Depotting

Discussion of rare things not covered in other fora.
Forum rules
Only taste in good post.
Post Reply
User avatar
Fonotec
Site Admin
Posts: 105
Joined: 03 Apr 2013, 03:13

Depotting

Post by Fonotec »

Some components have been "potted" by the manufacturer to conceal the circuit from would-be cloners. Not only are these encased completely, except for the pins. the interior air space has been claimed by foam rubber that must have been sprayed on before the lid was welded on, or was injected somewhow, and then allowed to dry inside. Depotting such a component is hard to do without ruining the part, but one can usually obtain the circuit and BOM (bill of materials) from such an undertaking.

[below: Original part (from 1970's electronics magazine)]

Image

[below: Potted sample that arrived DOA from an obso market vendor for $250/each]

Image


The job of cloning the hard-to-find, 6-bit, parallel output A/D converter*, used in THE LATHE for its Variable Pitch and Variable Depth cards (in order to keep the circuit original, in terms of, both, the voltages applied (up to + 9.5 vAC analog input before Zener shunt) and footprint (i.e., number and relative spacing of leads)), entailed cutting open the black plastic housing with a DREMEL rotary cutter and then tweezing out all the white foam silicone rubber stuffing. The package and stuffing both probably hastened the demise of the broken piece by causing heat to be trapped and to build, but, until recently, at least no one could tell what was inside... 0;

[below: depotted part (that had already stopped working) reveals components (uses two 4-bit binary counters (in 3-bit ripple-thru mode) a mil. spec. hex inverterº, and a voltage comparator op amp (LM710H))]

Image

[below: Newly cloned board just packed by Serif Sound with mostly new parts and some NOS [BOM]]

Image

[below: finished clone #000001; tested on Scully #656 VD card - works...]

Image


* one may be surprised that a 6 bit A/D converter is adequate resolution for a sound disc cutting lathe, but the converter is not used to digitize the audio being cut to disc. Rather, it is used to digitize the audio that is used as a side-chain, control signal in order to inform the variable pitch and depth computer how much and when to accelerate the base pitch of the feedscrew (its speed of rotation that corresponds to lateral motion of the cutting head carriage) in order to preclude new groove-turns overcutting existing ones, but without wasting space. The lathe computer is sensible to the audio's level, frequency, phase, and absolute polarity, so, compatible signals can be nestled closely, as "spooned" groove-turns (...'-turns', as in, it's just one long, inward spiraling groove of many turns).

The depth automation's reference audio, and that of the pitch automation's right-channel, is supplied to the on-board disk computer in advance of the audio that is supplied to the cutting head amps, so that the audio samples being responded to have enough time to be measured by the circuit and then acted on by the feed motor and depth coil in time for the desired carriage-, and saddle yoke-, displacement to be achieved at the time that the groove modulation of the program requires. In the case of THE LATHE, the delay is equivalent to the time required for one full revolution of the turntable (for 33 1/3 rpm, this is 1.8 seconds) because this amount of memory would allow the most space-savings, although decent groove packing has been seen in computers that only look ahead by 1/2 revolution, if they have a sufficiently long memory, because the left channel, being cut on the side of the groove that's closer to the inside of the disk, must be acted upon a full revolution of the turntable after it is recorded, in order to avoid overcutting into it. Therefore, the LS-76 disk computer also requires a copy of the Left Program audio - sent in real-time with it being cut, holding onto the voltage-measurements for one revolution before summing them with those of the right Advance pitch automation and the L/V pitch-expansion component of the depth automation expansions The bit stream in the shift registers is reconverted to analog just in time for the voltages to be applied to the base feed signal that will be sufficient to move the new orbit of groove out of the way of the old. In fact, the only thing the converter needs to measure is basso profundo and strong difference signals. Treble is negligible for feedscrew acceleration, and so is in-phase, quiet midrange, and so is extremely quiet bass, which is already reduced greatly by the RIAA disc recording equalization (i.e., ~ -20 dB @ 20 Hz...).

The cutter drive coils are velocity transducers, so they move in one direction until it's time to change directions. Therefore, loud bass wiggles, even after the RIAA recording curve is applied in the cutting amps, takes up about 4 times the space on a side that is taken up by the loud treble. The audio which enters the lathe's disc-cutting computer is still "flat" (i.e., sounds like the normal playback) so the disc-cutting computer circuit applies a soft facsimile of the steep RIAA (bass-boosting and treble-cutting) curve that gets applied to the groove modulating signal during cutting, leaving a bit more bass in than would be left by the cutting RIAA curve in order to compensate for the needs of the converters and to achieve the desired feed expansion "ballistics" (i.e. reaction time). That's why the 6 bit converter, which provides 6.02 x n - bits = 36.12 dB (plus 1.2 dB process gain) = 37 dB dynamic range, it is more than enough resolution for the requirements of the feed motor speed variations.


º ...and it must have at least the fourth inverter's output be "open collector."

{I happened to have two such custom hexes in stock, without knowing how unique they were in that regard, since I had purchased them from an ebay vintage component vendor, who probably didn't know that, on each chip, 4Y was like a 7405. It would have passed the test in most chip testers as a stock 5404, so that's how they were sold.. My hex inverter which reads, SN23808 SN5404J won't turn on the A/D converter. However, amazingly luckily, Model: SN23808J SN5404J will only make 4Y present a high impedance, but without "high" voltage," when 4A is "low." This is an essential difference from the standard, 5404 (or 7404), which makes all outputs high logic, with a high voltage. Due to the circuit design, only hex inverters with their 4th inverter, "open collector," will make the right waveform properly to regulate the parallel outputs of this particular A/D converter. I wondered why none of the 5404 chips in my inventory that passed the ABI ChipMaster test for a 5404 would make the clones work. Fortunately, I learned through DATEL that the part, 7405, is in their old schematics for the ADC-Econoverter, and I discovered that, indeed, it will work, also, with all outputs, "open collector." However, the DATEL potted converter which I depotted did not contain a part which read, 7405. It read, 9NO5/5404 DM. That might be a custom version of the 5404, but, with 4Y, o.c. (?)}
Post Reply