UK Vintage Radio Repair and Restoration Powered By Google Custom Search Vintage Radio and TV Service Data

Go Back   UK Vintage Radio Repair and Restoration Discussion Forum > Other Discussions > Homebrew Equipment

Notices

Homebrew Equipment A place to show, design and discuss the weird and wonderful electronic creations from the hands of individual members.

Reply
 
Thread Tools
Old 24th Nov 2022, 8:04 pm   #261
regenfreak
Heptode
 
Join Date: Nov 2018
Location: London SW16, UK.
Posts: 600
Default Re: 6-gang FM stereo tuner heads

The HP 3324A fails its self-test during every start-up. It shows the VCO error unlock at 30MHz...I guess it means the PLL cannot lock onto the clock ref frequency of 30MHz. It will be returned for refund. Oh well the HP 3324A is very bulky, I probably struggle to find space to put it, a glass half full.

My plan for two tone IP3 measurement has been thrown out of window. My intention was based on Marki's web site methodology:

two-fundamentals of 7MHz and 7.1MHz centred at 7.05MHz through two low-pass diplexers, 10db attenuators before the combiner, a bandpass diplexer before the ring diode mixer. The LO will be 17.75MHz and IF = 10.7MHz. The IF output bandpass diplexer (has to be relatively wideband, higher L/C ratio due to the two tones) is the most important one that it dumps all low and high-sided IMD energy through the 50 ohm loads, instead of reflecting back to the mixer and going into the spectrum analyzer's mixer. So 3-4 diplexers would be needed. It has become increasingly apparent that it is too much of a hassle and equipment intensive to do such a measurement.
Attached Thumbnails
Click image for larger version

Name:	IP3 mixer.jpg
Views:	10
Size:	41.2 KB
ID:	268671  

Last edited by regenfreak; 24th Nov 2022 at 8:19 pm.
regenfreak is offline   Reply With Quote
Old 24th Nov 2022, 8:31 pm   #262
Radio Wrangler
Moderator
 
Radio Wrangler's Avatar
 
Join Date: Mar 2012
Location: Fife, Scotland, UK.
Posts: 20,539
Default Re: 6-gang FM stereo tuner heads

It's a fractional-N frequency synthesiser. I'm not sure whether it's one of the analogue-phase-interpolation ones. Unless you are very comfortable with these things, I wouldn't recommend trying to repair at home.

I had the job of redesigning the API frac-N synth in the 3325A around a new LSI ASIC a number of years ago and while it's doable, getting all the spurs in-spec is a major undertaking. It may be that the 30MHz PLL is a straight forward one for the heterodyne band. The synth itself runs 30-51MHz and is mixed down to giv 0-21MHz

David
__________________
Can't afford the volcanic island yet, but the plans for my monorail and the goons' uniforms are done
Radio Wrangler is online now   Reply With Quote
Old 24th Nov 2022, 10:08 pm   #263
regenfreak
Heptode
 
Join Date: Nov 2018
Location: London SW16, UK.
Posts: 600
Default Re: 6-gang FM stereo tuner heads

Quote:
Originally Posted by Radio Wrangler View Post
It's a fractional-N frequency synthesiser. I'm not sure whether it's one of the analogue-phase-interpolation ones. Unless you are very comfortable with these things, I wouldn't recommend trying to repair at home.

I had the job of redesigning the API frac-N synth in the 3325A around a new LSI ASIC a number of years ago and while it's doable, getting all the spurs in-spec is a major undertaking. It may be that the 30MHz PLL is a straight forward one for the heterodyne band. The synth itself runs 30-51MHz and is mixed down to giv 0-21MHz

David
Yes I would poke it with a stick. I had the pleasure of skipping through 316 pages of user manual and 166 pages of service manual today.

I saw an article that it described the 3324A as " a low cost" unit and the list price was $3500 around 1989-90. Threw it another $765 you would get a built-in OCXO frequency reference. Luckily the seller has been quick to offer the return without a fuss..not all the sellers are easy-going like that.

How many hobbyists injured their lower backs after moving HP spectrum analyzers at home? I was very shocked to find out how massive the 3324A is.
regenfreak is offline   Reply With Quote
Old 24th Nov 2022, 10:15 pm   #264
G0HZU_JMR
Dekatron
 
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,012
Default Re: 6-gang FM stereo tuner heads

The Marki IMD test setup is very comprehensive, and I doubt many people use a setup with all those diplexers. The Marki setup is designed to measure the mixer in an ideal environment where there are no constructive/destructive port reflections anywhere. They need to avoid these reflections in order to put a valid result on their datasheet for that mixer. That's why they have that impressive test setup.

Even if you did replicate their setup, the IMD result you are really interested in is the result when the mixer is fitted in your receiver design and not when it is in Marki's ideal test rig. depending on your circuit, the mixer IP3 could be degraded quite a bit compared to the result achieved using the Marki setup.

Usually, the mixer input IP3 is about 12-15dB higher than the input P1dB for a regular diode ring mixer. The input P1dB is about 6dB lower than the LO drive level.

Therefore, a classic LO Level 7dBm mixer like the SBL-1 could be expected to have an input P1dB of +1dBm and an input IP3 of about +15dBm. However, if the mixer circuit has poor IF port match at the image frequency, then this IP3 measurement will often be degraded. It's also possible to deliberately mismatch the IF port at the image frequency in such a way that the image reflection is constructive. This might improve the conversion loss for example. Usually, a poorly terminated IF port makes the IP3 performance notably worse than expected.
__________________
Regards, Jeremy G0HZU
G0HZU_JMR is online now   Reply With Quote
Old 24th Nov 2022, 10:43 pm   #265
G0HZU_JMR
Dekatron
 
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,012
Default Re: 6-gang FM stereo tuner heads

The diagram below shows how I would normally set up an IP3 measurement. I'd start with 3dB attenuators as indicated and then increase them if required.

The setup below is usually good enough, unless trying for a measurement with very low IMD levels. I'd be happy to use the setup below to carry out an initial check on a mixer. With some mixers the LO drive level can make a big difference. It is also worth playing around with the RF tone levels and the RF and LO frequencies to see how the result changes.

A modern (strong) lab analyser won't need a diplexer at the IF port, but the internal attenuator needs to be set such that the analyser's internal IMD is at least 20dB lower than the mixer IMD. A good margin is needed here because the analyser also sees the image test tones as well as the wanted tones.
Attached Thumbnails
Click image for larger version

Name:	ip3_setup.gif
Views:	19
Size:	8.0 KB
ID:	268687  
__________________
Regards, Jeremy G0HZU
G0HZU_JMR is online now   Reply With Quote
Old 25th Nov 2022, 4:02 am   #266
Radio Wrangler
Moderator
 
Radio Wrangler's Avatar
 
Join Date: Mar 2012
Location: Fife, Scotland, UK.
Posts: 20,539
Default Re: 6-gang FM stereo tuner heads

OK, in the 3324A the 30MHz PLL is the frequency standard/heterodyne mixdown LO

30MHz loop unlocked could simply mean that the external 10 MHz frequency reference has been selected and you haven't connected a 10MHz standard to the unit. If so, the fault should clear if you switch to internal reference.

It also looks like the frac-N loop does self tests at 30MHz and 60MHz frequencies. One of these could fail due to drift in the frac-N loop VCO. Might be a simple tweak to fix, otherwise these frac-N systems scare people off and are full of special and selected parts. The make of one logic gate has around a 20dB influence on spur levels!

David
__________________
Can't afford the volcanic island yet, but the plans for my monorail and the goons' uniforms are done
Radio Wrangler is online now   Reply With Quote
Old 25th Nov 2022, 3:05 pm   #267
regenfreak
Heptode
 
Join Date: Nov 2018
Location: London SW16, UK.
Posts: 600
Default Re: 6-gang FM stereo tuner heads

Quote:
Originally Posted by G0HZU_JMR View Post
The diagram below shows how I would normally set up an IP3 measurement. I'd start with 3dB attenuators as indicated and then increase them if required.

The setup below is usually good enough, unless trying for a measurement with very low IMD levels. I'd be happy to use the setup below to carry out an initial check on a mixer. With some mixers the LO drive level can make a big difference. It is also worth playing around with the RF tone levels and the RF and LO frequencies to see how the result changes.

A modern (strong) lab analyser won't need a diplexer at the IF port, but the internal attenuator needs to be set such that the analyser's internal IMD is at least 20dB lower than the mixer IMD. A good margin is needed here because the analyser also sees the image test tones as well as the wanted tones.

Cheers. I assume the use of 3db attenuators is to improve port impedance matching rather than cross-talk isolation between two sign gen? I am kinda of thinking of using 10db attenuators. Secondly, this set-up is for amplifier measurement only? You can't measure a mixer like that, as the LO have to switch the ring diodes on.

I am making a comparison between HP8566B and SSA3021X Plus:

HP8566B:
TOI > +7dbm (5M-5.8G) . DANL < -134dbm (0 atten, 10Hz RBW, 1M-20Ghz) . phase noise -90dbc/Hz at 10KHz offset.

SSA3021X +:
TOP: +10 to + 18dbm. DANL, -161dbm/hz (0 atten at RBW 1 Hz). Phase noise, -98dbc/Hz at 10KHz

Unlike phase noise measurements, it is possible to measure the DUT having a TOI significantly higher than that of the spectrum analyzer. My understanding is that the measured 3rd-order IM products are relative to the noise floor level at a particular RBW instead of absolute measurements. Since the calculation of IP3 involves the difference between the fundamental and the average of the key troublemaking anti-social neighbours; (2f1-f2)-LO, (2f2-f1)-LO (or upconversion LO-(2f1-f2), LO-(2f2-F1) ), the contribution of IM products of the first-stage mixer of the spectrum analyzer is subtracted out. The IM products of the 2nd, 3rd and 4th mixers are irrelevant due to the up- and down conversation and the associated LP/bandpass filtering at microwave, VHF and Hf frequencies inside the spectrum analyzer. However, it does not mean that the IM products from the first mixer of spectrum analyzer can get away scot-free.

The use of attenuators at the spectrum analyzer input will reduce the level of the spectrum anlayzer's IM products so they are small compared with the IM products from the DUT. The gain slope of the 3rd order IM is 3, so for a 1db rise in mixer level, a 3db rise in 3rd order IM?

The DANL is important in TOI measurement. For DUT with high TOI, the measurement accuracy is severely limited by the low SNR at the noise floor of the spectrum analyzer, so it is really good when the RBW can go down to 1Hz. I read somewhere a spectrum analyzer can have zero a span, how is it even possible?


Quote:
Originally Posted by Radio Wrangler View Post
OK, in the 3324A the 30MHz PLL is the frequency standard/heterodyne mixdown LO

30MHz loop unlocked could simply mean that the external 10 MHz frequency reference has been selected and you haven't connected a 10MHz standard to the unit. If so, the fault should clear if you switch to internal reference.

It also looks like the frac-N loop does self tests at 30MHz and 60MHz frequencies. One of these could fail due to drift in the frac-N loop VCO. Might be a simple tweak to fix, otherwise these frac-N systems scare people off and are full of special and selected parts. The make of one logic gate has around a 20dB influence on spur levels!

David
The 3324A is packed and ready to go back. I wont take the risk to repair myself unless I got it dirt cheap.
regenfreak is offline   Reply With Quote
Old 25th Nov 2022, 4:12 pm   #268
G0HZU_JMR
Dekatron
 
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,012
Default Re: 6-gang FM stereo tuner heads

The IP3 setup I showed is for any DUT. If a mixer is the DUT, then an LO would obviously be required for the mixer, but this is really part of the DUT. It's worth experimenting with different LO drive levels for the mixer though.

Quote:
I am making a comparison between HP8566B and SSA3021X Plus:
HP8566B:
TOI > +7dbm (5M-5.8G) . DANL < -134dbm (0 atten, 10Hz RBW, 1M-20Ghz) . phase noise -90dbc/Hz at 10KHz offset.

SSA3021X +:
TOP: +10 to + 18dbm. DANL, -161dbm/hz (0 atten at RBW 1 Hz). Phase noise, -98dbc/Hz at 10KHz
I've got an old HP8566B here and the TOI for the 8655B mixer is typically +8dBm to +10dBm at VHF with 0dB attenuation. At VHF test frequencies, the DANL is typically -140dBm with 0dB attenuation and 10Hz RBW. This implies that the noise figure is about 24dB at VHF.

The HP8566B doesn't have great phase noise, partly because the IF is at about 3.6GHz on the lowest band. However, it is typically going to be about the same as the Siglent when measured at 10kHz offset at about -98dBc/Hz at VHF test frequencies. The HP 8566B phase noise will be much lower than the Siglent when measured at 100kHz offset. Probably 20dB lower because it uses a YIG oscillator. My 8566B is really an HP8566A that has had the factory conversion to a B. I think the later B versions had improved phase noise but not by much.

If I look at the datasheet for the Siglent SSA3021X, the DANL is -129dBm with a 10Hz RBW and 0dB attenuation. This implies that the noise figure is about 35dB. The TOI is +10dBm with the same setup conditions. The Siglent has a preamp that can improve the noise figure, but it will also degrade the TOI quite a lot when selected. It does look like the 8566B generally has a higher dynamic range than the Siglent.

Newer lab analysers are much better than the HP 8566B though. My Agilent E4440A analyser is about 15 years old and has -117dBc/Hz phase noise at 10kHz offset and the DANL is about -145dBm with 0dB attenuation at 10Hz RBW (no preamp selected). This implies that the noise figure is about 19dB. The mixer TOI is about 19dBm with 0dB attenuation. The modern PXA and UXA signal analysers are even better.
__________________
Regards, Jeremy G0HZU

Last edited by G0HZU_JMR; 25th Nov 2022 at 4:19 pm.
G0HZU_JMR is online now   Reply With Quote
Old 25th Nov 2022, 5:20 pm   #269
G0HZU_JMR
Dekatron
 
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,012
Default Re: 6-gang FM stereo tuner heads

Quote:
Since the calculation of IP3 involves the difference between the fundamental and the average of the key troublemaking anti-social neighbours; (2f1-f2)-LO, (2f2-f1)-LO (or upconversion LO-(2f1-f2), LO-(2f2-F1) ), the contribution of IM products of the first-stage mixer of the spectrum analyzer is subtracted out.
I'm not sure what you mean here, but in the simplistic case of the DUT being an RF amplifier, then the analyser IMD levels need to be about 20dB lower than the IMD levels of the amplifier if the aim is to keep the measurement uncertainty of the amplifier IMD below about +/-1dB. Otherwise, the analyser IMD tones can sum or cancel with the IMD tones of the DUT and cause uncertainty exceeding +/- 1dB.

If the input TOI of the analyser is the same as the output TOI of the DUT, then the DUT IMD and analyser IMD tones will be at the same level, and they generally sum together in the analyser and the analyser will then display the IMD levels about 6dB higher than the true IMD level of the DUT. The tones can also cancel (in theory) and the analyser might show much lower distortion than the true result. That's why the analyser IMD levels ideally need to be about 20dB lower than the IMD levels of the DUT.
__________________
Regards, Jeremy G0HZU
G0HZU_JMR is online now   Reply With Quote
Old 25th Nov 2022, 6:32 pm   #270
regenfreak
Heptode
 
Join Date: Nov 2018
Location: London SW16, UK.
Posts: 600
Default Re: 6-gang FM stereo tuner heads

Quote:
Originally Posted by G0HZU_JMR View Post
Quote:
Since the calculation of IP3 involves the difference between the fundamental and the average of the key troublemaking anti-social neighbours; (2f1-f2)-LO, (2f2-f1)-LO (or upconversion LO-(2f1-f2), LO-(2f2-F1) ), the contribution of IM products of the first-stage mixer of the spectrum analyzer is subtracted out.
I'm not sure what you mean here, but in the simplistic case of the DUT being an RF amplifier, then the analyser IMD levels need to be about 20dB lower than the IMD levels of the amplifier if the aim is to keep the measurement uncertainty of the amplifier IMD below about +/-1dB. Otherwise, the analyser IMD tones can sum or cancel with the IMD tones of the DUT and cause uncertainty exceeding +/- 1dB.

If the input TOI of the analyser is the same as the output TOI of the DUT, then the DUT IMD and analyser IMD tones will be at the same level, and they generally sum together in the analyser and the analyser will then display the IMD levels about 6dB higher than the true IMD level of the DUT. The tones can also cancel (in theory) and the analyser might show much lower distortion than the true result. That's why the analyser IMD levels ideally need to be about 20dB lower than the IMD levels of the DUT.
Sorry I was confused and wrong about the subtraction bit. To improve the DANL in IP3 measurement, it is necessary to switch on the pre-amp of the spectrum analyzer, so the displayed IMD readings also include the IMD of the spectrum analyzer's pre-amp. I have been struggling with the idea why not the spectrum analyzer's pre-amp and its mixer IMD do not mess up with the true IMD level of the DUT.

The IP3 is a fictitious figure of merit that linearizes something is really non-linear. At times, I get confused when I try to understand things underneath the surface. The definition of OIP3 is simple but i feel it is deceptive, the more i think about it, the more I realize it is complicated.

Constructive and destructive IMD interference is a completely new concept to me.

Earlier I got confused by these sentences from the link below:

The spectrum analyzer also produces internally generated, third-order distortion products at exactly the same frequencies as those generated by the DUT. These distortion products are relative to the mixer level, rather than the output level of the DUT. This distinction enables the measurement of a DUT with a significantly higher TOI than that of the spectrum analyzer.

https://www.mwrf.com/technologies/te...i-measurements

Last edited by regenfreak; 25th Nov 2022 at 6:59 pm.
regenfreak is offline   Reply With Quote
Old 25th Nov 2022, 8:00 pm   #271
regenfreak
Heptode
 
Join Date: Nov 2018
Location: London SW16, UK.
Posts: 600
Default Re: 6-gang FM stereo tuner heads

To illustrate how confusing and muddy it can get, the RS white paper in the following link states that the two tones at f1 and f2 of frequency apart delta_f (=f2-f1) will produce 3rd order IMD products at f1-delta _f and f1+delta_f. What the heck? Everywhere i have seen the 3rd order IMD should be at 2f1-f2 and 2f2-f1(attached):


https://scdn.rohde-schwarz.com/ur/pw...Distortion.pdf

Digging deeper, i get more nasty surprise, constellation diagram, IM interference etc. It is like opening the Russian Matryoshka Nesting Dolls, there are more Maths, abstraction and complications every time I unravel the inside, frying my brain.
Attached Thumbnails
Click image for larger version

Name:	3rd order IMD.jpg
Views:	6
Size:	59.2 KB
ID:	268723  
regenfreak is offline   Reply With Quote
Old 25th Nov 2022, 11:42 pm   #272
regenfreak
Heptode
 
Join Date: Nov 2018
Location: London SW16, UK.
Posts: 600
Default Re: 6-gang FM stereo tuner heads

I have just done a "filthy" TOI test of the SSA3021X+ without DUT like what it is described in the RS white paper. I dont have the LP filters for the 10MHz and 10.1MHz tones so I stick two 20 db attenuators to isolate two signal sources further. The 2nd harmonics for the FY6900 are to be -67dBC, and DG1022Z to be -51.65dBc at 10MHz fundamental.

I have tested the mixer of the SSA3021X, setting the built-in attenuation 30db, 20db, 10db and 0db at 10Hz RBW (10 sample averaging). The pre-amp is turned off in all cases except in the last attachment with 30db attenuator.

My two tones are not exactly the same level and you can see the TOI is higher with higher input level at the upper side. The results are very similar for attenuation from 30db to 10db. Only at 0db attenuation, the TOI degrades drastically. Strangely enough the pre-amp does not degrade the TOP with 30db attenuation( attachment 5).

I can try to repeat the test by making two LP filters for the frequency sources. I am trying to understand what is going on here and what are the limitations of the spectrum analyzer rather than treating it as a black box.

According to Siglent, the TOI is > 14db at 1GHz, with -20dbm two-tone, 0db atten, RBW=10khz.

I think there are mistakes in the RS white paper mentioned in my last post about the 3rd order IMD products at f1-delta _f and f1+delta_f.. But how could it be? They are the experts and not me. I struggle with the idea that when you have two phase uncorrelated sources enter the mixer, somehow there will be constructive and destructive interference. I can easily understand it in time domain when two tones of different frequencies can have constructive and destructive interference along the nodes, but I cannot visualize the same in the mixer within the frequency domain.
Attached Thumbnails
Click image for larger version

Name:	atten 30db.jpg
Views:	7
Size:	50.8 KB
ID:	268727   Click image for larger version

Name:	atten 20db.jpg
Views:	4
Size:	55.0 KB
ID:	268728   Click image for larger version

Name:	atten 10db.jpg
Views:	4
Size:	55.0 KB
ID:	268729   Click image for larger version

Name:	atten 0db.jpg
Views:	4
Size:	51.2 KB
ID:	268730   Click image for larger version

Name:	atten 30db preamp on.jpg
Views:	7
Size:	55.3 KB
ID:	268731  

regenfreak is offline   Reply With Quote
Old 26th Nov 2022, 1:32 am   #273
G0HZU_JMR
Dekatron
 
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,012
Default Re: 6-gang FM stereo tuner heads

I've not used the Siglent so I don't know what to expect from it. The same applies to your sig gens and combiner.

If we go back maybe 45 years, then I think there were various design rules in place for the design of a lab spectrum analyser. If you were to take the 1500MHz HP8568B apart, you would see that the front end is entirely passive up until after the second mixer. The first amplifier in the system is at the second IF.

I think this was done to try and maintain predictable distortion performance. In other words, the two-tone distortion performance at the front end should be dominated by the first mixer for fairly well spaced test tones. The second mixer will have similar TOI as the front end mixer but it is protected by the conversion loss of the first mixer and any additional loss in the IF1 post mixer attenuator and the first IF filter. This makes sure that it obeys the >-20dB IMD distortion in the second mixer compared to the distortion in the first mixer.

This makes these analysers easy to understand with fairly predictable distortion performance. Modern analysers like the Rigol and Siglent will have solid state switches and solid state attenuators ahead of the first mixer. There will also be an amplifier at the first IF. This means that all of these devices can (in theory at least) contribute a tiny amount to the overall distortion seen on the display.

It's fairly easy to predict the HP8568B TOI and noise floor performance. I first looked at the design of this analyser over 30 years ago. I've just had a quick go at designing something similar to the HP8568B front end using excel. I've used a RBW of 10kHz and assumed that the default 10dB attenuation is at the front end.

You can see that the spreadsheet predicts an input IP3 (TOI) of 23dBm, a noise floor of -100dBm and a spurious free dynamic range of 82dB. See also a chart from the user manual that shows that the HP8568B has a typical SFDR of 82dB at 10kHz RBW. This agees very well with the spreadsheet.

It's 30 years since I studied the HP8568B design in any detail, but I was tasked to do this at work so I could learn how the frequency planning was done and how the front-end design was optimised for low distortion. The line entries in the spreadsheet are done from memory but they should at least be in the right ballpark for the HP8568B up to the 2nd IF at least.
Attached Thumbnails
Click image for larger version

Name:	HP8568B DR.jpg
Views:	13
Size:	76.9 KB
ID:	268733   Click image for larger version

Name:	HP8568B DR82.jpg
Views:	9
Size:	92.8 KB
ID:	268734  
__________________
Regards, Jeremy G0HZU

Last edited by G0HZU_JMR; 26th Nov 2022 at 1:46 am.
G0HZU_JMR is online now   Reply With Quote
Old 26th Nov 2022, 2:01 am   #274
regenfreak
Heptode
 
Join Date: Nov 2018
Location: London SW16, UK.
Posts: 600
Default Re: 6-gang FM stereo tuner heads

Thanks i have watched the teardown of Siglent SSA3021x and SSA3032x in EEblog and Signal Path a few times, so i have a fair good idea of their design blocks. Both are identical in hardwares. Regarding the phase of spurs in a mixer, this video from Marki livestream explains it but the concepts are "confusing and like a black hole " as described by the speaker:

https://youtu.be/BvaOCW0rUYU

I have also seen a video about zero span from RS channel. I must go to bed.
regenfreak is offline   Reply With Quote
Old 26th Nov 2022, 11:21 am   #275
G0HZU_JMR
Dekatron
 
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,012
Default Re: 6-gang FM stereo tuner heads

Quote:
To illustrate how confusing and muddy it can get, the RS white paper in the following link states that the two tones at f1 and f2 of frequency apart delta_f (=f2-f1) will produce 3rd order IMD products at f1-delta _f and f1+delta_f. What the heck? Everywhere i have seen the 3rd order IMD should be at 2f1-f2 and 2f2-f1(attached):
The equations look different, but they should give the same result.

Quote:
I struggle with the idea that when you have two phase uncorrelated sources enter the mixer, somehow there will be constructive and destructive interference. I can easily understand it in time domain when two tones of different frequencies can have constructive and destructive interference along the nodes, but I cannot visualize the same in the mixer within the frequency domain.
Sig gen tones f1 and f2 enter the DUT and produce IMD tones at 2f1-f2 (IMDtoneA) and 2f2-f1 (IMDtoneB).


However, f1 and f2 then enter the analyser and the analyser will also produce its own 2f1-f2 (IMDtoneA) and 2f2-f1 (IMDtoneB) distortion tones.

IMDtoneA therefore gets injected into the analyser from the DUT and the analyser also generates its own IMDtoneA through its own internal distortion.

These two versions of IMDtoneA will often not have exactly the same phase. Usually in a spectrum analyser, they are very close to being in phase. This is especially the case if the IMD testing is done with closely spaced test tones f1 and f2.

If the two versions of IMDtoneA are the same amplitude and they are in phase then they will sum in voltage to give twice the voltage. This means the analyser will report the level of IMDtoneA to be about 6dB higher than the level of IMDtoneA in the DUT. This is seen on the blue curve on the right-hand side of the graph below.

It can happen that the two IMDtoneA terms are not in phase, so in theory at least, they can cancel, and the analyser will report the level of IMDtoneA to be lower than the level of IMDtoneA in the DUT. This is the red curve. The phase could be anywhere in-between these two cases.

If you look at the graph below, it shows the measurement uncertainty caused by the relative amplitude and phase of 'DUT IMDtoneA' and 'Analyser IMDtoneA' inside the analyser. The green arrow and text shows that it is desirable to keep the analyser generation of IMDtoneA to be at least 20dB lower that the DUT generation of IMDtoneA. This keeps the measurement uncertainty displayed on the analyser to below +/- 1dB. Most RF engineers will be well aware of this graph and it's a good idea for anyone who uses a spectrum analyser to be aware of it too.

The same graph can be found in app notes by HP/Agilent/Keysight for example.
Attached Thumbnails
Click image for larger version

Name:	RS_IMD_phase.jpg
Views:	11
Size:	68.2 KB
ID:	268741  
__________________
Regards, Jeremy G0HZU
G0HZU_JMR is online now   Reply With Quote
Old 26th Nov 2022, 11:39 am   #276
regenfreak
Heptode
 
Join Date: Nov 2018
Location: London SW16, UK.
Posts: 600
Default Re: 6-gang FM stereo tuner heads

Quote:
The equations look different, but they should give the same result.
Thanks I got it now:


IM3A = f1-(2f1-f2)=f2-f1

IM3B=(2f2-f1)- F2 = f2-f1

I am going to re-read the article.

I am going to make two LP diplexers...
regenfreak is offline   Reply With Quote
Old 26th Nov 2022, 5:55 pm   #277
regenfreak
Heptode
 
Join Date: Nov 2018
Location: London SW16, UK.
Posts: 600
Default Re: 6-gang FM stereo tuner heads

When I look at graphs with lots of straight lines, most articles never bother explain where the equations of straight lines coming from, how they were derived from the first principle, and what do they mean.

Then they drop phrases like "mixer level" and "DANL and IMD relative to mixer level" like bombs and never explain the meaning of "relative to ". For novice like me, it just creates a nebulous cloud of confusion. I interpret "mixer level" as the Pin the input power of the mixer but then he drops undefined symbol Pn in the article...and "relative to" meaning Pin -PIM3?

From the RS article,

PIM3 = 3Pin-2TOI

Pn is never defined in the article and they just throw it to your face:


Pn=(DANL + B -Pin)dB

d is the amplitude difference between internal and external IMD products.

Then I have found from other articles:

SFDR = 2/3(TOI-DANL)

max IM3 dynamic range = 2/3(DANL - TOI).

TOI = Pn -1/2PIM3

and the list goes on...

Ok these are all straight lines but understanding where they come from is not easy and end up taking them at face value.
regenfreak is offline   Reply With Quote
Old 26th Nov 2022, 7:58 pm   #278
Radio Wrangler
Moderator
 
Radio Wrangler's Avatar
 
Join Date: Mar 2012
Location: Fife, Scotland, UK.
Posts: 20,539
Default Re: 6-gang FM stereo tuner heads

By looking at TOI and SFDR, you're looking at a processed result of some measurements, often a good number of measurements.

Set up an intermodulation test and progressively step up the levels of the pair of tones. At each step, measure the levels of the various products. Also measure the level of the wanted output.

You can now plot them all on an X-axis of the applied tone level. Do all levels in dBm for convenience.

For a well-behaved device you will see that the third order products go up 3dB for every 1dB increase in the applied tones. Fifth order products 5dB, fourth order 4dB and you get the pattern. It's terribly convenient! Go high and you'll see the intermod products start to level off and eventually saturate, or something gets too hot and burns out. But look at lower levels and you get excellent straight lines. Project the straight section of the third order intermod beyond the levelling and saturation, and it intersects the straight line of the wanted output (projected also as necessary) Where they cross is your true TOI point. The linear section of the wanted output has a 1:1 slope (dB/dB slope) while you'll see the third order linear section has a 3dB/1dB slope. So the gap between the wanted and unwanted 3rd order product closes at 2dB per 1dB rise in applied tones.

The spurious free dynamic range is defined as the range of variation of input level between the noise floor, and the input level which makes the intermod products just come up to equality with the noise floor. So it's twice the difference between noise floor and TOI point. IN THEORY. You will only get this if the upper level is still within the nice linear region of the intermod level plot.

There are also badly behaved devices whose intermod plot lines are not linear, curving smoothly into saturation. Some can be decidedly kinky and even jumpy towards the top end. Things like crystal filters, some very high level mixers and some feedback amplifiers do these sorts of things. But, the lower slopes are usually well behaved and you can extrapolate to a TOI crossing from them. The results are valid and will predict low level behaviour. Doing the plot and seeing where things go silly will tell you what levels not to approach, and where the cosy TOI modelling breaks down.

So, why the convenient dB/dB slopes? They seem almost too good to be true and are reliable at low enough levels, and the factors are accurate, not just close.

Take a device and model its transfer function as a mathematical power series. The zeroeth power gives you DC output if you're bothered. The first power gives your wanted output, the square power gives you second order distortions, the cube power gives you third order, the fourth and so on

Differentiate this series to get the slopes and you see that as per normal calculus, the power of a term is multiplied into its coefficient. And there is the basis of those convenient scales for each order of product.

The bad behaviour at higher levels would be modelled properly if you took the time to make a complicated enough series to model high level behaviour in detail. No-one does, it's easy to take the simple view because we only want to use devices in their well-behaved region and aren't bothered just how bad the badly behaved region is.

So there you have it.

You need to do a full plot of the intermod product power so that you can see this in action and feel comfortable about it.

You need to do a full plot of a new device so you can check where it goes badly behaved and the small-signal assumptions and models fail.

Jeremy's mentioned the handy estimate of TOI versus 1dB compression point. It's pretty good on all well-behaved devices. On the bad guys it's an assumption that will have you.

There are other rough rules of thumb. Howard Swain was responsible for one relating relative levels to relative amounts of compression. Handy for spectrum analyser designers deciding where to pitch maximum operating levels and designing to meet an overall compression figure which is actually the budget for the total of several contributing stages. Noise figure instruments are touchier in this area than spectrum analysers so it's one reason why the purpose-designed NF instruments out perform general analysers with a software personality.

David
__________________
Can't afford the volcanic island yet, but the plans for my monorail and the goons' uniforms are done
Radio Wrangler is online now   Reply With Quote
Old 26th Nov 2022, 8:21 pm   #279
regenfreak
Heptode
 
Join Date: Nov 2018
Location: London SW16, UK.
Posts: 600
Default Re: 6-gang FM stereo tuner heads

Quote:
Radio Wrangle So, why the convenient dB/dB slopes? They seem almost too good to be true and are reliable at low enough levels, and the factors are accurate, not just close.

Take a device and model its transfer function as a mathematical power series. The zeroeth power gives you DC output if you're bothered. The first power gives your wanted output, the square power gives you second order distortions, the cube power gives you third order, the fourth and so on

Differentiate this series to get the slopes and you see that as per normal calculus, the power of a term is multiplied into its coefficient. And there is the basis of those convenient scales for each order of product.

The bad behaviour at higher levels would be modelled properly if you took the time to make a complicated enough series to model high level behaviour in detail. No-one does, it's easy to take the simple view because we only want to use devices in their well-behaved region and aren't bothered just how bad the badly behaved region is.

So there you have it.

You need to do a full plot of the intermod product power so that you can see this in action and feel comfortable about it.

You need to do a full plot of a new device so you can check where it goes badly behaved and the small-signal assumptions and models fail.

Jeremy's mentioned the handy estimate of TOI versus 1dB compression point. It's pretty good on all well-behaved devices. On the bad guys it's an assumption that will have you.

There are other rough rules of thumb. Howard Swain was responsible for one relating relative levels to relative amounts of compression. Handy for spectrum analyser designers deciding where to pitch maximum operating levels and designing to meet an overall compression figure which is actually the budget for the total of several contributing stages. Noise figure instruments are touchier in this area than spectrum analysers so it's one reason why the purpose-designed NF instruments out perform general analysers with a software personality.
Thank you. I think my problem is more to do with working out the origins and derivations of the linear equations from the first principles in the graphs.

I have attached 5 versions of the same graphs. Seemingly, they are the same but not quite. Some plot 1.distortion dbc, 2. signal-to-noise ratio dBc, 3 dynamic range db, 4. distortion products relative to mixer db.

It seems that

DANL = -1 x mixer level + y-intercept 1

IMD2 = 2 x mixer level + y-intercept 2

IMD3 = 3X mixer level + y-intercept 3

It will be very helpful if I can have these three equations with the unknown intercepts. None of the articles give you this information.

I can see where the straight lines for IMD2 and #IMD3 coming from. But none of the HP, Agilent and R&S articles explain why DANL is a linear function of mixer level with a negative slope of -1. It is not so obvious when i have tried to derive the proof of those straight lines and intercepts by looking at the geometries of the intersecting lines and triangles.

The maximum 2nd and 3rd order dynamic ranges are at the intercepts with the DANL line.
Attached Thumbnails
Click image for larger version

Name:	V 1 distortion relative to mixer level.jpg
Views:	8
Size:	67.3 KB
ID:	268779   Click image for larger version

Name:	V 2 dynamic range.jpg
Views:	6
Size:	62.6 KB
ID:	268780   Click image for larger version

Name:	V3 signal to noise ratio.jpg
Views:	8
Size:	69.5 KB
ID:	268781   Click image for larger version

Name:	V 4 dynamic range.jpg
Views:	8
Size:	34.6 KB
ID:	268782   Click image for larger version

Name:	V 5 distortion.jpg
Views:	6
Size:	60.5 KB
ID:	268783  


Last edited by regenfreak; 26th Nov 2022 at 8:36 pm.
regenfreak is offline   Reply With Quote
Old 26th Nov 2022, 9:54 pm   #280
G0HZU_JMR
Dekatron
 
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,012
Default Re: 6-gang FM stereo tuner heads

Quote:
I think my problem is more to do with working out the origins and derivations of the linear equations from the first principles in the graphs.
Another way to look at this is to forget about the TOI point and just inject two clean distortion free tones into the analyser. Set the tone level really low to start with. Choose a convenient RBW, say 10kHz and set the front-end attenuator to 10dB and the reference level to -5dBm. On your Siglent analyser, I think that the noise floor (DANL) would be at about -90dBm with these settings if you have the preamp turned off.

In theory, you should be able to keep increasing the input tone levels up to a point where the IMD tones would have the same power as the displayed -90dBm DANL. In practice, it will be difficult to judge where this happens, you could briefly select a narrower RBW to make it easier to see the -90dBm IMD tone level as it begins to compete with the DANL and become visible.

When you reach this point, the difference between the input test tone levels and the -90dBm DANL is the spurious free dynamic range with a 10kHz RBW.
You would obviously have to set the analyser to a span where the phase noise of the analyser doesn't mask the DANL. Maybe try a 5MHz span with 1MHz tone spacing to try and minimise the impact of the LO phase noise. Set the RBW to 10kHz. I'm guessing that your analyser phase noise will be -120dBc/Hz at 1MHz offset. 10kHz RBW is 40dBHz so you should be able to see about an 80dB dynamic range if there was no IMD. However, I think the Siglent will probably have a SFDR of about 73dB with 10kHz RBW because of the IMD it will generate.

Alternatively, if you know the input TOI of the analyser and the DANL from the datasheet, you can predict the SFDR with a reasonable amount of accuracy without having to do any hardware tests. Just use the equations you already have. The equations allow you to work backwards from the published TOI and noise floor data to work out the SFDR at a given RBW setting.

This generally works well for classic spectrum analysers that use a diode pair or diode ring mixer at the front end. I don't know if the Rigol and Siglent analysers will follow the same classic equations. There are solid state devices in the front end of these analysers and the linearity of these devices probably won't be as predictable. Therefore, the tests outlined above might not give the expected results.
__________________
Regards, Jeremy G0HZU
G0HZU_JMR is online now   Reply With Quote
Reply

Thread Tools



All times are GMT. The time now is 12:32 am.


All information and advice on this forum is subject to the WARNING AND DISCLAIMER located at https://www.vintage-radio.net/rules.html.
Failure to heed this warning may result in death or serious injury to yourself and/or others.


Powered by vBulletin®
Copyright ©2000 - 2022, vBulletin Solutions, Inc.
Copyright ©2002 - 2021, Paul Stenning.