[Collins] More on the Collins synthesizer.

[Dr. Gerald N. Johnson]

On Sun, 2006-10-15 at 00:34 -0400, kiyoinc at attglobal.net wrote:
> I found one kit on the internet that sorta does what I want but it was 
> way up there, over $200.
> 
> It did a lot more than I want, which is simply to replace the crystal 
> bank with a DDS VFO that steps in 200 kHz increments and displays the 
> frequency of the band segment.  14.2  14.4  14.6.
> 
> So it's only three digits.  Also the step would have to be programmed 
> for the S-Line PTO + IF.
> 
> The DDS designers are thinking VFO replacement.  Well, what's the point 
> of that when you have a good PTO that reads to 1 kHz or better.

To a DDS, step sizes are mostly arbitrary. Its no harder to step 100 Hz
than to step 200 KHz always landing on .x55 KHz that the S-line needs.
> 
> The area that really needs the work is the front end oscillator.
> 
> I do have a Collins crystal bank, it's almost complete. I will be 
> ordering a switch from Glen when I figure out which model I want.

Lloyd Winter, group head on at least some of the S-lines and my boss at
Collins told me the 75S3B with crystal pack was a far better receiver
than the 51S1. He'd been involved with both.
> 
> What I really want is a DDS device that will turn my 75S-1 into a 51S-1, 
> a real band cruiser, and do it for $50 or $100.  I think it could be done.

I don't think it can be done for $100. The big problem with DDS is the
bunch of spurs that abound that come from the sampling process in
creating the output frequency. Fundamentally you need a clock several
times the highest desired output frequency. Then you have to watch out
for fold back and harmonic fold back. Those put spurs right close to
your desired output and are impossible to filter. Say you have a 100 MHz
clock and desire 24 MHz output. Easy enough... but 24 x 3 = 72 and 100 -
72 is 28 so you have a spur at 28 MHz. But it gets worse. Suppose you
want 24.9 MHz. 3 x 24.9 = 74.7 MHz, sampling puts that also at 25.3 MHz.
Or even worse, 24.999 MHz. 3 x 24.999 MHz = 74.997 MHz, sampling puts
that spur at 25.003 MHz, just 4 KHz from the desired output. The main
solution for that problem is proper selection of clock frequency, which
often tends to be HIGH. There was an article in one of the HF or
microwave journals a month or two ago on a way to cut those spurs by
programming the controller microprocessor to produce components on the
spur frequency but out of phase. Not something to be accomplished in a
$2 PIC chip micro. Getting a good wide range (6 to 32 MHz IS wide range)
DDS with decent spurious outputs is not at all trivial.

> 
> de ah6gi/4

-- 
73, Jerry, K0CQ, Technical Advisor to the CRA
All content copyright Dr. Gerald N. Johnson, electrical engineer