Connecting VCOs for generating RF signals 2

[mrkenneth]

I have been looking for something that can generate a RF signal from around 1 GHz to 2.5 GHz, and found that Maxim makes voltage-controlled oscillators that can be used to generate a small spectrum of frequencies. (e.g. MAX2622 and MAX2750)
Can I just connect these directly to some antennas for testing without any amplifiers or buffers? I only need around -20 dBm to -10 dBm at 50-ohm, and the datasheets says that the power output is -5 dBm.


Now, I would like to connect three of these VCOs together on one circuit, so that I can select from three preset frequencies. Would connecting all of the outputs together, and powering one of the ICs be okay? Thus at any given moment, only one IC will have a voltage present at the Vcc pin. Like this:

Would the RF output from the VCO that is powered damage the other two VCOs that are off? Conversely, will the other two VCOs that are not powered on affect the operation of the one VCO that is on?


Thanks in advance!


P.S. are VCOs the simplest solution to generating a simple RF signal?

http://www.mrkenneth.com

 

[itsmoked]

I wouldn't do it. The output of the active unit will drive the unpowered inputs causing you headaches.

If you are just making *one* just use three antennas.

If you need other methods look for a high speed crosspoint switch or something. High frew relays.

 

[mrkenneth]

Thank you for the reply.

I can only use the same antenna, at different frequencies, to test the antenna's frequency response, among other things.

What if the switch is at the output of the VCOs? The only SP3T switches I can find are through-hole, so I am not sure whether the RF frequencies will like it...

Are there switches suitable for passing RF?

Each VCO by itself would be fine, I assume?

http://www.mrkenneth.com

 

[itsmoked]

http://rocky.digikey.com/WebLib/C&K Components/Web Data/CRD Series.pdf

Is a rotary SMT hex switch. Digikey has a bunch of diff types.

You can pick three unique settings ignore the rest.

"Are there switches suitable for passing RF?"
I can't say. If I was doing the tests you are I would still turn off the neighboring oscs.

 

[nbucska]

If U R using VCO, why not make a PLL with a reference frequency -- i.e. a fr. synthetizer and use only one VCO?

<nbucska@pcperipherals DOT com> subj: eng-tips
read FAQ240-1032

 

[mrkenneth]

Itsmoked, would the SMT switches you posted be okay for passing the high frequencies?

The reason why I would rather not use PLLs is because of the added complexity. I will only need to make one or two of these circuits, and will be soldering all of these SMDs by hand. So leadless packages are not possible, and the fewer pins, the better.

I just looked at the datasheet again (

http://pdfserv.maxim-ic.com/en/ds/MAX2622-MAX2624.pdf

and

http://pdfserv.maxim-ic.com/en/ds/MAX2750-MAX2752.pdf

), and it says that it has an output buffer:

Would it be enough to connect a resistor in series with the RF output before joining them together?

Sorry, I am new to the RF field.

---------------------------------------------------------
Operation Radiation:

http://www.mrkenneth.com

 

[itsmoked]

I would think not mrkenneth. Running these frequencies requires matching impedances carefully!! Just adding resistors is not a good idea.

If I, (knowing the little of your application u have mentioned), were to do this here's what I would do.

This thing CANNOT be jammed together on some perfboard you found laying around.

I would design a simple board with all the connectors required to hook the unit up easily and lots of surface ground plane that surrounds the entire circuit. letting out only the power and the output. SEE ## BELOW The board houses offer prototype deals that are pretty reasonable but have recently gone from about a three board limit to about 10 boards. Go with one of these deals and get ten boards. Build your (what 3 systems?) out of three separate boards each with no failures due to sketchy paralleling. Then just switch boards to switch freqs.



## This is so you can make a small box out of brass shim stock, (buy at OSH), that would cover the entire circuit except for the ground and trace carrying out to the antenna connector. And the ground and *choked* +V to it's connector. You then solder the brass box to the board ground covering the circuit.

 

[mrkenneth]

I will be ordering double-sided FR4 boards for this along with some other circuits. The ICs are too small for anything homemade. The local place I am getting the boards fabricated charges $1.00 USD per square inch with a $50.00 minimum order. Is that a fair price?

I guess I should just make three different boards and connect the antenna to the right board when required. I just thought of connecting them together after I finished designing the one-VCO board. Combining them would allow me to switch frequencies more easily, and save me the expense of two SMA connectors. (Are there any places that sell female PCB SMA connectors for less than what Digikey charges?)

Do you mean the small box as a Faraday cage? Because the circuit I have designed is under one square inch, I was thinking that I can just stack a one-sided copper-clad board on the component side and "seal" the two boards with some copper tape. Is that alright?

Thanks in advance!

---------------------------------------------------------
Operation Radiation:

http://www.mrkenneth.com

 

[itsmoked]

Yes that price sounds fine.

Gesh, I get a paperback catalog every month form somewhere that I promptly trash... I don't remember the name!!! But all they sell is *JUST* RF connectors at nice prices. Someone here might know.. It's like newsprint but about 4X5 inches and an inch thick. Anybody??

Yes a Faraday cage. The copper tape will prolly work. Tho you may need to put a spot o' solder on an edge if the adhesive insulates.

 

[itsmoked]

What are you doing this for??

 

[mrkenneth]

Does the company to which you refer ship to Canada at reasonable rates? I am looking for alternate suppliers to Digikey, and of course, eBay.

I believe the adhesive is conductive, but I will check again to make sure.

This is for a project to investigate the effects of the ellipticity ratio of a circular disc antennas.

---------------------------------------------------------
Operation Radiation:

http://www.mrkenneth.com

 

[itsmoked]

Interesting investigation. Good luck with it.

As for the company... I still can't remember the name so I can't comment on it's shipping abilities.

If you google ' "rf connectors" ' you will get over 100,000 hits and 40 sponsored links. This is/means companies of all types home run, thru Amphenol. You should have little trouble finding what you need shippable to Inuvik, Northwest Territories! (As long as the Mackenzie river is still frozen.)

 

[sdmays]

'itsmoked', are you referring to "Pasternack Enterprises"?

 

[VEBill]

"...referring to 'Pasternack Enterprises'?"

Can't be, he wrote "...at nice prices." Unless US$24+ for a one-foot long BNC to BNC cable is considered 'nice' these days.

 

[itsmoked]

Hot dang! You got it sdmays!! Thanks for the help there.

Oh ugh... you're correct VE1BLL that price sucks! Well the last thing I looked at was deemed resonably priced must have been 10 years ago... (In my defense I did say I throw them all in the trash, I haven't opened one in years.) ;0

In their defense... There probably isn't an RF connector made that they don't catalog.

 

[mrkenneth]

The prices do look quite high, some even higher than Digikey!

I have also looked through a few results from Google, and the prices do not look too appealing either...

I guess back to eBay I go!

Thanks everybody!

---------------------------------------------------------
Operation Radiation:

http://www.mrkenneth.com

 

[jsolar]

You have wire OR'ed an RF connection. At your frequencies you need to have impedence control and matching. The vendor does not specify the output Z when not powered, so you do not have the information to make a decision. Past work has shown that non powered ckt.'s can still look like diods. Also, your wire OR is a digital design, not an RF design. You would need a couopler or a switch (as mentioned before). Also,you donot show the rest of your PLL ckt. Do you intend to run the power supply noise onto the extremely sensitive voltage control port? If you can live with the open loop error and drift, you should still at least have precision signal sources and isolation between parts. It would be much simpler if you did your switching at DC and used only one VCO. You can switch a voltage source into the pat much easier than you can mess with the RF output. Analog switches are very simple and cheap and easy to use.

 

[sdmays]

MrKenneth, ItSmoked mentioned impedance matching on June 19th. I haven't seen any posts since then regarding impedance matching, but Motorola (Freescale,...whatever) has an App Note in which you might be interested: AN267, "Matching Network Designs with Computer Solutions". For four different impedance matching networks, the necessary reactances are calculated per network element to match a given output amplifier to a given 50 ohm load. You might or might not find that helpful, but I have wearing my copy of the App Note out...

 

[itsmoked]

That IS a nice article sdmays.