Spectrum Analyser arrives today and worried about input protection

[trevwhite]

Hi All

My SSA3021x arrives today and reading through all the posts on here I can find got me more and more worried about protecting the input port.

The spec says the port is protected for 50V and 30dB. This seems a lot but then I read that a mobile phone near the device can flag overload.

I looked at the rfshop here in the uk and they have DC blocks and attenuators. If I buy a 30dB attenuator, that would suggest I am protecting the unit from a lot of input. Do I still need a DC block as well?

I think 30dB seems quite high really but I am new to SA use and so needs some advice.

Trev

[borjam]

The spec says the port is protected for 50V and 30dB. This seems a lot but then I read that a mobile phone near the device can flag overload.

I looked at the rfshop here in the uk and they have DC blocks and attenuators. If I buy a 30dB attenuator, that would suggest I am protecting the unit from a lot of input. Do I still need a DC block as well?

I think 30dB seems quite high really but I am new to SA use and so needs some advice.

The RF Explorer guys sell a limiter which, as far as I know, can help.

https://www.amazon.com/Power-Limiter-for-RF-Explorer/dp/B00IF8N1BO

I keep it attached to the antenna of mine whenever I use it.

[nugglix]

The spec says the port is protected for 50V and 30dB. This seems a lot but then I read that a mobile phone near the device can flag overload.

Hmmm...
Couldn't it be that protection and measurement range are two distinct things?
The input can withstand the 50V/30dB, but the machine can't measure it and giving the message.

Would be nice if someone could explain/verify that.

[technogeeky]

Hi All

My SSA3021x arrives today and reading through all the posts on here I can find got me more and more worried about protecting the input port.

The spec says the port is protected for 50V and 30dB. This seems a lot but then I read that a mobile phone near the device can flag overload.

I looked at the rfshop here in the uk and they have DC blocks and attenuators. If I buy a 30dB attenuator, that would suggest I am protecting the unit from a lot of input. Do I still need a DC block as well?

I think 30dB seems quite high really but I am new to SA use and so needs some advice.

Trev

You only need a DC block if you plan to put more than 50V DC into the unit. Most of the time (not always, but most of the time) people who use DC blocks have inputs which can't take any DC voltage whatsoever.

As for an attenuator, it's not really a big deal. As you learn about and use your spectrum analyzer, you will probably find yourself buying various kinds of attenuators. Some experimental setups call for putting an attenuator before and after a cable. Sometimes they just call for putting an attenuator at the end of a long run of cable. Some user's guides call for using two identical attenuators (say, two 10dB pads in a row). Eventually, you'll have a grab bag of attenuators (and possibly other, more sophisticated attenuation equipment) at your disposal. So it's not too important to worry about it. You're going to want to make sure in general you don't hook up very powerful equipment to your SA anyway.

There are a pair of guidelines which are more important to me than either of those two things:

1. Get a pair of sacrificial connectors or adapters.
2. Decide on a standard connector type, adapt to it from Type N, and buy all of your equipment (connectors, attenuators, switches, etc) accordingly.

In my case, this meant buying:

• 2x Type N to SMA adapters of good quality (I got some 18 GHz ones, and my specan is only rated to ~ 3 GHz)
• SMA to SMA adapters of every kind
• SMA to everything else adapters
• SMA attenuators
• parts to make DIY SMA cables

In other words, in my case, I immediately convert from Type N to SMA, and I do everything in SMA. SMA is a little bit less robust as a connector, but everything is a great deal cheaper.

Cheers,
-tg

[technogeeky]

The spec says the port is protected for 50V and 30dB. This seems a lot but then I read that a mobile phone near the device can flag overload.

Hmmm...
Couldn't it be that protection and measurement range are two distinct things?
The input can withstand the 50V/30dB, but the machine can't measure it and giving the message.

Would be nice if someone could explain/verify that.

Yes, usually equipment basically has three ratings. I'll list them along with the specs of the Siglent in question:

• 30dB maximum reference level
• 33dB maximum input sustained, 3 minutes, assuming > 20dB attenuation already^[1]
• ??dB maximum instantaneous input, causing shutdown of input

The 3dB extra is an acknowledgement of the fact that most/all instruments allow the signal to be a little above the reference without causing linearity or spurious tone problems.

The actual hard limit of the input is not listed, but I suspect it's probably a little above 33db, say, 35 or 36dB. On my instrument, if you put this kind of power in, it will immediately warn you that such an event has happened, and require you to restart the instrument to use it again.

^[1]: This attenuation can come from the internal attenuator (0~51 dB in 1 dB steps) or from the external pad, or both.

[tautech]

Hi All

My SSA3021x arrives today and reading through all the posts on here I can find got me more and more worried about protecting the input port.

The spec says the port is protected for 50V and 30dB. This seems a lot but then I read that a mobile phone near the device can flag overload.

My post ?
Usage case was with a 50mm dia coax loop....poor mans near field probe while looking at various instrument emissions. I had zero attenuation internal or external and some might say it was user error.
Dragged the loop past my body when moving it to another position just as the phone in my shirt pocket decided to handshake with the local cell tower.
More coincidence than anything else but it's good to be safe. Some small lessons learnt. 

Follow the good advice thus far. 

[rf-loop]

Hi All

My SSA3021x arrives today and reading through all the posts on here I can find got me more and more worried about protecting the input port.

The spec says the port is protected for 50V and 30dB. This seems a lot but then I read that a mobile phone near the device can flag overload.

I looked at the rfshop here in the uk and they have DC blocks and attenuators. If I buy a 30dB attenuator, that would suggest I am protecting the unit from a lot of input. Do I still need a DC block as well?

I think 30dB seems quite high really but I am new to SA use and so needs some advice.

Trev

Maximum input level is 30dBm (1W) (do not mix dB and dBm)
This is specified for MAX 3minutes duration and when internal attenuator is 20dB or more.
Damage level is 33dBm (2W)
Also it need note that these all are given for 50ohm system.

These can calculate easy with this
http://www.analog.com/en/design-center/interactive-design-tools/dbconvert.html

And it need also know that this is also if SA is power off!  These are maximums to input connector without exceptions.

And there is NOT overload protection! (not like example some Tek oscilloscope what disconnect 50ohm if overload)

If you work in situations where is danger for over stress SA input just use external protection installed to input N connector.

From Dave Jones tear down images:


Left C10 is first what receive all from N connector center tap. After then start sensitive circuits, directly.

[bd139]

I really wish they’d have a test/measure switch on SA’s. Test will give you a power reading and if it’s overloading the input it will warn loudly and refuse to measure.

[TK]

I really wish they’d have a test/measure switch on SA’s. Test will give you a power reading and if it’s overloading the input it will warn loudly and refuse to measure.

Then you need a DC block and attenuator to protect the test circuit

[Performa01]

Hi All

My SSA3021x arrives today and reading through all the posts on here I can find got me more and more worried about protecting the input port.

The spec says the port is protected for 50V and 30dB. This seems a lot but then I read that a mobile phone near the device can flag overload.

I looked at the rfshop here in the uk and they have DC blocks and attenuators. If I buy a 30dB attenuator, that would suggest I am protecting the unit from a lot of input. Do I still need a DC block as well?

I think 30dB seems quite high really but I am new to SA use and so needs some advice.

Trev

If the input can stand 50VDC, this is a strong hint that the instrument has a DC block already built in. No need for an external one.

And be careful, even if you think 50V might not be enough, just adding a capacitor (which a DC block essentially is) with a higher voltage rating only provides the expected protection if its capacitance is much lower than the internal one. This in turn means that the external DC block must have a lower bandwidth limit of at least 100kHz in order to be absolutely safe. AFAIK the lower bandwidth limit of the SSA3000X is 9kHz with the internal DC block.

It is quite obvious that you need a power attenuator if you want to measure high power sources, such as power amplifiers. Power attenuators are quite expensive though.

For general protection during normal use, I recommend a 10dB or even 20dB attenuator connected directly to the SA input all the time – except for these rare cases, where the maximum sensitivity is required.

Power limiters are also an option, but they will almost inevitably degrade the input characteristics of the instrument – particularly at higher frequencies – and might even add some distortion to the input signal. So I would not trust them blindly for accurate measurements.

[tautech]

If the input can stand 50VDC, this is a strong hint that the instrument has a DC block already built in. No need for an external one.

Yep, it's C10 in the pic rf-loop posted above.

[orin]

It is quite obvious that you need a power attenuator if you want to measure high power sources, such as power amplifiers. Power attenuators are quite expensive though.

Or use a directional coupler rated for the power source and feed the power source into a dummy load via the coupler.  'tis what I did to keep an 8568B safe from a 250W amplifier output.  I also used a 20dB attenuator at the SA input.  The 8568B has a internal 10dB attenuator by default and they make it difficult to disable, so in effect, I had 30dB attenuation on the input, as well as the 20dB coupling factor making a total of 50dB attenuation. 

Of course, you need a coupler rated for the power, but I was working at HF and an Elecraft CP1 did the trick.

For general protection during normal use, I recommend a 10dB or even 20dB attenuator connected directly to the SA input all the time – except for these rare cases, where the maximum sensitivity is required.

I don't usually bother for normal signals and rely on the built in 10dB, but I cannot recommend such a practice.  It doesn't do any harm to leave a 20dB attenuator on your SA input - you can always remove it if you are absolutely certain the input signal is 'safe'.

[technogeeky]

It is quite obvious that you need a power attenuator if you want to measure high power sources, such as power amplifiers. Power attenuators are quite expensive though.

Or use a directional coupler rated for the power source and feed the power source into a dummy load via the coupler.  'tis what I did to keep an 8568B safe from a 250W amplifier output.  I also used a 20dB attenuator at the SA input.  The 8568B has a internal 10dB attenuator by default and they make it difficult to disable, so in effect, I had 30dB attenuation on the input, as well as the 20dB coupling factor making a total of 50dB attenuation. 

Of course, you need a coupler rated for the power, but I was working at HF and an Elecraft CP1 did the trick.

For general protection during normal use, I recommend a 10dB or even 20dB attenuator connected directly to the SA input all the time – except for these rare cases, where the maximum sensitivity is required.

I don't usually bother for normal signals and rely on the built in 10dB, but I cannot recommend such a practice.  It doesn't do any harm to leave a 20dB attenuator on your SA input - you can always remove it if you are absolutely certain the input signal is 'safe'.

IMHO, the main harm of any of these things being constantly attached to a SA is potential physical damage from bending, falling, etc.

[Performa01]

Or use a directional coupler rated for the power source and feed the power source into a dummy load via the coupler.  'tis what I did to keep an 8568B safe from a 250W amplifier output.  I also used a 20dB attenuator at the SA input.  The 8568B has a internal 10dB attenuator by default and they make it difficult to disable, so in effect, I had 30dB attenuation on the input, as well as the 20dB coupling factor making a total of 50dB attenuation. 

Of course, you need a coupler rated for the power, but I was working at HF and an Elecraft CP1 did the trick.

I did not mention this on purpose…

Firstly, we still need a properly rated dummy load and this is not cheap either.

Secondly, no directional coupler will cover the frequency range 9kHz ~ 3.2GHz.
The Elecraft CP1 is only rated 1MHz ~ 30MHz, and even within its specified bandwidth I would not expect any directional coupler to have a completely flat frequency response.

Thirdly, even if the coupler was perfect, the dummy load will have a frequency dependant standing wave ratio, which causes amplitude errors in conjunction with the directional coupler.

As a conclusion, we have to deal with the frequency response of the entire system (dummy load + directional coupler) and frequency range is rather limited, yet both parts together might cost more than an appropriate 40dB power divider, which has one single accuracy specification of e.g. +/- 0.6dB from DC to 3GHz and handles 100W average, 10kW peak (PE7385-40).

Of course, for hobby use we often have to make do with less optimal (and considerably less costly) solutions, but especially a coupler like the Elecraft CP1 is a good example to demonstrate the tight limitations:
1.   You cannot use it for frequencies below 1MHz, whereas the resistive power divider goes down to DC.
2.   When characterizing a PA output, we want to measure at least up to the 5th harmonic. So the maximum TX frequency can only be 6MHz (or 10, if we make do with 3rd harmonic measurement).

[orin]

Or use a directional coupler rated for the power source and feed the power source into a dummy load via the coupler.  'tis what I did to keep an 8568B safe from a 250W amplifier output.  I also used a 20dB attenuator at the SA input.  The 8568B has a internal 10dB attenuator by default and they make it difficult to disable, so in effect, I had 30dB attenuation on the input, as well as the 20dB coupling factor making a total of 50dB attenuation. 

Of course, you need a coupler rated for the power, but I was working at HF and an Elecraft CP1 did the trick.

I did not mention this on purpose…

Firstly, we still need a properly rated dummy load and this is not cheap either.

Secondly, no directional coupler will cover the frequency range 9kHz ~ 3.2GHz.
The Elecraft CP1 is only rated 1MHz ~ 30MHz, and even within its specified bandwidth I would not expect any directional coupler to have a completely flat frequency response.

Thirdly, even if the coupler was perfect, the dummy load will have a frequency dependant standing wave ratio, which causes amplitude errors in conjunction with the directional coupler.

As a conclusion, we have to deal with the frequency response of the entire system (dummy load + directional coupler) and frequency range is rather limited, yet both parts together might cost more than an appropriate 40dB power divider, which has one single accuracy specification of e.g. +/- 0.6dB from DC to 3GHz and handles 100W average, 10kW peak (PE7385-40).

Of course, for hobby use we often have to make do with less optimal (and considerably less costly) solutions, but especially a coupler like the Elecraft CP1 is a good example to demonstrate the tight limitations:
1.   You cannot use it for frequencies below 1MHz, whereas the resistive power divider goes down to DC.
2.   When characterizing a PA output, we want to measure at least up to the 5th harmonic. So the maximum TX frequency can only be 6MHz (or 10, if we make do with 3rd harmonic measurement).

I did measure the CP1 on a VNWA, but unfortunately, I didn't keep any results other than a 1-30MHz plot (attached, somehow I managed to title it CT1 rather than CP1).  I'll measure it again if I get a chance.  I went to this method after a Bird 8325 30dB 500W attenuator decided to change its attenuation from 30.2dB to 38dB(!) after a few minutes at 200W and yes, there was oil in it.

Getting an attenuator that can take 250W continuous is going to be _expensive_!  Though I did eventually get another 8325 that appears to be unused for $280.  Even so, it's only spec'd to 500MHz.

Realistically, the amp was going to work into an HF antenna, the coupler is the only way I'd be able to measure anything once it's on the air.  Measuring into the dummy load was just a sanity check... and a check that said this amp wasn't going on the air yet.

FWIW, this setup revealed that a certain QRP rig would generate 200kHz AM modulation from it's SWR foldback with an SWR of only 1.5!  I only found that out when I used a frequency span of 1MHz around the carrier!

[simone.pignatti]

Hi trevwhite, you can sort your spectrum analyzer's accessories by your self, here are interesting kit to take as an example:

- RF ATTENUATOR SET
https://www.tekbox.net/images/documents/testequipment/tbas1/TBAS1_Manual.pdf

- SMA ATTENUATOR/TERMINATION SET
https://www.tekbox.net/images/documents/testequipment/tbas2/TBAS2_Manual.pdf

- COAXIAL ADAPTER SET
https://www.tekbox.net/images/documents/testequipment/tbcas1/TBCAS1_Manual.pdf

[Alex P]