reflecting excessive RF energy back into the router, for example. This may be very different from the typical scenario in a home/business, where the users are just far away from the router (or many walls in between), resulting in lower RF levels being seen by both the router and the user, and a lower S/N ratio.

As an analogy, if I were tasked to discover how adverse conditions might affect the fuel efficiency (mpg or km/L) of a vehicle, I might test it in cold weather, up hill, head winds, etc. But it wouldn't make sense to flatten all four tires, and expect to get anything meaningful from that. Your Coke can may be much like flattening all four tires. Too far from real-world conditions to translate in any meaningful way.

I suggest you step back and think about what it is you are trying to accomplish.

We don't have the luxury of mimicking real world conditions. Some imagination and improvisation will be required.

EDIT:

To add: What we mostly are testing are devices (POS machines) that will be connected through this router to the Verizon network. However, we are curious about how much traffic they can handle. We will be running a lot of tests on the registers themselves, including transactions, but in particular reports that will create a lot of traffic.

I am sorry if this is not precise, but we just want to lower the signal - i.e. have "less bars" so we can see how our tests perform when the signal is impeded somehow. It is really a bit of a side project - not the major focus of our testing, so we don't have to be official about it.

-Dave K

 

[NTL2009]

...

I am sorry if this is not precise, but we just want to lower the signal - i.e. have "less bars" so we can see how our tests perform when the signal is impeded somehow. It is really a bit of a side project - not the major focus of our testing, so we don't have to be official about it.

-Dave K

Just go down the hall, put some distance/walls between your devices and the router? Use 'bars' as your guide. That seems pretty 'real world' (though not very precise) to me. But maybe good enough to get a general feel for signal level versus throughput issues.

 

[dkotschessaa]

Just go down the hall, put some distance/walls between your devices and the router? Use 'bars' as your guide. That seems pretty 'real world' (though not very precise) to me. But maybe good enough to get a general feel for signal level versus throughput issues.

My apologies for not saying explicitly or not clearly in previous post: The devices are wired to the router. The router is 4G. We want to impede the router's traffic to the 4G network.

-Dave K

 

[rbelli1]

Unplugging the antenna briefly during operation may be your best bet. The throughput for the weakest signal will likely be hundreds of kilo-bits per second. Your POS systems should not need to sustain anywhere near that even in multiples.

An easy way to do that is get some TNC or RP-TNC adaptors. RP-SMA->TNC--TNC->RPSMA. Make sure the adaptors you get are for the same impedance as the system you have. Likely 50Ohm. Or get the wrong one. That will give you some extra difficulty.

Cut off the female threaded part of the TNC coupler and you have a semi-secure connection that can be cycled fast. TNC is essentially threaded BNC. Use a rotary tool. Diagonal pliers will not cut those easily even though they are brass. Yours may be steel.

BoB