Depends on the SNR and how far out of focus. If there's enough actual signal there to allow you do do Fienup phase retrieval, you can do some surprising things. If you're far enough out of focus, the optical transfer function (*) will have zeros in it (real zeros, not complex zeros) so information is indeed lost.
Cheers
Phil Hobbs
(*) The OTF isn't a real transfer function, but don't get me started.
-- 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
Sorta. Camera shake and subject motion can be helped quite a lot, but intensity-only images don't contain enough info for real deconvolution (i.e. complex-valued filters in the frequency domain) unless you can do phase retrieval on them.
If you have interferometric data, you can do numerically anything you can do in real space with lenses and such, such as refocusing. It's when you take the mod squared of the fields by shining them on an incoherent detector that you lose that ability.
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
Jeff,
A LISN isolates measurements from Mains interference - standard 50ohm outputs present noise generated (ideally) by DUTs connected to the stabilized impedance power output terminals.
While the 50ohm outputs are HF pass filtered, they do not protect instrumentation from signal transients or surges on the line, or those created by connecting or disconnecting the power line.
You need a proper attenuator/limiter before the input of an SA.
RL
This particular one also won't tolerate *any* DC at all at that input. Any idea what protective device I could use to protect from that which won't give rise to BWL constraints or whatnot?
Yep. I was more worried about the OP using a direct probe across the power line. Get the hot and neutral backwards, and that puts 240VAC on the spectrum analyzer case.
I agree that the LISN (line impedance stabilization network) does not officially provide any instrument protection. However, the receiver or spectrum analyzer is usually capacitively coupled through about
0.1uf, which does help. That's about 30K reactance at 50Hz which into a 50 ohm load will reduce the 240V line voltage to about 0.4V, which should be adequate protection. See the various LISN schematics:However, you're correct that it won't stop power line spikes from blowing out the spectrum analyzer. On the LISN that I built, I installed several MOVs (metal oxide varistor) across the line as a precaution. I think I put a small fuse on the measurement port, but I don't recall. My LISN was fairly crude and ugly as it was only intended to work up to 30MHz. I was going to build something like these: or buy the originals: but instead found some that I could borrow.
I don't think it's necessary to protect against inductive spikes caused by connecting or disconnecting the power line. I usually wire everything together and then flip the power switch. I don't pull the plug on a live and powered device. Although I've blown up some expensive test equipment with transmitters, I haven't had any problems during EMC testing.
-- Jeff Liebermann jeffl@cruzio.com 150 Felker St #D http://www.LearnByDestroying.com Santa Cruz CA 95060 http://802.11junk.com Skype: JeffLiebermann AE6KS 831-336-2558
I don't have to bother with half of that fiddling about. I've taken snaps and uploaded them cropped and enhanced many *hundreds* of times over for my ebay listings and it takes me < 20 seconds per pic (because I've done it so often, basically)
For zero DC, you want a shunt inductor, or transformer, of course. You get the BWL because you WANT it, to have zero DC, and the whatnot is a free side effect! Enjoy!
On a sunny day (Thu, 25 May 2017 17:43:52 -0000 (UTC)) it happened Cursitor Doom wrote in :
Series capacitor and 2 diodes anti-paralel, but which diodes...
mains >--||---------------> instrument | | --- --- \ / / \ --- --- | | /// ///
On top of that since a few weeks I have VDRs and a sparkgap... Some kV spikes seem common here if the power company is at it. One killed my laundy dryer, had to replace the LNK304PN chip in it.
A transmission line transformer with back-to-back Schottky diodes across the secondary would be a good start.
For low frequency stuff I have a little Pomona box with a big cap in series and two Schottkies in shunt. Just don't reduce the attenuation on the SA input--that 10 dB is what makes the external diodes take the hit instead of the internal ones.
I still disconnect it at the SA before making any serious voltage change on the input.
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
It will handle a bit more than "any". The HP 8566B data sheet on Pg 14: proclaims that the input attenuator will tolerate up to 100ma DC input (1/2 watt) before ruining your day.
RF input 100 Hz to 22 GHz, precision type-N female connector, dc-coupled Maximum input level ac Continuous power: +30 dBm from 50 ohm source Mixer protected by diode limiter, 100 Hz-2.5 GHz
Incidentally, the HP 8566B mostly sold on eBay for between $400 and $700 with $100 to $200 shipping:
-- Jeff Liebermann jeffl@cruzio.com 150 Felker St #D http://www.LearnByDestroying.com Santa Cruz CA 95060 http://802.11junk.com Skype: JeffLiebermann AE6KS 831-336-2558
Congrats. Just one problem. A leaky IF section will not display what appear to be genuine FM broadcast signals. To verify this, just tune the SA front end frequency in either direction. If the signal moves on the display, it's coming in via the RF section. If stays fixed and does NOT move, it's coming in via the IF section.
-- Jeff Liebermann jeffl@cruzio.com 150 Felker St #D http://www.LearnByDestroying.com Santa Cruz CA 95060 http://802.11junk.com Skype: JeffLiebermann AE6KS 831-336-2558
y what
dThe 8566 has an unusual frequency plan that makes this not quite true, iiuc . That's also why there's no tracking generator for it.
Cheers
Phil Hobbs
Might be somewhat related that my Trendnet PowerPlug AV setup that used to work for me for years to carry ethernet over the power wires in my house stopped working a few weeks ago. Now it only works for much shorter distances, so I know the adapters are still working. Too bad because the setup was used by IP cameras that were just useless with WiFi connection.
I looked for a block diagram of the HP 8566 but couldn't find one. So, I assumed that it had a fixed IF frequency. I'll download the full manual later, and see how it really works.
There is a do it thyself tracking generator for the HP 8566:
"A 2.5-GHz Tracking Generator for the HP 8566A/B Spectrum Analyzer"
"Making a Tracking Generator for the HP 8566B Spectrum Analyzer"
-- Jeff Liebermann jeffl@cruzio.com 150 Felker St #D http://www.LearnByDestroying.com Santa Cruz CA 95060 http://802.11junk.com Skype: JeffLiebermann AE6KS 831-336-2558
Slightly newer version at
-- john, KE5FX
I have another HP SA here which has PSU-style banana plug terminals for inputs! Must be for audio testing, I haven't taken a close look yet.
I already have a free-standing HP VNA somewhere. 'Only' goes up to
1.3Ghz, but that's *plenty* for my purposes.
That's only if the attenuation is set high enough to protect the diodes.
The 8568 has an AC-coupled input, so as long as you don't have nasty transients, the input mixer is much safer.
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
Well, let's see what I can blow up in the HP 8566B A6 assembly. Full searchable service manual (23MB): Extracted from the block diagram on PDF Pg41:
Looks like I have the options of destroying the: input attenuator (A6A2) RF switch (A6A1) Diode limiter (A6A14) 1st converter mixer (A6A6) If I set the input attenuator to straight through (0dB atten) it's effectively out of the circuit. There's no series resistance so the diode limiter is going to take the hit. In theory, the diode limiter is suppose to protect the following 1st converter mixer circuitry, but with my luck, an accident always destroys the most expensive component first. So, the diode limiter would need to be sacrificed for a DC coupled transient hit.
I don't know why HP chose to DC couple the HP 8566B. My guess(tm) is that with a frequency response going down to 100Hz, an input coupling capacitor would need to be 60uF or larger into 50 ohms, which is impractical.
After blowing up considerable test equipment with various transmitters, I tend to be a bit paranoid with input protection. Most of my signal sources have reverse power protection modules or add-ons installed. For the spectrum analyzers, I use a 10dB attenuator at all times. I prefer to blow up a cheap 50 ohm attenuator, instead of an expensive spectrum analyzer input attenuator. I only remove it when I'm absolutely sure the test source is clean and free of surprises. So far, my luck is holding.
Spec sheet for HP 8568B: Looks like it's not AC coupled down to 100Hz, but instead quits at 100 KHz for AC coupling. Looking at Pg16, I find that it has two RF inputs:
RF inputs 100 Hz to 1500 MHz, 50 ohm dc coupled (BNC fused) 100 kHz to 1500 MHz, 50 ohm ac coupled (type N)
Max input level ac: +30 dBm (1 watt) continuous power:
When I tried to find the HP 8568 data sheet on the Agilent web pile: the download link delivered the HP 85685 preselector data sheet instead. Sigh. Still, it's interesting to see what it does for input protection: On Pg 6, it has a fuse on the RF input connector and is rated to
+30dBm (1 watt).You really shouldn't have brought up the topic. I'm now lusting after an HP 8568B and searching eBay for prices. About $750 including shipping for both sections, cables, and claimed to be working. Very tempting.
-- Jeff Liebermann jeffl@cruzio.com 150 Felker St #D http://www.LearnByDestroying.com Santa Cruz CA 95060 http://802.11junk.com Skype: JeffLiebermann AE6KS 831-336-2558
Model number please?
My crystal ball deduces that you have an oscilloscope mainframe, with some manner of HP spectrum analyzer plugins. As I vaguely recall, the banana jacks were either for scope probe calibration, or DC power to an external powered probe. My HP 141T has those labeled "CAL 10V", "1V", and "GND". Note that not all versions of the HP 141T have these jacks.
If you happen to have an HP 141T mainframe, and you want to do audio, the HP 8556A plugin should work. 0 - 30KHz and 0 - 300KHz ranges. I have a broken one that I probably should fix, some day, maybe...
-- Jeff Liebermann jeffl@cruzio.com 150 Felker St #D http://www.LearnByDestroying.com Santa Cruz CA 95060 http://802.11junk.com Skype: JeffLiebermann AE6KS 831-336-2558
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.