Author Topic: 50 dB Power Attenuator Construction  (Read 2607 times)

0 Members and 1 Guest are viewing this topic.

Offline German_EETopic starter

  • Super Contributor
  • ***
  • Posts: 2399
  • Country: de
50 dB Power Attenuator Construction
« on: June 29, 2018, 06:53:39 pm »
I want to construct a 50 dB attenuator that will operate at 200W maximum for long periods. I have two 250W 50 ohm RF resistors and the required resistor at the center of the 'pi' is 7850 ohms. I also have two very large heatsinks. Now I need some help.

To make this attenuator as useful as possible I want it to operate as high as the 70 cm band as a minimum and operation at 23cm would be nice. How do I make an attenuator that is a broadband device? Is it just a question of joining the input and output with a 50 ohm line on a PCB strip and distributing the 7850 ohms uniformly along its length? What about the heatsink, do I need to machine out a trough from input to output and put the line in it?

Note that the heatsinks are 15 x 15cm and 3cm thick, heat will not be a problem, especially when I join the heatsinks face to face
Should you find yourself in a chronically leaking boat, energy devoted to changing vessels is likely to be more productive than energy devoted to patching leaks.

Warren Buffett
 

Offline KJDS

  • Super Contributor
  • ***
  • Posts: 2442
  • Country: gb
    • my website holding page
Re: 50 dB Power Attenuator Construction
« Reply #1 on: June 29, 2018, 07:07:28 pm »
How much capacitance will be across the series resistor?

If it's about 1pF, that's 367ohms in parallel with your 7850 ohm resistor.
You might, just might, with a bit of luck, get it down to 0.05pF, which is still a smaller impedance than your resistor.

You'll be far better off aiming for a 30dB pad followed by a 20dB pad.

Offline T3sl4co1l

  • Super Contributor
  • ***
  • Posts: 21606
  • Country: us
  • Expert, Analog Electronics, PCB Layout, EMC
    • Seven Transistor Labs
Re: 50 dB Power Attenuator Construction
« Reply #2 on: June 29, 2018, 07:38:23 pm »
Think in terms of transmission line segments.  You want constant-loss TLs to get the power absorption, then whatever's left at the end is your output tap.

A 7850 ohm TL will be very high aspect ratio indeed.  Enough so that, even with a very thin resistance wire (but still big enough to handle what little power it needs to), you won't be able to get an aspect ratio that large without encountering higher modes in that space.

So yeah, staged is a good idea. :)

Even better, do, say, 3 then 23.5 then 23.5dB, so the first stage absorbs half the power, second absorbs almost all the rest, and the final stage can be reasonable.  Unfortunately this won't use any resistors you have...

Tim
Seven Transistor Labs, LLC
Electronic design, from concept to prototype.
Bringing a project to life?  Send me a message!
 

Offline dmills

  • Super Contributor
  • ***
  • Posts: 2093
  • Country: gb
Re: 50 dB Power Attenuator Construction
« Reply #3 on: June 30, 2018, 12:10:05 pm »
50dB?

High power 50 ohm right across the input socket, with a tap at maybe -20dB, followed by a 30dB Pi or whatever.

Advantage is you only need one high power resistor, everything else is 2W or below, and the 450 ohms or so of the first stage tap divider is large enough that it should not upset the 50 ohm match too badly.

Screen the 30dB bit well, in fact I would probably make that a separate little in line attenuator.

Don't forget you can tune out the series cap with a shunt cap.

73 Dan.
 

Offline awallin

  • Frequent Contributor
  • **
  • Posts: 694
Re: 50 dB Power Attenuator Construction
« Reply #4 on: June 30, 2018, 04:30:42 pm »
FWIW the construction of a 500W 50R RF-dummyload I've seen uses a copper-plate/bar close to the resistor(s) and then a finned aluminium heatsink with an optional fan for forced-air cooling.
So I guess a flat copper-plate as a heat-spreader from the resistors to the larger heatsink is a common design pattern.
 

Offline wkb

  • Frequent Contributor
  • **
  • Posts: 905
  • Country: nl
Re: 50 dB Power Attenuator Construction
« Reply #5 on: July 16, 2018, 08:31:16 pm »
FWIW the construction of a 500W 50R RF-dummyload I've seen uses a copper-plate/bar close to the resistor(s) and then a finned aluminium heatsink with an optional fan for forced-air cooling.
So I guess a flat copper-plate as a heat-spreader from the resistors to the larger heatsink is a common design pattern.

It is in high-power RF poweramps that use RF power FETs.
 

Offline Wirehead

  • Regular Contributor
  • *
  • Posts: 177
  • Country: be
    • Wirehead.be
Re: 50 dB Power Attenuator Construction
« Reply #6 on: July 17, 2018, 05:52:24 am »
What's the goal? Using the attenuated output as an input to something? Would a sampling device not be easier?
"to remain static is to lose ground"
 

Offline mc172

  • Frequent Contributor
  • **
  • Posts: 487
  • Country: gb
Re: 50 dB Power Attenuator Construction
« Reply #7 on: July 17, 2018, 10:20:50 pm »
I would use the resistors to build a 300 Watt load (good luck getting it to handle 500!) and build a 30 dB directional coupler, and buy a 20 dB attenuator. The equipment will be far more useful to you than a 500 W (read 300) attenuator.

Why the need for 50 dB out of interest? Most gear can handle between 0.5 and 1 Watt on the input.
 

Offline dave_k

  • Frequent Contributor
  • **
  • Posts: 285
  • Country: au
Re: 50 dB Power Attenuator Construction
« Reply #8 on: July 18, 2018, 03:27:52 am »
Use a directional coupler if you want to do it correctly.
 

Online Wolfgang

  • Super Contributor
  • ***
  • Posts: 1773
  • Country: de
  • Its great if it finally works !
    • Electronic Projects for Fun
Re: 50 dB Power Attenuator Construction
« Reply #9 on: July 20, 2018, 12:57:50 am »
IMHO, you have bad luck here. A multi-Kiloohm resistor at 23cm is a no-no. Just think of some stray capacitances in the sub-pF range and their impedances at these frequencies; even 0.5pF of shunt capacitance will be only 255Ohms, a lot less than what you need by orders of magnitude.

Two remedies:

Buy a reasonably priced 20dB coaxial attenuator with 200W and a small coaxial one for 30dB (cost: ca. 200€, up to 3GHz). I would do that.

Or: Get a ceramic attenuator instead of just a 50Ohm resistor. Values above 30dB are rare, though. You will need 2 stages. Then, you still have to make a (low loss) PCB for it. Tricky.
 

Offline mc172

  • Frequent Contributor
  • **
  • Posts: 487
  • Country: gb
Re: 50 dB Power Attenuator Construction
« Reply #10 on: July 20, 2018, 11:20:19 am »
You shouldn't need low loss PCB for 1 GHz.
 

Online Wolfgang

  • Super Contributor
  • ***
  • Posts: 1773
  • Country: de
  • Its great if it finally works !
    • Electronic Projects for Fun
Re: 50 dB Power Attenuator Construction
« Reply #11 on: July 20, 2018, 02:53:24 pm »
... depends on the power level. FR4 has a nasty smell when it gets hot.  :)
 

Offline mc172

  • Frequent Contributor
  • **
  • Posts: 487
  • Country: gb
Re: 50 dB Power Attenuator Construction
« Reply #12 on: July 20, 2018, 02:58:09 pm »
Just because something is running at high power doesn't mean it's OK to get so hot that the PCB starts to smoke. Horrible smell when overheated is not a fault of the dielectric material, it's bad design.
 

Online Wolfgang

  • Super Contributor
  • ***
  • Posts: 1773
  • Country: de
  • Its great if it finally works !
    • Electronic Projects for Fun
Re: 50 dB Power Attenuator Construction
« Reply #13 on: July 20, 2018, 04:25:04 pm »
200W, FR4, 1GHz - thats marginal  :)
 

Offline mc172

  • Frequent Contributor
  • **
  • Posts: 487
  • Country: gb
Re: 50 dB Power Attenuator Construction
« Reply #14 on: July 21, 2018, 09:34:49 pm »
I'm still baffled as to why the dielectric material has anything to do with it. 1/2 oz copper, yeah, that'd be marginal since you'd struggle to get anywhere near 50 Ohms with the trace widths required, but the dielectric material really doesn't matter.
 

Online Wolfgang

  • Super Contributor
  • ***
  • Posts: 1773
  • Country: de
  • Its great if it finally works !
    • Electronic Projects for Fun
Re: 50 dB Power Attenuator Construction
« Reply #15 on: July 21, 2018, 11:49:24 pm »
Not true. What kills a bad UHF or SHF material at high power or field strength levels is the loss tangent of the dielectric. Guess why all commercial SHF high power circuits (e.g. mobile phone station finals, ...) are built on Rogers or better materials, even if the copper patterns are the same or at least comparable.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf