Bench journal · opened 18 September 2024

Journal

Short notes from the two rooms: repairs, things we got wrong, measurements worth keeping. Written mostly by whoever was holding the screwdriver.

  1. Capstan bearing, wire transport 2

    The spare wire transport developed a once-per-revolution tick audible on a quiet passage, about −52 dBFS, which is below the noise of any wire we have ever transferred but enough to annoy us. Bearing is a 6 mm bore flanged type and the replacement is on order. Until it arrives all wire jobs run on transport 1 only, so wire quotes went from 12 to 18 business days. repair

  2. A quarter-track tape that was really half-track

    Job 26-0431 came in labelled as a 1973 band rehearsal, four tracks. The box was right about the year and wrong about everything else: it was a half-track recording made on a machine that had been converted, so the quarter-track head picked up half of each track and everything sounded phasey and thin. Caught it in the first ten seconds by the channel correlation sitting near unity when it should not have. Re-ran on Deck 1 with the half-track block. Lesson is old and we relearn it twice a year: trust the tape, not the box. transfer

  3. Quarterly figures

    Weighted flutter: Deck 1 0.04 %, Deck 2 0.07 %, Deck 3 0.06 %, wire 0.09 %, cassette A 0.11 %, cassette B 0.12 %. Deck 2 has drifted from 0.05 % in June and the belt looks glazed; ordered, fitting it in November when the queue is shorter. All distortion figures at operating level under 0.6 %. measurement

  4. File specification revision 4

    Added the P2, P3 pass suffixes after a client with sixteen lacquer sides asked, reasonably, why our extra stylus passes disappeared from the delivery. They used to live in a working folder and get deleted. Now they ship when the pass was useful, and the log says which one we chose and why. process

  5. Vinegar smell, 36 reels, one afternoon of triage

    A church basement collection arrived with eleven of thirty-six reels clearly hydrolysing. We stopped the intake queue and transferred those eleven first, out of order, over four days. Two of them lost a short span to curl that no amount of tension juggling recovered; both are noted in the condition report with footage marks. The other twenty-five were fine and waited their turn. If you have acetate tape in a basement, this is the paragraph to read twice. intake

  6. We oversold hiss reduction and had to walk it back

    A client asked for «as clean as you can make it» on a 1988 live recording and we took the broadband reduction to about 11 dB. It sounded impressive on the near-fields and awful on headphones - the hall tail pumped and the hi-hat sizzled. We redid the restored file at 5 dB and sent both, with an apology and no second invoice. The 6 dB ceiling on the restoration page dates from that week. mistake

  7. Priya joins, cataloguing becomes a real step

    Until now the item sheet was whatever the transfer engineer could type between reels. Priya started on the first of the month and the sheets immediately got longer and more useful: spoken dates, names, which of three people is talking. Our average job gained about forty minutes and lost most of its follow-up email. people

  8. Head amplifier for the wire transport

    Built our own, because the stock electronics were a 1947 design with a hum field we could not get under −40 dBFS no matter where we rotated the chassis. Discrete input stage, flat to 15 kHz, transformer out. Hum on a blank wire went to −71 dBFS. Schematic lives on the wall in room 2 and we will photograph it for anyone who asks. build

  9. Why this journal exists

    Three clients in one month asked variations of the same question: what actually happens to my tape while it is at your place. Answering it properly in email took longer than writing it down once. So: notes, dated, including the ones where we got it wrong, because a lab that only publishes its good weeks is not telling you anything. note