Page 1 of 1


PostPosted: 17 Nov 2015, 03:22
by Fonotec
I noticed too much land at 50 Hz @ +0 VU (ref: 1 kHz = 7 cm/sec peak lateral velocity), even though 200 Hz and higher were being cut with only 20 % land between closest excursions at +0 VU, and nestled, where phase permits. 50 Hz groove modulation with automated expansion was leaving 50% land (i.e., 100% groove width wasted space between each 50 Hz lateral-modulated groove)! So, after consulting some op amp theory YouTubes and looking at the LS-76 schematics, I decided to try to modify the 3 dB point of the automation circuit's reference audio bass-boosting RC network. The values of the components in the drawing showed that the first RC network around the first op amp would have a 3 dB point of either 50, or 38, cycles per second, depending on whether a 100 k Ohm auxiliary resistor was in series with the mandatory 316 k Ohm resistor. Whichever it was, I thought I could reduce the low bass-caused land waste by using a resistor of a higher value than was already packed, since it would lower the frequency of the existing 3 dB point, and, therefore, 50 Hz logical expansions would be more finely spaced, due to 50 Hz getting less of a boost than before by RC network, VP R3A. I wondered if the auxiliary resistor had not yet been packed or if it possibly had been deleted. After looking at the components on the board, I realized that the auxiliary film resistor, R3A (100 k Ohm), was already soldered into its through-holes, allowing it to be used in series with the 316 k Ohm film resistor (a 0.01 µF film capacitor is in parallel around them, and this network is right in front of the non-inverting input of a µA748 op amp.). But R3A was nevertheless not being used by the filter due to its having a bypass jumper attached across both legs on both of these 100 k Ohm film resistors.

R3A.jpg (155.71 KiB) Viewed 3186 times

When calibrating the sensitivity of the Variable Pitch control signal inputs, 1 kHz is fed to the card at +4 dBm, which is equivalent to +1.228 volts (RMS) AC. Removing the bypass jumpers from R3A (on both audio channels) apparently moved the 3 dB point down to ~ 38 Hz, as expected, because it made the boosted gain at 50 Hz go down, from 3 volts AC (when the bypass jumper was in place), to about 2.6 volts AC (when it was removed), which causes the feed motor to expand noticeably less than before in response to 50 Hz. So, it tightened up the low bass expansions from ~ 50% land (before the deletion) to around ~ 33 % land, after R3A was restored! - a savings of ~ 15 LPI at my typical Pop/Rock pitch settings, while leaving the midrange groove spacing at 20 % land. That's potentially an extra 1.5 minutes at +0 VU per LP side, just from lifting a jumper on both Variable Pitch channels.

{It should be noted that the low frequency time constant of the RIAA emphasis/de-emphasis network is 3180 µSec, which is approximately 50 Hz, also. However, the main reason for the lathe automation's reference audio input circuit to include a boost around 50 Hz is that cutting heads cause such increased stylus displacement in the presence of low frequency program signals, compared to that of high frequencies, that even after the RIAA emphasis filter, which attenuates low bass by 20 dB, is applied to the drive coil modulation signal, the magnitude of low bass groove modulation ends up being still 4 times wider at wiggling than that of high treble groove modulation.}

In the drawing (from the manual, shown, above), I added the green arrows and also the boxed frequency values, next to the RC networks that straddle the right channel's first op amp. The networks are, of course, identical, for both channels. I'm thinking that 38 Hz is the 3 dB point of the bass boost that is established by these input networks, 2.7 kHz, created in front of the inverting input, is the stop band (?), since higher frequencies could be cut at base pitch (without expansion), and 9.5 Hz, in the negative feedback loop, is a HPF, set below the lowest audio signal likely to be modulating the groove.

There is no mention of this optional modification to the pitch automation in the lathe manual. But the schematic notation for the resistor has the letter, A, after the number, which probably means "auxiliary," which was my clue. Glad I tried it, since only a high quality external equalizer could have accomplished the same result, but that would be "ungainly." );

* On the LS-76, the pitch and depth automation circuit boards, each, take two audio inputs for reference. Ideal pitch expansion (when the base pitch increases, temporarily, to a lower number of lines per inch) is calculated from peak voltage measurements, and their relative timing, during each revolution, using of a copy of the Left Program signal and a copy of the Right Advance signal. The peaks are stored 25 times per turntable revolution and are acted upon every 1/5 revolution. By boosting the bass, the larger expansions will be in response to lower frequency content, as well as the broadband loudness, of the program.

Width is always twice the Depth, due to the right triangle shape of the jewel tip now commonly used. In the days of shellac 78s, a cutting stylus with an included angle of about 110 degrees, rather than 90, was used. This enabled a wide groove to be cut without having to drag the tip as deep into the blank material as would, nowadays, with the 90 degree included angle, be required. Depth is calculated from the difference voltage created by summing a copy of the Left Advance signal with a (separate) copy of the Right Advance signal, but out of phase with each other. When the program is identical on both channels, the difference signal will be 0 Volts, due to the signals nulling, causing the groove width to remain constant, at twice the base depth used. When program is different on one channel than the other, the difference signal needs to be large enough to invoke the necessary depth increase so as to preclude (stylus tip) lift off during impressing, and the Depth circuit will also, then, increase the Feed sufficiently to preclude overcuts, caused by increased groove width from the increased cutting stylus depth. This is referred to as "L/V" and signifies vertical pitch expansion.

The three Advance audio channels (Left, Right, and a Right copy) arrive at the lathe one turntable revolution before the Program signals arrive at the cutting amplifier, since the result of the audio voltage comparisons is done by semiconductors and has to be completed and ready for servo motor control in synchrony with the groove modulation. The inner groove wall (Left channel) lateral excursions and average groove width variation don't need to be acted upon by the feed expansion circuit until the following revolution of the turntable. So a copy of the Left Program channel arrives, in real time, with the cut, but, it doesn't get its peak values acted upon until the subsequent turntable revolution.

420 k

PostPosted: 12 Dec 2015, 01:35
by Fonotec
Today, I realized increased space savings by changing the 316 kOhm film resistor (R3) in the Left (Program) and Right (Advance) audio paths used by the Variable Pitch circuit with a 420 kOhm film resistor by Dale, for each. Now, the total resistance in the bass-boosting RC network in front of the non-inverting input of A10 and A1 is 520 kOhm, so, with the bass boosting time constant, now, at 5200 µs (i.e., '30.607...' c.p.s.), 50 Hz expands about 10 LPI less wastefully than when the 3-dB point was recently moved only to 4200 µs (i.e., '37.89...' c.p.s. (from ~ 50 Hz)). After re-calibration, I tried cutting again a side that I had the honor originally to cut last year for Bootsy, since it would make a good obstacle course for the mod, as it's very dynamic and has a loud kick drum. The groove-packing is now extremely tight - like a Zuma or a Compudisk. Side plays very well and was cut 2 dB hotter than before. Love that 1976 disk computer - now improved, again...

420 k - A-OK

PostPosted: 13 Dec 2015, 12:08
by Fonotec
Yesterday, I changed the two "R3" film resistors on the Variable Pitch card for LS-76 #660. So, that card now also has 520 kOhms total resistance in its bass-boosting RC networks, just like on the Variable Pitch card for LS-76 #656. The disk computer of #660 was being tested on #656, and all went well and showed a similar improvement in the efficiency of its expansions to that which was seen on #656 when using its own disk computer with the lowered corner frequency on the Variable Pitch automation audio bass ramp. So, it seems to be a valid mod for this lathe. If you operate an LS-76 (Hi Tor, Pete, and Deroy!), you might benefit also from swapping out the 316 kOhm resistor (VP_R3) for a 420 k... (has to be film (precision tolerance and stable value when heated) and can be 1/4 or 1/8 Watt). The exact value of the new resistors is not crucial, but it's important to cherry-pick a pair with as close a match in resistance (to each other) as possible, since they are affecting the lateral excursions towards the inner and outer groove walls, every half-cycle. The ones I matched yesterday were showing about 419.5 kOhms, each.