[NLRS] power measurement

[Gerald]

On Fri, 2006-02-10 at 09:28 -0600, John P. Toscano wrote:
> 
> Ken Boston wrote:
> 
> > Bottom line, an all-in-one power meter for 1-24 gig, power rating 10
> > milliwatts to a KW probably does not exist. (unless there is some super
> > expensive top of the line meter out there that does this)  You will end
> > up with a collection of a decent , older power meter with some sensor
> > heads, power attenuators, and directional couplers that cover the bands
> > that you need them for.  Cruise the swap meets, conferences and E-Bay,
> > and expect to spend some money.
> 
> Hi, Ken, and everyone else contributing to the thread.
> 
> I didn't really expect a single unit to do it all, from milliwatts to 
> kilowatts, but was hoping for one meter, a manageable number of slugs or 
> plug-in sensors, and if necessary, a small number of attenuators or 
> directional couplers.  My thought was that the meter would be the most 
> expensive part, but looking at the price of sensors, that doesn't 
> necessarily seem to be the case, especially if it will take more than 
> one or two to do the job.  The Marconi/Aeroflex 6960B that I originally 
> asked about covers 30 KHz to 4.2 GHz and measures -30 dBm (1 microwatt) 
> to +20 dBm (100 mW), but they make other sensors for the 10 MHz to 
> either 20 GHz or 46 GHz range, and with just two of the right ones, you 
> could measure between -70 dBm and +35 dBm.  But the chances of finding 
> just those two sensors in working condition for less than the price of a 
> new mortgage on my house? slim and none.  Likewise, an HP/Agilent 432 
> with a type 8481B sensor spans 0 to +44 dBm at 10 MHz to 18 GHz, so that 
> would come awfully close to being a perfect combination for my needs.

And while the 432 and 8481B may cover the range, its the least
expensive, but the heads are the most tender. All I have seen have been
limited to +10 dBm (on the display and by the power handling of the
head), none to +44. +44 is at least 20 dB about the point of
destruction. Rated with good calibration to 18 GHz, it probably works
with less predictability and poorer SWR at 24 GHz since the head works
on heat. The meter heats the thermistors in the head, and adjusts the
applied heat downward to maintain resistance and indicates the reduced
power level from the meter. There is nothing inherently frequency
sensitive except the fineness of head RF circuit construction. But to
get the sluggish response (compared to a diode detector) as fast as it
is takes really tiny thermistors that are fragile, easily damaged by DC
spikes, too much RF, too much humidity, and too much shock.
>  
> (Not quite up to 24 GHz, but who knows when or if I'll get there, and 
> not quite up to the full power I will have on 2304 and 3456, but one 
> good high-power -10 dB attenuator or -10 to -20 dB directional coupler 
> would fix that.)
> 
> So, I've been scanning eBay for attenuators and directional couplers. 
> It is easy to find -10dB attenuators but most of them seem to be rated 
> only to 1 or 2 watts at 0 to 2200 MHz, but the problem is that an 
> attenuator is needed the most when there's more power than only 1 or 2 
> watts which needs to be sampled.  I've not yet had success finding an 
> attenuator that has a high enough power rating and a wide enough 
> frequency rating, but I will keep on looking.  I suppose in the worst 
> case, if I found a 100 watt -10dB attenuator that was, for example, only 
> rated to 2200 MHz, I could have someone sweep it to the higher 
> frequencies and develop a "calibration curve" for the actual attenuation 
> at the higher (out-of-rating) frequencies.

The few Narda power attenuators I've seen or acquired at hamfests have
proven to have been over done and so aren't worth their weight in scrap
metal.
> 
> Most eBay listings I've seen so far for directional couplers don't seem 
> to mention anything about power ratings.  I guess from the theory of how 
> it is supposed to work, you put RF in one end, and 1% (-20dB) comes out 
> the sampling port and the other 99% comes out the output end, and there 
> is minimal power to actually dissipate within the device itself.  So am 
> I correct in assuming that I could probably put 100 watts into it, have 
> a dummy load or antenna on the output end, sample my 1 watt from the 
> sampling port (which now is within the power range of one or more 
> inexpensive -10dB attenuators as needed), and not worry about blowing up 
> the directional coupler itself?

Waveguide directional coupler power is limited by the waveguide and the
connectors. Strip line directional couplers (like the many coaxial with
SMA connectors from Narda) can't handle so much power.
> 
> Thanks for all the input, and the kind offers of help with setting up my 
> systems.  I will definitely be looking forward to the day (hopefully 
> soon) when I get things assembled to the point where I'll be ready for 
> using some of that help.
> 
> 73 de W0JT

Now as a practical matter for amateur applications, you aren't likely
above 1 GHz to be creating power approaching FCC limits so laboratory
precision in checking that power isn't very critical so the relative
indications of diode voltmeters (and you don't demand sensitivity below
1 mW) can do all you really need. Diode probes in coax connectors can
have wide bandwidths and with capacitive dividers before them (or
resistive attenuators) can indicate relative power reliably at low cost
even if bought new. You mostly need to tune for maximum output and you
are on the air.
-- 
73, Jerry, K0CQ
Dr. Gerald N. Johnson, electrical engineer
All content copyright, Dr. Gerald N. Johnson