what do you think the amplifier circuit for EMC monopole e-field antennas is ?

what do you think the amplifier circuit for EMC monopole e-field antennas is ?
Posted by coppercone2 on 15 Aug, 2018 01:04
http://www.ets-lindgren.com/datasheet/antennas/rods/4010/401001

The standards and manual provide information on what the response is, but what kind of topology is good to use here as a design start point?

http://www.ets-lindgren.com/sites/etsauthor/ProductsManuals/Antennas/399792%20D%203301C%20Manual.pdf
http://everyspec.com/MIL-STD/MIL-STD-0300-0499/MIL-STD-461_8678/

#1 Reply
Posted by hwalker on 15 Aug, 2018 01:39
I worked in EMC for a number of years and the common amplifier design for EMC monopoles (usually 10 kHz - 30 MHz) was a high impedance FET amplifier. Matter of fact, most of the repair problems with the antenna was caused by blowing out the FET input due to not grounding the input properly during handling.

Herb

#2 Reply
Posted by coppercone2 on 15 Aug, 2018 19:19
Do you directly connect the counterposie and antenna directly to the input or do you still use some kind of PIP or IPI network between the antenna and the amplifier? I don't see how you could match something to the input of a amplifer input impedance if it is High-Z.

What kind of input impedance and bias current would I look for in an op-amp? Teraohms? Gigaohms? I don't even know how stable the input impedance of such high Z amps is honestly.

I figure a op-amp would be a good choice? Or JFET/opamp hybrid?

Also the standard agrees with the frequency range you say and it says the antenna is 41 inches with a counterposie of 60cm per side.

Based on what you are saying a good option would be to put a normally shorted relay on the antenna so you disable it when you power it off.

Can you still use coax to connect an antenna like that? The impedance mismatch seems crazy. I think if its not impedance sensitive you would use a shielded twisted pair?

I am guessing the parastics of any high value resistors at these ranges are so big that you could not use one to make a fixed stable input impedance in combination with the amplifier where the amplifier Z is >> then the pull down resistor which is still high enough Z for the antenna? Or some kind of active circuit the regulate the input impedance? Using a active element in a impedance sensitive location as a input impedance seems odd for metrological purposes.

Any idea what the impedance of the antenna might be close to so I can do some rules of thumb to see what kind of mismatch difference there is if something starts drifting in such terrifically high ranges?

I did not have this concern with my active loop antenna since the input is differential and the input impedance of the amplifier should track each other, so I overlooked it and no one seems to do anything about it anyway. I am guessing the leaked (in anechoic) and atmospheric noise is so great that it does not matter and its stable as a saint for all intents and purposes?

#3 Reply
Posted by hwalker on 15 Aug, 2018 22:27
From the 41 inch and counterpoise references, it sounds like you are referring to the MIL-Std-461 test set-up. The counterpoise is an extension of the ground-plane and up to MIL-Std-461E was connected to the metal test table by a 1 meter long sheet of copper or aluminum. The antenna was spec'd down to 30 Hz for MIL-Std testing and therefore the 41" rod was dc coupled to the input of the FET amplifier. Our lab used the EMCO 3301B where the hi-Z FET input simulated the hi-Z e-field of free space. I don't have my old references anymore, but to match the 50 ohm receiver input, there must have been a 50 ohm matching network after the FET amplifier. I've been retired for a few years now and always have trouble locating my old notes when I need to refer to them these days. I once pulled apart a "not economical to repair" unit and was actually surprised at the simplicity. Since the unit was only designed to work up to 100 MHz at most, the layout looked to be basically a MPF102 FET followed by an op-amp for low output impedance matching. There was also a switched filter to attenuate powerline noise in-case you didn't need to go down to 30 Hz. There is a specific calibration procedure that gives you antenna factors at each frequency for the 41" telescoping rod. If you run into a EMCO 3301B on eBay, that would be your best bet for getting the antenna without paying over a $1,000 to EMCO.

I'll see if I can dig up my old test notes if you think it might help.

-Herb

#4 Reply
Posted by coppercone2 on 15 Aug, 2018 23:12
All the filters were after the signal was buffered right?

Did not realize the rod was telescoping.

#5 Reply
Posted by coppercone2 on 16 Aug, 2018 19:03
oh boy these things have input impedance and suceptance plots instead of what I expected. Not sure if I ever saw a opamp that had acceptance and admittance graphs. Then again I never did much at these frequencies. I want to see if I can get it done with a single part rather then a hybrid amplifier.

Looks like the input impedance is something like 1.3 Megaohms @ 30MHz, at least for the real term

Did you have any protection diodes?

#6 Reply
Posted by IconicPCB on 16 Aug, 2018 21:55
Consider using a dual gate mosfet in cascode configuration.

And You should use inverse parallel low capacitance diodes as protection devices.

#7 Reply
Posted by coppercone2 on 16 Aug, 2018 22:20
I am looking at DGM, I never heard of them before.

Am I right with the math here? The transistor has specs

admittance : 0.7 millimillihos, 1.4Megaohms
conductance : 0.5 millimillihos, 2Megaohms

Antenna impedance ~ 36.8 ohms (1/4 wave dipole)

VSWR : 114906
Mismatch Attenuation : 44.583 dB

I just used ~33MHz as a test point.

https://chemandy.com/calculators/return-loss-and-mismatch-calculator.htm

So I have attenuation of 141x at this frequency

If I do it 10MHz less, at ~23MHz,
Admittance: 0.45 mmHos, 2222222 Ohms
Conductance: 0.3 mmHos, 3333333 Ohms,

Antenna impedance ~ 36.8 ohms (1/4 wave dipole)

VSWR : 196256
Mismatch Attenuation : 46.908

So over a 10MHz range the impedance of the op-amp causes a mismatch error of ~2.4dB, and it appears to be linear

is my understanding that the V/M on the antenna basically has a ripple/slope error (if you trust the graph being so strait) of 2.4dB on the 22-32 MHz band, which needs to be corrected with a lookup table thats sloped with a coefficent correction of times ~140 to negate the loss? Is any kind of impedance matching network between the antenna and the buffer possible to try to make the losses less? 40dB is alot of gain at these frequency ranges, the GBW of the amplifier would need to be in the GHz, if you wanted 40dB with 30MHz you would need a gain bandwidth of 4.2GHz, so you likely need to use multiple stages which would increase the noise by a fair bit, I think 4GHz GBW is basically the fastest amplifier that LT/AD makes, and I am worried about using it, a more reasonable current mode amplifier with 1.2GHz bandwidth means you would need to use 5 of them, I figure the quiescent current would be high if you want to make it battery powered like those devices. I guess its not mandatory to bring it up to 40dB because instruments can handle alot less, but I am curious as how these devices are used.

Where do I go from here to find out what volts/meter on the antenna (this is the unit they seem to use) translates to output of the amplifier, so I can set the gain stages? Some kind of impedance match would make it less noisy because there would be less amplified gain required right?

#8 Reply
Posted by hwalker on 16 Aug, 2018 23:42
Still looking for my old test notes. I actually knew the engineer that designed the original EMCO3301B and received a proto-type unit from him for being one of the beta testers (many moons ago). Check on ETS website (emctest.com ??) and see if you can download their EMCO3301B operational manual. It has a lot of valuable information on both the design considerations and antenna factor calibration procedure. It also provides low frequency roll-off curves for the switchable filters and 10 and 30 dB switchable attenuators ( I forgot to mention the attenuators). I have also attached a history of the 41 inch rod that you might find interesting.

-Herb

41 inch rod history.pdf

#9 Reply
Posted by coppercone2 on 16 Aug, 2018 23:45
fasctinating, btw I just realized the unit was MMHOS not MMMHOS (I thought it was just HOS, but its actually MHOS), so my above calculations are off by alot, milli instead of micro. Will redo them in a bit after I read that. I don't like M units anymore.

#10 Reply
Posted by coppercone2 on 20 Aug, 2018 18:52
Do you have any more historical documents like this one relating to electromagnetic mil/commercial standards? I find it completely fascinating and I think having that stuff around would lead to less frustration and more understanding when it means spending money on compliance.

Some of the standards are fairly paranoid/obvious like not having munitions fire wires bundled with other cable runs (the milspec wiring standards I read through), but this document was rather illuminating.

#11 Reply
Posted by hwalker on 21 Aug, 2018 02:44
Most of the reference documents I have are either paper copy or electronic documents that are far to large to share on the forum. If you any interest in specific test procedures and I happen to have something of interest I 'd be happy to share. In case you were not able to download the active rod manuals I previously suggested, I have attached them to this post. They should answer most of your questions regarding how the antenna factors are calibrated. The active rod can be calibrated in a well equipped lab, unlike other EMC antennas which generally must be calibrated at an open area test site.

-Herb

3301B.pdf

Active Monopole Antenna.pdf

#12 Reply
Posted by hwalker on 21 Aug, 2018 23:47
And a write-up by David Weston on the monopole calibration...

1mMono.pdf

#13 Reply
Posted by coppercone2 on 09 Sep, 2018 15:04
Did you ever find yourself in a situation where you found the telescoping to be useful for power reduction (and not just transport)? (as a sort of attenuator as it says)?

I think I have a old stainless steel threaded antenna I found in the street some where but I would need to cut it down, it does not telescope however. Think it was for CB the top was bent anyway.

I also need to make sure it satisfies the thickness for the 10pF (but that actually makes me question it too, the equation in the document where you can use capacitance to solve for diameter of the antenna is for a cylinder,

but the telescoping antenna is actually a kind of cone shape since it gets thinner.. any idea on the error allowed in diameter to deviate from the 10pF, since some how they use a telescoping one?) I did not bother doing the calculation yet because I was confused.

I also don't know much about what makes a good telescoping antenna (in terms of quality).

#14 Reply
Posted by hwalker on 10 Sep, 2018 23:41
I've used both telescoping and fixed length 41 inch elements on the same active rod antenna without any perceptible difference in the emissions profile of the equipment under test. I actually always assumed that the diameter specification referred to the base of the active rod rather than the rod element although I can't find any data in my references that show calculations that confirm that. My experience is that the 10 pf specification is affected more by improper packaging and element length than by the diameter of the element. You must ensure the element is isolated from the case containing the electronics or you will introduce loading effect errors. This true for both the antenna port and the calibration port - if separate from the antenna port. The 41 inch length of the element includes the threaded portion. Use the following MIL-STD verification procedure to perform a sanity check of your calibration factors after your prototype is completely assembled.
--------------------------------------------------------------------------------------------------------------------------------------

MIL-STD – 461E RE102 Active Rod Antenna Low, Median and High Pre-test Injections:

The RE102 requirement for performing the three active rod antenna injection calibrations is accomplished by performing the following steps:

The rod antenna is set up as normal with the 41” (104 cm) rod removed.

A signal generator, tuned to the calibration frequency, is coupled to the antenna rod threaded input jack through a 10pf capacitor loaded with a 50ohm load. A “T” connector, with a 50ohm load is satisfactory.

Set the signal generator injection level based on the following calculation:

LIMIT (at the injection frequency)
-6dB (calibration injection goal of 6dB below limit)
-6dB (rod correction factor)
-Antenna factor (at injection frequency)
= Injection Level (signal generator level setting)

For Example:

      34 dBuV/m (limit)
      -6dB
      28 dBuV/m (target level for calibration)
      -6dB (rod removal correction factor)
      22dBuV/m
      -4.5dBuV/m (antenna factor)
      17.5dBuV injection level

4.   Perform injections at high, medium, and low frequencies of the active rod band.

- Herb