The thoroughly modernized Altec 1567A tube mixer
Due to circuit simplicity and component size, vintage gear is more easily serviced than most newer equipment, which is one reason I’m constantly revisiting the topic. In the process of increasing functionality and complexity, miniaturization has made visual analysis—and heat dissipation—of audio equipment more challenging. Yet some of the same basic problems plague all gear regardless of its place in the evolutionary time line.
I recently wrapped up work on an Altec 1567A (a 4-input vacuum-tube mixer), an Eventide Omnipressor (VCA compressor/expander), a pair of Sony PCM-7030 DAT recorders and a Neumann SM 69 tube mic (a stereo version of the M 269, which is essentially the European version of a U67). This odd audio collection spans some 50 years’ worth of electro-mechanical evolution. As viewed from the tech’s perspective, the internal real estate ranges from luxuriously spacious—Altec and Eventide—to tediously cramped, as in the case of the SM 69 and the Sony DAT.
Such restoration/repair tasks necessitate a steady hand, a good light and assisted vision, like reading glasses. Working near mic capsules and DAT heads requires a hyper-awareness of their vulnerability, so I’ve made protective sleeves and covers to keep capsules and heads safe. A speck of hot solder or even flux landing on a capsule diaphragm will leave its mark as a crater or a hole. An accidental screwdriver slip can puncture a diaphragm, scratch a head drum or, worse, break off the head chip. Such potentially costly mistakes emphasize the need for preventive measures.
Inside the SM 69, most of the components are soldered to two circular circuit boards, including the vacuum tubes. For that reason, only known, good tubes should be installed because the circuit boards won’t tolerate repeated tinkering. Many early circuits boards are similarly intolerant. I recommend Chem-wik Lite de-soldering braid (Chemtronics Part #10-100L, MCM Order #21-325).
Once the PCM-7030 transports were rebuilt, I relied heavily on comparative analysis by swapping transport and circuit boards, and checking system default parameters until both units behaved identically. The last anomaly to be resolved was the length of time each PCM-7030 required to power up; a nearly 30-second difference implied that a future failure was right around the corner. The problem was localized to the 5-volt power supply used by the logic circuitry (various microprocessors that deal with transport control, front panel display and user presets).
Of the two styles of power supplies, the linear type was found in conventional vintage gear like guitar amps, recording consoles and, with the exception of the Stephens, most analog tape machines. Switch-mode supplies, which have always been a key component in computers and cell phones, have been migrating into linear’s territory. Switch-mode supplies are not only more efficient at converting AC to DC, but are also smaller and lighter. An Apple laptop’s external supply/charger delivers 18 VDC at 4 amps if the line voltage is anywhere in between 100 and 240 volts AC. This is called regulation. The equivalent linear supply would be at least the size of a U87 storage case, be much heavier and require the user to set jumpers or switches to accommodate various line voltages.
The PCM-7030 has multiple linear supplies and one switch-mode supply for Logic that conveniently hangs off the rear of the unit and for which no documentation is provided. There was no obvious heat fatigue. However, on a whim—and a hunch based on a similar switch-mode problem found in an Alesis ADAT PSU—I swapped out a few low-voltage caps around the regulator and—voila!—not only did the two units now power up the same, but a third dead supply was resurrected. (Google “uc2844 application notes” for a typical schematic.)
Electrolytic capacitors tend to be the problem children of the passive-component family. They are used in power supplies and in-between amplifier stages. Fortunately, switch-mode power supply requirements have forced component manufacturers to improve these mission-critical parts. Look for brand-name capacitors (Panasonic, Nichicon) with low ESR (Equivalent Series Resistance), low leakage and long life at high temperatures (2,000 hours at 105 degrees C).
Tip: As with guitar amps, switch-mode supplies can have hazardous voltages inside, even when unplugged. Always unplug the power cord and discharge the main reservoir capacitor before doing any work.
DID THIS EVER REALLY WORK WELL?
This question applies pretty much equally to the Eventide Omnipressor and the Altec 1567A mixer projects. The Omnipressor was one of the earliest VCA signal processors—circa 1975—existing at the same time as Allison’s Memories Little Helper automation system and dbx noise reduction. In fact, versions of the Omnipressor used a dbx VCA.
As on all VCA-based processors, the signal path on the Omnipressor is very short: Input and output op amps interface the VCA to the outside world. These can be replaced or upgraded as needed. However, there are LM-301 op amps on the gain and attenuation limit controls that should not be upgraded, because they are in the sidechain wired as comparators. VCA circuits from that era typically have a distortion null adjustment; optimization via a spectrum analyzer is easier than it is if you are using a distortion analyzer alone.
A pair of LEDs can assist you in centering the processing around the hard knee, which is recommended, especially with the Omnipressor servicing as it can be wild and unruly. The 50-microAmp mechanical meter is failure-prone and the circuitry is a bit unusual—three different DC signals represent input, gain reduction and output levels. In gain-reduction mode, the meter should be centered.
ON TO THE ALTEC
The Altec 1567A was sent all the way from Iceland, and it was clearly worse off than when it began the journey. The octal-socketed transformers had fallen out in transit, rolled around inside the unit and damaged the filter caps. I can’t overemphasize careful packing emphatically enough. This includes separating tubes and other vulnerable un-pluggables into their own protective packaging.
This tube mixer has four transformer- balanced mic inputs, one unbalanced input and a balanced output with master bass and treble controls. Optimizing headroom and signal-to-noise is no picnic, but the 1567A was used in a wide range of applications, from sports events and other remote broadcasts to the original Woodstock music festival. Once the power supply and interstage capacitors were replaced, I brought this late-’50s mixer into this century. A 1U front panel was added with XLR I/Os, plus input pads, polarity reverse, a voltage tripler (to generate phantom power) and a few internal circuit mods.
The four input transformers each feed one-half of a 12AX7 voltage gain stage, followed by mixer level controls that passively sum together (via resistors). The common cathode 12AX7 was converted to cathode follower for current gain rather than voltage gain, allowing mixer levels to be turned up higher without fear of overload, improving both signal to noise and EQ network drive. To recover the necessary makeup gain, the feedback loop on the 6CG7 output amp was made adjustable via the master level control, varying distortion in the process but in a way that was far more controllably euphonic than the stock version.
There are times when the technician’s goal is simply to fix what’s broken. Other times, accumulated knowledge can be applied to make preventive repairs and “improvements.” In this case, trying to preserve desirable artifacts while minimizing the annoying noises required a mixture of both art and science.
Visit Eddie’s cave for visual and sonic samples.