Measuring with the RF Explorer

[Dannyboy]

Hi, complete RF beginner here. I am using the RF Explorer http://micro.arocholl.com/ to check the power level on these 433Mhz radios: http://www.nordfield.com/wireless-serial-adapter

The datasheet for the radios say that the radios has a power level of 50mW but when I measure the signal with the RF Explorer I can only measure around -48dbm which equals to only 0.000016mW.
I am using the 6db RF Power Limiter and have compensated for this in the RF Explorer.

Is it normal that I cannot measure the exact power level or is my RF Explorer broken?

[ElectroIrradiator]

Is it normal that I cannot measure the exact power level or is my RF Explorer broken?

You may have fried it. According to the specifications the absolute maximum input power is +5 dBm, and you used a 6 dB attenuator. That puts the limit at 11 dBm, which is barely more than 10 mW. So the input may be toast, as you poured 50mW into it.

'Absolute maximum' specifications frequently mean 'you don' really want to do this (but the widget may survive)'. Going four times over ought to give grief.

[Dannyboy]

Damn it..I was afraid of that. What kind of power limiter (is it called an attenuator?) do I need to use if I want to measure 50mW? and how about if I want to measure 500mW?

Is it worth getting the RF Explorer repaired or is it cheaper just to buy a new (at $129)?

[ElectroIrradiator]

No idea if it is worth trying to repair it, but the first question I can help with.

First, dBm is power measured on a dB scale, compared to one mW.

So 1mW is 10*log10(1/1) = 0 dBm.
50 mW is 10 * log10(50/1) = 17 dBm.
500 mW is then 27 dBm.
5 mW, the limit, is 7 dBm.

If you want to play it reasonably safe, then I wouldn't put more than 1 mW into it, 0 dBm. That makes it easy to calculate how many dB of attenuation you need at the least, 17 dB for 50 mW, 27 for 500mW. 20 dB and 30 dB attenuators are standard, and gives a nice margin of safety of a further 3 dB.

Note that many RF coaxial attenuators, particularly the small ones with SMA connectors, have strict power limits. Some may be damaged by as little as your 500 mW. However there are specialist types, which can take quite a lot of power. Additionally high power models may be unidirectional in the sense, that you must connect the input power to a particular connector.

[Fraser]

A golden rule of direct connection RF power measurement is to always insert adequate RF attenuation to protect the test equipment. As has been stated, it looks like you fried the Explorer's front end. It is a dedicated commercial radio module that will need to be replaced. Put it down to experience and be glad that it was not a US$30K Rohde & Schwarz analyser that you fried !  They have a $5K fixed repair fee    A lesson well learned I feel sure.

Ariel is a great guy and I feel certain that he will assist you in repairing your RF Explorer. The damaged components are very cheap to replace.
Ariel also sells a limiter unit that contains a 6dB attenuator and will limit a +30dBm input signal to just +4dBm, so protecting the Explorer . Its $45 but worth every penny and may be used to protect other RF equipment as well.

http://www.seeedstudio.com/depot/power-limiter-p-1444.html?cPath=63_64

Email Ariel via the RF Explorer web site and ask what the repair options are. You can also just buy one of the RF expansion modules that is sold on seeed studio for the Explorer. The blown front end is then obsolete and you use the expansion board instead. It may be an easier and more economic solution. You will NOT have damaged anything beyond the little Rf front end module so expansion boards will still work fine.

I have two RF Explorers. One is a 2.4G model plus RFEMWSUB1G expansion module, and the other is a WSUB1G model plus WSUB3G exoansion module 

The RF Explorer is a very versatile little unit with very easy frequency coverage expansion.

http://www.seeedstudio.com/depot/rfemwsub1g-module-p-909.html?cPath=63_64

Or upgrade to the top model by adding the 3G expansion board. This module already contains integral protection up to +30dBm

http://www.seeedstudio.com/depot/rfem-wsub3g-module-p-1264.html?cPath=63_64


Fraser

[Dannyboy]

Thanks a lot everybody, that is very good information for a beginner like me.

Im just confused about one thing now. Aurora writes above that:

Ariel also sells a limiter unit that contains a 6dB attenuator and will limit a +30dBm input signal to just +4dBm, so protecting the Explorer . Its $45 but worth every penny and may be used to protect other RF equipment as well.

This is the actual power limiter I used and still my RF Explorer got fried, even it says that it should be protected for up to +30dbm (1W). So I guess my question is: What is the difference between a power limiter and an attenuator? It appears as this 6db attenuator is not sufficient to protect the RF Explorer when meassuring a 50mW signal. As ElectroIrradiator mentions above I need a 25db attenuator. So why does it say that the Power limiter will protect the RF Explorer for up to +30dbm?

[Fraser]

Ahhh all is clear !

A Power limiter is in fact a pair of very fast acting limiter diodes. Once the RF level applied to them exceeds the threshold point, they conduct and present the input RF source with a resistance of around 0.2 Ohms. They are very different to attenuators as they are a protection device that heavily clamps the signal rather than attenuated it. The diodes can cope with 1 Watt continuous RF and they dissipate most of it as heat.

The low readings you were seeing are due to that fact that the RF Limiter was activated and its effectively clamped you signal level.

The good news is that you RF Explorer and RF limiter should have survived this event. Well done for using the limiter, its saved your RF Explorer !

To measure higher powers than the RF Explorer can cope with, you should add attenuators of the correct values to the input of the limiter. The limiter will add 6dB attenuation as you know but this can be compensated for in the RF Explorer menu to achieve correct dBm readings.

I hope this makes sense. If not, do a Google on Agilent RF limiters and you should find an excellent guide to their operation.

I have added the application note for you to save you searching.
Fraser

[ElectroIrradiator]

Wouldn't the limiter clamp the peak voltage to roughly the forward voltage drop of the diodes? If they are high speed Schottky diodes, then the voltage would be about +/- 0.4V or so. That would present itself as a +2 dBm signal to the SA, which it ought to be able to register if it is operational?

[Fraser]

These are not your average diodes, they are designed to be RF limiters. Please read the Explanation by HP in the previous post that I have made. The diodes present a very low RESISTANCE to the 50 Ohm feed carrying the incoming signal. These do not work like an AC clipper that clips to 0.6V.

[WarSim]

Wouldn't the limiter clamp the peak voltage to roughly the forward voltage drop of the diodes? If they are high speed Schottky diodes, then the voltage would be about +/- 0.4V or so. That would present itself as a +2 dBm signal to the SA, which it ought to be able to register if it is operational?

Schottky diodes would not be fast enough.  The diode would be some thing much faster like a pin diode or even a bare junction.  And when you activate a limiter(a protection device) don't expect a calibrated output.  You would have allot more loss in a descaler (or rescaler).
The sensing section is not completely voltage dependant.  When you give an SA a non sine there is some error involved. 

[G0HZU]

If the analyser is based on some kind of wireless chipset (eg IQ demodulator)  then I'd be surprised if the real damage level was as low as +5dBm. Also, if you were using the limiter/pad then you should be OK in terms of damage protection from a 50mW signal because the limiter will protect the analyser.

However, you can't feed +17dBm (50mW) into the limiter/pad and expect to make a useful measurement because the signal will get limited plus you need to keep the RF level into the analyser a bit below 0dBm to keep it in the linear region.

So you need to add an extra attenuator, eg 20dB.

Note:
With your current setup, the +17dBm going into the attenuator/limiter would probably still produce >0dBm at the output of the limiter so should give a huge signal on the analyser if it is undamaged. Therefore, are you sure you have the analyser set to the same frequency as the transmitter? Maybe you are seeing a spurious term on the analyser rather than the main signal.

Note that the little analyser probably isn't the right tool to use for measuring RF power with any accuracy because the spec sheet says it has accuracy of +/- 3dB.
At 433MHz and 50mW power level you could make a simple RF detector that would be much more accurate and it would only cost about $1 to make.

I guess it's unfair to be too critical of a spectrum analyser that only costs $129 but I'd imagine that that analyser will also have a very poor instantaneous dynamic range. Probably less than 60dB on a very basic test and also there will be limitations wrt image rejection. There's a distinct lack of performance related data in the specifications so it really is aimed at the hobbyist who probably won't ever understand or care about the limitations of the device.

[ElectroIrradiator]

Schottky diodes are not fast enough?!

The review article specifically states that the PIN diode clamps takes several nano-seconds to activate. "Low Cost Surface Mount Power Limiters", page 2.

I'm heading for bed. as I am clearly missing something obvious here. From the two articles provided, it doesn't immediately seem clear to me that PIN diode RF clamps can work the way the pair of you seem to think they do.

There will be RF coming out, even when the diode(s) have a low ON resistance, and that RF output should be noticeable. Additionally, both the low cost limiter described in the article just mentioned, and the official Agilent made units, doesn't do much of anything until the power reaches 15-20 dBm. That is their limiter level, it doesn't magically drop back down to zero, once the limiter activates. Figure 25 and 26, page 10 in the "Low Cost" article, and figure 6 and 7, page 7 in the Agilent datasheet.

Even if different diodes are used in the unit we consider here, giving a lower clamping level, then the output would still not be reduced to zero, once the clams activates.

[G0HZU]

Schottky diodes are not fast enough?!

The review article specifically states that the PIN diode clamps takes several nano-seconds to activate. "Low Cost Surface Mount Power Limiters", page 2.

I'm heading for bed. as I am clearly missing something obvious here. From the two articles provided, it doesn't immediately seem clear to me that PIN diode RF clamps can work the way the pair of you seem to think they do.

There will be RF coming out, even when the diode(s) have a low ON resistance, and that RF output should be noticeable. Additionally, both the low cost limiter described in the article just mentioned, and the official Agilent made units, doesn't do much of anything until the power reaches 15-20 dBm. That is their limiter level, it doesn't magically drop back down to zero, once the limiter activates. Figure 25 and 26, page 10 in the "Low Cost" article, and figure 6 and 7, page 7 in the Agilent datasheet.

Even if different diodes are used in the unit we consider here, giving a lower clamping level, then the output would still not be reduced to zero, once the clams activates.

Yes, you can use either PIN diodes or Schottky diodes for limiter design. Also you are correct in that there will be a leakage level from the limiter and this will typically be well over +10dBm if driven hard. So a relatively large signal will still exit the limiter as you suggested.

I suspect that the RF Explorer limiter will have some attenuation nested within the limiter in order for it to keep the leakage below +5dBm at the output and that is why it is marketed as a 6dB attenuator + limiter.

FWIW I've designed various limiters over the years and I usually use PIN diodes that are designed for limiting. I tend to not use Schottky diodes because I don't think they have the same leakage vs power handling performance but then I may be out of date on this. I usually design for 5W to 10W power handling. But there are plenty of 1W rated Schottky based power limiters available

[Fraser]

Another useful datasheet on RF Limiters is attached.

The RF Explorer Limiter contains PIN Limiter Diodes designed to perform well at low power levels. The maximum leakage past them is around 10dBm. There is a capacitor preceeding the limiter diodes and a 6dB attenuator following them. The maximum output level when 30dBm is applied to the input is +4dBm.

It is correct to say that the RF Explorer should have read around +4dBm. The question that should be asked is ...does the RF explorer respond to background RF and show signals on its screen (without limiter attached) or is it deaf ? If it shows plenty of life when viewing the background it is unlikely that the receiver front end is fried. You should then consider the following possibilities:

1. Your test cable is faulty
2. The Limiter is faulty.
3. The Transmitter is not producing the specified RF output (faulty or RF set to minimum ?)
4. The transmitter is protecting its RF output stage when the limiter activates. The limiter causes a large reflection of RF power towards the source.
5. The GFSK modulation is confusing the RF Explorer ?
6. Have you access to a signal source of less than +5dBm that you can safely attach the RF Explorer to and check its input ? Even a car RF keyfob transmission may be good enough to assess the RF Explorer sensitivity, or not, as the case may be. Car RF Keyfobs tend to radiate around 100uW.
7. At least one RF explorer has suffered deafness due to a faulty input SMD capacitor (cracked ?)

Fraser

[WarSim]

It is correct to say that the RF Explorer should have read around +4dBm. The question that should be asked is ...does the RF explorer respond to background RF and show signals on its screen (without limiter attached) or is it deaf ? If it shows plenty of life when viewing the background it is unlikely that the receiver front end is fried. You should then consider the following possibilities:

Yes if you exactly match their test frequency and power.  The -6db spec is not a simple spec, the actual spec is a chart spanning the entire intended frequency band.  Also it is accepted that anytime a safety device like this is used the non protecting calibration should be recharged.  Also the protecting rating is never considered a calibrated spec.  The values given are to inform you how the protection device will react when subject to excessive power.  It is like asking if you can use the spec on a MOV as a regulator
The limiter is a protection device.  If you exceed it's limits it should sacrifice it's self for the sake of the instrument and its damage could be small or catastrophic.

You have to evaluate the 'background' based on the instruments spec.  If you have the input correctly terminated and you floor is an order of magnitude too low or too high, expect input damaged.  If you want to very roughly check a cable check your flor un terminated and terminated.  If you are not in a cage you should a similar response.  I

It is also likely you have not accurately spec'd your cable, simple specs have the greatest variance of all.  If the cable is a calibrated cable it should have a chart just like any other RF device. (splitter, detector, attenuator, limiter, recycler, coupler etc.)   If it isn't you will need to derive the specs required by using known sources. 

[Dannyboy]

Thanks a lot for all the information, now everything makes better sense. I will get a 25db attenuator and take som new measurements.
I have attached a screenshot of what the measurement currently looks like for a 915Mhz radio and a 500mW radio that Im currently testing. It will be exiting to see if I can meassure the correct power level once I put the attenuator on.

[Fraser]

For info the receiver used in the RF Explorer Sub1G is the Si4432.

http://code.google.com/p/rfexplorer/wiki/RF_Explorer_schematics

Here is the link to the datasheets etc.

http://www.silabs.com/products/wireless/EZRadioPRO/Pages/si44303132.aspx

http://www.silabs.com/products/wireless/EZRadioPRO/Pages/si44303132.aspx

[KJDS]

Once a PIN diode starts to conduct then it will continue to conduct for quite some time, known as the minority carrier lifetime. When it is conducting then it has a low resistance, probably only about 2 ohms, which is a significant shunt in a 50 ohm circuit, however with the reactive element of the PIN also at play then unless it is tuned to a specific frequency then a single PIN diode will only provide about 20dB of attenuation so a very large signal can still damage something protected by a limiter, though that's unlikely to happen in this case.

When a limiter is conducting then almost all of the power from the transmitter is reflected back and will damage a lot of transmitters.

I'd recommend that a suitable attenuator is put on the output of any transmitter being tested. Should anyone want one, then I do have some 100W 30dB attenuators for sale, PM me for details.

[G0HZU]

For info the receiver used in the RF Explorer Sub1G is the Si4432.

http://code.google.com/p/rfexplorer/wiki/RF_Explorer_schematics

Here is the link to the datasheets etc.

Thanks.

It looks like all of the RF and DSP is contained in the 4432

It also looks like it uses the 8 bit RSSI as the detector output for the analyser because this agrees with the 0.5dB resolution capability.
This is a pretty neat chip but there are lots of clues in the 4432 datasheet as to why this analyser will have poor RF performance compared to a regular analyser.

However, because the price of the analyser is so remarkably low and it is so tiny, it becomes really hard to criticise it in terms of performance vs cost vs size

You can often buy small RF attenuators on ebay (eg 2W rated with SMA connectors) so if you only want to measure 0.5W then maybe this is a cheap option. However, secondhand low power attenuators are always a risky purchase unless the vendor says they are fully tested. Many become erratic if they have led a hard life and they can still look shiny and 'new' on the outside.