Radically rebuilding a PM2000 input strip to more closely match the great API 312 microphone preamp and 553 equalizer.
Yamaha quite obviously wanted their PM2000 to sound like the API 1604 consoles from the 1970’s. Many of the components are very similar. Last year I rebuilt a PM2000 and made a few modifications. That project left quite a few extra parts lying around my workroom, mostly input channel strips and transformers. I believe the most valuable parts on these circuit boards are the discrete opamps (NE80100 & NE80200) and the parametric EQ switch section. I had previously built a bunch of API 312 microphone preamp copies using the GA80980 input transformer, NE80200 discrete opamp, and GA80990 output transformer along with a portion of the original circuit board. I swapped around a few components and changed some values to get very close to the original API schematic. Those turned out sounding quite nice. Might as well try my hand at crafting a matching API 553 style parametric EQ based around the NE80100 opamp, already existing on the same circuit board as the mic preamp. I’ve built a few LCR filter sets to replace the Gyrator filters on the PM2000. Many engineers find the LCR filters more musical than the IC based Gyrator filter. I assume this is due to the extra harmonics produced by the high quality inductors and film capacitors in the LCR filter set. George Philbrick devised a great circuit in the 1940’s that has been used many times over the years by other engineers. The LCR filters are balanced between the positive and negative inputs of the opamp. The same technique was used in the famed Pultec circuit.
In order not to screw up vintage parts, I wanted to plan out my steps to avoid reworking or worse, destroying any of them. First I printed out copies of the schematics for the PM2000 input channel, API 312 preamp & 553 parametric EQ. I compared and contrasted the circuits to determine which parts were needed, which parts I could use, which parts I could move and which parts I needed to purchase new. After a few long nights interpolating two very different approaches to schematic drawing, I finally found a way to recreate the API circuits while recycling the PM2000 circuit boards. For the mic preamp, Yamaha created a nested pad/preamp/booster circuit. The input gain knob is an 11 position, 5 layer wafer switch. At minimum gain, the signal passes through a pad and the second opamp only (HA1457). At maximum gain, the pad is removed and both the discrete opamp (36db) and the monolithic opamp (22db) combine with the input transformer (6db) to provide 64db of gain. The API 312 is a much simpler circuit. One discrete opamp for gain, transformer in and out, with no capacitors in the signal path, resulting in 58db of gain. Moving on to the the PM2000 EQ, it is also similar to the API version, but again, the 553 is simpler, with similar components. The biggest difference on the PM2000 version is that the filter sets are positioned between the opamp output and the negative input. The API 553 positions the filter sets between the positive and negative inputs to the discrete opamp, removing the need for a DC blocking cap on the output. To keep the layout the same as the 1604, I would need to use an interstage transformer (600:600) between the mic pre and EQ. That eliminates the need for yet another coupling capacitor and adds a little more color.
I used MS Paint to edit the original schematics and PCB layout to match the circuit I sketched up. Moving things around so that the schematic still looked vintage, but the circuit was as close to the API circuits. There are a few resistor values in important positions that I wanted to experiment with. I wanted to be able to overdrive the interstage transformer if necessary, for saturation and effect. That means plenty of gain is needed from the NE80200. Out of that into another GA80990 transformer, a recycled PM2000 FB meter and XLR out.
Before I could get started, I needed a power supply. Digging in the junk drawer revealed a nice Danish toroidal transformer that would work for the +/-24VDC. I will figure out +48VDC for microphone phantom power later. Add a bridge rectifier, a few electrolytic capacitors and voila! power.
The most expensive parts of most EQ circuits are the controls. Wafer switches and “W” taper potentiometers are hard to find. But thankfully Yamaha did a great job engineering a compact, robust set of controls for four bands of EQ each with five selectable frequencies. They even have little circuit boards for each band to interface the filters with. Ground wires, concentric knobs, shorting resistors, all ready to use.
I prepped the Input PCB by stripping it of all components not necessary for the API circuits. Basically I would need both discrete opamps and the +/-24vdc power sections for them. Resistor values were chosen to rearrange the mic pre and EQ to match the API circuits. I reused the existing copper traces to recreate the API circuits. The reclaimed PCB real estate came in handy to mount the input transformer and LCR filter sets. I decided to leave the clipping indicator circuit just in case.
The preamp circuit was pretty simple. Creative use of existing copper traces determined the layout. The new EQ section is deceptively simple. Requiring just a few components. Again reusing the existing copper. The interstage and output transformers are too heavy to mount on the circuit board. A few jumpers made the necessary interconnections between the mic pre, interstage transformer, EQ section and output transformer. The filters were the biggest unknown for me. I wanted to distribute the filter frequencies to cover a wide range of choices, avoiding large gaps. The high frequency filter sets originally chosen by Yamaha were a good place to start. So I left those in place. The high mid and low mid filter sets were chosen based on inductor values available from Mouser while keeping the frequency spacing even. The low frequency filter sets turned out to be a harder problem. Finding the proper value inductors without a custom order proved difficult. I decided to work out the circuit using the high, hi-mid and lo-mid filter sets. I expected the low frequency filter sets would work themselves out somehow.
Adding a few TRS cables allowed me to measure and compare the signals created by the circuit. I used Rational Acoustics Smaart software and the Roland Octacapture USB interface to measure the signal at 4 points in the signal path. I was quite please when it fired right up. Injecting pink noise and cycling thru the filter sets produced a nice set of transfer function graphs. I swapped a few capacitor values to move around their frequency center and arrived at the frequencies shown below. I will need to adjust the filter resistors to try and match the gains more closely. I’ll figure out the low frequency section and how to case it soon. For now, it’s time to find the right inductors.