Re: Modifying the 3.3V DDS oscillator supply [1 Attachment]


I’ve replaced the 30 MHz oscillators on several of the AD9851 modules and had no trouble removing the old oscillators using my hot air soldering station.  I didn’t replace them because of a failure, just to avoid one.




From: PHSNA@... [mailto:PHSNA@...]
Sent: Tuesday, April 04, 2017 3:07 PM
To: PHSNA@...
Subject: [PHSNA] Modifying the 3.3V DDS oscillator supply [1 Attachment]



[Attachment(s) from anthonyjaques94@... [PHSNA] included below]

         Modifying DDS modules with a 3.3V oscilator

   Recent DDS modules are fitted with a smaller surface mounted
crystal oscillator module than earlier (tin can) units. It has
been reported in several places that these are intended for 3.3
Volt operation. Running them with a 5 volt supply can make them
run hot, and could cause failure.
   Some manufacturers do make these to operated with either 3.3
or 5 Volts supplies, and in some cases they are marked as such.
But usually it is impossible to find out just what a particular
unit is intended for.
   It is possible to run the whole module at the lower voltage of
course, and the DDS chip will probably function. But it may not
operate at the 125MHz (or 180MHz) clock frequency with only a 3.3
Volt supply.
   A better way is to supply the Oscillator module with its own
supply, while leaving the DDS chip on 5 Volts.
   Several sources suggest using two diodes in series to drop most
of the excess, and for every module that I have tried this works.

   So far this is just a rehash of material seen elsewhere, However
I have seen no suggestions just how to do this physically. So here
is how I did it:
   (See the attached image.)
   There is a print track between the two Vcc pins that supplies
every component in the row between them (except R6) with a takeoff
between C12 and C10 for the DDS. The track left of C10 only
supplies C9, C8 and the oscillator.
   I cut the track between C9 and C10 (at the end nearest the
oscillator module!). This is fiddly, but anyone who can do the LPF
mod on these boards should not have any problem.
   I then connected two 1N41 48 diodes across the Vcc pins to restore
a reduced oscillator supply.
   Just one thing to be aware of; In the "Type 2" PHSNA that I
built there is no supply to the left most Vcc pin of the module.
But in other projects there may be. It needs to be disconnected
or the whole modification achieves nothing!

   On a related topic:  It is easy to find the earlier "tin can" types
of oscillator, (Ebay) and these could be used to replace a failed
surface mounted oscillator - except for one thing; I have found it
impossible to remove the surface mount oscillators! I suspect that
they are stuck down with an epoxy resin.
   What  finished up doing was to grind the original oscillator down
until the "works" vanished and then mounting the new module over the
top of the old one.
   Ok. It's crude. but it works, does restore a failed unit, and
fixes the 5V\ 3.3V problem at the same time.

Tony   (G3PTD).


Join to automatically receive all group messages.