Close

Cavity V5 measurements

A project log for EMDrive/satellite

Developing a small fuelless microwave thruster

paul-kocylaPaul Kocyla 01/03/2017 at 21:195 Comments

Here is the setup:

The EMDdrive V5 board is connected to the cavity´s feed antenna (lambda/4 stub).

A shorted antenna is used as feedback port which is connected to a power sensor.

First, I measured the amplifier´s performance by connecting it directly to the power sensor.
Here are the results:

At over 23.5 GHz the power is over 100mW - hmmmm it´s a 1000mW amplifier, so that´s not that good, the board is probably not perfectly designed, cables and connectors have losses. The sensor has a 20dB attenuator (compensated calculations for that) and a SMP to 2.92mm adapter, they will probably also cause some losses. So let´s say it´s around 150mW at the working frequency range.

Now comes the interesting part: The feedback port is connected to the sensor, and frequency sweeps are performed. It´s basically a scalar network analysis.
First, I left the cavity opened - this means the big endplate was not connected.
Here is the result - flat, almost no feedback. That´s not surprising, should be like this:

Look what happens when the big endplate is attached:

We get three main peaks. Two strong and one weak - and some smaller artefacts.

I assume the small disturbances are caused by the antennae destroying the optimal shape of the cavity.

When you look at the power level, it seems that the feedback antenna is sucking all the power out off the cavity, not good, it was probably too long - but better starting off too long than too short. So I cut the feedback antenna shorter, from around 1.7 mm to 1mm (approximately). Here´s the result:

That´s better. The feedback antenna now sucks 40mW instead of 100mW. That´s better, but still too much.

Now I cut the antenna to the ground, it´s just a pin in the hole, but the result seems ok:

Note that the V/div is now 20.0mV instead of 200mV, so it´s sucking just 10mW now.

Probably some room for improvement here - let´s see, but that´s acceptable.

BTW before the trolls cry again because of missing axis labels: The two last pictures have same labeling than the third last - I´m just too tired to insert them - have a newborn baby now and a full time job: X:FRQ sweep, Y: power 10mV <=> 1mW

Discussions

willemstaal wrote 01/09/2017 at 12:23 point

Congrats for both your babies, Paul!  

90 % loss means someting is leaking away.  How "clean" is the silver? 

i mean, does the silver have a first grade (925) quality?  

and did you use silver connections? (use silver for the connections and wires to the oscilloscope as well!) 

  Are you sure? yes | no

Paul Kocyla wrote 01/09/2017 at 16:39 point

The oscilloscope is connected to the output of the power-meter, it´s just a display.
The power loss is happening on the EMDrive-Board and the cables. Although everything is rated for 26GHz and routed carefully, the PCB material is not for HF and slight errors in routing can also attenuate these high frequencies. I made changes in the board already, will try to make it better.

  Are you sure? yes | no

willemstaal wrote 01/10/2017 at 10:51 point

May i suggest something different? 

Why not use a insulated shieldingcan around your pcb and ground this outside of the unit?  (i hope you have room for that) My guess your frustrum act now as a shield for the RF interference, and you don't want this  to happen.. 

  Are you sure? yes | no

max.kargl wrote 01/09/2017 at 11:00 point

Congrats for becoming father and best wishes for the new year, also for your project !

  Are you sure? yes | no

Paul Kocyla wrote 01/09/2017 at 16:37 point

Thanx :)

  Are you sure? yes | no