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] Printed Log-periodic dipole array: optimization (openEMS) and balanced-unbalanced matching

Status
Not open for further replies.
oberstet

oberstet

Newbie level 6
Joined
Feb 1, 2024
Messages
12
Helped
0
Reputation
0
Reaction score
0
Trophy points
1
Activity points
193
Hi all,

I'm new here, also a RF newbie, and this is my first post, so pls bare with me;)

I'm looking for help with 2 specific questions rgd a Printed LPDA I am designing using openEMS.

The PLPDA goals are:

* 400 MHz - 2.8 GHz
* -10 to -20 dB S11 over the whole range
* 50 Ohm unbalanced connect
* Aisler HD 2L PCB

I'll post the current status and EM simulation results below, but to prefix the 2 mentioned questions I have are:

1)
The PLPDA (obviously) has a balanced output, and that output is already designed to ~50 Ohm, but I need to convert that to _unbalanced_ 50 Ohm.

I have learned that I need a "Balun", and that the easiest thing is to go for a "Lumped Balun" (thanks Volker on openEMS! https://github.com/thliebig/openEMS-Project/discussions/178#discussioncomment-8181172).

Essentially SMD 2 inductors and 2 capacitors.

That sounds great, but e.g. when using the https://www.changpuak.ch/electronics/LumpedBalunCalculator.php Volker hinted at, _what frequency_ do I need to enter?

It is 1:1 (50 Ohm in and out) since I designed the antenna for that, but I have to enter _one_ frequency, whereas in my design the antenna should work over a 1:7 frequency range from 400 MHz to 2.8 GHz.

2)
My current WIP results are posted below, I've posted some using different parameters, and while I've identified all relevant model parameters (<50 floats), before going the "last mile" of fine balancing and tuning those, there is one aspect left I am puzzled / wondering about:

What can be seen is that I can achieve broadband operation in a lower ~1GHz range (400-1300MHz), and in a higher range (1800-300MHz), BUT in the mid-range ~1300-1800MHz, there is a big hill (bad S11).

Why? I mean what could be the reason?

I should note: all sizing and geometry parameters of the model are completely set using formulas in a systematic and uniform way in all cases.

I am hence wondering _where_ that hill comes from, there must be some aspect of the geometry or whatever I am not yet aware of.

Any helps or hints or comments (pro or con) would be highly appreciated!

Thanks/Cheers,

/Tobias

----

Couple of results / current status

note that modeling / geometry parameters vary e.g. even number of dipoles etc

frequency/dipoles are using "equal tempering" but with a "triple-frequency per-octave", just different "number of half-tones per-octave" ... I have no idea whether this kind of "electrochromatic scaling" is something usually done, but for me it made sense




screenshot-2024-02-03-1.png



screenshot-2024-02-03-2.png
screenshot-2024-02-03-4.png
Bildschirmfoto vom 2024-02-03 03-43-50.png
Bildschirmfoto vom 2024-02-03 03-15-25.png
 

Sort by date Sort by votes
Upvote 0 Downvote
thank you very much! this is obviously crucial.

it sounded too good with the "just a bunch of SMDs" approach. I thought I could avoid going yet another round on a road completely new for me (this transmission line transformer thing .. I have never done RF before;)

oh well. I'll figure it out;)
 

Upvote 0 Downvote
BigBoss said:
Single Stage L-C Balun cannot provide this bandwidth. Instead, you can use LTCC Balun for this bandwidth.
You can select an appropriate Balun for your application.
Click to expand...

This sounds great!!

If I can avoid digging into yet another area new to me (transmission line balun), that would be highly welcome!

I've searched Mouser, and it appears only 2 part might fit:

1. Mini-Circuits NCS1.5-232+
- 400 MHz to 2.3 GHz
-1:1.5 Ratio
-RF Power 2W

2. Mini-Circuits TCW2-2700+
- 500 MHz to 2.7 GHz
- 1:2 Ratio
- RF Power Input 0.5W max.

---

both of these parts are not 1:1 .. and I designed for 50 ohm ... so do I need to redesign/tune for

1:1.5 => design for 75 ohm
1:2 => design for 100 ohm

right?

---

I just noticed, the part you linked

Macom MABA-010125-TC1113

is indeed 1:1!

and it's 4.5 - 3000 MHz. cool!

so my only itch is the 250 mW ..

---

rgd max . power .. I planned for a max RF transmit of 5 W .. I might be adjust the goal to 2 W .. but 0.5 W is really a bit low.

would a transmission line based balun allow me to "easily" do 5 W?

I'm not sure ... _if_ I _could_ get such a thing designed anyways;)

---

I've seen some plots rgd "phasing" in those part's datasheets.

that triggered me. because (I should have mentioned), I want to use _two_ instances of the PLPDA antenna fixed positioned into a 2-Element Phased Array.

ultimately, I want to be able to do DoA finding and such on any frequency within 400 mhz - 2.8 ghz

now, with phased antenna arrays, the antenna element positioning to each other and the phasing is crucial.

but that's a different phase thing, right?

I mean, as far as I can read above datasheets, that phasing is about phase shifts depending on 2 different frequencies?

----

thank you so much for helping a RF newbie! =)
 

Upvote 0 Downvote
Upvote 0 Downvote
vfone said:
I don't say that is not a good idea to use a balun in a Log Periodic antenna, but most of the log periodic antennas I've seen do not use a balun, and connect the coax cable directly to the antenna.

Click to expand...

I might understand stuff wrongly, or incompletely, I'm really a RF newbie (I have a ton know-how/experience in all kinds of software tech).

_Connecting_ a coax to a dipole is of course possible, but

- it will use the coax cable to transport a balanced signal (for which it is not designed)
- it will make the coax shield part of one arm of the dipole array
- not using a balun to make the balanced dipole array output into an unbalanced signal before sending it down any wire .. seems like a bad idea?
- if the coax shield has any connection to GND downstream (eg LNA etc board), it will make all the downstream GND part of _one_ of the two dipole arms

as said, I might be pretty wrong, not sure.

if you have an argument of _why_ the use of a coax in such a way might be a working solution for dipoles or dipole array or any balanced antenna, I would be highly interested!!

I do want to learn;)
 

Upvote 0 Downvote
BigBoss said:
Single Stage L-C Balun cannot provide this bandwidth. Instead, you can use LTCC Balun for this bandwidth.
You can select an appropriate Balun for your application.
Click to expand...

Can I use a 1:1 LTCC Balun of _any configuration_ to match a balanced 50 ohm dipole to an unbalanced 50 ohm line?

After searching parts, I now realize that there are at least a couple of different configurations, eg

1706977351545.png
1706977360720.png
1706977366596.png
1706977373553.png
 

Upvote 0 Downvote
oberstet said:
I just noticed, the part you linked

Macom MABA-010125-TC1113

is indeed 1:1!

and it's 4.5 - 3000 MHz. cool!

so my only itch is the 250 mW ..

---

rgd max . power .. I planned for a max RF transmit of 5 W .. I might be adjust the goal to 2 W .. but 0.5 W is really a bit low.

would a transmission line based balun allow me to "easily" do 5 W?

I'm not sure ... _if_ I _could_ get such a thing designed anyways;)
Click to expand...
You didn't mention the Output Power.
Now the scenario is changed.
Of course MACOM' Balun cannot handle 5W. 2W can be used for a short period but not continuously.
You have to design/find High Power Balun with connectorized model. But I should warn you these models are very expensive.

Connectorized Products : Baluns

Marki baluns are hand tuned to offer superior amplitude balance, phase balance, and common mode rejection ratio up to 67 GHz. When selecting a test and measurement balun, important specifications to consider include frequency coverage, excess insertion loss, and isolation between balanced ports.
markimicrowave.com markimicrowave.com
The Power Handling specifications of these Baluns have not been mentioned, you can ask by support channel.
 
Upvote 0 Downvote
BigBoss said:
You didn't mention the Output Power.
Now the scenario is changed.
Of course MACOM' Balun cannot handle 5W. 2W can be used for a short period but not continuously.
You have to design/find High Power Balun with connectorized model. But I should warn you these models are very expensive.

Connectorized Products : Baluns

Marki baluns are hand tuned to offer superior amplitude balance, phase balance, and common mode rejection ratio up to 67 GHz. When selecting a test and measurement balun, important specifications to consider include frequency coverage, excess insertion loss, and isolation between balanced ports.
markimicrowave.com markimicrowave.com
The Power Handling specifications of these Baluns have not been mentioned, you can ask by support channel.
Click to expand...

Thank you so much for pointing me to Marki! I never heard of before, but it looks great and I've added the company to my docs / pot. parts suppliers.

also sorry for not making power requirements explicit from the beginning:

* the absolute minimum RF transmit power I need is _500mW ERP for each of 2 LPDA arrays_ - as I want to support standard unlicensed PMR446 - and using Aisler HD 2L (preferred) or Aisler HD 4L (if absolutely required) PCB stack.
* additionally it would be _nice_ to have at least an option to also support "a couple of W" - eg for HAM ops - like 4W or 5W. but I now recognize this should probably only be available in a product variant with different HW/PCB. so I am back to absolute minimum is 500mW as a starter

-------

The Macom MABA-010125-TC1113 only has 250mW.

The Mini-Circuits TCW2-2700+ has 500mW, but is 1:2, and frequency starts at 500 MHz and PMR446 is at 446 MHz.

The Mini-Circuits NCS1.5-232+ has 2W, but is 1:1.5, and frequency tops out at 2.3 GHz and WiFi is at 2.4 GHz (and WiFi 2.4 is besides PMR446 also mandatory).

The Mini-Circuits TTC1-33W+ has 500mW, is 1:1 and supports my whole range 400-2800 MHz (and more).

-------

So ATM, the Mini-Circuits TTC1-33W+ is my candidate for the "normal power" entry product version. Not for a "high power / HAM" version, for which Marki you pointed to might be exactly the right choice.

My remaining worry about the Mini-Circuits TTC1-33W+ is: I don't know if it's _configuration_ can be used for this dipole balanced-unbalanced thing, because I don't really understand / have clues about transformers and all that.

I just need the "bits" for I and Q, the rest is software and I'm at home again;)

All the stuff before the bits is what is new for me, but I hope I'll be able to figure it out, thanks a lot for helping a RF dummy/newbie!!
 

Upvote 0 Downvote
oberstet said:
What can be seen is that I can achieve broadband operation in a lower ~1GHz range (400-1300MHz), and in a higher range (1800-300MHz), BUT in the mid-range ~1300-1800MHz, there is a big hill (bad S11).

Why? I mean what could be the reason?
Click to expand...

I found out the reason!

I'll dump it here maybe it's of use for others, and then will close this issue for now as "resolved" - thank you all for your kind help and tips!! Blocker removed.

So the reason was pretty much unexpected - for me;)

Early on I was recommended to design the LPDA for 50 ohms so no impedance matching is needed. Makes sense, also the LPDA boom needs one common impedance anyways.

Then I decided to use a 2-layer PCB .. price reasons, and why would I need more?

**Turns out, a 50 ohm trace on that 2L PCB buildup requires exactly 2.9 mm width.**

Further I found out that Lambda/60 is a quite good trace width for the dipoles .. a compromise between bandwidth and size/resonance.

Now the problem is, for my frequency range in the upper end (1.5 ghz and higher):

**Lambda/60 < 2.9 mm (!)**

Which turned out to be a problem .. there is a resistance when the current wants to go from the middle boom into one of such higher frequency dipoles when the width is smaller .. so I clamped the width at min. 2.9 mm

However, an apparently necessary consequence of that is "loss of bandwidth" ... and this is what "makes the hills"!

I guess I have 2 options .. which I will explore coming days .. a) use a 4L stack and make use of that to shrink the boom width .. or b) use a finer grained frequency division in the higher end ... maybe I can cheat the effect;) otherwise, $$ cost to swallow.

Anyways, here is a picture, I then stop.

1707005674578.png
 

Upvote 0 Downvote
Status
Not open for further replies.

Similar threads

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top