Re: {Collins} 32S-3 XTAL LO frequency high across all bands, excessively at 15M and above

With vacation, a heavy travel schedule for work, and other distractions, progressing only slowly here.  Just as a brief re-cap (no pun intended), using the same set of xtals in my 32S-3 and 75S-3, the measured 32S-3 LO frequencies are all higher than the 75S-3 frequencies.  The delta is higher at the higher frequencies, but exceeds 1.5 kHz at the higher frequencies and for the same xtals.  In the 75S-3, all the xtals are within what I consider to be an acceptable nominal range (worst case band was 800 Hz high), but in the 32S-3 all of the xtals were at least 1.4 kHz high.  Clearly the xtals are not the problem since these were the same in both measurement sets.  Testing and measurement of other related LO components in the 32S-3 showed all to be operating normally and within spec (including swapping several NOS LO tubes)

I have measured the capacitance across the 2E/28B xtal socket on both my 75S-3 and 32S-3 with an Agilent U1732A LCR meter and found the following: the 75S-3 capacitance measures ~16pF, and 32S-3 capacitance measures ~13pF.  Not much of a difference.  To reassure myself the measurements were valid, I immediately measured with the same short test leads a brand new 12pF  5% silver mica, and it measured 12.1pF.  Both Richard Knoppow and Bill Carnes had independently indicated Collins designed these rigs for 30pF parallel capacitance at the xtal sockets, so I'd appreciate their comments here, as well as any measurement feedback from others who have measured this capacitance.  It strikes me as odd that the 32S-3 has the higher parallel capacitance, but also the higher LO frequencies.

As stated previously, I connected a 3 pF cap between V12 pin 1 (LO control grid) and ground on the 32S-3, and the LO output for the 28.5-28.7 MHz band dropped 1.2 kHz, which put the LO frequency within 120 Hz of its nominal output frequency.  LO output drive dropped by only 0.5 dB.

I'm reaching a point of diminishing returns here, so barring any insights from this group, I'm going to hone in on an appropriate value of C (V12 pin 1 LO control grid to ground)  to bring all 32S-3 bands within a reasonable LO frequency tolerance.  I would be curious if anyone else has measured the LO output across xmtr/rcvr pairs using the same xtals in both rigs, and if so, what you found.


----- Original Message ----- 
  From: Gaylord Hart 
  To: Gaylord Hart ; Collins Listserve 
  Sent: Monday, June 17, 2013 2:08 AM
  Subject: Re: 32S-3 XTAL LO frequency high across all bands, excesively at 15M and above

  Here's a brief update.

  I've been traveling a lot for work so progress is slow.  I'm still gathering data, but here's what I've found so far.  I'll do a broader write-up for the list when I finish gathering measurements.

  I measured the capacitance across the crystal socket with an Agilent U1732A LCR meter.  I get ~16 pF.  I trust the meter within a couple of pF.  It has short test leads and is calibration with the leads attached.  I need to do this measurement on my 75S-3 as well since the same XTALs provide a very different LO frequency when installed in the 32S-3 and 75S-3.

  All measurements noted below were with the band switch set for 28.5-28.7 MHz.

  I paralleled 150 pF C across C105 (82 pF on schematic), and this dropped the LO frequency 1 kHz, but also dropped the LO output significantly.  Not the source of the frequency delta.

  Substituted XTAL LO tube with NOS and substitute NOS tubes.  Worst-case frequency delta was under 200 Hz, so this is not the source of the problem, either.

  Also, I confirmed tuning the plate-side trimmers of the LO has no impact on LO output frequency.

  I then connected a 3 pF cap between V12 pin 1 (control grid) and ground, and the LO output dropped 1.2 kHz, which put the LO frequency close to nominal output.  So the issue seems to be load capacitance on the XTAL, grid side of V12.  This seems very sensitive to me.  Would appreciate comments here.  I need to measure LO frequency on all bands with a cap paralleled here, but that awaits another available time slot. 

    ----- Original Message ----- 
    From: Gaylord Hart 
    To: Collins Listserve 
    Sent: Sunday, June 02, 2013 2:35 AM
    Subject: 32S-3 XTAL LO frequency high across all bands, excesively at 15M and above

    After several weeks of Service Bulletin upgrades, repairs, and tune-ups I finally got my 75S-3 / 32S-3 combo on the air.  The 75S-3 re-furb went without a hitch, but the 32S-3 was a different story.  Basically everything on the 32S-3 functioned, but had low drive levels, a parasitic, and other minor issues.  Also, this 32S-3 was a 1965 unit with no upgrades, and some factory original downgrades from previous SB's that were later "un-installed" with superseding SB's from Collins.  I owe a great deal of gratitude to Bob Jefferis (KF6BC) of this group, who has restored several 32S-3's and who very patiently walked me through all the changes in schematics over the years as well as provided measurements and other insights to guide me through the process.  I can't tell you how much time he saved me.  

    As part of my close-out for any rig I re-furb, I capture voltage and resistance readings for every pin for every tube as future reference points, and I also capture key reference points for operating parameters (local LO and PTO frequencies and output levels, RX sensitivity levels, XMTR power out levels, etc.).  As part of this process, I measured the XTAL LO frequencies on the 32S-3 and discovered several of these at and above 15M to be too high in my judgment--well over 1.5 kHz high.  I thought these original XTALs were bad, and since I needed to order 10M XTALs to fill the sockets anyway for a full range, I ordered new 15M XTALS, too, all from ICM, so these are all new manufacture.  BTW, ICM charges just under $24 for these custom crystals, and they deliver in about 3 weeks.

    To my chagrin, the new ICM XTALs were all high, too.  So I decided to test all my Collins XTALs (original 75S-3, 32S-3, and KWM-2 XTALs plus the new ICM XTALs I purchased) using my 75S-3 and 32S-3 as the test units for measuring the XTAL operating frequencies.  What I found surprised me.  In general, at 80M the 32S-3 XTAL LO frequency output was 300 Hz higher than the SAME XTAL measured in the 75S-3, but this increased to over 1.3 KHz at 15M and above.  If anyone is interested in the original spreadsheet, let me know off-line and I'll send it to you.

    I have excellent XTAL LO performance on my 75S-3 (WC delta from nominal = 804 Hz high across all bands), but my 32S-3 is over 2 kHz high WC.  Given that the same XTALs are OK in the 75S-3, but are too high on the 32S-3, I suspect the 32S-3 has some "issues" in the XTAL LO.  I'm just beginning to dig into this, but I'm hoping someone in the group can provide some insight here.  Of course, I'm sure I can find a way to load down the 32S-3 LO XTALs with extra C to lower the LO frequency, but I prefer to keep my rig "original" if possible.  BTW, I replaced the LO oscillator tube in the 32S-3 with an NOS tube as a first step, but this only raised the frequency higher.

    Gaylord Hart

This archive was generated by a fusion of Pipermail (Mailman edition) and MHonArc.