Gentlemen:

I have an application where I just need a 920 MHz fixed sine wave oscillator output. A square wave would be acceptable also, since I could just filter it into a sine wave. I have used many oscillators in the past, but never at this high a frequency. No one seems to carry crystals that high, though they do have oscillators. The oscillators they have are all differential-pair output. I'm used to oscillators that you just apply voltage to and you get a wave out.

I tried one of these oscillators with an LVDS output. The only LVDS receivers I can find at that frequency are monstrous ones like a SN65LVDS314. I tried a high-speed comparator like an AD8561, but I get no output at all.

I also tried a high-frequency VCO and set the "tune" input with a voltage divider, which apparently doesn't work either. I even found a sample circuit on the Maxim website for the exact chip and it was nearly identical to what I had:



But when I connect it to 3V, nothing happens. The trace on the scope doesn't even flinch...totally dead. I tried two different MAX2623 chips and neither worked. I double-checked all my wiring. I must be doing something stupid with all this.

Isn't there a simple way to get a usable 920 MHz sine wave?

Don

 

No. There is no simple way to get anything to beat up and down that fast. The output is AC coupled and the impedance is 50 Ω. Are you using a 50 Ω 1:1 probe. The data sheet doesn't say what the output levels are but I expect them to be pretty small. It does say that output power is -3dBm, so it should be visible with an appropriate probe and a scope that can handle the frequency. I'm not sure why but it says the tune input must be from a low impedance source, which a resistor divider .. is not. Another question -- how did you fabricate this circuit?

 

Gentlemen:

I have an application where I just need a 920 MHz fixed sine wave oscillator output. A square wave would be acceptable also, since I could just filter it into a sine wave. I have used many oscillators in the past, but never at this high a frequency. No one seems to carry crystals that high, though they do have oscillators. The oscillators they have are all differential-pair output. I'm used to oscillators that you just apply voltage to and you get a wave out.

I tried one of these oscillators with an LVDS output. The only LVDS receivers I can find at that frequency are monstrous ones like a SN65LVDS314. I tried a high-speed comparator like an AD8561, but I get no output at all.

I also tried a high-frequency VCO and set the "tune" input with a voltage divider, which apparently doesn't work either. I even found a sample circuit on the Maxim website for the exact chip and it was nearly identical to what I had:

View attachment 207035

But when I connect it to 3V, nothing happens. The trace on the scope doesn't even flinch...totally dead. I tried two different MAX2623 chips and neither worked. I double-checked all my wiring. I must be doing something stupid with all this.

Isn't there a simple way to get a usable 920 MHz sine wave?

Don

Hi Don,

Why not use a 61.3333 MHz overtone crystal oscillator, quintuple it to 306.6667 MHz and then triple that to 920 MHz using a stage of amplification after each multiplication?

- Nandu

 

It should be possible to start with a 61.3333 MHz overtone crystal oscillator, quintuple it to 306.6667 MHz and then triple that to 920 MHz using a stage of amplification after each multiplication.

- Nandu

I didn't think anybody had come up with a 900 MHz fundamental crystal since the last time looked about a decade or so ago.

 

No. There is no simple way to get anything to beat up and down that fast. The output is AC coupled and the impedance is 50 Ω. Are you using a 50 Ω 1:1 probe. The data sheet doesn't say what the output levels are but I expect them to be pretty small. It does say that output power is -3dBm, so it should be visible with an appropriate probe and a scope that can handle the frequency. I'm not sure why but it says the tune input must be from a low impedance source, which a resistor divider .. is not. Another question -- how did you fabricate this circuit?

Here is my circuit:



My scope is an older Tektronix 2440 digital at 500 Mbps. I thought it might show up as a blur like on my old analog, but maybe the signal is outstripping the digital entirely.

They show an antenna connected directly to the output, so it must have some output power. You'd think I'd see *something*. I'm set on 500 mV per hatch and 2 ns per hatch in the time domain...not the slightest flinch.

Don

 

I didn't think anybody had come up with a 900 MHz fundamental crystal since the last time looked about a decade or so ago.

For sure, Papabravo!

I've taken the liberty to modify my earlier post.

Nandu.

 

Here is my circuit:

View attachment 207076

My scope is an older Tektronix 2440 digital at 500 Mbps. I thought it might show up as a blur like on my old analog, but maybe the signal is outstripping the digital entirely.

They show an antenna connected directly to the output, so it must have some output power. You'd think I'd see *something*. I'm set on 500 mV per hatch and 2 ns per hatch in the time domain...not the slightest flinch.

Don

Seeing as how a wavelength @ 900 MHz is about 13" -- don't you think those component leads are just...a bit long? Depending what your scope probe looks like impedance wise you may or may not see anything. Since it is not an analog scope I'm not sure I can predict what you will see, or if you will see anything at all. You might be better off with a general coverage receiver that goes up to 2-3 GHz.

Icom IC-8600
https://www.icomamerica.com/en/products/amateur/receivers/r8600/default.aspx

 

RE: ""Isn't there a simple way to get a usable 920 MHz sine wave?""
Why not to use any serial or parallel oscillators circuitry?
a) Clapp circuit will easily produce a kilovolt-scale but milliamps if aim is the plasma. Colpittz circuit will produce the half Vcc scale voltage with about 1 kOhm output impedance, however it can be diminished until 50 Ohms if the aim is transmitter. Royer circ will produce a larger power for the transmitter aim. etc etc etc.
b) Material basis for UHF are BFU910x (Ft=90 000 MHz), BFP840, BFU725, BFP520 and many more however this moment obsolete and their substitute list in most of stores is clearly (must be bolded) IDIOTIC. The competitors have 90 MHz instead of GHz there are said. So in coronaepoch I struck in absence of most important transistors thus I bought Farnell the mch4015, it isnt bad at all, especially taking in account the voltage and current. Of course the BF998 is never-ending balsam for soul, it is available as well.
c) The resonator tank itself may be a choice between coaxial stub or bifilary line stub, or sandwich line stub (inner type, outer type etc etc). Normal coil plus cap there are senseless.
d) RE:""I didn't think anybody had come up with a 900 MHz fundamental crystal"" - if the freq stability is the most piercy point, then Vačkar circ is the solution capable to compete within 10^-5...10^-6 stability what is normal kingdom of any non-thermostated crystall. By the way, it has unbelievable good (low) phase noises.

 

Hello,

I think you will not be able to measure the 920 MHz on a 300 MHz oscilloscope.
http://w140.com/tekwiki/wiki/2440

Also the use of a breadboard for frequencies higher as 1 MHz is a no-no.
The capacities of the strips in the breadboard will destroy the signal.

Bertus

 

Bertus: You mean 1 GHz not 1 MHz probably. Normal FR4 material with an ease may work at 30 MHz and some more special sorts even 300 MHz. However for 2,4 only rare surprizingly stable circuits may operate well at FR4, there one of good options are Rogers Duroid materials (3 choices). By the way, capacity of strip-lines some cases are even much wanted effect when it is sandwich-line.

 

Hello,

@Janis59 , In post #5 a picture is shown with a breadboard.
The module may made of good RF pcb material, but the strips of the breadboard will have a high capacity for the high frequency.

Bertus

 

Aaah, I didnt read the previous mails thus answered for general case.

 

So, while we're on the subject...any tips on how to get a look at such a wave without spending $15,000?

Don

 

So, while we're on the subject...any tips on how to get a look at such a wave without spending $15,000?

Don

I already gave one suggestion and that was a communication receiver like the Icom IC-8600. As an SDR with I-Q outputs a spectrum display would not be a heavy lift. You could try to borrow something if you know people. The other option is to rent the equipment; the cost should be in the range of 1-3%/month or $150 to $450 per month on $15,000. At those price levels you need a powerful reason to fool around with such things.

 

Actually, what the benefit gives any LOOKING on the wave? One may need to know the voltage, current, SWR, % of harmonics - all those are measurable for cheap.

 

Well, I put a .1uF cap on the output and put an AC voltmeter on it, set on millivolts, and got nothing: 28 mV in plain air; 28 mV on the output. I'm pretty sure there's no wave. It seems unlikely that both chips are bad. I can have a special surface-mount PCB made to test, but if my circuit is wrong, I'm wasting time. I was hoping there was an obvious error.

Don

 

Your test equipment is just not up to the job your trying to do.
The scope, has been covered,
The plug board, is going to add a significant reactance on the output of the chip, probably enough to stop it working.
You have tried to measure with a DVM,
your DVM will not respond to MHz, the leads are just to long , let alone the DVM capability.

It is possible to measure an RF waveform with a DVM.
You have to use a buffer and rectifier circuit that works at the frequency you want, an IN4001 will not do .
then your meter is measuring DC, which it can do, and your measuring the Resultant from the RF amplitude.

Up at these frequencies, the bulk effect of components is less than at lower frequencies.
e.g. a cable has significant capacitance and inductance,
resistors and capacitors become inductive,
an induct or has significant capacitance.

I have seen "circuits" made with out Capacitors or conductors, just bent track, and shapes,,,

You need to start thinking about things like smith charts, power matching not impedance matching,

A very different world.

Just putting you finger near a circuit at these frequencies can stop an oscillator oscillating, and make an amplifier start oscillating..

To get an idea how different, read through this
https://cds.cern.ch/record/1407402/files/p223.pdf

 

Or, you can just make a very simple radio frequency detector. You can use a DVM as the meter. The diode should be a fast small signal diode, (not a power supply rectifier) 1N914, 1N916, 1N4148, 1N60 (either type Germanium or Schottky), 1N5711, etc. Use a ceramic capacitor for the .001 uf shown in the schematic.

Use the coil to search around the oscillator and see whether there is any detectable RF.