really nice Rigol scope

I don't see why it's only an option since it takes relatively little to get it from the signal generated in the analyzer.

Ahh, I guess I don't know how the spectrum analyzer works. I was thinking it did an FFT. But it does a frequency sweep?

The SRS770 FFT spectrum analyzer has this 'chirp' source that puts out a bunch of sine waves centered on each 'bin' of the FFT spectrum. It's a lot faster at getting a filter response than just a noise source. (No wasted spectrum)

George H.

ed text -

It's a $1000 option (20% or 25% more depending on model)

Maybe because it allows them to advertise a lower starting price.

Nice gizmo. Pity about the -80 dBc phase noise at 10 kHz. :(

Cheers

Phil Hobbs

I like it. In the past I was liked Tek scopes (and HP/Agilent spectrum analyzers) better but I think the 3000X beats the current Tek crop. The generator is a pretty nice feature, too. Tek loses there.

I'll have to try it. I haven't explored too deeply, yet, but what I've used I like.

I haven't used a LeCroy for *years*. Then, it's UI was terrible. Never wanted to deal with that again. In fact, I don't remember having anything other than a Tek or HP/Aligent scope at work, since college. ;-)

Tek's new MDO has a mode a lot like that with the spectrum analyzer. They have a time plot of the spectrum (frequency on the horizontal, time-vertical, and color=magnitude). It's really slick, but even better is the time correlation between the time and frequency domains (digital is too easy ;-). OTOH, 28K is steep for bench test equipment.

;-)

We do a lot of time delay and RMS jitter measurements, and have been using old HP 5370 counters, which have 20 ps single-shot resolution and about 30 ps RMS jitter. They are getting hard to maintain, and Agilent finally came out with a new design that does 20 ps.

So we bought one. But they really screwed it up. On the 5370, you could push a couple of buttons and select a sample size over which it would measure time interval, and continuously report the mean and RMS deviation... perfect. The 53230A can't be made to do this simple function. We battled through the complex menus looking for a way, and finally gave up and called them. Confirmed: it won't do it.

(Incidentally, the 5370 managed to process 2000 samples per second using a 6800 CPU! That was 8-bit n-channel, no hardware multiply, 1 MHz clock!)

You can get a histogram of time delays, but it pre-averages shots (reporting femtosecond bins, totally absurd) and the statistics build forever once you start it running. That's doubly useless if you want to see jitter change as you adjust something, which is a really fundamental thing to do.

The only way to do useful jitter measurement is to talk to it remotely, extract single-shot time interval measurements, and do the math yourself on a computer.

What a step backwards!

At least it comes standard with a TCXO. On their older counters, they wanted an extra $1000 for that; it probably costs them about $15.

Exactly the same thing they screwed up on the counter. The programmers must be mathematicians who don't know much about what people really need.

Don't buy a 53230A unless they fix this.

Absolutely! That measurement is fundamental.

It is a pain to have to

I'm going to get another Rigol scope for home. And maybe another one for the cabin in the mountains. I already have power supplies and DVM/thermocouple things both places. A big bench supply makes an excellent car battery charger.

I have 4, LG-532 Lambda power supplies at home base..

Those are 0..5 amps, 0..40 volt digital read out.

I have all four set up for operation at the bench, makes for easy multiple voltage references when testing things. :)

Of course, I do have a couple of heavy duty supplies when needed.

Jamie

Don't care.

Nothing you have to say is worth plowing through four pages of untrimmed quotes to read.

nb

spectrum

asking

I have to find some big supplies. I'll probably need a 16V, 40A supply, soon. One of the other guys was having trouble getting a low enough output impedance to bring up one of the power supplies (25V, 16A boost) and had to resort to a car battery.

Another thing I'm looking for is a good four-quadrant supply, about 0-6V at perhaps 10A.

Then don't. Stop whining, kill the thread, and move along.

Just goes to show the asshole you are, otherwise, you wouldn't of bother to reply if you didn't care to comment. Which means you must of actually thought what I had to say was of some interest. It's nice to think I have fan club out there. ;)

In other words, if it does not interest you, don't bother to read it or reply to it. Of course, you could be another parasite in the group?

Seems to be a lot of insignificant enigmas here lately.

Jamie

Yer obviously too stupid to get the point. No surprise.

nb

Wise to put a big diode in series with the output. Stops the battery trying to discharge into the PSU, if the power goes off. Some PSUs won't stand that.

Doubly necessary, if the PSU has overvoltage crowbar protection that might go faulty ;-(

Effectively a swept receiver. Triple, quadruple, or higher order conversion, with switchable final IF bandwidth, and logarithmic response. Keeping all the various LO and IF frequencies out of where they don't belong is a major contributor to cost. Dynamic range in excess of 100dB. Resolution bandwidths down to below 30Hz are commonly available, even whilst sweeping over several MHz.

Spectrum analyzer users should be familiar with using slow sweeps at narrow RBW, already.

Accurately leveled (or should be).

Not *that* hard to implement, derived from the analyzer first LO.

And still, you are replying.

Need I say any more about your problem?

Jamie

That thing on the right side of your screen is call a "scroll bar." Try it.