1. Home
  2. Electronics Design
  3. more AD7793 fun.
  4. Page 3

more AD7793 fun.

Only if it's comparable in size to 25 mV.

Let's plug it into the Taylor series and see.

I(V) = exp(Vf/25mV) = I(V_0) exp(dV/25mV)

= I(V_0) * (1 + u + u**2/2! + u**3/3! + ....)

where u = dV/25mV. Taking just the normalized perturbation X,

X = I(V)/I(V_0)-1, we get

X = u + u**2/2! + u**3/3! + ....

If u = A sin(wt), then up to second order,

X = A sin(wt) + A**2(1-cos(2wt))/(2*2!) + ....)

where I've used the usual identity for sin**2.

All the odd order terms are at harmonics of w, but the even order terms also contribute DC, i.e. rectification. You can't get a DC component without an even-order nonlinearity, due to the behaviour of powers of sin(x).

In particular, the linear term A sin(wt) is just reproduced as an RF signal and doesn't do anything to the other things. (First order terms obey superposition, in other words.)

Thus the rectified contribution is

X_dc = A**2/4 + O(A**4).

(It really is quadratic for small signals.)

I'm entirely in agreement about your interference measurements themselves, of course, but it needs more RF than you think.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
ElectroOptical Innovations LLC 
Optics, Electro-optics, Photonics, Analog Electronics 

160 North State Road #203 
Briarcliff Manor NY 10510 

hobbs at electrooptical dot net 
http://electrooptical.net
Reply to
Phil Hobbs
Loading thread data ...

One can Spice a diode detector, but I'm guessing that opamp models do not handle RFI rectification very well, if at all.

--

John Larkin         Highland Technology, Inc 
picosecond timing   precision measurement  

jlarkin att highlandtechnology dott com 
http://www.highlandtechnology.com
Reply to
John Larkin

I'd be very surprised if a behavioural model did a good job of that, for sure.

OTOH it wouldn't be rocket science to do a simple SPICE op amp out of

2N3904s and see.

I've spent a significant fraction of my career trying to make rectifying junctions, so I've been through the theory a lot of times.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
ElectroOptical Innovations LLC 
Optics, Electro-optics, Photonics, Analog Electronics 

160 North State Road #203 
Briarcliff Manor NY 10510 

hobbs at electrooptical dot net 
http://electrooptical.net
Reply to
Phil Hobbs

Fun thread.. mix of theory and practice. I'm talking a bit over my head, but can we put some numbers on the electric field strength (from a cell phone or something.) and guess an antenna size (d) small wrt the wavelength.. and get a voltage that is E*d... So (guessing at numbers) 100 mV/m with a 1 mm length of antenna gives

100uV of RF. (Is that at all close to the ball park?)

George H.

Reply to
George Herold

The presence of the antenna shorts out the field in its vicinity, so it isn 't quite that effective when the antenna is electrically very short.

Cheers

Phil Hobbs

Reply to
Phil Hobbs

And I don't doubt your math :-)

The problem is, I could really make opamps rail with just a few hundred uV of RF into the input (when clients had a hard time believeing the effect). It can't really be simulated. Or maybe with some expensive parasitic extraction tool and fancy simulator but not with SPICE behavioral.

--
Regards, Joerg 

http://www.analogconsultants.com/
Reply to
Joerg

_Hundreds_ of microvolts, I believe. That's roughly the geometric mean of 1 uV and 25 mV. ;)

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
ElectroOptical Innovations LLC 
Optics, Electro-optics, Photonics, Analog Electronics 

160 North State Road #203 
Briarcliff Manor NY 10510 

hobbs at electrooptical dot net 
http://electrooptical.net
Reply to
Phil Hobbs

"Analog Devices" should change its name to "Anal Vices".

Reply to
bitrex

It might be possible for the input diff pair to have actual gain at RF, with the second stage doing the rectifying. That would be more likely at low RF frequencies.

I should remember to hold my cell phone close to any analog-circuit breadboards or first articles. But I've found that RF suceptability tends to happen at fairly sharp resonances, so a sweep would be needed to find the bad spots.

I should get one of those surfboard-type wideband antennas and get more systematic about testing products. I have a few old HP and GR tube-type RF generators that have nice twirly frequency-set knobs and lots of power.

--

John Larkin         Highland Technology, Inc 

lunatic fringe electronics
Reply to
John Larkin

That's a bit graphic. How about Analog Disasters.

--

John Larkin         Highland Technology, Inc 

lunatic fringe electronics
Reply to
John Larkin

That's where it railed. Tens of uV were enough to cause grief in the ADC results. It is a phenomenon that sometimes frustrates design teams to no end.

--
Regards, Joerg 

http://www.analogconsultants.com/
Reply to
Joerg

Possible, although not too likely at 2GHz depending on resistances used around the first pair.

Absolutely. Sometimes all is well with cell phones and then in the field it isn't because the maintenance folks use portable radios at a very different frequency. Also happened: Disturbance only on one particular day. Turned out that was the day when the military held air exercises with low passes of fighter jets. Really low. With their Radar turned on ...

But they'd have to go to at least 3GHz or so. I have a Signalhound TG44A track gen for that, plus a near field probe kit. That usually spots such RF interference problems. It even has a sniffing dog logo on there.

--
Regards, Joerg 

http://www.analogconsultants.com/
Reply to
Joerg

You originally said "microvolts". Those are pretty different numbers with a quadratic effect.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
ElectroOptical Innovations LLC 
Optics, Electro-optics, Photonics, Analog Electronics 

160 North State Road #203 
Briarcliff Manor NY 10510 

hobbs at electrooptical dot net 
http://electrooptical.net
Reply to
Phil Hobbs

I mostly care about suceptability. Anything that I radiate is someone else's problem!

--

John Larkin         Highland Technology, Inc 
picosecond timing   precision measurement  

jlarkin att highlandtechnology dott com 
http://www.highlandtechnology.com
Reply to
John Larkin

It does not short out the field. It supplies energy to the connected load.

Reply to
John S

Susceptibility is what I measure using the TG44A. Its brother SA44B does the radiated part because my stuff often has to pass more stringent EMC than just class B.

--
Regards, Joerg 

http://www.analogconsultants.com/
Reply to
Joerg

Microvolts do show up in the data, just not as hard. But they do show up when you are dealing with systems like the one on the table right now which contains a 22-bit ADC.

--
Regards, Joerg 

http://www.analogconsultants.com/
Reply to
Joerg

Really? If that's the case, a 1-mm piece of #10 copper wire (3.2 microohms), placed in a 1 V/m field, ought to draw a current of

1 mV/m * 1 mm / 3.2 uOhm = 300 A.

which is a dissipation of

(300A)**2 * 3.2uOhm =~ 300 mW

Pretty hot for a volume of

pi * 1.27 mm **2 * 1 mm = 0.005 cm**3.

Or maybe not. ;)

(Antennas that are some reasonable fraction of a wavelength long do much better.)

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
ElectroOptical Innovations LLC 
Optics, Electro-optics, Photonics, Analog Electronics 

160 North State Road #203 
Briarcliff Manor NY 10510 

hobbs at electrooptical dot net 
http://electrooptical.net
Reply to
Phil Hobbs

Really? If that's the case, a 1-mm piece of #10 copper wire (3.2 microohms), placed in a 1 V/m field, ought to draw a current of

1 mV/m * 1 mm / 3.2 uOhm = 300 A.

which is a dissipation of

(300A)**2 * 3.2uOhm =~ 300 mW

Pretty hot for a volume of

pi * 1.27 mm **2 * 1 mm = 0.005 cm**3.

Or maybe not. ;)

(Antennas that are some reasonable fraction of a wavelength long do much better.)

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
ElectroOptical Innovations LLC 
Optics, Electro-optics, Photonics, Analog Electronics 

160 North State Road #203 
Briarcliff Manor NY 10510 

hobbs at electrooptical dot net 
http://electrooptical.net
Reply to
Phil Hobbs

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.