DIY (not)tracking generator for the spectrum analyzer

[Miti]

Some time ago I was looking for a (cheap) solution to test and tune some RF filters. Having a SA without tracking generator I thought, how hard can it be to make a simple tracking generator. Not hard at all, I found out, if your SA has an LO output, typically the 1st LO.
http://www.kerrywong.com/2016/01/10/building-a-tracking-generator/  and some videos on youtube.
But my SA doesn't have it so, that's it...or is it?
I also have one of these relatively cheap windfreak RF generators
https://windfreaktech.com/product/usb-rf-signal-generator/

Well, not exactly that one but an older version of that one. This generator can do freqency sweep.
So I thought, what happens if I set the generator to do a sweep and enable the MAX HOLD on the SA.
Below is the response of a band pass filter. The first picture is the straight through connection between the generator and the SA, pretty flat, there's no need for normalization. The second picture is the filter response. Pretty sweet...
Sure you'll have to wait 2 - 3 minutes, it is not easy to tune a filter with it but it is doable.

[Miti]

And no, the SA doesn't have to be an HP, it can be an RF Explorer. I have an RF Explorer 3G Combo. Unfortunately the 15-2700MHz input is too slow but if your filter is within the 240-960MHz, it is doable.

[hendorog]

Nice work - however the max hold is a well known method.

What is your SA? It looks like a big old HP of some description, I'm guessing you probably could get an LO out of it if you tried as they are usually modular and internally connected by cables. So an unequal splitter or a directional coupler in the right place might do the trick.

[Miti]

Nice work - however the max hold is a well known method.

What is your SA? It looks like a big old HP of some description, I'm guessing you probably could get an LO out of it if you tried as they are usually modular and internally connected by cables. So an unequal splitter or a directional coupler in the right place might do the trick.

I couldn't find practical results using this method, all I found was the real tracking generator method using the 1st LO. Could you post some links?
However, I don't intend to patent it, I don't doubt that somebody thought about it before, I just wanted to show practical results that I thought are better than expected.
It is an oldie but goldie HP8594E.

[Performa01]

I couldn't find practical results using this method, all I found was the real tracking generator method using the 1st LO. Could you post some links?
However, I don't intend to patent it, I don't doubt that somebody thought about it before, I just wanted to show practical results that I thought are better than expected.

It is nothing new or even unusual indeed and can be utilized even with a DSO that provides peak hold for its FFT analysis.

The 2nd screenshot in this posting shows the result of using the sweep generator + peak hold method for measuring an 455kHz IF filter:

https://www.eevblog.com/forum/testgear/siglent-sds1104x-e-in-depth-review/msg1449277/#msg1449277

Here the method has been used to verify the low frequency response of a scope frontend (1st screenshot):

https://www.eevblog.com/forum/testgear/siglent-sds1104x-e-in-depth-review/msg1454076/#msg1454076

[sorenkir]

Comparable method on a R&S CMU200 with the internal signal generator:
https://www.eevblog.com/forum/testgear/cmu200-almost-tracking-generator/
Michel.

[hendorog]

Nice work - however the max hold is a well known method.

What is your SA? It looks like a big old HP of some description, I'm guessing you probably could get an LO out of it if you tried as they are usually modular and internally connected by cables. So an unequal splitter or a directional coupler in the right place might do the trick.

I couldn't find practical results using this method, all I found was the real tracking generator method using the 1st LO. Could you post some links?
However, I don't intend to patent it, I don't doubt that somebody thought about it before, I just wanted to show practical results that I thought are better than expected.
It is an oldie but goldie HP8594E.

I'm probably biased, as on my Signalhound SA it's a very common thing to do, due to its non-standard architecture. Because of the way it normally works it sees a changing input signal as a spurious signal and removes it, when the span is greater than 250kHz. So Max hold + some time is often required when the input signal is sweeping or hopping around when spurious detect is turned on.

It looks like there is an option 009 for the 8594 which adds an LO distribution amp and an external LO output. Not too difficult to add that yourself, or bodge something in - well, when compared to building a tracking generator

Looks like 5086-7744 is the LODA (LO Distribution Amplifier) part number:
https://www.ebay.com/sch/i.html?_from=R40&_trksid=p2047675.m570.l1313.TR0.TRC2.A0.H0.XHP+5086-7744.TRS5&_nkw=HP+5086-7744&_sacat=0

Edit: Block diagram attached from service manual
https://www.keysight.com/upload/cmc_upload/All/08590-90316.pdf

[jpb]

Thanks for this - it is useful to see practical results.
I have been wondering about the effectiveness of this approach as I am looking to get a SA, I already have a 0-3GHz RF generator and though I'd prefer to have a tracking generator there are more options open (and less money to pay) in getting a SA without the tracking generator.
For instance, with the latest R&S "low end" spectrum analyser (the 1000 or 1500) the tracking generator adds another 1000 euro and that is without all its functionality enabled.
More particularly, SAs for sale on ebay often lack a tracking generator but are otherwise well specced.

[tautech]

Thanks for this - it is useful to see practical results.
I have been wondering about the effectiveness of this approach as I am looking to get a SA, I already have a 0-3GHz RF generator and though I'd prefer to have a tracking generator there are more options open (and less money to pay) in getting a SA without the tracking generator.

FYI the SSA3kX models now come with TG option free whereas it was previously an additional $237.