[SOLVED] Spectral purity of a signal generator

[hafrse]

Hi,

Have purchased a used Anritsu 68347B signal genrator, watching a trace on my SA at for example 2Ghz with RBW 3 Hz , span 200 Hz, I can see adjacent spurior signals at about 50 hz spacing, but att RBW 300Hz looks ok.

Attached two pictures taken att diffenerent RBW:

One at RBW 3 Hz VBW 3hz span 200hz and
One at RBW 300Hz VBW 300Hz span 20Khz.

Is that within specifications or I have a problem with the signal generator ?

Thnaks

George

 

[BigBoss]

Your SA is probably mixing due to high input level.

 

[E-design]

What is your mains frequency, 50 Hz?

 

[hafrse]

Hi,

My mains frequency is 50 hz, do I suspect power supply caps ?

Thanks

 

[hafrse]

Found the problem, it was from the new aftermarket Vellman fan I installed in this unit, I need to find a good 120x120x32/38 fan...
Thanks!

 

[FvM]

Bad power supply would give 100 Hz rather than 50 Hz side band spacing. Also the modulation is apparently not exactly 50 Hz. Did you check with a marker?

Found the problem, it was from the new aftermarket Vellman fan I installed in this unit.

I guess driven by a BLDC motor.

 

[hafrse]

Or, is there any way to filer out the electrical noise from the fan ? it is a 24V 0.2A fan.
Thanks

- - - Updated - - -

Bad power supply would give 100 Hz rather than 50 Hz side band spacing. Also the modulation is apparently not exactly 50 Hz. Did you check with a marker?


I guess driven by a BLDC motor.

Thanks, how can I know what motor type if I want to buy a correct one?

 

[FvM]

A series choke can help. But are you sure that the problem is electrical noise? It can be rotating magnetical field as well.

Modern DC fans have always BLDC drives. They are generating a lot of electrical noise and magnetical disturbances, too. I remember a case where a 2,5" HDD drive mounted near to a DC fan lost most of it's data throughput when the fan was running, apparently caused by magnetical disturbances. The problem could be fixed by a magnetic shield.

 

[hafrse]

A series choke can help. But are you sure that the problem is electrical noise? It can be rotating magnetical field as well.

Modern DC fans have always BLDC drives. They are generating a lot of electrical noise and magnetical disturbances, too. I remember a case where a 2,5" HDD drive mounted near to a DC fan lost most of it's data throughput when the fan was running, apparently caused by magnetical disturbances. The problem could be fixed by a magnetic shield.

I am not sure if it is pure electrical, will take away the fan and connect it from outside the instrumnent. The problem is that the fan is neer the YIG and the old fan (made by Papst) had a circular metal plate glued to the back side of the motor, I took it away and did the same with the new fan, see the picture

 

[biff44]

yes, the old time fans actually had some very nice low noise and vibration isolation technology in them. It is not easy to find a good replacement nowadays.

 

[volker@muehlhaus]

the YIG

YIG frequency is controlled by magnetic field, so magnetic coupling from the rotating fan is a possibility.

 

[D.A.(Tony)Stewart]

Hi,

My mains frequency is 50 hz, do I suspect power supply caps ?

Thanks

BLDC fans unless exactly 1500 or 3000 RPM will not be your 50 Hz problem.

It will be either radiated from poor shielding or ground connections or conducted from poor filtering such as aged ESR in large electrolytics. Consider replacing all with high quality caps with low ESR, high temp types for supply filters.

Only you can tell by changing the method of grounding or isolation from magnetic lines of fields of 60Hz

If it were Caps , I would expect 100 Hz not 50, as all diode bridges multiply the frequency plus harmonics.

hint... So look for poor conducted or radiated grounds or power cables. A line filter is a good add-on if there are radiated or conducted noise. Or lack of common ground in measurements or poor shielding....

 

[hafrse]

BLDC fans unless exactly 1500 or 3000 RPM will not be your 50 Hz problem.

It will be either radiated from poor shielding or ground connections or conducted from poor filtering such as aged ESR in large electrolytics. Consider replacing all with high quality caps with low ESR, high temp types for supply filters.

Only you can tell by changing the method of grounding or isolation from magnetic lines of fields of 60Hz

If it were Caps , I would expect 100 Hz not 50, as all diode bridges multiply the frequency plus harmonics.

hint... So look for poor conducted or radiated grounds or power cables. A line filter is a good add-on if there are radiated or conducted noise. Or lack of common ground in measurements or poor shielding....

Thanks for the information, in my situation if I disconnect the fan supply cable, the spurs dissapears, so it is the fan which is radiating noise (magnetic or electrical) . Later this evening I will take the fan outside the instrument and connect the supply cable to the fan to check if the spurs are still visible on my SA . I will report that later.

Please see the picture I have in my earlier post where the fan and the YIG are very neer each other.
Thanks

 

[hafrse]

Done test with the fan outside the instument, still som little noise when neer the instrument, about 20 cm away , no noise observed.
so the fan is radiating mechanical / electrical / magnetic noise vibration

 

[N1UL]

This still is surprising as modern DDS and PLL based generators have or should have enough PLL loop bandwidth to avoid/correct this problem.

Two way radios with PLL and build-in loudspeakers have long solved this problem and a fan should be much less complicated. If it is conducted rather then radiated interference then it seems a filter problem.

 

[hafrse]

This still is surprising as modern DDS and PLL based generators have or should have enough PLL loop bandwidth to avoid/correct this problem.

Two way radios with PLL and build-in loudspeakers have long solved this problem and a fan should be much less complicated. If it is conducted rather then radiated interference then it seems a filter problem.

Original fan is a Papst Model 4341 24V 5W, no interference what ever with it except that it sounds very high because of worn bearings, The new Vellman fan is a crap

 

[N1UL]

Happy you solved it !

 

[E-design]

Welcome back on the forum Ulrich! I always appreciated your vast knowledge and comments on crystal oscillators.

 

[biff44]

This still is surprising as modern DDS and PLL based generators have or should have enough PLL loop bandwidth to avoid/correct this problem.

Two way radios with PLL and build-in loudspeakers have long solved this problem and a fan should be much less complicated. If it is conducted rather then radiated interference then it seems a filter problem.

having enough bandwidth, and actually having enough open loop gain to correct vibration noise to maybe -90 dBc, are too very different things!