Author Topic: Need implement RF gain equalizer (create uptilt in MMIC-based amplifier stage)  (Read 1806 times)

0 Members and 1 Guest are viewing this topic.

Offline wb0gazTopic starter

  • Frequent Contributor
  • **
  • Posts: 304
I did some searching on eevblog forum and on internet and didn't find anything that I felt I could use as a starting point, so please forgive if this is FAQ or otherwise obvious?

I need implement a MMIC RF gain stage that has some "uptilt" in gain (that is, somewhat more gain at higher frequencies, less at lower frequencies.) This sort of stage is common in cable TV line amplifiers (that compensate for tilt in coax segments).

From what I have been able to find, a T-type resistive attenator with some parallel capacitance on the input and output sections and series inductance on the mid-point section is a way these can be built. What would be great would be an online calculator (or failing that, a calculation sequence) that would yield a nominally 50-ohm port impedance with a transfer function that has some desired degree of uptilt.

Any suggestion/pointers?

Thanks!
 

Offline A.Z.

  • Super Contributor
  • ***
  • Posts: 1125
  • Country: it
  • if you can't measure it, you can't improve it
    • QRPer forum
maybe I misundestood but... won't an high pass filter achieve what you're asking ?
 

Offline wb0gazTopic starter

  • Frequent Contributor
  • **
  • Posts: 304
A high pass filter is the correct concept, however, the db/octave or db/decade tilt in the MMIC is lower than db/octave or db/decade tilt in a high pass filter. For example, an ideal high pass filter has 100% loss until "knee" frequency, then 0% loss. In the MMIC amplifier, the loss increases slowly across the whole specified frequency range, by some factor that is unique to the individual device type. So, equalizer needs to be designed to match this slow increase in loss (decrease in gain) with a equivalent-but-opposite effect. That is why a calculator or design tool would be helpful.

I have seen some designs (which use resistor/capacitor and resistor/inductor pairs to create very low Q, this I think is the general concept, but detail implementation method is still not seen, but I am sure exists.
 

Offline radar_macgyver

  • Frequent Contributor
  • **
  • Posts: 777
  • Country: us
You could use a gain slope equalizer MMIC.

https://www.minicircuits.com/appdoc/AN60-106.html

MCL makes a few equalizer components, up to 6 GHz. Other MMIC companies (Miller, Marki) make higher frequency components.
 

Offline miniant

  • Newbie
  • Posts: 7
  • Country: ro
Or you can build it yourself depending on your needs. You can easily simulate it with the RFSim 99 software.
https://www.changpuak.ch/electronics/Equalizer.php
 

Offline RFDx

  • Regular Contributor
  • *
  • Posts: 221
  • Country: de
What would be great would be an online calculator (or failing that, a calculation sequence) that would yield a nominally 50-ohm port impedance with a transfer function that has some desired degree of uptilt.

Matthew A. Morgan (2017): Reflectionless Filters, Artech House, Boston/London.

On p.103 you can find a synthesis procedure for pos./neg. slope bridged-T equalizers. Compared to the more economic HP version, the BP configuration has better slope linearity and it actually reaches min. loss at the upper design frequency.
 

Offline wb0gazTopic starter

  • Frequent Contributor
  • **
  • Posts: 304
Excellent, and thanks for the specific and detailed referrals!

I am now awaiting PCB from chinese fab (so hopefully gets here "in time"), then will be ready for hands-on experimentation - the readings above are very much appreciated.

Also, I did find the MCL inline equalizer components; hoping to DIY as this is personal use and the learning experience is >50% of the objective.

Thanks again!
 

Offline A.Z.

  • Super Contributor
  • ***
  • Posts: 1125
  • Country: it
  • if you can't measure it, you can't improve it
    • QRPer forum
A high pass filter is the correct concept, however, the db/octave or db/decade tilt in the MMIC is lower than db/octave or db/decade tilt in a high pass filter. For example, an ideal high pass filter has 100% loss until "knee" frequency, then 0% loss. In the MMIC amplifier, the loss increases slowly across the whole specified frequency range, by some factor that is unique to the individual device type. So, equalizer needs to be designed to match this slow increase in loss (decrease in gain) with a equivalent-but-opposite effect. That is why a calculator or design tool would be helpful.

I have seen some designs (which use resistor/capacitor and resistor/inductor pairs to create very low Q, this I think is the general concept, but detail implementation method is still not seen, but I am sure exists.

give this tool

https://markimicrowave.com/technical-resources/tools/lc-filter-design-tool/

a try

HTH




 

Offline wb0gazTopic starter

  • Frequent Contributor
  • **
  • Posts: 304
Thanks, A.Z. -

I've used that tool from time to time before it was picked up by Marki, however, I've not seen the additional documentation (that appears before you click "Load Tool" before. The references to LTSPICE and QUCS are both encouraging (this task may well afford the motivation/opportunity to learn one of those two tools, although as I recall SPICE was available on the university CDC 6400 mainframe in 1975 when I was in undergraduate EE program.

The suggestion is much appreciated!

Dave
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf