ee books

Generally being off by a factor implies multiplication. Being off by a factor of two means multiply by two. In this case here, the error is some fixed constant rather than a multiplier.

Or am I interpreting you you incorrectly?

Whom are you replying to? At least some of the folks who ignore you might be doing it because of the lack of context in your posts. :)

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

Thanks!

No, I haven't run across any documents that talk about the E5500 file struc ture, but then I haven't gone looking for any. I did post a beta update re cently

to import result files from t he older HP 3048A system, either emulated in HTBasic or running Kent Falken stein's VB6 port. If the E5500 software can export legacy HP 3048A result files in little-endian BDAT format, that might be one way to get the data i nto TimeLab.

Failing that, send me a file from the E5500 and I'll see if I can do anythi ng with it when I have some time. I don't have any E5500 gear myself.

I can't speak to the Gigatronics block diagram, but my 83651B sweeper uses a 2-7.5 GHz YIG oscillator followed by a YIG-tuned SRD multiplier.

There are 50 GHz YTOs (e.g. Avantek AV-33050) but good luck finding one. M ultiplying and/or mixing would be a safer bet.

Yep, the HPJ issues are the best starting places to understand all of the o ld-school iron.

What's important is the reason for the high noise level, though. The M/N s ynth is responsible for most of that noise (see chapter 11 of www.hpmemory. org/an/pdf/an_283-3.pdf for more on that).

Ideally, the M/N synth wouldn't have been tunable at all; they would have j ust multiplied 400 MHz up to the desired YTO harmonic from 2-6 GHz and used the offset synthesizer to bridge the resulting gaps. A swept synthesizer with the necessary coverage was impractical at the time, hence the 20-30 MH z range of the existing one. But that's where your DDS would come in.

No prob, should be a fun and educational project!

-- john, KE5FX

I don't have one either but I'm making my own unit and I need to find a way to emulate something that your TimeLab will accept. I downloaded a lot of files related to the HP 3048 and related equipment. Maybe one of them will describe the GPIB interface well enough to get the data format.

I did a quick scan - most vendors stop at 40GHz. The 83651B is the only one I found that goes to 50GHz.

[...]

A lot of companies seem to use them. They talk about features and specifications that probably can only be obtained with a DDS. One company proudly explains that the spurs always fall below the phase noise shelf so they are not a problem. That says a lot about their phase noise. Another talks about phase continuity when changing frequency. I think DDS is the only method that can do that.

The AD9912 phase noise seems to be quite a bit better than most, especially in the 20MHz to 50MHz range. I'm hoping the first cut will be usable as a base reference and work from there. I have some tricks to reduce the phase noise from the YIG and the phase-frequency detector in the pll and it will be interesting to see how well they work.

Thanks again for your help.

It's being off by a an integer multiple of some constant

In the case where the result fits the integer multiple is 0

--
umop apisdn 


--- news://freenews.netfront.net/ - complaints: news@netfront.net ---

Yes, you are right, it's not straight-forward to come up with a criterion. It all depends on the process that creates the signal to be filtered. If you cannot bound that in some way, this trick won't work.

Jeroen Belleman

OK, I think I comprehend now. It just appeared to me at first as if this was being presented as some sort of magical shortcut, instead of an optimization that might turn out to be more trouble than it was worth.