At the low end the models are very, very good. Above 10MHz you need to really get imaginative about modeling the parasitics, which include the cabling *and* the environment (It's surprising the effect a shielded room can have on your readings)
Note at the low end to be sure to include energetic magnetic fields, like those from a wire supplying 20A spike through some such switching. Those magnetic fields inject a lot of signal elsewhere. Note also if you do anything active *ALWAYS* go back and test with a sweeping generator and an SA. Look for out of bound fundamentals generating inbound spurious tones. ...caught me off guard once. Never noticed the effect using the tracking generator option, duh! - Robert - -----Original Message----- From: Jim Eichner <jeich...@statpower.com> To: 'EMC-PSTC - forum' <emc-p...@majordomo.ieee.org> List-Post: emc-pstc@listserv.ieee.org Date: Wednesday, September 15, 1999 10:09 PM Subject: Re: Sweeping components and filters > >Thanks, Robert and Ed, for your replies. I have included them below, >along with my original request. > >I'm guessing by the low number of responses, that not many people do >this, which I guess isn't that surprising considering I too have been >getting by without. Still, I'd appreciate any further input people >have, particularly regarding methods that extend up into the 10's or >100's of MHz. > >Ed's suggestion to use a sweeper with our SA is an interesting one. It >hadn't occurred to me that I don't need to synch up the two instruments, >but can just let the display fill in. If Ed or anyone else can add a >catalogue number or two to the list of mfr's (Wiltron, Wavetek, and HP) >that would save me reinventing the wheel. Also, Ed, is this method >useable over a wide frequency range, into the 100-200 MHz neighbourhood? > >Robert: You say you sweep the components, then model the filter, then >sweep the filter and compare to the simulation. How good is the >correlation between the model and the network analyzer sweep? How good >is the correlation between either and the real-life performance of the >filter when you drop it into your noisy circuit? > >Finally, the idea of upgrading my SA to add in the tracking generator is >still kicking around my head. Any input people have regarding the >usefulness of the SA/TG combo for sweeping filters would be appreciated, >along with any other things that people use the combo for. > >Thanks for your help, > > >Jim Eichner >> Senior Regulatory Compliance Engineer >Statpower Technologies Corporation >jeich...@statpower.com >http://www.statpower.com >> phone: (604) 420-4820 ext. 2162 >> fax: (604) 420-1591 >Any opinions expressed are those of my invisible friend, who really >exists. Honest. > > >-----Original Message----- >From: owner-emc-p...@majordomo.ieee.org >[mailto:owner-emc-p...@majordomo.ieee.org] On Behalf Of Price, Ed >Sent: Friday, September 10 1999 02:21 >To: 'Jim Eichner'; 'EMC-PSTC - forum' >Subject: RE: Sweeping components and filters > >Jim: > >OK, here's my 2 Euro's opinion! > >If you are starting from scratch, then buy a used HP-141T mainframe with >an >8552 and 8556 plug-ins. This will give you a spectrum analyzer with >analog >memory, covering 20Hz to 300 KHz. And it has a tracking generator to >provide >the excitation. Later you can buy some other plug-ins which will give >you a >pre-compliance tester for CE & RE. Here in the USA, you can get this for >under $1000. > >OR > >If you have a spectrum analyzer with a memory, you can use a sweeper as >the >source. Wiltron and Wavetek and HP sweepers can be obtained on the >surplus >market for several hundred dollars. No need for any sync between the >two. >Set the sweep ranges for both, and let them run for a while until you >fill >in the spectrum of interest. > >You can kludge up other combos with sweepers and scopes/RF VTVM's etc, >but I >don't recommend it. > >Hey, why not do it all with software? I have a project (on very low >priority) to use a PC's sound card as both an excitation source and a >spectrum analyzer. I bought the Spectra Plus software for $300, then am >using a Media Vision Pro sound card (also have a Turtle Beach Fiji with >better S/N ratio) to generate a sweep and track/display the response. >The >only down-side to this is that you need to pad the output and input >ports >with 50 ohms (to duplicate the way OTHER filters are almost always >tested). >And, the frequency range is limited to about 30Hz to 22 KHz (because the >sound cards sample at 44 KHz). > >Regards, > >Ed > >-----Original Message----- >From: Robert Macy [SMTP:m...@california.com] >Sent: Monday, September 13, 1999 2:09 PM >To: Jim Eichner; 'EMC-PSTC - forum' >Subject: Re: Sweeping components and filters > >I use that HP Network analyzer that goes to 500MHz. On a component by >component basis. Then shove that all into a PSpice circuit which allows >me >to add parasitics, etc. > >Then I can play games with waveforms, risetimes, components, etc and >pretty >much predict what the filter will do. > >Sometimes as a confirmation, I'll then put the whole filter back into >the >network analyzer and see how it compares. Performance versus frequency. > >It is true that different noise sources inside the unit have different >characteristic impedances and as with any filter, you need to match >those >impedances to what you're doing in order to maximize the performance of >your >filter. > >In otherwords, if this is a custom filter, make it match the >requirement, >don't make a generic one. > > >- >Robert - > > >> -----Original Message----- >> From: Jim Eichner [SMTP:jeich...@statpower.com] >> Sent: Thursday, September 09, 1999 4:03 PM >> To: 'EMC-PSTC - forum' >> Subject: Sweeping components and filters >> >> >> Greetings all: >> >> I have long wished that I had an instrument for sweeping filters and >> components that I am working on, to determine where I can expect the >> usefulness to start rolling off, where and if there are resonances, >> etc. >> Several options have crossed my mind, including adding a tracking >> generator to my spectrum analyzer, buying an impedance analyzer, using >> a >> network analyzer, etc. >> >> I'd appreciate any experiences, advice, and recommendations the group >> has to offer. I'm looking for input on cost, effectiveness, usable >> frequency range, methods, issues like minimizing (or nulling?) >> parasitics, effects of impedance mis-match (the instrument may be >> 50ohms >> but there's no way the circuit being filtered will be!), etc. >> >> I'll gladly collate the replies and send them back to the group once >> the >> discussion's done. Please reply to the forum, not to me directly, so >> we >> get some dialogue going. >> >> Thanks >> >> Jim Eichner >> > Senior Regulatory Compliance Engineer >> Statpower Technologies Corporation >> jeich...@statpower.com >> http://www.statpower.com >> Any opinions expressed are those of my invisible friend, who really >> exists. Honest. >> >> >> --------- >> This message is coming from the emc-pstc discussion list. >> To cancel your subscription, send mail to majord...@ieee.org >> with the single line: "unsubscribe emc-pstc" (without the >> quotes). For help, send mail to ed.pr...@cubic.com, >> jim_bac...@monarch.com, ri...@sdd.hp.com, or >> roger.volgst...@compaq.com (the list administrators). >> > >--------- >This message is coming from the emc-pstc discussion list. >To cancel your subscription, send mail to majord...@ieee.org >with the single line: "unsubscribe emc-pstc" (without the >quotes). For help, send mail to ed.pr...@cubic.com, >jim_bac...@monarch.com, ri...@sdd.hp.com, or >roger.volgst...@compaq.com (the list administrators). > > --------- This message is coming from the emc-pstc discussion list. To cancel your subscription, send mail to majord...@ieee.org with the single line: "unsubscribe emc-pstc" (without the quotes). For help, send mail to ed.pr...@cubic.com, jim_bac...@monarch.com, ri...@sdd.hp.com, or roger.volgst...@compaq.com (the list administrators).