The only part of the 3586 frequency chain that is not locked to the 10 MHz reference is the BFO. Two 2 MHz range crystals are divided down to serve as injection oscillators at the final very low frequency (around 20 kHz?) IF.
While the BFO frequency might be off by a Hz or so, in the units I've had the audio note drift during operation is negligible. I don't remember the details, but the required frequencies weren't particularly straightforward to synthesize, and even for use in frequency measuring tests the crystals were stable enough that it wasn't worth the effort to replace them.
John
----
toggle quoted message
Show quoted text
On 11/11/24 00:08, Chuck Harris via groups.io wrote:
As I recall, one of the interesting things about the 3586ABC is
with all of its synthesizer capability, it uses separate crystal
oscillators to handle internal mixers. This is a known accuracy
flaw in an otherwise beautiful instrument.
My recollection is that with a little hardware acumen, one can
build a board to recreate their frequencies all phase locked to
the main reference, and drastically improve the frequency
performance.
For your purposes, look at the various crystal oscillators in the
system... one of them is off frequency, or not warmed up.
-Chuck Harris
On Sun, 10 Nov 2024 20:12:58 +0000 "Bill Riches via groups.io"
<bill.riches@...> wrote:
My HP 3586's counter reads about 24 Hz high on any input frequency.
The rear panel fx output reads whatever is input to the front panel.
It passes all TF tests with the exception of TF-13 - 3-2 (counter).
I swapped A22 from a parts donor and that fails the test also. ( That
A22 could be bad also!)
Any suggestions>
73,
Bill Riches, WA2DVU
Cape May, NJ
