Measurement Testing

Mjr · Site Admin Joined: 13 Jul 2005 Posts: 294

Robert noted:

My lab is thinking ahead towards the 6GHz requirements of 61000-4-3,
and I'm curious what equipment, specifically the amplifier(s), are
being used for this test.

We currently have an AR 60w amp that, unfortunately, only goes to
3GHz. I wish they had bought the model that goes up to 4.2 GHz since
AR has several choices that cover from 4.2 to at least 6GHz.

We also need a signal generator, and have pretty much decided to get
the R&S SMT. We have a horn antenna already that covers that range.
I know we will need a different power sensor for our HP power meter,
and a directional coupler. We will need a better cable inside the
chamber, and I am thinking a Utiflex cable from Microcoax. We have
the Holaday HI6005 field probe.

The amplifier is the big question I have, but I would like to hear
what other equipment everyone is using.

Mjr · Site Admin Joined: 13 Jul 2005 Posts: 294

Ken responds:

TWTAs have lots of bandwidth. If yours stops at 3 GHz, that means beyond 3 GHz you get less than 60 Watts output. But the drop-off is slow, which is why they make multi-band TWTAs in the first place. As long as it is coax output you should measure just how much power you get at 6 GHz, relative to what you need. While the amount of power you need depends on the precise antenna you use, a quick trip to ets/lindgren and inspection of the ubiquitous 3115 antenna shows you ought to be able to get by with 15 Watts or less at the 10 V/m 80% AM level, if you bolt the antenna directly to the amplifier or otherwise minimize cable losses. And that's at three meters.
I don't know what they do with 61000-4-3 above 1 GHz, but technically there is no need to illuminate a 1.5 m square at these frequencies. If you can get away with the antenna at one meter separation of antenna and EUT and multiple scans to cover the entire test set-up - technically justified but increases test time - then you need less than 2 Watts. Well worth checking into.

Mjr · Site Admin Joined: 13 Jul 2005 Posts: 294

Robert notes:

The existing amp is solid-state, not TWT. I'm not sure what the drop-off is after 3GHz, but I will check.

One problem I am finding with broadband TWT amps is the harmonic content at the low end of the frequency range. It is quite possible the 2nd harmonic is as strong as the fundamental. (Most of the TWTs that I find cover one octave).

I am already planning on moving the amp to eliminate as much cable as possible. I may even rearrange the amps in the rack to put the high freq amp down close to the chamber bulkhead fitting. I can eliminate the under-floor cable completely if I do all of that (about 5m of heliax cable). The last option would be to place the high-freq amp inside the chamber. I'm thinking of everything to reduce loss, those high-freq amps get expensive really quick as you go up in power (a squared function, it appears!!!).

Painting the EUT in multiple passes is an option. Most of the EUTs that we test will fit in 1 x 0.5 m area, so we may calibrate to that size uniform field, and for the rare cases we have anything larger then will will make multiple sweeps.

Mjr · Site Admin Joined: 13 Jul 2005 Posts: 294

Ken notes:

If it is solid-state, then you won't find the extended usable bandwidth.
You are also correct that TWTAs have very high 2nd & 3rd harmonics. With a TWTA, it is often helpful to have (an) octave band filter(s) to knock down the harmonics.

If I am paying kilobucks for ten or twenty Watts to cover one microwave
octave, I am powerfully motivated to make a direct connection between
transmit horn and amplifier output. Since the horn is directional, the
amplifier directly behind it won't affect the antenna pattern. If there is
too much loss between signal generator and amplifier that way, a line-drive type amplifier to get you from -20 dBm to 0 dBm is much less expensive than an amplifier that gets you from 0.1 Watt to 10 Watts.

Mjr · Site Admin Joined: 13 Jul 2005 Posts: 294

Bob replies:

I believe I will not have too much problem with the signal from the sig-gen to the amp. Most amps have a 0dBm input max, the sig-gen will have a +13 dBm output (but I will have to keep it below +8dBm to leave room for 80% AM). I think I will be able to run a low-loss cable from outside the chamber and not have to do anything else special. If I have to, I will locate the sig-gen with the amplifier rack, I will already have GPIB control going there for the power meter anyway.

If the amp rack must go inside the chamber, then so be it. I have a small 19" rack that I can put the amp and power meter in. I also have two gpib-coax extenders that I can use to carry gpib control into the chamber.

BTW: Someone at AR, regarding our 3GHz amp, noted: one should be able to extend the top end to 4GHz, but not at full power. Should be fine for most uses.