EMRFD Message Archive 14899

Message Date From Subject
14899 2018-07-01 05:40:07 jwolczanski Speaking of square waves
Myself and KC4GIA are building a receiver designed by VK3XU*.  The mixer is a 74HC4066N driven by a pair of out-of-phase square waves created by an LO driving a 74HC04.  IF is a 4MHz ladder filter.

The square waves, if they can be called that, are really nasty looking.  Yeah, they're "square-ish" with a lot of spikes/ripple.  Is this a problem?

I am reminded of how a nice sine wave looks after being fed to a diode ring mixer....but that doesn't seem to create problems.  Is that the case here?

I can upload the O-scope display if necessary.

*The receiver is from "Radio Projects for the Amateur - Volume 3" by Drew Diamond, VK4XU.  This is a straight forward super-heterodyne design.  Drew's schematics are rendered very nicely with all the information one needs (pin-outs, coils, etc.) neatly drawn within the schematic.

Jerry W
Warrenton, VA

14900 2018-07-01 06:01:21 Tayloe, Dan (Noki... Re: Speaking of square waves
I have found in the past that a digital mixer like this excells at producing numerous very high level mixing products.  

I once used an an exclusive or gate to mix two signals.  On a spectrum analizer, I would see a "picket fence" of mixing products that were much more numerious, much higher level and fall off much more slowly compared to other more analog methods.

I suppose this could be a good thing under some unique set of circumstances such as trying to lock a VCO to some odd high order combination of n*f1 +/- m*f2  However, for a normal superhet receiver, it leaves the output open to a lot of birdies and unexpected out-of-band received signals.

Front end filtering does only so much to reject really strong out-of-band signals such as shortwave broadcast stations.

- Dan, N7VE

Sent from my Galaxy Tab® A

-------- Original message --------
14901 2018-07-01 07:47:01 kb1gmx Re: Speaking of square waves
The 4066 is the slow version of the current vogue FST part.

If the ripples are only a few percent of the tops its not an issue as these operate as switches on or off.
However they do have to have to be square or the system produces mroe even or odd harmonics and resulting mix products.

Usually the ringing you see is the problem of getting the wave signal from the source to the mixer
and the wire/PCB trace must behave as a transmission line as the ringing is basically reflections
from poor terminations.

14902 2018-07-01 07:53:11 kb1gmx Re: Speaking of square waves
In general sampling mixers and switching mixers have to recognize Nyquist and aliasing.
Also the fast chopping action is a great source of harmonics.  Look at a raw DBM with 
a spec analyzer, it can be very horrifying.

The key to keeping out the unwanted is not only good filters but, the right ones, and
respecting the overload behaviour of the device/system in use.   Some cases a small 
amount of attenuation can clean up something that seems terrible. 


14903 2018-07-01 11:57:01 jwolczanski Re: Speaking of square waves
Thanks all...

I'm going to try a few of the things suggested in this website:

I'll report back!

Thanks again
Jerry W
Warrenton, VA

14905 2018-07-02 13:34:51 jwolczanski Re: Speaking of square waves
So, I've been told that attempting to look at these square waves using my old
Tektronix (35MHz bandwidth), might be the problem - never occurred to me!
Thanks Mark (KC4GIA)

Nick (WA5BDU) also pointed out that, in this application, the important thing is
whether the points where they transition from logic 0 to logic 1 and the reverse
is correct and consistent.

Thanks all!
Jerry W
Warrenton, VA
14907 2018-07-03 05:23:08 i7swx Re: Speaking of square waves
Hi, I have been using LVDS Differential receiver, FIN1002, to drive the FSA3157 H-Mode Mixer. It is a 3.3V ic with TTL output with a maximum frequency betwwwn 200 and over 300MHz (400Mbs).
It is expecting a balanced input but it can work unbalanced too.
Certanly it has not a 50-50 but it is quite near.

If I do recall, there are/were a couple specified at 600Mbs.


14914 2018-07-04 06:56:07 jwolczanski Re: Speaking of square waves
As it turns out, the square waves I was trying to generate were probably fine all along.  At Nick Kennedy's prompting, I built a pair of scope probes with lengths of small diameter coax that had a 450 ohm series resistor at the end.  The ground was a short piece of wire around the braid.

At the O-scope end, again with Nicks advice, I used a BNC connector with one branch connected to a 50 ohm load and the other branch, my home-constructed scope probe. 

I uploaded a photo of the two outputs

Looking at these outputs with the "normal" scope probes shows a very different picture.

Thanks all!
Jerry W
Warrenton, VA