how to drive four srd diode

[mohammadaz]

Hi
i have a 100 MHz OCXO with a phase noise of -160 dBc/Hz @ 1KHz offset.
I need to have four copy of this reference with powers in order of 27 dBm.
for distribution, is it better to use a distribution amplifier or a simple power divider is enough?
how much degrades the phase noise a simple power divider with good insertion loss.
if discrete distribution amplifier is recommended, can anyone suggest me any structure or schematic?
after divider or any distribution method a mediume power amplifier with very low additive phase noise is required. Can anyone suggest me a part number? Doese its additive phase noise or flicker noise has benn mentioned in the datasheet?

Thanks alot.

 

[biff44]

the noise floor of the world is -174 dBm/hz. lets say your oscillator has +14 dBm out. If you pass it thru a 4 way power divider, you would loos 6 dB of power, so the 4 signals would be at +8 dBm each. 160 dB/hz below +8 dBm would be -152 dBm/Hz, or almost 20 dB above the noise floor of the world. So your phase noise at 1 KHz offset would be preserved. You could then use four separate half watt amps to boost the four signals up to +27 dBm power level each

Phase noise further out, like at 1 MHz offset, might be degraded though.

 

[mohammadaz]

Ok, thanks.
so, a simple divider is enough for me.
whould you please introduce a part number for half watt amplifier?
the output power of my ocxo is about 10 dBm.
the amplifier should have a gain of 17 dB to reach this power without going to saturation. I searched hittite, triquint, rfmd,...
but could not find such a part. I am looking from among SiC amplifiers to have not far fliker corner.
thanks again for clear answer.

 

[D.A.(Tony)Stewart]

Hi
i have a 100 MHz OCXO with a phase noise of -160 dBc/Hz @ 1KHz offset.
I need to have four copy of this reference with powers in order of 27 dBm.
for distribution,...
Thanks alot.

Sounds like the Vectron OX-041 low-phase noise option on the SC cut OCXO... But that is only,5.5dBm typ, 8dBm max

In any case you need 4x 50Ohm outputs that equates to -6 dB split @27dBm per channel which implies an input splitter with 4 separate 23dB gain with 1W per channel or one 4W amp with 6dB loss on the output splitter.

You will also need vibration isolation to guarantee <0.1g vibration to prevent vibration induced phase noise and clean power to ensure the amps do not add PS induced phase noise. MiniCircuits is one such source. also all cables must be semi-rigid unless you are in a Benign environment.

I first used vectron AT cut OCXO's in the mid 70's and even though they were ruggedized, there was always at least one axis with more vibration sensitivity.


Why so much power for Step Recovery Diodes? and what if they trigger at different thresholds and trigger poor return loss?

 

[mohammadaz]

Tnx dear skyguy
in fact i want to use srd diodes as x22, x23, x24 and x25 multipliers.
i have designed the input and output matching networks for srd diodes. They were measured with network analizer and the results were satisfactory. I have used four different microstrip quarter-wave transformer followed by four different saw filters as output resonators. But when I test theme as multiplier, the output phase noise were quite bad!
As a matter of fact, when I tuned the input level of srd, for input power in order of 26 dBm, the output phase noise and efficiency were good.
my srd diodes are from MA/COM company (MAVR series). My OCXO IS AOCJY-100 from ABRACON. Manufacturer.
which type of amplifier should be used in each branch of divider to have low fliker noise?
Can you suggest me a SMD part number?

 

[biff44]

jeez, you might have mentioned that first! :laugh:

 

[D.A.(Tony)Stewart]

Tnx dear skyguy
in fact i want to use srd diodes as x22, x23, x24 and x25 multipliers.
i have designed the input and output matching networks for srd diodes. They were measured with network analizer and the results were satisfactory. I have used four different microstrip quarter-wave transformer followed by four different saw filters as output resonators. But when I test theme as multiplier, the output phase noise were quite bad!
As a matter of fact, when I tuned the input level of srd, for input power in order of 26 dBm, the output phase noise and efficiency were good.
my srd diodes are from MA/COM company (MAVR series). My OCXO IS AOCJY-100 from ABRACON. Manufacturer.
which type of amplifier should be used in each branch of divider to have low fliker noise?
Can you suggest me a SMD part number?

I figured you were using some nonlinear load. Of course you realize phase noise increases with harmonics by the order of the harmonic.

I have no 3Ghz multiplier diode preference, but would look at Gunn Effect and any other avalanche mode.

Phase noise reduction occurs with higher Q resonant cavites at harmonics.


What noise density/ offset do you get?

 

[mohammadaz]

I measure a phase noise of -127 dBc/Hz at 2.5 GHz when the input level is around 26 dBm @ 1KHz offset.
in fact the output phase noise is quite sensitive to the input level. due to this, I'm afraid of operating temperature range (-40 +80)!
theoretically, the phase noise should be around -132 dBc/Hz.
I use Rohde&Schwarz source signal analyzer to measure the phase noise. according to its operations manual, it seems I reached the measurement limit of the device.
I measure the phase noise of OCXO abuot -155 dBc/Hz, which according to the datasheet, should be around -160 dBc/Hz.
but the increased phase noise may be due to additive phase noise of multiplier, amplifiers and...
So, Im working on multiplier and drive amplifiers.
does any one have any suggestion?

 

[D.A.(Tony)Stewart]

So you have 27dB noise gain with harmonic multiple of 25. What did U expect? 14dB?

Naturally , an unbiased SRD bridge has a higher threshold/sensitivity for phase noise.

 

[mohammadaz]

I use a single shunt srd, not a bridge.
At first i tried a biased srd, my single srd were DC biased with a voltage around 0.5 V.
biased srd needs lower power for drive, but it was quite sensetive to bias voltage.
i encountered some stability issues! Small changes to bias voltage leads to unstability.
The best stability was obtained in self biased mode. With an input of 28 up to 28 dBm, the diode is stable, and phase noise is acceptable for me.
Now, i want to feed four SRDs all together. So, i need a reletively high power amplifier to amplify the reference signal. I am afraid of phase noise degradation due to additive phase noise of such an amplifier. My explicit question is that wich type of amplifier should i use???
Is there any difference between GaAs, SiC or lnGaP HBT amplifiers?
the phase noise of ocxo is quite low, and improper selection of the first amplifier can degrade it.
does any one hav such an experience to introduce me a part number?
Thank you so much.

 

[biff44]

a silicon bipolar amplifier will be the best for additive phase noise at 100 MHz, HBT is probably 2nd best. GaAs the worst

 

[mohammadaz]

Really tanx dear biff
whould you please introduce me a reference about this topic?