Advertisement

Blog

Very Wideband Amplifiers Are Difficult No Longer

One of my recent basement lab projects involved designing and building an antenna amplifier for the omnidirectional FM antenna on my roof. This is one of those dual dipole antennas with one dipole set at 90° to the other dipole. The antenna has no special directionality or gain, but it gets better reception than the antenna on most radios — a 1-meter length of hookup wire dangling from the back. It's up higher (that's good) and away from pesky sources of radiation in my house (that's even better).

I wanted to make up for losses incurred in the coaxial cable feed-line. I also planned to run the coax to a four-way splitter for multiple receivers around the house, so there's more incurred losses.

Since I needed to amplify the standard FM broadcast band (88MHz to 108MHz), I considered an RF line amplifier from Radio Shack/Best Buy/Walmart. Not much fun in that. I also considered building an amplifier with one or two, or maybe three, small signal RF transistors. That might work out — but my guess is that I don't recall enough about RF circuit design to build a wideband (untuned) RF amplifier. I'd probably end up with an RF oscillator running at who-knows-what frequency. Probably all of them.

But then I saw some info on a newly released device from Linear Technology, the LTC6430-15. This device is intended as an ADC driver (hmm… not quite right):

A typical circuit for driving a high-speed ADC. (Source: Linear Technology)

A typical circuit for driving a high-speed ADC.
(Source: Linear Technology)

Or it can function nicely as an IF amplifier (closer to my app):

A typical circuit that could be part of an IF amplifier; the input and output transformers would be tuned. (Source: Linear Technology)

A typical circuit that could be part of an IF amplifier; the input and output transformers would be tuned.
(Source: Linear Technology)

Or it can function as a wideband RF amplifier (that's me!):

Now we're getting somewhere - 75Ω in and out. (Source: Linear Technology)

Now we're getting somewhere – 75Ω in and out.
(Source: Linear Technology)

It's like these guys knew just what I was planning on building. The data sheet says this version (with the 75Ω input and output) can operate from 40MHZ to 1000MHz, making it not only good enough for the FM band, but also suitable for CATV distribution in case I want to send that around my house.

A complete unit. Just add RF. (Source: Linear Technology)

A complete unit. Just add RF.
(Source: Linear Technology)

A quick look at the features from the data sheet tells why it performs so well. The noise figure is 3dB at 240MHz; it has a bandwidth of 20MHz to 2GHz; gain is just over 15dB; and 1nV per √Hz total input noise. Oh, and it's unconditionally stable, rather than oscillating at all frequencies, as my design probably would. I plan on procuring one of the eval boards and running my own test to see how it works.

The only thing I may need to add will be some extra clamping diodes at the input, output, and power supply connections to prevent damage in the event of nearby lightning strikes. It's important to note that realistically, you can only prevent damage in the case of nearby lightning strikes. When the strike is real close, the protection becomes rather ineffective.

6 comments on “Very Wideband Amplifiers Are Difficult No Longer

  1. Michael Dunn
    February 21, 2013

    One thing I noticed when looking at parts of this class for an ADC driver a few years ago, is that you likely want a part with MUCH higher bandwidth than you might think. 

    My BW was ~50MHz, but parts in that range had quite high distortion at the top of their bandpass. Using a part with ~10x the needed BW gave acceptable distortion in my passband.

  2. Brad Albing
    February 21, 2013

    Makes sense – at the higher frequencies, the gain drops (the nature of gain-BW product). So, w/ less margin between open loop gain and closed loop gain, disortion increases.

  3. eafpres
    February 22, 2013

    @Brad–is there any risk of over-driving the front ends of your FM receiver or cable tuner?

  4. RedDerek
    February 22, 2013

    Though it may not make much difference, any reason the output ground connection was reversed in the schematics? Input side had the antenna on the positive notated side of the balun. The output shows the output on the negative side.

  5. Brad Albing
    February 22, 2013

    I believe that is done (intentially) so that the output is out of phase w/ the input so as to reduce the likelihood of any inadvertant feedback causes spurious oscillation.

  6. Brad Albing
    February 26, 2013

    There is – altho' risk manifests as intermodualtion distortion or spurious response of the receiver. So if there is obvious distortion, I can just pad down the signal going to the receiver. In my case, gain of the amplifier and loss due to the 4-way splitter should even out pretty well.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.