lathe brains restoration #3

THE LATHE, aka: LS-76, and the LJ-10 and LJ-12 tape machines
Post Reply
User avatar
Site Admin
Posts: 87
Joined: 03 Apr 2013, 03:13

lathe brains restoration #3

Post by Fonotec » 04 May 2018, 00:14

lathe brains restoration #3 - a refurbishment diary from October, 2017
Last Fall, after a year of advance notice, a complete set of LS-76 (THE LATHE) circuit boards in their own card cage arrived from Nigeria, by way of AUS ...
left_cage.jpg (28.55 KiB) Viewed 678 times
right_cage.jpg (43.3 KiB) Viewed 678 times
...thanks to AussieOzborn's wish to part with the lathe that accompanied his newly-acquired set of vintage Ortofon GO 741 amps (which were recently extensively modified for use with a Neumann SX- cutter).
Ortofons.jpg (46.96 KiB) Viewed 678 times
The feedscrew on that lathe (#662) was completely corroded, and there was no tt platter or carriage...
feedscrew.jpg (44.48 KiB) Viewed 678 times
There was no turntable motor. The tt servo drive board was in the first slot of the card cage, but it would only be of use to someone who had the correct diameter capstan sleeve (for the right motor's shaft) and the correct diameter turntable platter (since those on original LS-76s are 17", while later ones have 16" platters), which would be driving a seamless Mylar belt around its recessed rim...

Interestingly, it was an updated version of the tt servo drive board that did not appear in either of the manuals I have. It might well be the last circuit that L. J. Scully Manufacturing made. It appears to be based on the original tt servo drive which I cloned. But it uses some newer parts, such as a microcontroller and PROMs. Also, the crystal for the master clock has a much higher resonance frequency - 3.45 MHz, instead of 432 kHz.
4_70901.jpg (97.82 KiB) Viewed 661 times
There was also no power supply or even snakes from the power supply to the card cage (having chassis-mount multi-pin Winchester connectors of the MRE type), and the Control Panel was clearly 'rotted one note'.

The recent dismantling of Rhythm Shack's #654 which I heard about nearly too late, nevertheless resulted in some otherwise unobtainable wood paneling and the original mic boom arm becoming available to help restore #660 which wanted those very items. Similarly, the dismantling of the Nigerian LS-76 had already begun...
ChipJar.jpg (64.35 KiB) Viewed 678 times, I only wanted the card cage and the cards within - plus the strobe circuit board and the 7-segment l. e. d. LPI digital display circuit board from the Control Panel.
strobe_board.jpg (31.42 KiB) Viewed 659 times
Display_Board.jpg (59.8 KiB) Viewed 678 times
I've done a lot of work refurbishing those boards and have stored them in the wall safe at the studio, along with the other spare set of circuit boards from #660 - which presently only needs to use the tt servo drive board, as it is merely a test jig for mods that I might try to add onto the main lathe, #656, and also for QC'ing nickel stamper 'mothers', since it's my only other tt that can accept a 14" record. It's fitted with a Shure DJ cart. (Shure is now leaving the phono products industry!)



Saturday, October 21, 2017

Yesterday, I worked on the [Feed] Servo board. It has a large, black metal heatsink (shaped like a grill cheese sandwich) to cool the Torque Systems PA-112 hybrid op amp. I have this version of the Servo board for Scully #656 and #660, so, having a third, makes a welcome spare. Had it been the newer version of the Servo, it would be somewhat on its own, having a slightly-different pinout for the Apex PA-10 hybrid op amp... It’s nice that that op amp is still available on ebay, but the PA-112 was made by the same company that made the feed motor (and tt motor, incidentally), so, it’s definitely appropriate and utilizes DC voltage feedback from the motor tachometer.

This Servo board also forms the first candidate from LS-76 lathe brains restoration #3 for cloning, which entails completely denuding the through holes of their components and having the board scanned into a computer’s memory for Gerber File creation - files which can be used by outfits such as Sunstone Circuits (in Oregon) to make new, identical, drilled, plated, and masked PCBs for packing with new and NOS components, having brand new traces and edge contacts to help keep the old machine running well for many years.

The Scully feedscrew can't truly be cloned at this time because Scully made their own burnishing machine that would deburr each newly cast feedscrew for ~ 40 hours, each, running it back and forth, slowly removing any small irregularity.

All of the digital signal processing within the LS-76 disk computer is discrete, through-hole technology, although it includes ADCs, DACs, shift registers, one shots, flip flops, and hex inverters… …variable modulus counters, AND, NAND, and multiplexer… ICs, all, using the +5V TTL protocol in the 74xx, and 93xx, packages (and 54xx and 74Cxx versions, as well).


The Servo board, as all the others, and the card cage, itself, was first 'aired', outside on a warm, dry Autumnal afternoon, and also blown with compressed nitrogen from a K bottle with an air-spray gun on the sidewalk, on cardboard, in front of the studio. Vocoder sings: ‘Throw down some cardboard and spin on your head…’ (Pretty Tony 12" LP: 'Fix it in the Mix')

The cage was also dust-blown.

A lot of dust was removed by the gun-blasts, but a lot remained, cleaving to each package, often, only on one of its small surfaces…

So, the Servo board was rinsed with 99% pure isopropyl in a sink. The residue was gently brushed with the shop-toothbrush, using just its weight for cleaning force and, in some places, a little more… Then the board was rinsed, again, with the iso.

Three tantalum capacitors on the Servo board each had one leg completely rusted through. Those caps were replaced with precision film ones (22 µF 15V) from studio shop inventory, The several µA748 eight-leg op amps were removed from their sockets. The sockets were brushed with the little De-Oxit nail-polishing brush. The De-Oxit residue on the contacts of the sockets was further brushed around with the more rigid bristles of the shop-toothbrush, using little more than its own weight as down-force during the chemical cleaning. (Wire brushes should be avoided when restoring old electronics, because component legs and contacts that are gold-plated have a very thin phase which can be scraped though easily by metal bristles.)

Some of the original op amps had legs broken off and stuck in the sockets from having rusted through. Some leg stumps had to be removed from the sockets by using a bench microscope, a sewing needle within a hilt (as skewer), and tweezers. A vintage mic lamp was shined on the workpiece.

Other op amps were intact and received a PASS in the ABI Linear Master. I reused the clean-looking ones that PASSed and, of course, discarded the others. I replaced each of those with new chips from the studio shop inventory.

I tested every diode - all were still dropping Volts, the one way, and blocking current, the other, quite well.

I measured the capacitance of the other caps, in situ. All seemed plausible - none had exploded or even bulged.
All transistor legs presented with the appropriate voltage-dropping and current-blocking (PNP, NPN) when the DMM was in diode-check mode.

Also, today, I worked on the Display Processor board. I cleaned it with the iso rinse and shop-toothbrush as well. I tested each chip by removing it from its socket and using either the Linear Master or the Chip Master.

The three 9602 (dual, retriggerable, resettable, monostable multivibrators) on the top row had heavily rusted legs, so I didn’t even bother testing them. I cleaned their sockets, as above, and packed them with new old stock chips from studio inventory.

[If needed, I can test the 9602 IC, but not in the ABI units, which don’t have that model in their libraries. The Heathkit IT-7400 chip tester, however, can 'play' a 9602 chip and a good one will always make the same pattern of blinking lights when you press its SHIFT button, provided you have the pin switches in a certain position.]

The Display Processor board has four PROMs that are socketed, like the other DIP and ceramic chips. The PROM used is a fuse-link type, having the Ni-Cr link matrix design. Most PROMs are programmed destructively by blowing selected fuses, Whereas, with the specific variety of Ni-Cr link matrix used in these Signetics model 82S126 'fuse-link' PROMs, nothing is 'blown' - rather, the burner 'fuses' a link across selected addresses and refrains from fusing links, elsewhere... Fused links become logical 1. Unfused ones remain at logical 0.

Sockets 5, 6, 16, and 20 have such a PROM with a unique code for each. They help convert the signals in the Display Processor into lighting instructions for the 7-segment l.e.d. lamps on the Control Panel which show the lines per inch reading digitally (e.g., 400 LPI).

Several of the bottoms of the legs of PROM: CM6 broke off upon removal due to heavy rusting, but the tops might be weldable with wire (for the purposes of cloning via the DATA I/O). It would only have to work long enough to read its programming into memory.

The three, other PROMs (CM5, CM16, and CM20) seem to be in good shape and will be the first candidates for cloning the PROMs for studio inventory. If they work in the QC lathe, then they’ll work in the cutting lathe. I already have the fuse-link variety of NOS blank Signetics PROM chips at the studio.

If CM6 can’t be read, but the process is proven to work for the other PROMs from #662, then I’ll be willing to put 'S6' (as it is called in older-version LS-76 brains) from #660 in the DATA I/O and make a clone for it that way. The third set of lathe brains for testing and cloning greatly lowers the apparent risk, so, something irrevocable can hopefully be avoided with these examples of Unobtainia.


Sunday, October, 22

I worked on the Fixed Pitch board, today. I rinsed it with 99% isopropyl and gently brushed the board and sockets. The sockets were also brushed with De-Oxit. Chips - tested. ...some, discarded, and replaced with new old stock, others kept, due to looking clean and getting a PASS in the checker.



I worked on the General Logic board, today. Same procedure as above. Several chips failed the ABI test, but some PASSed. Mil-spec AND gates were installed in the appropriate sockets. All transistor legs showed the correct voltage drop when the DMM was in diode-check mode. All diodes also tested as good.



I worked on the Variable Depth board, today. It has a daughter board as well as the main board which have the letters, A and B, after the model number - A, for the daughter board, and B, for the main… Same procedures as above. Several chips failed the ABI test, and others had such corroded legs that some of them broke off during careful removal. I soaked the wells of the sockets with DeOxit and gently brushed them with the shop-toothbrush. Then, using the bench mic for viewing, I skewered out the rusted pin-stumps using a sewing needle within a hilt. I had to replace all six of the 7496 5-bit shift registers, a 9602 'one shot', and several uA748 op amps. Overall, it still looked pretty good, however. All transistors caused the appropriate voltage drops when the DMM was in diode-check mode. A medium sized 470 µF electrolytic cap was replaced with a 500 µF cap from inventory. (All three sets of lathe brains, so far, had a 470 µF cap, originally, although the schematic shows 500…)



I worked on the Variable Pitch daughter board, today, and started on the main board. Most of the uA748 op amps on the daughter board needed to be replaced - had rusted legs and would fail. Some skewering was needed in a few of the sockets to remove a broken off piece of leg. I also worked on the card cage, itself - cleaning a lot of the dirt. One screw is refusing to unscrew.



I worked on the Variable Pitch main board, today, and also rinsed off the new version TT servo drive board (PN: 4-70901) with Techspray (99% iso).
All of the 7496 5-bit shift registers had to be replaced, as well as both 7493 binary counters. Fortunately, the TTL device inventory at Lot 33 is robust.


Friday, October 27

I worked on the Display Processor board and the PROM for A6 (with 'CM6' on the package) which was on that board. Most of the socket holes were ok, but several held onto their guest chips’ heavily-corroded legs when they were slowly removed, after rinse and gentle toothbrush-agitation with 99% iso and then oxidation-dissolving treatment via a 24-hour De-Oxit soak, followed by resoak, with gentle shop-toothbrush agitation and followup soak…

{Watched a video about how to use a Dremel (mini) rotary grinding stone to remove the corner of package material of a chip and reveal the top of the pin that’s horizontal. If one fluxes and tins that bare spot of pin which was revealed (quickly - under 2 seconds’ dwell time with hot iron), one can put the chip in a machine tooled (16-pin) socket, stick a piece of solid core wire (24 AWG) in the hole of its socket (since it’s not filled with a pin) and bend the wire over, on top of the exposed, tinned length of (once-)‘inner pin,’ and solder them together. Leave chip, thereafter, in the screw machined socket.
Might work at least for cloning that chip in the DATA I/O.

…also searched fora for advice on removing sockets. A couple of the sockets on the Display Processor board look as if they should be replaced. The best idea seemed to be to add new solder to each existing through hole weld (since they’re old and dry) before removing the socket by applying heat to its captured legs with the suction gun (or braid). Some had luck with using a long, thick soldering iron tip and heating an entire row of the socket’s pins at once, prying it out, as soon as those solders liquefied, with the thumbnail of one hand… Again, those pads had been newly re-soldered, first, to assist in wetting the old (40-year-), dried, factory solders.}


Monday, Oct. 30

Put some machine tooled, gold plated contact, 16-pin DIP sockets in my ebay cart - which already had some 7496, 7404, 7408, and 7493 items for replenishing some of the chip inventory used in repopulating the boards from #662. Will add 14-pin ones, and 8-s.


...a few weeks ago (March of 2018), inspired by the video I had seen last October, I went ahead and put the amputee Display Processor PROM for socket #6 in one of the new, 16-pin, screw-machine IC sockets I recently ordered and soldered solid-core 24 AWG [Ag] wires as jumpers between the exposed tops of the broken ends of several pins and the socket holes associated with them.
C6_PROM.jpg (69.63 KiB) Viewed 666 times
Continuity was confirmed with the Fluke DMM. Next I'll need to fire up the old DATA I/O 29B PROM burner to try to store its memory in volatile RAM and then 'write' that onto a new-old-stock blank (fuse-link) Signetics PROM.

It was important to 'stock up' on this back-up set of 'brains' (for either #656 or #660) , since the Scully factory has been closed for decades and no one who specializes in repairing lathes in the U. S. is willing to work on these mid-70s-era TTL circuits.

Post Reply