As you may have noticed from some of my other Instructables, I have a passion for audio. I am also a DIY guy going way back. When I needed four more channels of microphone preamplifiers to expand my USB audio interface, I knew it was a DIY project.
Several years ago, I bought a Focusrite USB audio interface. It has four mic preamps and four-line level inputs along with some digital inputs. It is a great piece of hardware and met my needs. That was until I built a bunch of microphones. So, I set out to resolve this discrepancy. Thus, the SSM2019 Four Channel Mic Preamp was born!
I had a few design goals for this project.
- It would be as simple as possible and use a minimum of components.
- It would have phantom power to allow me to use all the Pimped Alice microphones I have built.
- It would have a high impedance (Hi-Z) input on each channel for piezo transducers, a future project of mine. This would be an easy add if the case and power supply were already part of the main project.
- It would have pro audio specs: clean, low distortion and low noise. As good or better than the existing preamps in my Focusrite interface.
Step 1: The Design
I started studying what was already out there. I am very familiar with analog design and had my eye on the SSM2019, having previously used its older cousin, the now obsolete SSM2017. The SSM2019 is available in an 8 pin DIP package, which means it can be easily bread boarded. I came across some fantastic information on microphone preamplifier design from That Corp. (See the reference section) Unfortunately, all of their specific preamplifier chips are small surface mount packages. And, the specs are only marginally better than the SSM2019. I do applaud them for their knowledge sharing and design information. The specs on the SSM2019 are fantastic and like most audio operational amplifiers these days, will exceed the rest of the signal chain for performance. I used two fixed gain stages with a potentiometer allowing adjustment of the signal between them. This keeps the design simple and eliminates the need for challenging to find parts; such as antilog potentiometers and multi contact switches with unique resistor values. It also keeps THD + noise well below .01%
During my design process, I had an epiphany on phantom power. Most people think of 48 Volts as the “standard”. This goes way back and was important when the phantom power voltage was used to bias the capsule for condenser microphones. Currently, most condenser microphones use phantom power to make a stable lower voltage source. They use a Zener internally to generate 6-12VDC. That voltage is used to run the internal electronics and to generate a higher voltage to polarize the capsule. This is actually the best way to do this. You get a nice stable capsule voltage which can be higher than 48V if needed. The phantom power spec for microphones call out 48V, 24V and 12V. Each uses different values of coupling resistors. 48V uses 6.81K, 24V with 1.2K and 12V uses 680 Ohm. In essence, phantom power is needed to get a certain amount of power to the microphone. My epiphany was this: The voltage needs to be high enough for the internal 12V Zener to function. If I used the +15V available in my project and the appropriate coupling resistor value, it should work just fine. This actually solves two other problems. First is not needing a separate power supply just for phantom power. Second, and more important to my design is simplicity. By keeping the phantom power voltage at or less than the supply voltage for the SSM2019, we eliminate a lot of extra circuitry that is needed for protection. The guys at That Corp presented two papers at AES entitled “The Phantom Menace” and “The 48V Phantom Menace Returns”. These specifically deal with the challenges of having a 47-100uF capacitor charged to 48V in a circuit. Shorting that out accidentally can cause a lot of issues. Energy stored in the capacitor is function of voltage squared so just by going from 48V to 15V we lower the stored energy by a factor of 10. We also prevent a voltage above the supply voltage on any of the signal input pins of the SSM2019. Read That Corps design guide for examples of what is needed to make a preamp bullet proof.
Just to be transparent, I started this project thinking I was going to use 24VDC phantom power and then in the process of troubleshooting the power supply, came up with the idea of using the +15 already available. Initially I put the power supply inside the preamp case. This caused multiple hum and buzzing problems. I ended up with the bulk of the power supply in an external case with just the voltage regulators in the case. The end result is a very quiet preamp that is on par if not better than the internal ones in my Focusrite interface. Design goal #4 achieved!
Let’s look at the circuit and see what is happening. The SSM2019 block in the blue rectangle is main circuit. The two 820 Ohm resistors couple in the phantom power from the light green area where the toggle switch applies +15 to the 47uF capacitor via a 47 Ohm resistor. Both 820 Ohm resistors are on the “+” side of 47uF coupling capacitors that bring in the microphone signal. On the other side of the coupling capacitors are two 2.2K resistors that tie the other side of the capacitors to ground and keep the inputs to the SSM2019 at a DC ground potential. The data sheet shows 10K but mentions they should be as low as possible to minimize noise. I picked 2.2K to be lower but not greatly affect the input impedance of the whole circuit. The 330 Ohm resistor sets the gain of the SSM2019 to +30db. I picked this value as it provides the minimum gain that I would need. With this gain and +/-15V supply rails clipping should not be an issue. The 200pf Capacitor across the input pins are for EMI/RF protection for the SSM2019. This is right off the data sheet for RF protection. There are also two 470pf capacitors at the XLR jack for RF protection. On the signal input side, we have a DPDT toggle switch acting as our phase select switch. I wanted to be able to use a piezo contact pickup on a guitar (or other acoustic instruments) while simultaneously using a microphone. This allows for phase reversal of the microphone if needed. If it weren’t for that, I would have eliminated it as most recording programs allows you to invert phase post recording. The output of the SSM2019 goes to a 10K potentiometer for level adjustment to the next stage.
Now on to the high impedance side. In the red rectangle, we have a classic non-inverting buffer based on one section of an OPA2134 dual op amp. This is my favorite op amp for audio. Very low noise and distortion. Similar to the SSM2019, it won’t be the weakest link in the signal chain. The .01uF capacitor couples the signal in from the ¼” input jack. The 1M resistor provided a ground reference. Interestingly, the noise of the 1M resistor can be heard by turning the level of the high Z input all the way up. However, when a Piezo pick up is connected, the capacitance of the piezo pickup forms an RC filter with the 1M resistor. That knocks the noise way down (and it’s not bad in the first place.) From the output of the op amp, we go to a 10K potentiometer for final level adjustment.
The final section of the circuit is the final gain stage summing amplifier built around the second section of the OPA2134 op amp. See the green rectangle in the illustrations. This is an inverting stage with the gain set by the ratio of the 22K resistor and the 2.2K resistor(s) giving us a gain of 10 or +20dB. The 47pf capacitor across the 22K resistor is for stability and RF protection. The 10K potentiometers are linear. Which means that when the wiper moves across the range of rotation, the resistance from the starting point varies linearly with change in rotation. In the middle, you get 5K to either end. However, we hear differently. We hear logarithmically. Which is why decibels (dB) are used to measure sound levels. By using a 10K linear potentiometer feeding a 2.2K resistor, we achieve a level change that sounds way more natural. The op amp keeps the inverting input at a virtual ground. For AC signals, the 2.2K resistor is tied to the virtual ground. The halfway point of rotation is about -12dB attenuation with the last eighth of rotation only 1.2db of difference. This feels much smoother than a lot of other preamplifiers where the pot is changing the gain of the preamp. It works better than pre-amps that have a gain adjust potentiometer. Usually the last bit of increase causes a quick bump in the final gain and a bit of noticeable noise. The Focusrite responds this way. Mine does not. The signal is coupled out of the op amp via a 47 Ohm resistor. This protects the op amp and keeps it stable when driving a long cable run should you need to do that. One final thing for the two IC chips. These are both high bandwidth high gain devices. They must have good power supply bypassing with .1uF capacitors mounted close to the supply pins. This prevents weird things from happening and keeps them nice and stable.
To sum it all up, there are two fixed gain stages, a 30dB and 20dB for a total gain of 50dB. The level adjustment is made by varying the signal level between the two gain stages. There is also a high impedance input available on each channel that is perfect for piezo pickups and other instruments (guitar and bass) that need a bit of level adjustment prior to recording. All with very low distortion and noise. Phantom power is 15VDC which should work with most modern condenser microphones. One notable exception is the Neumann U87 Ai. That microphone is my pride and joy. Internally it has a 33V Zener for an intermediary power supply. For me that is not as issue as my Focusrite has 48V phantom power. All the rest of mine work just fine.
The Power Supply:
The power supply is an old school classic design. It uses a center tapped transformer, a bridge rectifier and two large filter capacitors. The transformer is 24VAC center tapped. Meaning we can ground the center tap and get 12VAC from each leg. Wait – aren’t we using +/- 15VDC? How does this work? There are two things happening: First the 12VAC is an RMS value. For a sine wave, the peak voltage is 1.4X higher (technically the square root of two) so that gives a peak of 17volts. Second the transformer is rated to supply 12VAC at full load. Which means at light load (and this circuit is not using a lot of power) we have an even higher voltage. All this results in about 18VDC available to the voltage rectifiers. We are using 7815 and 7915 linear voltage regulators and I picked ones from National Japan Radio that are plastic cased. This means you don’t need an insulator between the regulator and the case when mounting them. Initially I built the power supply internal to the mic pre-amp case. That didn’t work out too well as I had some hum and buzzing, all related to how close my transformer was to the internal microphone wiring. I ended up putting the transformer, rectifier, and large filter caps in a separate box. I used a 4 terminal XLR connector I had in the parts bin to bring the unregulated DC into the main case where the regulators are mounted close to the main circuit board. As mentioned earlier, initially I was going to use 24VDC for Phantom power and ended up not doing that thus simplifying my circuit and getting rid of the 24V regulator (and a higher voltage transformer!)
Step 2: Construction: the Case
If you haven’t noticed yet, my paint scheme and labeling are pretty funky. My kid was doing a school project and we had the three colors of spray paint available so on a whim I used all three. Then I got the idea to just hand paint the labeling with yellow enamel and a small brush. Pretty much the only one in the world that looks like this! I got my case from Tanner Electronics in Dallas, a surplus store. I found it on line at Mouser and other places. It is Hammond P/N 1456PL3. You may want to label it and paint it differently, that is up to you!
Step 3: Construction: Circuit Board
I built the circuit on a prototyping breadboard. First building one channel to ensure the design worked as expected. Then built the other three channels. See photo 1 and 2 for the layout. My OPA2134’s are from Burr Brown, which was acquired by TI in 2000. I bought 100 of these back in the day and still have a few. Notice the .1uF bypass caps all mounted on the underside of the board. These are important for stability of the IC chips.
Step 4: Construction: Front Panel Jacks and Controls:
Front Panel Jacks and Controls:
Depending on your case choice your layout may vary. I used Switchcraft panel mount ¼” jacks that will connect the front panel to ground. To minimize ground loops, connect the ground of the XLR jack (Pin-1) with the shortest length possible to the front panel. For my layout, I connected them to the ground lead of the “Hi Z” input jacks. I prewired the phase reversal switches by cross connecting the two outer connections of the Double Pole Double Throw (DPDT) switch. Then the microphone input from the XLR will go to the center leads and one of the outer connections to the circuit board. This way when the switch position is changed, the phase reverses. Before mounting the XLR jacks, solder on the two 470pf capacitors for RF/EMI shielding. This makes it much easier later! Mount the potentiometers on the front panel. I used a small sharpie or other marker to label things on the inside panel to help with connections later. And to remind me which lug of the potentiometers should be connected to ground. Then connect all the ground connections for the pots together using a common uninsulated bare wire. Later that connection will run to the common ground point.
Step 5: Construction: Internal Wiring
For the microphone signal wires, I twisted 22gauge wires together and connected the input XLR jacks to the phase select toggle switches. Twisting them together minimizes any stray EMI and RF. In theory, internal to the metal case we shouldn’t have any, as everything in this project is pure analog circuitry. Don’t worry about the phase specifically yet. Be consistent in how all the channels are wired. We will figure out in testing which position of the switch will be “normal” and which one is reverse.
For the rest of the audio wiring I used single conductor shielded and connected the shield to ground at one end only. This keeps our signals shielded and prevents ground loops. I had a roll of 26-gauge shielded Type “E” wire that I got surplus from Skycraft in Orlando a long time ago. There are vendors that sell it online or you can use a different single conductor shielded. For each connection, I prepared a length of it with the shield exposed on one end and the other just the center conductor. I put some heat shrink over the shield on the non-connected end to insulate it. See the photo’s. Work methodically and connect one thing at a time. I then tie wrapped each group of four wires together to keep things as neat as possible.
Step 6: Construction: Power Supply
I built my supply in a smaller project box. There is ONE thing you must do to make this safe and meet code. You must have a fuse on the primary of the transformer. I used an in-line fuse holder with a ¼ amp fuse. That will blow if the transformer draws more than 25W, which it should not. This whole thing uses at most 2W with four mics connected.
Prepare the voltage regulators before mounting to the panel by soldering on the two filter capacitors, 10uF for the input and .1uF on the output. I also attached input wires to them to prevent confusion later. Remember: The 7815 and 7915 are wired differently. See the data sheets for pin numbering and connections. After everything is mounted, it is time to make all the internal connections.
Power and Ground Connections:
I used color coded wire to connect the DC power leads to the circuit board. All the ground connections run back to one connection point in the project case. This is a typical “Star” grounding scheme. Because I had already built the power supply internally. I still had two large filter capacitors internal to the case. I kept these and used them for the incoming DC power. I already had a power switch in the case (DPDT) and I used that to switch the +/- unregulated DC power to the regulators. I directly connected the ground wire.
Once all the connections are complete, take a break and come back later to check everything! This is the most critical step.
I recommend that you test the power supply and ensure that the polarities are right and you have +15VDC and -15VDC from the regulators before connecting them to the circuit board. I mounted two LEDs on my panel to show that there was power. You don’t have to do this but it is a nice add. You will need a current limiting resistor in series with each LED. A 680 Ohm to 1K will work just fine.
Step 7: Construction: Patch Cables
This part could be a separate Instructable. To make this usable, you need to connect all four channels to the line inputs of the Focusrite interface. I plan on having them right next to each other so I needed four short patch cables. I found some great single conductor cable that was sturdy and not expensive at Redco. They also have good ¼” plugs. The cable has an outer copper braided shield and a conductive plastic inner shield. That has to be removed when making the patch cables. See the photo sequence for my cable assembly method. I like to take the shield and wrap it around the ground connection of the ¼” jack then solder it. This makes the cable quite sturdy. Although you should always unplug a patch cable by holding the connector, accidents happen sometimes. This method helps.
Step 8: Testing and Use
Testing and use:
The first thing we need to do is determine the polarity of the phase switches. To do this you will need two identical microphones. Which I am assuming you have, or you wouldn’t need a four-channel pre-amp! Connect one to a Focusrite mic pre-amp input and the other to channel one of the four channel mic-pre. Pan both to center. Hold the microphones close to each other and talk sing or hum while moving your mouth past the two microphones. Headphones really help with this part. You should not hear a null or dip in the output if the mics are in phase with each other. Switch the phase of the mic and repeat. If they are out of phase, you will hear a null or dip in level. You should be able to tell really quickly which position is in phase and out of phase.
I noticed with the level pot about half way I get nominal gain for my mics and that matches roughly where I normally set the Focusrite pre-amp gain knob to about 1-2 O’clock. Interestingly the spec on the Focusrite is up to 50dB of gain. When I have it turned all the way up (with no mic connected) I get a slight hiss. It is just a bit louder than my SSM2019 based preamp. I do not have elaborate test equipment available. However, I do have lot of experience in both the studio and live sound and this preamp is a top performer.
For the Hi-Z inputs, I soldered a Piezo Disc to a 1/4" jack and verified that everything works and the gain range is correct. I plan on testing this on an acoustic guitar in the near future.
I am excited about having a full eight channels of mic inputs available for recording. I have a couple MS microphones and 8 of my Pimped Alice microphones. This will let me experiment with different mic placements at the same time. It also opens the door for a project I have wanted to try for a long time – an Ambisonic microphone. One with four internal capsules intended to capture surround sound and multidirectional sound.
Stay tuned for several more microphone Instructables!
Step 9: References
These are a wealth of information for analog audio, mic preamp design and proper grounding for audio circuitry.
Data Sheet SSM2019
Data Sheet OPA2134
Phantom Power Wikipedia
That Corp“Phantom Menace”
That Corp Designing Microphone Preamps
Whitlock Audio Grounding, Whitlock
Rane “note 151”: Grounding and Shielding