Breaking Input Stage of Spectrum Analyser with Radio TX

[Bicurico]

Hello,

This is bugging me and I don't want to break my pr cious spectrum analysers.

Imagine I have i.e. a Silent SSA3021X. It is switched off.

Could the input stage break, if I take a radio, imagine a Baofeng UV-5R and transmit with the provided antenna? Both devices are less than 1m apart and the Baofeng transmits with 5W, some models even go up to 8W.

Is this safe? Is so, is it safe because the power diminishes squared to the distance? Does it make a difference to have the SA switched off?

If this would be a problem, what if your shack is close to a DVB-T or 4G/5G transmitter? Wouldn't they transmit with considerablyore power?

[radiolistener]

Is this safe?

It depends on spectrum analyzer protection circuit. But in many cases transmitter can damage your device. For example several days ago my acquaintance damaged two WiFi routers and PC server just because he tested power amplifier (something like 1 kW) for short wave transmitter in the same house...

As I remember, spectrum analyzers usually have absolute maximum rating about +20..+30 dBm which is 0.1-1 W.
Baofeng UV5R has 5 W power, so it can damage it.

[Bicurico]

Note that the input is not connected by cable.
What if my next door neighbour has a Baofeng and transmits 5W from the other side of the wall?

Somehow this intreagues me...

[xrunner]

Note that the input is not connected by cable.
What if my next door neighbour has a Baofeng and transmits 5W from the other side of the wall?

Somehow this intreagues me...

Well let's do a simple calculation and see what range we are in. Let's try these paramaters:

Tansmitter power
5W = +37 dBm

Free space calculation (simple version)

Distance = 1 m

Antenna gain of both TX and RX = 0 dBi

F = 146 MHz

Free space loss = 16 dBm
(use free space loss calculators available on internet)

So we have:

+37 - 16 = +21 dBm at input to analyzer

Many analyzers have a Max input of +20 dBm

However, your analyzer input does not have an antenna connected to it, so it won't have an antenna gain of even 0 dBi, it will be a lot lower than that. So I say no it won't harm your analyzer.

For an experiment, you would have to make a test rig shielded box with an N connector as the input with NO antenna connected, and a 50 ohm R on the inside, and be able to measure the power actually received by this passive setup a certain distance from to emitter to find out what you would get. Be an interesting test though.

[Wallace Gasiewicz]

I test lots of radios, but mostly into a load. You can see something on a SA even with no connection and the SA input "open"You can put an few wraps of wire on the outside of a transmitting coax and get a signal into your SA for spectrum analysis.
I have not damaged anything yet.If you are concerned put a 50 ohm load on the input of your test equipment to protect it. The load does not have to be good quality, so cheap ones can serve this purpose and provide a dust cover for the BNC. (or whatever)
We did analyze a Baofeng with an antenna connected to a SA and analyzed the Baofeng 2 meter spectrum emissions.  Several results with several transmitters showed spurious emissions that are way above FCC specifications..We did not damage the SA.

[Bicurico]

Thanks for the replies so far and sorry for the delay in answering back.

I never used my radios in TX close to where I have my spectrum analysers and they do have a 50/75 Ohm terminator on the input.

Still, I wonder how fragile the input stage really is: I mean a cell phone is transmitting, too. And many shacks have proper RX amplifiers and antennas installed.

Another question: what is the difference in using i.e. three 10dB attenuators in series or a load with a big cooler?

[xrunner]

Thanks for the replies so far and sorry for the delay in answering back.

No problem.

I never used my radios in TX close to where I have my spectrum analysers and they do have a 50/75 Ohm terminator on the input.

Still, I wonder how fragile the input stage really is: I mean a cell phone is transmitting, too. And many shacks have proper RX amplifiers and antennas installed.

What do you mean by "fragile"? A spectrum analyzer will state the Max input power like I said earlier. My Rigol DSA 815 has a Max input of +20 dBm (100 mW). That's how "fragile" the input is. If you mean something else by using the word "fragile" please say so.

Another question: what is the difference in using i.e. three 10dB attenuators in series or a load with a big cooler?

An rf load (dummy load) or termination is going to absorb all the power input to it and not reflect any back (ideally).

Three 10 dB attenuators in series is not a "dummy load" in the sense you seem to be thinking. They will still need to be terminated in a 50 ohm load to absorb all the power left over (which will be 30 dB lower than the input).

For example, if you input +50 dBm (100 W) into three 10 dB attenuators, what happens? Well the first attenuator needs to be rated for 100W. It will drop the power out to +40 dBm (10W), Then the second 10 dB att. will need to be rated for 10 W (of course a higher rating is fine too). It will drop the power to +30 dBm (1W). Then the last 10 dB att. will need to be rated for 1W and will drop the power to +20 dBm (100 mW).

Well then you have 100 mW of power that needs to see a 50 ohm load (or the proper impedance you are working with). You can't just leave the output open because three 10 dB attenuators in series is a 30 dB attenuator - not a dummy load.

[FAXFIXR]

My main hobby is restoring old tube transmitters and tube type RF Amplifiers... so I've keyed up a lot of amplifiers and transceivers on my bench with no damage to my SA or any other gear on the bench. I do use a rather large dummy load in the corner of the shop or sometimes do on the air testing, but I've never had any issues with computers, WiFi or Analyzers, which are all in close proximity.

I wonder if you could just keep a 50 ohm load or even an attenuator connected to the input of your SA for peace of mind?

[Bicurico]

Hi,

This time I am sitting on my computer, so hopefully I will have less mistakes and better description of what I am asking.

I already have 50 ohm terminators (some are in fact 75 ohm, as I mostly am interested in TV broadcast and reception) on most/all input ports of my spectrum analyzers (2x R&S CRTU, 2x R&S CMU200, 1x HP8594E, 1x SSA3021X). On the broken units (1x R&S CMU200, 2x HP8594E) and on the mandatory small devices (TinySA, etc.), I don't have any protection. Everything is stored/placed in my shack within 3 square meters.

The reasoning of this thread (and my worries) is: what can happen if I sit in on my desk (with all devices at 1m-2m distance) and start transmitting on cheap chinese radios, which claim up to 8W power?

According to xrunners calculations, there is in fact reason to worry, but it is borderline: sure enough, the SA's don't have an antenna connected to the input port, so little power should go in. And I could of course just move away, when I want to hit the TX button.

But thinking about this, led me to imagine other scenarios:

Imagine you ordered a brand new SVA1032. Its input is specified at +30dBm = 1 Watt.

Now the DHL guy, who has your package in the back of his van, but somehow close to his CB antenna is broadcasting to his colleagues that he will be delivering this last package now. He is transmitting 4 Watts. Could he break the packaged SVA1032 by doing so?

What if a taxi is stopped next to the DHL van and is transmitting by radio that he just dropped of the passenger. His antenna is half a meter from the SVA1032 in the DHL van...

I know that the van has a metal shield, but for the sake of it, imagine it is sunny Australia and the van is like an open pickup truck.

Could the SVA1032 be "endangered"?

When I mentioned "fragile", I was wondering that a switched off SA should be imune to RF radiation next to it. The input stage should be in a metallic shielded box and a relay should cut off whatever could come in through the open RF input.

I know, perhaps I am just paranoid...

Dummy load: I know what that is and it wasn't what I meant.

I meant this: https://pt.aliexpress.com/item/1005006031464052.html as opposed to this: https://pt.aliexpress.com/item/4000412423002.html

But meanwhile I understood: I forgot that the attenuation is dependant of the Watts it is attenuating and there needs to be a cooler for higher Watt rating.

Thanks again for your input.

[2N3055]

If you want to be sure, simply buy a metallic connector cover.
It will shield input from any coupling, protect it from dust and protect
threads on the connector from mechanical damage.
So they are good idea anyways.

[xrunner]

The reasoning of this thread (and my worries) is: what can happen if I sit in on my desk (with all devices at 1m-2m distance) and start transmitting on cheap chinese radios, which claim up to 8W power?

According to xrunners calculations, there is in fact reason to worry, but it is borderline: sure enough, the SA's don't have an antenna connected to the input port, so little power should go in. And I could of course just move away, when I want to hit the TX button.

But thinking about this, led me to imagine other scenarios:

Imagine you ordered a brand new SVA1032. Its input is specified at +30dBm = 1 Watt.

Now the DHL guy, who has your package in the back of his van, but somehow close to his CB antenna is broadcasting to his colleagues that he will be delivering this last package now. He is transmitting 4 Watts. Could he break the packaged SVA1032 by doing so?

What if a taxi is stopped next to the DHL van and is transmitting by radio that he just dropped of the passenger. His antenna is half a meter from the SVA1032 in the DHL van...

I know that the van has a metal shield, but for the sake of it, imagine it is sunny Australia and the van is like an open pickup truck.

Could the SVA1032 be "endangered"?

I'm thinking of an experiment that will (hopefully) put your mind at ease, but I need to check some things out today. I will post the results here ...

When I mentioned "fragile", I was wondering that a switched off SA should be imune to RF radiation next to it. The input stage should be in a metallic shielded box and a relay should cut off whatever could come in through the open RF input.

I know, perhaps I am just paranoid...

I understand, but let's see what my experiment shows.

Dummy load: I know what that is and it wasn't what I meant.

I meant this: https://pt.aliexpress.com/item/1005006031464052.html as opposed to this: https://pt.aliexpress.com/item/4000412423002.html

Mmmm ... one model is a 10W att. with SMA connectors "as opposed to this" the others are 2 W att. with N connectors. What's your point? Is it the fact that the 10 W has cooling fins and the others do not? It's a design choice made by the manufacturer. The design choice is "what cooling system is required to dissipate X amount of power in a certain size package. What enclosure do we need to put the attenuation components in?"

In other words, the one with cooling fins could have been made larger without cooling fins if they wanted to put it in a bigger package that would dissipate the heat by virtue of it's larger thermal mass.

And by the same token, the 2 W designs with no cooling fins could have been made smaller if they wanted to put the attenuation element in a package with cooling fins. But then the N connectors might be bigger than the attenuation package so what would be the point?

The point is "cooling fins" do not tell you anything about what power something can dissipate. It just tells you, for the package the manufacturer chose, they needed cooling fins to dissipate the Max power rating of the device without a fan. You see, it all comes down to design choices - lowest cost, good performance, package size, etc.

Anyway, let me try some experiments and I'll be back ...

[xrunner]

For you Bicurico

The test: What we want to know is, using the N female connector as the "antenna", how much power is received at a measurement device behind the connector. It could be a spectrum analyzer, a power meter, or a receiver. It doesn't matter what is receiving the power, because it all depends on how good of an antenna the bare N connectors is.

I have a Baofeng UV-S9X3, a 3 band (146, 220, 440 MHz) HT. It puts out on high power ~4.7W on 146 MHz, 3.2 W on 220 MHz, and 4 W on 445 MHz. This is my test transmitter.

I will use a Boonton 4220 power meter to measure the power at the "output" of the N connector. The sensor can measure down to -60 dBm. As I said above, it doesn't matter what measures it as long as it is accurate.

First, is the power meter working correctly? The output of the Baofeng should be ~5 W or +37 dBm (depending on the charge state of the battery). I put 40 dB of attenuation in front of the power sensor to prevent damage. I compensated the power meter reading for this in the dB reference offset. I read +37.59 dBm or 5.7 W which is acceptable. But this is with a system connected with coax. What happens if we just use the antenna of the HT?

I removed the attenuators from the system and set the reference back to 0 dB, and screwed on a female N connector. This is our input - our "antenna", just like the spectrum analyzer would have with nothing connected to it. Now I want to transmit from the HT using it's antenna and see how close I can get to the N connector before reaching a danger level for damage to the sensor.

I started across the room transmitting and walked slowly to the sensor. It quickly became obvious that there was nothing to worry about, certainly at any distance greater than a hand's length. Then I continued slowly until I had the end of the HT antenna jammed up into the end of the N connector. This is as much coupling as I could achieve in any orientation around the N connector.

The Max coupling I could get gave -2.27 dBm (631 uW) for the 145 MHz band and -2.22 (600 uW) for the 450 MHz band. This is nowhere near the damage level of my power sensor (or spectrum analyzer of +20 dBm (100 mW)). So you see, the output of the HT is +37 dBm and the best coupling with the antenna stuck in the end on the N connector gives -2.2 dBm. One might think that this would be dangerous before this experiment, but even as it stands, the loss is +37 down to -2.2; a loss of 39.2 dB.

So we were checking whether a spectrum analyzer with an open N connector can accidentally be damaged by rf. What if the N connector has an antenna on it then? I put one on and then transmitted at 5 W and of course the antenna works like a charm. I could have easily exceeded the damage level of the power sensor of +24 dBm by moving just a bit closer than you see in the picture.

OK Bicurio, are you going to sleep better? Perhaps you're still not sure? Just don't keep an antenna on your spectrum analyzer. 

[cncjerry]

How many people remove all their cables all the time?  Brings up a good point, don't go chasing DX with your 2kw amp while cables are attached to sensitive inputs.

[Bicurico]

Thank you very much.
That really helps ease my mind.
I realized that I lack sensibility to evaluate signal loss over air and distance.
Sorry for the short answer, I am away. I will try to replicate some of this and have much more confidence now

Kind regards,
Vitor

[xrunner]

Thank you very much.

No problem I found it interesting as well. Never had a reason to anything like that.

That really helps ease my mind.
I realized that I lack sensibility to evaluate signal loss over air and distance.
Sorry for the short answer, I am away. I will try to replicate some of this and have much more confidence now

Kind regards,
Vitor

I'd be interested in your results if you duplicate them, or if they are different.

[joeqsmith]

Another question: what is the difference in using i.e. three 10dB attenuators in series or a load with a big cooler?

...
Well then you have 100 mW of power that needs to see a 50 ohm load (or the proper impedance you are working with). You can't just leave the output open because three 10 dB attenuators in series is a 30 dB attenuator - not a dummy load.

Guessing even two series 10dB attenuators, unterminated is would be fine in most cases.  Return loss would be 40dB which is pretty decent.   

***
Just for a quick comparison,  I used the LiteVNA to sweep from 1.5 - 500MHz.  Performed a SOLT using the ideal model and my sorted mini-circuits load.    Measured the return loss of the load supplied with the LiteVNA (black).   Replaced the load with a 10dB Midwest Microwave attenuator that was left unterminated (brown).  Measured roughly 22dB.   Added a second 10dB attenuator in series and left it them unterminated.  Return loss is now comparable with the load.

[Bicurico]

So, I finally got some short break to test it myself.

First test was with my cheapest device, the TinySA:



I tested outside, far enough from my other more valuable devices, but there was no problem at all.

I even dared to attach an antenna do the TinySA.

Feeling comfortable, I repeated the test inside, on my Rohde&Schwarz CMU200:



Again, no issues, I removed the attenuator and termination and still got a marginal signal.

This cleared my mind, it is safe to play with my radios next to the spectrum analysers, as long as they don't have a matching antenna.

Thank you all for your inputs.

Kind regards,
Vitor

[Wallace Gasiewicz]

You said you had 75 ohm Terminators, if these are just Loads you can check the performance of your VNA quickly by using a 75 ohm load on the 50 ohm output.   
 The resulting VSWR should be 1.5 to 1. Of course lots of cheap loads do not handle high frequency well. But a "quick check" none the less.

[pdenisowski]

I have an amateur radio station which I regularly operate with 100W on HF and up to 100W on VHF.  I also have numerous high end spectrum analyzers, network analyzers, etc. located < 10 meters from the antennas.  I am completely unconcerned with damaging my instruments and have never had any instruments damaged by OTA signals.

And although I can't go into any details in a public forum, I have often used some of our (R&S) spec ans to measure over-the-signals from very high power transmitters in very close proximity and have never seen an instrument damaged by this.  (Overloaded, yes, but not physically damaged)

As some other folks have mentioned, you can use the free space path loss equation to estimate the power at the spec an input if you're concerned.

That said, I have seen plenty of people blow out the front end of an analyzer by directly connecting (with a cable) something they shouldn't have been directly connecting without a sufficiently large pad

[CatalinaWOW]

The first attenuator in the string does not dissipate 100W. If it did there would be no power transmitted to the next stage.  A nit since it does dissipate 90W.   This does not require a chiller, but may benefit from a good finned heat sink.  It would otherwise heat to temperatures comparable to a 100W incandescent bulb and could burn your skin.   Obviously the first attenuator must be rated for greater than 90 watts which will likely mean it has the aforementioned fins.

[xrunner]

The first attenuator in the string does not dissipate 100W.

I stated it has to be rated to accept 100W of power - to withstand 100W of power - to not destroy itself upon accepting an input of 100W. That's what the rating is for.

[joeqsmith]

The first attenuator in the string does not dissipate 100W.

I stated it has to be rated to accept 100W of power - to withstand 100W of power - to not destroy itself upon accepting an input of 100W. That's what the rating is for.

He's just pointing out for your 100W example that the first 10dB attenuator  drops the signal from 50dB to 40dB as you pointed out, meaning it dissipates 90W.  This is all it needs to be rated for.   That or I missed it.

[xrunner]

He's just pointing out for your 100W example that the first 10dB attenuator  drops the signal from 50dB to 40dB as you pointed out, meaning it dissipates 90W.  This is all it needs to be rated for.   That or I missed it.

You missed it, but I don't understand why or how, or if you are just wanting to spend more time on the forum. Maybe you are teasing me? I know you know all this, perhaps you are testing me? Ah that's it.

Just research what the power rating of an attenuator is. It isn't how much it dissipates when X watts is applied at the INPUT. It's how much power it is rated for at the INPUT - the Max INPUT power. The power rating of an attenuator is the Max INPUT power it can take. How many other ways can I form the same statement?

I mean, there are thousands of documents and tutorials out there, or just look at any manufacturer's data sheet, as I've gone to the trouble of attached below. Look at the power specification. It's for a 10 dB, 100 W part. It can take no more than 100 W INPUT. The same thing would be true for a 100 W, 3 dB att. It's rated for Max input power. That's how all of them are done.

So no, if you start out with 100 W and want to drop the power with an attenuator, the first one in series (or only one) must be rated for 100W, no matter how much the dB att. is. There's really no more to say about it.

[joeqsmith]

Just research what the power rating of an attenuator is. It isn't how much it dissipates when X watts is applied at the INPUT. It's how much power it is rated for at the INPUT - the Max INPUT power. The power rating of an attenuator is the Max INPUT power it can take. How many other ways can I form the same statement?

So not a function of the dissipated power but instead the breakdown voltage?

https://birdrf.zendesk.com/hc/en-us/articles/360041312394-Can-you-use-two-25W-attenuators-in-series-for-a-50W-input-signal-

Bird FAQs Attenuators
Can you use two 25W attenuators in series for a 50W input signal?

Unfortunately no. The first attenuator in the series needs to be rated for the 50W input power. Otherwise the attenuator will be damaged. Depending on the attenuation value of the first, the second attenuator can have a rating of less than 50W.

For example, if you have a 50W signal and need a total reduction of 6 dB, you can use a 50W 3dB attenuator at the input and then a 25W 3dB attenuator after to the 50W attenuator.

***
Now I figured you would call me out for calling out your post on using un-terminated attenuators for a load.

[xrunner]

So not a function of the dissipated power but instead the breakdown voltage?

https://birdrf.zendesk.com/hc/en-us/articles/360041312394-Can-you-use-two-25W-attenuators-in-series-for-a-50W-input-signal-

Say we have a 100 W signal, and we want to drop it by 10 dB, so we have a 10 W output. OK then 90 W will be dissipated. But think about it. How does the power dissipation work inside the attenuator? Does it just happen at a single point inside, or does it happen in a gradual fashion, taking into account what resistive elements are inside?

It doesn't simply happen at a point, so if 100 W goes into the input connector, what happens? The first element is going to see 100 W. So it must be able to handle 100 W (yes it's a certain voltage of course) at the input.

It doesn't matter whether it's a 1 dB attenuator or a 40 dB attenuator, it has to be rated for 100 W at the input if that is the Max input power. If the Max input power is less, buy a lower rated att.

Now I figured you would call me out for calling out your post on using un-terminated attenuators for a load.

Oh yea I know what's going on. 

I've been on this forum for a long time and I know when to move along ...