Cheap RF signal generator any good?

[Merlysys]

Looking for a simple RF signal generator to test antennas in the 2.4-2.5 Ghz.
I came across this cheap unit. It doesn't have the buttons and displays of the typical unit as it is USB controlled. I don't need the accuracy, just the ability to create an RF signal in the desired range (yes I will keep power output to lowest)

If this unit not a good buy what basic unit would you recommend?

Specs
Output frequency range 138MHz-4.4GHz,
Stepping: 1kHz
Output power (for reference): 150M-3.2DBM; 250M-3.4DBM, 500M,-3.4DBM; 750M-2.8DBM, 1000M-3.4DBM;
1500M-3.6DBM; 2000M-4.0DBM; 2500M-2.8DBM; 3000M-1.8DBM; 3500M-3.6DBM
Spectrum of the input frequency range of 138M-4.4GHz
Frequency Accuracy (after calibration): +-1k @ 1GHz
Low-noise input signal amplitude:-70DBM @ 0.5G

http://www.ebay.ca/itm/138MHz-4-4GHz-USB-SMA-Source-Signal-Generator-Simple-Spectrum-Analyzer-138M-4-4G-/351805484236?hash=item51e93daccc:g:4D0AAOSwU-pXqX7W

[tggzzz]

Harmonics can be an issue.

Cheap RF sig gens often have a PLL/VCO which is divided down digitally to get the lower frequencies. The consequence is that the high frequencies are a "reasonable" approximation to a sine wave, but lower ones are a square wave. Whether that matters depends on what you are trying to measure, how you are trying to measure it, what is in the UUT, and what extra filtering you could add to the testbench.

I suspect, but have not checked, the unit you refer to has a PLL/VCO operating 2.2-4.4GHz.

[uncle_bob]

Looking for a simple RF signal generator to test antennas in the 2.4-2.5 Ghz.
I came across this cheap unit. It doesn't have the buttons and displays of the typical unit as it is USB controlled. I don't need the accuracy, just the ability to create an RF signal in the desired range (yes I will keep power output to lowest)

If this unit not a good buy what basic unit would you recommend?

Specs
Output frequency range 138MHz-4.4GHz,
Stepping: 1kHz
Output power (for reference): 150M-3.2DBM; 250M-3.4DBM, 500M,-3.4DBM; 750M-2.8DBM, 1000M-3.4DBM;
1500M-3.6DBM; 2000M-4.0DBM; 2500M-2.8DBM; 3000M-1.8DBM; 3500M-3.6DBM
Spectrum of the input frequency range of 138M-4.4GHz
Frequency Accuracy (after calibration): +-1k @ 1GHz
Low-noise input signal amplitude:-70DBM @ 0.5G

http://www.ebay.ca/itm/138MHz-4-4GHz-USB-SMA-Source-Signal-Generator-Simple-Spectrum-Analyzer-138M-4-4G-/351805484236?hash=item51e93daccc:g:4D0AAOSwU-pXqX7W

Hi

Test antennas how?

A network analyzer (which this is not) is a much more common thing to test antennas with.

Bob

[Kalvin]

One can test antennas with a signal generator and a simple SWR bridge. Something like these:

http://pe2er.nl/wifiswr/
http://www.w1ghz.org/QEX/A_UHF+_VSWR_Bridge.pdf

[uncle_bob]

One can test antennas with a signal generator and a simple SWR bridge. Something like these:

http://pe2er.nl/wifiswr/
http://www.w1ghz.org/QEX/A_UHF+_VSWR_Bridge.pdf

Hi

Neither one of those are anything you are likely to have just sitting around. They also aren't all that easy to put together unless you have microwave PCB material running around. Once you get them "close" getting them reasonably well calibrated is an interesting process. The point being that the SWR (or return loss) bridge tacked on to a signal generator like the one in the original post can easily cost more than the generator he's looking at.

Bob

[Kalvin]

The FR4 is usable up to few GHz.

[uncle_bob]

The FR4 is usable up to few GHz.

Hi

There is a difference between "usable" and "suitable". You might get away with FR4 if you really know what you have (all the characteristics of that batch). For making a piece of test gear, a proper microwave laminate is a much better choice. Both of the lined devices in the previous post are done on some sort of microwave material for that very reason.

Again - this is not to say you *can't* do it. It will likely take a couple of passes and some fairly good analysis. You also will be counting on the board fab to not switch things around while you are doing your experiments.

Bob

[Kalvin]

The FR4 is usable up to few GHz.

Hi

There is a difference between "usable" and "suitable". You might get away with FR4 if you really know what you have (all the characteristics of that batch). For making a piece of test gear, a proper microwave laminate is a much better choice. Both of the lined devices in the previous post are done on some sort of microwave material for that very reason.

Again - this is not to say you *can't* do it. It will likely take a couple of passes and some fairly good analysis. You also will be counting on the board fab to not switch things around while you are doing your experiments.

Bob

Yes, I agree on that of course. But if one doesn't have money to buy decent gear, one typically needs to go DIY unless one can borrow the stuff.

[Kalvin]

Thanks for the idea. Just placed an order for a similar device with the output frequency range 35MHz-4.4GHz. I am not sure what to expect, though

[bitseeker]

There are several version of these on AliExpress. Some with more passive components and one with a touch-screen UI. Looking forward to seeing what yours ends up like.

[Bicurico]

Hi.
The cheap SKA spectrum analyzer will only use input or output connector at a time.  Not both simultaneously!
I explain this in my blog.  See web link in my profile.
Regards

[Kalvin]

Hi.
The cheap SKA spectrum analyzer will only use input or output connector at a time.  Not both simultaneously!
I explain this in my blog.  See web link in my profile.
Regards

Thanks! I wasn't 100% sure what I was paying for. But in the spirit of exploring new gadgets I decided to go for this one. I wasn't 100% sure whether this can transmit and receive at the same time, but thanks for your nice review, it is now clear that the device can either transmit OR receive - not both at the same time.

For antenna testing simultaneous transmitting and receiving is not a must as one can use a VSWR bridge to check antenna tuning which requires only the transmitter to work.

It seems that if one wants to use this device as a spectrum analyzer, one can use this device as a receiver but one needs to have an external signal generator. That's why some sellers are offering those wideband noise generators with this gadgets which can be used as poorman's signal source.

I do already have a signal generator "35 M ~ 4.4 GHz signal source ADF4351", so in principle I could use this as a tracking signal generator with the signal source/receiver I just bought. In order to get a decent frequency performance I need to modify those two designs so that they will use the same good quality, oven compensated external oscillator for the timebase at least.

The output signal from the signal source/receiver and the ADF4351 contains quite a lot of harmonics. This may or may not be a problem. The ADF4351 datasheet http://www.analog.com/media/en/technical-documentation/data-sheets/ADF4351.pdf states that the VCO fundamental frequency is from 2200 MHz to 4400 MHz. I am not sure, but it might imply that at that frequency range the signal quality is decent without too much of harmonics. At the lower frequencies which are generated by a divisors the signal will contain more harmonics. For better signal output quality one can use filtering or use some other signal generator/architecture.

[hendorog]

One can test antennas with a signal generator and a simple SWR bridge. Something like these:

http://pe2er.nl/wifiswr/
http://www.w1ghz.org/QEX/A_UHF+_VSWR_Bridge.pdf

Hi

Neither one of those are anything you are likely to have just sitting around. They also aren't all that easy to put together unless you have microwave PCB material running around. Once you get them "close" getting them reasonably well calibrated is an interesting process. The point being that the SWR (or return loss) bridge tacked on to a signal generator like the one in the original post can easily cost more than the generator he's looking at.

Bob

Looks like one of them is just hand cut epoxy, the other mentions teflon. The epoxy one doesn't seem to worry too much about using 50ohm microstrip either. I assume the idea is to build as small as possible to minimise effects of the substrate. Given they are using 0805 components and large connectors, it wouldn't be hard to make it even smaller.

I expect these high frequency resistors would help too:
http://www.vishay.com/docs/60093/fcseries.pdf

Another cheap(ish) option is to get a directional coupler from eBay for the frequency of interest.

[uncle_bob]

The FR4 is usable up to few GHz.

Hi

There is a difference between "usable" and "suitable". You might get away with FR4 if you really know what you have (all the characteristics of that batch). For making a piece of test gear, a proper microwave laminate is a much better choice. Both of the lined devices in the previous post are done on some sort of microwave material for that very reason.

Again - this is not to say you *can't* do it. It will likely take a couple of passes and some fairly good analysis. You also will be counting on the board fab to not switch things around while you are doing your experiments.

Bob

Yes, I agree on that of course. But if one doesn't have money to buy decent gear, one typically needs to go DIY unless one can borrow the stuff.

Hi

Or one shops on eBay and pays next to nothing for pre-built directional couplers .... which will not work with this gizmo very well.

I'm suggesting that one look at the whole cost of getting this or that up and running rather than buying the first piece and going ... what next?

Bob