high performance SSB system design technical questions

[neazoi]

sw3ora

Hello all,
I design a high performance ssb system for SW that I could also use for upconversion to 10GHz.
I am using the filter method with an IF of 500KHz.
I have two identical ssb filters of 501.650KHz, BW 2.7KHz, which are ok for ssb voice operation (300Hz-3KHz of audio signal).
My carrier is at 500KHz.
I use mc1496 for DSB modulator which gives -65dB carrier suppression @500KHz.
I do not know the exach response of the ssb filter but it should suppress the carrier even more, as it is outside of its limits.

1. Now the question is, can I use both filters in series to suppress even more the unwanted carrier?
Well I guess I can, but if the filter has a curve response, if I connect them in series then the output will become narrower (sharper). If from the other side the crystal filter has a more square-like response, the output will not change too much but it will filter better the unwanted carrier and sideband. Are my considerations right?


2. I consider especially phase noise, so I do not use a mixer for upconversion, but a Schottky diode doubler to get lots of even harmonics through the SW bands (ok sw is noisy by itself but I need the system for 10GHz upconversion too). Then I can keep the wanted harmonic by just using a single crystal in series, to reject others. If even harmonics considered for 500KHz, then I will have one harmonic every 500KHz. Are my considerations right?

3. what is approximately the bandwidth of a standard quartz crystal? How many KHz?


Thank you very much, these are a bit more technical questions but I cannot find many RF forums to ask.

Konstantinos Giannopoulos (SW3ORA)
Computer and Telecommunications Engineer
Director of the Greek Microwave Group (www.microwave.gr)

 

[vfone]

electromechanical filter 500khz

If the two IF filters that you are going to put in series have the same central frequency and the same bandwidth, you will get better side band suppression, than using only one filter.
In your case side band suppression is the main job of these filters.
Carrier suppression is affected by the filter, but this rejection is mainly done by the balanced mixer MC1496.

If the filters placed in series have slightly different central frequencies, will result a different bandwidth response (usually larger), different side band rejection, and probably different in-band ripple.
In-band ripple and group delay of the SSB filters are very important factors for a high-quality transmission. I would recommend to use only one filter, if this provides enough rejection for the side band.

The bandwidth of a standard quartz crystal is between few hundreds of Hertz and few kHz, depending by the load and by the topology of the adjacent circuits.

 

[neazoi]

technical questions for ssb

Thank you.
The filters are identical (same center frequency and bandwidth, even same brand and model).

So it is recommended to use one filter even if the filters are identical?

Group delay will have an effect in ssb phase noise?

 

[vfone]

ssb filter

Yes, if the filters are identical you can place two in series. Beware to the matching network between then.

I remember years ago that were available some electromechanical filters on 500 kHz for SSB communications. These filters used 9 resonators inside and provides up to 80dB side band rejection.
Collins still makes SSB mechanical filters, but on 455kHz:
https://www.rockwellcollins.com/about/additionalproducts/collinsfilters/app_notes/index.html

Group delay is a linear distortion and will not affect the SSB phase noise.

 

[neazoi]

ssb crystal filter technical

The filters I have are these: **broken link removed**
These have four crystals inside, and two variable shielded coils, one at the input and one at the output. I cannot find any specs.
Are these the type of filters you are reffering to as electromechanical?
Cheers

 

[vfone]

high performance ssb

No, this is a crystal filter and should be fine for your application.
I don’t think you need to put two in series, only one shall be fine.

Electromechanical filters use metal resonators inside (as you see in the Collins
application note below):
https://www.rockwellcollins.com/abo.../collinsfilters/mechanical_filters/index.html

and this is a 500kHz mechanical filter:

**broken link removed**

 

[neazoi]

ssb filter method

I got my filter specs! (victory)

My filter has a shape factor 6dB/60dB of 1:3.3
The BW (-6dB) is 2.7KHz and on -60dB is 9KHz max

1. What will the attenuation be on the LSB? The data states an attenuation of -10dB min at 500KHz (which is my carrier.

2. Finally, the I/O impedance is 600ohm, the insertion loss is 8dBmax and the max power in is 10mW. Would I need a buffer amplifier between if I connect both of them in series? I suppose I would need an amplifier at the main output of the series filters?

 

[vfone]

high performance ssb audio filter

Is important to adjust the carrier point (placed on the filter skirt), for the best audio frequency expansion (more audio octaves), and of course for maximum rejection of the unwanted side band.
More octaves means if you start for example from 200Hz up to 2.7kHz you get approximately 3.5 octaves.

Theoretically you can get better than 40dB rejection for the unwanted side band (depends where you place the carrier).

Whatever I will recommend for carrier adjustment, nothing compares with your final manual tuning to get these performances. In parallel with lab test equipment, as oscilloscope and spectrum analyzer, you need to use a selective receiver to verify the voice quality and its frequency response.

Your mentioned insertion loss and impedances are typical for this kind of crystal filters.
Yes, you need an amplifier to compensate the loss when filters are placed in series.