CATV line amps as RF power amps?

I've been tempted to use them as mega-wideband pulse amps, because of specs and cost. But I haven't got around to testing them in time domain apps.

John

datasheet say 75 Ohm, not so surprising since it is for tv

example:

-Lasse

Oops. I didn't see that line just above each table. I looked again after you pointed it out. Don't know how I could have missed it. Thanks.

The ones I've seen specified the total intermodulation distortion with a multi-tone test -- i.e., an answer to the question "how many channels, and how tightly must the channel amplitudes must be controlled?". And yes, they want to keep the distortion down.

I would expect that power dissipation would be an issue, and the limiting factor. The only way to know for sure would be to either find a friendly apps engineer and ask, take some apart and look, or just build some amps and see how well they survive.

They are push pull.

They are 75 ohms. The top of page 5 says: ZL = 75 Ohms. A block diagram would just be a couple cascaded P-P stages and a power pin.

I think they won't last long. I used to repair CATV line & trunk amplifers. They are picky about impedance matching, bypassing and circuit layout. 60 dBmV output is 1 volt at 75 ohms.

If you have a decent internet connection, here is some useful CATV data. It is a little over 12 MB download.

The CATV amp modules I've used were hybrids on an aluminum heat spreader, and epoxy filled.

should be fine as long as you keep the baseplate temp under control and don't drive the input so hard that the supply current increases significantly.

someone else already mentioned they are 75 Ohms. Mark

The way they're constructed, it shouldn't be a problem to bolt the device on to a good-sized heatsink.

Yes, I'll try it out. I wanted informed opinions to see if there's some factor I hadn't thought of that makes it a complete waste of time. At least, the sourcing problem I continually face in this place is not an issue. These devices are locally available at ~$7, as against the ~$40 at Digikey and $60 at the Indian branch of Farnell.

Yes, the mechanical design makes it very convenient to bolt them on to a heatsink.

I don't have a lot of experience in those things, but I know the basic principles involved. This will be a good excuse to build up some experience. They cost only ~$7 US apiece here, so it's not a big deal if I fry a couple of them in the process.

I'll see if I can get around 1W output at a single frequency.

Thanks. It took just a few minutes even with my 512kb/s connection. I used to have a 2Mb/s line - the fastest available here, but my kids came home from college and the 2.5GB monthly cap was nowhere near enough. So I switched to the lower speed with unlimited usage. I'm getting a separate 2Mb/s account for my own use.

Yes, monitoring the supply current is a good idea. Thanks.

As I said earlier, I don't know how I could have missed seeing that part.

Five and a half-and-a-bit-more watts max power consumption. What's your load look like?

They're intended to boost a CATV signal downstream of a splitter to just strong enough that a TV can see it clearly, right? That's not what I'd call a "power" app, but MMMV...

Active cooling worthwhile?

Built-in pass/notch filtering? Suppose you need to boost a CW signal

Mechanical sensitivity aka microphonics? Portable app? "Automotive qualified products".

Substrate diodes? ;>)

Mark L. Fergerson

They're specced at 60 dbmV which is 1V output. 1V into 75 ohms is a tiny fraction of a watt. Up to that point they are wonderfully clean and yes, I also wonder how much more output can be gotten beyond that spec.

I would presume the hybrid module has some internal push-pull magnetics and at some point below the watt level, I would hazard the guess that these would be the limiting factor.

For many purposes where power output at the watt level rather than cleanliness at much lower levels is what's important, you'll do much better with a far earlier CATV-inspired part, the 2N3866, in a narrowband circuit.

Tim.

Response flat within +/-0.2dB 40-860MHz. Return losses given for octave frequency blocks of 40-80MHz, 80-160MHz, 160-320MHz, etc. Does that sound like there are band-pass filters and inter-channel notches?

Come ON!