the 32 mhz is too close to the 30 mhz LPF filter's cut-off. the can
oscillators are quite noisy too. i think we can expend a few transistors
and keep it simple : use the Si5351 (if you want) or the Si570 (if you
insist). either way, the rest of the circuit will remain the same.
- f
On Thu, Jun 2, 2016 at 12:40 PM, Ben Aupperlee <beninturkye@xxxxxxxxx>
wrote:
Hello everybody,
To solve this problem, I am thinking of my old HW101 from Heathkit, there
we had as input filter a tuned preselector, consisting of two tuned coils
and a tube inbetween, that was a good way of making life easy for the first
mixer, a homebrew diode ring. This preselector was used during transmitting
as the bandfilter to get rid of the unwanted sideband. I think if anything
else fails we could think of a modern version of this preselector,
unfortunately, it means a second (analog) tuning knop on the front panel.
(Unless we can figure-out a way to let the processor do this extra
tuning).....
As i think of this, i do have an old module from an earlier trx using
pindiodes to switch-in 7 bandpass filters, this might do, if i change the 6
meterband into 60. I still have the schematic somewhere . (4 and 10 MHz
away is far enough for even the sloppiest filter)
Another way is using one fixed oscillator block and two si570's or do some
more calculations and measurings and come-up with a number of low and
hi-pass filters.
My bet is for now on the si5351 using 2 out of 3, with an SMD oscillator
block of 32 MHz, using 42 MHz as first IF. And the tx bandfilter.
Regards Ben PA9B (in Turkije)
-------- Origineel bericht --------
Van:Ashhar Farhan
Verzonden:Thu, 02 Jun 2016 08:55:03 +0300
Aan:minima@xxxxxxxxxxxxx
Onderwerp:[minima] Re: looking for HF1
The ideal case, of course, is that we use the Si5351 for all the three
oscillators. let me take one at a time
1. 1st oscillator : this is essential as the 1st oscillator needs to be
tunable. There is a subtle catch : if there is a leakage of the any other
oscillator into this, the output will certainly show up in the transmitted
signal. we can live with a bad receiver, we will be arrested for a bad
transmitter. so, this is legal stuff.
2. 2nd oscillator : I chose 45 MHz IF frequency as it is easily handled by
conventional HF methods and it is sufficiently above the HF range to be
cut-off with the usual LPF filtering. The oscillator needs to be around 55
MHz. The trouble is that overtone oscillators are very particular about
being tuned. i am looking at ways of avoiding critical tuning. but one main
reason that we should look at using Si5351 for this application is that the
overtone oscillator will be a no-brainer.
3. BFO : the bfo needs to be tunable over a small range to allow for the
upper and lower sideband selection. at present we are using a VCXO to do
this. having Si5351 do this will make things really simple.
now, here is the catch, BFO, if it leaks into the VFO (1st oscillator)
will show up as a strong carrier in the transmit output as it is inside the
front-end LPF's passband. The 55 MHz oscillator, when leaked into the VFO
will beat with the oscillator and produce a response that is consistenly 10
MHz away from the signal. In fact, I faced this in the PCB that I designed
for the HF1. It works like this : let's assume that we are tuned to 14 MHz,
our VFO would be at 45+14 = 59 MHz. This 59 MHz will beat with the leaked
55 Mhz and produce a transmission spur at 59-55 = 4 MHz.
So, that is the whole problem set. over to you guys now...
- f
On Thu, Jun 2, 2016 at 10:48 AM, Ben Aupperlee <beninturkye@xxxxxxxxx>
wrote:
Hello Joe (and the whole bunch),
We can go for 2 out of 3, with possebly buffering or so, as there are
little SMD blocks containing Xtal and oscillator chip, virtually a one
component Xtal oscillator that can be useful for osc nr 2. Conrad.com sels
them for 1 to 3 Euro, single item price. They only do not have the exact
frequencies of 55 or 35 MHz, but I hope there are outlets that are better
sorted.
But whatever we do here, it will not interfere with the construction and
layout of the motherboard , so that can go on. As long as we agree on the
interface with the motherboard, and stick to that.
I see this oscillator sub-board as a rectangular board, standing on one
of the long sides, not higher than 2 or 3 cm so it fits inside a shielding
box, with some pins in the motherboard, 3 pins for signal and close by 3
ground pins, then somewhere on one end of the board power supply pins. The
clock and data lines comes directly from the digital board, mounted on the
front panel and housing the processor and LCD. And should be shielded from
the analog part of the set as this whole digital circuit radiates the 16
MHz clock frequency.
That is the reason I like to split the whole set in an analog and digital
part. (With a solid shield inbetween)
Somebody might contact the si manufacturer and ask if they have an
application note for this chip, explaning how to prevent or get around this
crosstalk.
I am interested to hear what Ashhar will think of our idear to use the
si5351, as I remember well, he rejected this chip some year ago because of
the crosstalk.
Regards Ben PA9B (in Turkije)
-------- Origineel bericht --------
Van:Joe Rocci
Verzonden:Thu, 02 Jun 2016 03:18:46 +0300
Aan:minima@xxxxxxxxxxxxx
Onderwerp:[minima] Re: looking for HF1
Ben
I don't have a dog in the fight, so I'll defer the modularity and
component-type selection to you. Sounds like an interesting version
though. I did want to point out my personal experience with the Si5351.
It's usable, but not spectacular. Having said that, I have a phasing-method
rig on the bench that uses it and I'm happy with the performance so far.
Numbers to follow some day.
Joe
Sent from my Verizon Wireless 4G LTE Tablet
-------- Original message --------
From: Ben Aupperlee <beninturkye@xxxxxxxxx>
Date: 06/01/2016 7:23 PM (GMT-05:00)
To: minima@xxxxxxxxxxxxx
Subject: [minima] Re: looking for HF1
Hi Joe,
Yes, but we came from a situation where we had an X-Tal oscillator for
nr2, and the idea was: replace it with a chip and save a lot of board space
an discreet components. On the other hand this might solve the crosstalk
problem.
Are we going to place a lot discrete components on that small SMD board,
or will we put them on the motherboard and only use the smd board for osc 1
and 3.
Groetjes Ben (in Turkije)
-------- Origineel bericht --------
Van:Joe Rocci
Verzonden:Thu, 02 Jun 2016 02:03:59 +0300
Aan:minima@xxxxxxxxxxxxx
Onderwerp:[minima] Re: looking for HF1
Ben
Oscillator #2 doesn't have to be variable, or even very accurate. A
simple crystal oscillator will do and you can still do passband tuning. I
suggest this because the Si5351 will probably have horrible crosstalk
spurious if all 3 oscillators are running. It's much better if you just run
#0 and #2 and keep the output traces well isolated from each other. In the.
Si570 approach, this eliminates an expensive 3rd part.
Joe
Sent from my Verizon Wireless 4G LTE Tablet
-------- Original message --------
From: Ben Aupperlee <beninturkye@xxxxxxxxx>
Date: 06/01/2016 6:45 PM (GMT-05:00)
To: minima@xxxxxxxxxxxxx
Subject: [minima] Re: looking for HF1
Hello Ashhar,
I had some exchange of idears about SMD and modules, so here a new
version with a new drawing, the old .gif is also used in this temp.htm text.
Regards Ben PA9B (in Turkije)
