Continue to Site

Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

Register Log in

[SOLVED] 20KW dual directional coupler

Status
Not open for further replies.
A

ahsoopk

Member level 5
Joined
Mar 12, 2010
Messages
81
Helped
0
Reputation
0
Reaction score
0
Trophy points
1,286
Activity points
2,103
Dear all,

I have been given a task to design a dual directional coupler that can handle 20KW and should be microstrip based design. its S-band and i have no idea how to go about it and what calculations are required to make sure microstrip lines dont evaporate.

you help and support will be deeply appreciated.

thanks in advance
 

The question doesn't make sense without accompanying information.

What's the involved coaxial respectively waveguide interface? Frequency range? CW or pulsed power, duty cycle?

It's relative simple to design a directional coupler tapping a coaxial air line or a waveguide. These days, you have an EM solver to test the design.
 

Dear FvM,

many thanks for the reply, information is given below,

N-type input and i need to confirm the output connector type.
frequency range is 2.6 - 3.1GHz
system type is Pulsed
Duty cycle is 10%

I have no experience in this high power and would really appreciated if you could kindly guide me in the right direction.

Many thanks in advance
 

The power level is considerably above N connector limits.
 

20 kw peak or cw?

If it were me, I would design it instead with a stripline circuit using bonded boards. This should keep it from arcing over.
 

20KW@ 50 ohms = V ^2 / 50, V^2 = 1,000,000, so V = 1000, so the current is 20A. This looks too much for the stripline, So build a rectanglar box, > 1/2 wave long, configure it to have a inner conductor with the N types on the end, so the through line looks like 50 ohm. Mount two pairs of BNC connectors on opposite walls at 1/4 wave length apart. Mount your stripline between the BNCs, change the lines relative in/out position to get the require coupling. Two pairs?, so you can monitor the forward and backward power at the same time, else use one.
Like the amateur one seen here without the diodes :- https://en.wikipedia.org/wiki/SWR_meter
Frank
Frank
 
Last edited:
chuckey said:
20KW@ 50 ohms = V ^2 / 50, V^2 = 1,000,000, so V = 1000, so the current is 20A. This looks too much for the stripline, So build a rectanglar box, > 1/2 wave long, configure it to have a inner conductor with the N types on the end, so the through line looks like 50 ohm. Mount two pairs of BNC connectors on opposite walls at 1/4 wave length apart. Mount your stripline between the BNCs, change the lines relative in/out position to get the require coupling. Two pairs?, so you can monitor the forward and backward power at the same time, else use one.
Like the amateur one seen here without the diodes :- https://en.wikipedia.org/wiki/SWR_meter
Frank
Frank
Click to expand...


Dear Frank,

Sorry for amateur question but can you please explain type of through line, will it be coaxial cable, microstrip, stripline etc. same goes for the coupling sides i.e. connected to BNC what type of lines will they be and one last thing for my understanding, will this be edge coupled??

Many thanks for your kind help and support.

regards,

- - - Updated - - -

biff44 said:
20 kw peak or cw?

If it were me, I would design it instead with a stripline circuit using bonded boards. This should keep it from arcing over.
Click to expand...

20kw is peak power. when you said bonded boards does it mean design in stripline rather than microstrip, if that is the case then will having a thick substrate help?
 

I don't think any substrate can carry this power level in any case.
Waveguides would be a solution.
 

Agreed, waveguides are the practical choice. The problem isn't heat so much as dielectric breakdown, so whether the power is pulsed or CW doesn't really matter. You might be able to find specialty cables and connectors which can do it, but a standard N type will not handle 1000V.
 

You might be able to find specialty cables and connectors which can do it, but a standard N type will not handle 1000V.
Click to expand...
Neither it will handle the current at GHz frequencies. Either the peak power and average power handled by the N connector will be lower, or a different interface has to be used. But apart from the connector problem, an airline design as shown with the SWR meter in post #6 can do. It's probably reasonable to implement the sense transmission lines as microstrip, at least if the coupler attenuation is respectively high. Making the power transmission line as micro stripline doesn't seem reasonable because it must be quite large to avoid partial discharge at the edges. A round center connector in air is much better in terms of voltage strength.
 

Through line should be sized so you can mount your N types on the ends. This then sets the cross section of the device. Then depending on the thickness of the material used, will give you the cavity cross section. i.e. you could use copper water pipe, a block of aluminium with a channel machined out of it (like HP would do), A folded aluminium box with blocks fitted in the end for the N types. Once you have this then you have to select the size of the inner conductor to make the through impedance an accurate 50 ohms. But remember the 20 A and 1000V so it must all be of robust construction with decent connections to the Ns and sufficient clearance for the voltage rating. Try Googling VSWR meters , I am sure that you will find some mechanical arrangements.
Frank
 

chuckey said:
Through line should be sized so you can mount your N types on the ends. This then sets the cross section of the device. Then depending on the thickness of the material used, will give you the cavity cross section. i.e. you could use copper water pipe, a block of aluminium with a channel machined out of it (like HP would do), A folded aluminium box with blocks fitted in the end for the N types. Once you have this then you have to select the size of the inner conductor to make the through impedance an accurate 50 ohms. But remember the 20 A and 1000V so it must all be of robust construction with decent connections to the Ns and sufficient clearance for the voltage rating. Try Googling VSWR meters , I am sure that you will find some mechanical arrangements.
Frank
Click to expand...

Hi Frank,

are you suggesting a design approach given in link below

http://www.w1ghz.org/antbook/conf/High-Power_Directional_Couplers_with_Excellent_Performance.pdf

regards,
 

No obvious at a glance why they used a ribbon instead of a circular conductor, but it can surely work. Impedance of a coaxial transmission line is a simple hand calculation, you need a FEM tool to calculate the ribbon impedance.
 

FvM said:
No obvious at a glance why they used a ribbon instead of a circular conductor, but it can surely work. Impedance of a coaxial transmission line is a simple hand calculation, you need a FEM tool to calculate the ribbon impedance.
Click to expand...

I have Microwave Office that I can use for impedance calculation using Air as dielectric. But you do agree if a copper ribbon of 2/4mm is used in the same manner as this person has used should be able to handle the power, however how much gap should be left between the ground and the copper i.e. dielectric (Air) so that it doesn't arc.
 

Sband wave length ~100mm (3 GHZ). so the centre line is rather wide (36mm) with a high VSWR for the small version, but I think its a very good start. The problem with SHF is to keep it flowing in the correct direction, it will try to find other resonant modes all the time, with a resulting poor VSWR and high power loss.
Frank
 

That's indeed a good start, but he says they're only tested up to 500W. At 20kW I would expect to get corona discharge at the corners of the ribbon. A circular conductor with a large diameter would be a better bet. Might be worth trying to varnish the conductor.
 

mtwieg said:
That's indeed a good start, but he says they're only tested up to 500W. At 20kW I would expect to get corona discharge at the corners of the ribbon. A circular conductor with a large diameter would be a better bet. Might be worth trying to varnish the conductor.
Click to expand...

Hi Just checked circular coaxial cable on micro-coax's website and it seems like it can handle more than 1400W, link is given below. if I use flat ribbon type copper i can calculate the 50 Ohm track dimension using TxLine from AWR using Air as dielectric but if I use circular copper conductor then how am I going to calculate the 50 Ohm characteristic of this circular rod using Air as dielectric. I can see the finish line now, need one last push...

**broken link removed**

Many thanks all of you for such a tremendous help.
 

Hi all,

can you guys kindly confirm that i can use formulas from the following link to calculate the Z0=50 of a circular copper rod in air dielectric.

https://www.mantaro.com/resources/impedance_calculator.htm

if i use outer diameter as cavity wall to wall distance but what happens at the point where I am coupling because i need 60dB coupling so quarter wavelength lines have to be quite away from the main line that shall increase the cavity width i.e. impedance characteristics at the point.. I hope it makes sense what I am trying to say.
 

Status
Not open for further replies.

Similar threads

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top