ATTEN AT5005 Spectrum Analyzer?

[Mechatrommer]

googling around for spectrum analyzer, i found ATTEN AT5005 Spectrum Analyzer 150K - 500MHz USD430 in ebay http://cgi.ebay.com.my/ATTEN-AT5005-Spectrum-Analyzer-150-KHz-500-MHz-/270655175572?pt=LH_DefaultDomain_0&hash=item3f044e3b94. anyone with knowledge of this item? i need something for my 433MHz RF design/tuning project.

[alm]

No experience, but what's the point of a 500MHz specan for 433MHz? You can't even see the first overtone, so you might as well a frequency counter if you just want to determine the fundamental. The price does seem very low for spectrum analyzer, even used ones are usually much more expensive.

[Mechatrommer]

whats the overtone? any good reference? i want to maximize the rf transmitting power, so when making pcb for the 433mhz rf, i can see if my trace is too long, my inductors and capacitors value are not right etc, so i can make adjustment to the circuit to avoid to drift too much from 433mhz, is the atten 5005 is not enough for this? should i consider harmonics freq as well? or this overtone? sorry if this sounds too newbie, as this is my first attempt to design such a thing. so i'm looking for appropriate device/tester. from what i know/search, its the spectrum analyzer, and this atten, is something closest to affordability, even its still on the high side... for my budget.

[Mechatrommer]

i have another search, portable protek 3201 Sale $1,320.00 http://www.testequipmentdepot.com/protek/analyzers/3201.htm?gclid=CIWsjNqZrKYCFcV66wod0wJ8MA

[Mechatrommer]

the 1ghz version of atten is AT5011 US $797.88 http://cgi.ebay.com.my/ATTEN-AT5011-A-spectrum-analyzer-Tracking-Generator-1G-/300404563383?pt=BI_Oscilloscopes&hash=item45f181ddb7

[Floyo]

The trick with RF is that the amplifiers used are usually class C amplifiers. This means that the signal gets distorted rather badly. An advantage of class C amplification is that the efficiency is high.

http://en.wikipedia.org/wiki/Class_C_amplifier#Class_C

With distortion come harmonics, or overtones. These overtones are multiples of the fundamental freq. (433Mhz). The goal is to have the least amount of harmonics in the output signal so you do not interfere with communications in other bands. A rule of thumb is that your spectrum analyzer can display the 10th harmonic of the input signal in this case 4,33Ghz. You might get away with less but 2Ghz bandwith is probably the least you'll have to have.

http://en.wikipedia.org/wiki/Harmonic

[jahonen]

Those spectrum analyzers are ok if you just want to see if the thing you are working with transmits anything, in approximately right band. And of course, you can see the relative change of output power when tweaking the antenna etc.
But if you later want to do other more specific measurements (like spurious outputs, adjacent channel leakage etc.), I guess that you will be disappointed at at that point. I'm not even sure that you can even determine the carrier frequency accurately enough for the ISM-band radios, as they are relatively narrow-band.
What if you later want to debug those 2.4 GHz radios?

Regards,
Janne

[Zad]

If you are simply optimising the antenna, cable etc, then what you need is just a simple standard receiver module. On many of them you can pick up an AGC or similar received signal strength indication and output it directly to a multimeter. Sure if you are making amplifiers (and you shouldn't be, the whole point of this band is that it is standardised low power) then you need a spectrum analyser that goes into multi-GHz, but I can't see that it is useful beyond being a toy in this instance.

The spectrum analyser above will be based on a standard TV tuner front end from someone like Philips. Radio amateurs have been hacking this sort of stuff together for years, it doesn't cost much (financially anyway, time is another matter).

[Mechatrommer]

well, maybe i guess, designing a serious rf from scratch is not for hobbiest? the ghz SA cost tens of thousands!
this is my picture taken a while back. what i want to do is to transfer all the components (rf transmitter) on the blue pcb (left), into my own made pcb, so i can put it all in one nice pcb with my own mcu and those pots. at the "right below" is the rf receiver. my anticipated problem will be the transmitter efficiency (with my own pcb), i need a way to monitor this. as Zad's post, we can monitor it on the reciever, but, but.... how? the agc?

i've done some range efficiency test (the practical one haha), what i did is hook the receiver up the dso, to catch any incoming digital data from rf, the data distortion will tell somehow the range efficiency, and i.. with my transmitter will move away at some distance and press the transmit button, and run back to the dso and see the triggered digital distortion. kind of funny for me if i have to do like >100meters range test running back and forth.

[qno]

Checkout


http://www.hameg.com/h013_hms3000.0.html

Its is available well under 10k.

Try to get one with a tracking generator

[Mechatrommer]

Its is available well under 10k.
Try to get one with a tracking generator

<10K? you are kidding me. you are lucky i still can take that joke!

HMS3000 Hameg HMS3000 3 GHz Spectrum Analyzer  $4,944.00
HMS3010 Hameg HMS3010 3 GHz Spectrum Analyzer with Tracking Generator  $5,828.00

[joelby]

I don't think that a spectrum analyser won't really help you improve antenna and matching circuit designs - you'll only be able to compare the output level of different circuits, one at a time, or perhaps in real time by tweaking variable capacitors and inductors, but you could do that with a cheap RF field strength meter or fast oscilloscope. To actually analyse the performance of matching networks, which will tell you which components you need to adjust or add, you will want to put a bit of money aside for a vector network analyser.

[Mechatrommer]

thanx for the advices so far, will keep my brain working. well... i found this cheapy field strength meter, from the price, maybe just suitable for me? http://cgi.ebay.com.my/EMF-RF-Radiation-Field-Strength-Power-Meter-300M-3GHz-S-/370470455561?pt=LH_DefaultDomain_0&hash=item5641c28909 usd32.

[Mechatrommer]

off topic already!

[joelby]

thanx for the advices so far, will keep my brain working. well... i found this cheapy field strength meter, from the price, maybe just suitable for me?

The 5 dB resolution isn't too exciting!

[slburris]

well, maybe i guess, designing a serious rf from scratch is not for hobbiest? the ghz SA cost tens of thousands!

Try: http://www.scottyspectrumanalyzer.com/

Up to 3Ghz in various modes.  Join the Yahoo group, there are lots of
helpful people there.

Scott

[Mechatrommer]

Try: http://www.scottyspectrumanalyzer.com/
Scott

good luck to me, with my 433mhz circuit mod and ghz vna building. thanx for the link.

[Mechatrommer]

come to think of it. isnt say a 500MHz spectrum analyzer will show the harmonics as well on the screen maybe up to the tenth? i mean isnt it implicitly/automatically means that a 500MHz SA will also have a built in circuit/capable to analyze the upper harmonics and show on the screen? and the said BW is only for the carrier signal? i have this thought after playing with 100MHz rigol FFT and it showed up to GHz frequency?

[alm]

come to think of it. isnt say a 500MHz spectrum analyzer will show the harmonics as well on the screen maybe up to the tenth? i mean isnt it implicitly/automatically means that a 500MHz SA will also have a built in circuit/capable to analyze the upper harmonics and show on the screen? and the said BW is only for the carrier signal?

What makes you believe this, except wishful thinking? Why would they sell a 500MHz specan that is capable of 5GHz? Why do most commercial spectrum analyzers have a bandwidth of well beyond 1GHz and often even well beyond 10GHz, even though most of the carrier signals are below 3GHz? Even the HP 141T from the seventies goes up to 18GHz with the right RF unit, and their first spectrum analyzer in the sixties went up to 2GHz.

The frequency response would be very attenuated, which would make it useless for measuring the magnitude of the harmonics. I think the local oscillator is a fairly hard upper limit (just like sample rate), the IF won't pass any signal if f_sig >> f_LO. The low-pass filter, designed to filter images, will further attenuate the signal.

i have this thought after playing with 100MHz rigol FFT and it showed up to GHz frequency?

By definition, a 1GS/s sample rate can't reproduce frequencies beyond 500MHz, anything above will produce aliasing (which is why you need a proper low-pass filter), and show up as frequencies below 500MHz. If the FFT shows higher frequencies, the FFT implementation is buggy. There would also be severy attenuation at 1GHz, and I doubt your probing setup will pass much signal at 1GHz.

[Wim_L]

i have this thought after playing with 100MHz rigol FFT and it showed up to GHz frequency?

By definition, a 1GS/s sample rate can't reproduce frequencies beyond 500MHz, anything above will produce aliasing (which is why you need a proper low-pass filter), and show up as frequencies below 500MHz. If the FFT shows higher frequencies, the FFT implementation is buggy. There would also be severy attenuation at 1GHz, and I doubt your probing setup will pass much signal at 1GHz.

That's not really true. The FFT will show higher frequencies, they just won't be useful. The FFT of a sampled real signal of, say, 1024 points, will also be 1024 points long. However, the upper half of that range will consist of folded frequencies only, so if you're looking at a magnitude display, it will look like a mirror image of the lower half. There's no useful information in there. Most sensible instrument makers will leave that "information" out by showing only the first half of the FFT output, but strictly speaking it's not an error. Merely useless duplication.

[Mechatrommer]

ok, i'll follow nyquist now. even at 2xBW sampling doesnt make me feel safe enough. maybe those GHz on rigol is just definitely a crap. i was just hoping... there is "analog circuit" inside that can do up to GHz FFT stuff (bypassing the digital/sampling side).

[Wim_L]

Nyquist is for the absolute best case. If you are sampling a signal that has absolutely no frequency components at all above a frequency f, you can reconstruct it completely by sampling at tiny little bit faster than 2 times f (exactly two times is just a little bit too slow). Provided that your sampler has infinite accuracy and no noise at all, of course. Oh and you should acquire an infinite amount of samples too...

For real measurements, you have to oversample a bit more