--- In [email protected], "Walter" <[EMAIL PROTECTED]> wrote:
>
> 
> > Ugly, but it worked.
> 
> Immediate checking: everything ok. 
> 
> Now what? 
> 
> Somebody here who has a working Elektor board with proper 
> mirror rejection? MAy we see a screenshot?
> If so we could conclude that not ALL 
> boards are defunct. 
> 
> Best regards, 
> 
> Walter
>
 
I Walter and the group,

I am sorry for your problem. I will try to guess where the problems
(s) may be... a guess.

I have not the Elektor SDR but I have downloaded the article and 
looked at the circuit diagram.

For such a design the best IC would have been the FST3125 or 3126. 
The simple reasons are: 

1) the Ton and Toff times are 10 times faster thar the one of 
74HC4066 (Ton = 11nS & Toff 13ns versus 1 to 1.5nS of FST3125 and 
0.5/0.5nS on the Pericom PI5C3125).
Ton and Toff differences will have a big impact on the image 
rejection, plus limit on frequency coverage with repsectable image 
rejection. This even if the 74AC74 digital quadrature generator will 
deliver all four phases with a good quadrature. Due to thes 
diofference in timing the quadrature signal should be modified in 
phase to "correct" the switching of switches.

2) The Ron of HC4066 is at least 5 times bigger than FST (4-5 ohms) 
this will permit lower conversion losses but also will help on a 
better lower difference on the switches Ron, reflecting in a more 
balanced amplitude.

You could replace the 74HC4066 (IC2)with the FST3126 to have the same 
gating as the 4066. You maybe able to adapt the pins connections.

You will need a scope to look at the amplitude of audio signals at 
IC5. If amplitude between the audio I & Q signals is different you 
will have a low rejection or nothing at all. Also if the phase of 
these sinals is in quadrature.

If you like to experiment with the FST3253 is OK, you will need only 
LO signal 0 and 90 degrees but the performances maybe lower that with 
FST3126.

Good fault finding.

73

Gian
I7SWX
 


Reply via email to