Last update 10/8/07

TapeOp 2004 Conference DIY Workshop Information

Belky and Fox, crashed after the week in New Orleans.  Thanks to everyone at Tape Op for making it another great time.  Also, thanks to all the people who signed up for the workshop; because of the success, the folks at Tape Op are already talking about expanding it for next year, keep your irons warm!

MY PROTOTYPES: MICPRE, OVERDRIVE

The two boxes above show a micpre, and an overdrive, made with the 2xHJFP board we built in the DIY workshop, I had these with me in the class.  Both of these applications are given in the handouts you received.  If you lost them, these are also given in the Sept/Oct 2003 issue (#37) of Tape Op magazine.

These pictures show the inside of the micpre.  The basic layout is the transformer/DI is wired into the first gain stage, the output of the first gain stage goes into the pot (white/green), the output of the pot goes to the input of the second gain stage (yellow/green, the green wires are ground and are redundant), and the output jack is wired to the output of the second gain stage.  The two gain stages on the board are label "IN1", "OUT1", "IN2", "OUT2".  The two stages are identical electrically, however the layout is slightly different.  This board is the REV G board, which a few of you might have (we went though all the parts and boards I brought, the REV G boards were extra I had in my shop, that ended up getting used up as well).  The only difference is the REV H boards have a different footprint for the LM78M20 voltage regulator.  On the REV G board above, I used the proto board section in the middle for the regulator circuit (shown in the above right picture).  If you use a wallwart (Mouser part number is given in your instructions, along with the mating power entry jack), I recommend using the regulator otherwise you will hear a faint but annoying noise floor.  With the regulator, its very quiet.

WIRING THE POWER:

It is very important not to reverse the polarity to the board when wiring in the power, if you do, it will torch the whole thing.  If the regulator is used, the power is wired into the regulator as shown with the red/green twisted pair in the picture below (note, in some of the other pictures, red/green was also used for other things).  You can also see in the pictures how the power entry jack is wired.

The center pin on the LM78M20 (+20v fixed regulator) is normally connected to ground.  On the REV H boards, you will see the center pin connects to a proto strip in the center section.  For the regulator to operate properly, this pin must be jumpered to ground.  This is easily done by jumpering the proto trace to the ground hole right next to it, shown in the picture above, far right where the metal pokers is indicating.  There is another pad on the board labeled "+24", use this if you are not using the LM78M20 regulator, and are providing clean DC power from a DC power supply.  Do not wire the wallwart into the pad labeled "+24" because this is actually the output of the voltage regulator.

Pictures of the overdrive circuit, with volume controls after each gain stage (i.e. pre/post), and a tone control on the input.  The tone control is very similar to what is in a guitar, a 500k AT pot and a 0.047uF cap in series, wired from the input jack to ground.  Teh layout for this is given in the handouts.

INCREASING THE REGULATORS OUTPUT VOLTAGE

Increasing the output voltage of the regulator has two positive effects, first it increases the dynamic range of the circuit, and second it lowers the power dissipation in the regulator.  To increase the output voltage of the regulator from 20vdc to 22vdc, two additional resistors are added as shown in the above right picture.  Connect a 2.2k 1/4w resistor from the center pin of the regulator to the output of the regulator, and a 100ohm 1/4w resistor (in the picture above its a 1/2w)  from the center pin to ground.  The jumper to ground should be removed. With the wallwart I used, it puts out about 27vdc under load.  The LM78M20 requires at least 3 volts DC from the input to the output to properly regulate (this is called the drop out voltage), which means if its run at 22vdc, it must get at least 25vdc in.  In practice, I have found it best to have at least 4 volts across these devices for proper operation.  Also, adding a cap to the ground pin reduces the regulators noise output (if its floated as described).  The power dissipation in the regulator is the current flowing through it, multiplied by the voltage from the input to the output.  The power dissipation in the regulator is calculated for both 20vdc, and 22vdc.  These are approximate values, each circuit will vary slightly in its current consumption. 

@ 20VDC out, P = (0.06A) * (27-20) = 0.42W

@ 22VDC out, P = (0.06A) * (27-22) = 0.3W

Approximately 30% reduction in power dissipation, with a 10% increase in headroom.

NOTES ON TRANSFORMERS FOR THE MICPRE APPLICATION

If you want something really nice, then either Jensen or Lundahl make superb tranformers.  IDeally look for something that is designed as a microphone input transformer, with a turns ration between 1:4, and 1:8.  The higher the turns ratio, the more gain it will have on the front end.  If you use a transformer with an 1:8 or higher, you should build a pad into it (shown in the application notes, and Tape Op issue #37).  Jensen also provides great application notes for their parts.

In a couple months from now, Hampton will be the authorized distributor for Altran, the company making our audio and power transformers.  These parts have outstanding specs, and will be very affordable.  Altran makes a wide array of magnetic devices, we will be distributing only the Audio components we have designed.  Currently, we have the following transformers in production, which will be available to the public soon:

1:4 / 1:8 Microphone input transformer in Mu metal can

1:5 / 1:10 Microphone input transformer in Mu metal can

1:5 Vacuum tube line output

1:1 Line to Line Bifilar

NOTE ON THE 1N4007 DIODE

I built two of the Rev H boards up before coming to the conference, but I didn't open the bags of parts for security reasons.  As it turns out, these diodes had larger diameter leads on them than they have in the past (anyone who is in manufacturing of electronics, will tell you these things happen all the time).  As a result the part wouldn't fit in the holes.  The good news is its a protection dioide from the input to the output of the regulator, to prevent the output from ever being more positive than the input.  Most regulators have these internal, so its redundant, and not really needed.  I have gotten in the habit over the years of adding these in anyway, in case a different device were used that didn't have one.

HEATSINKS ON THE TRANSISTORS

Ideally the ZTX603 and ZTX653 should have heatsinks.  I have over a dozen of these 2xjfp boards in various prototype eq's. mixers, and stomp pedals, none of which have heatsinks, and not one has failed.  If operating normally, the two transistors will feel warm, but not burning hot.  If you think yours are running too hot, you probably have a ZTX653 with too high of a gain.  You can remedy this easily by increasing the 68k resistor going to its base R3 and R10, to 100k or higher.  At a point you will begin to clip one half of the waveform (you will hear distortion), so there is a point of diminishing return.  If you can squeeze in small heatsinks on these parts (they are a TO-92 package size), its not a bad idea.

GENERAL COMMENTS ABOUT THE CIRCUIT,

First, if you look back to the original article I wrote in Tape Op of a DIY tube circuit, you will notice the circuit layout of the JFET gain stage is near identical; this was intentional.  Since this is for the purpose of learning, I thought people would better understand it if it was similar to the one I had already given, and that electrically they work similar, right down to the number of active devices.  So I am using solid state devices, in a layout that is more like a tube circuit.  Because solid state devices are so cheap, engineers have gotten in the habit of designing elaborate amplifiers with complicated current sources, using both AC and DC feedback to stabilize them, and eliminate for the most part the need to select components, and generally look better on the bench.  These designs are also far more producible, the design as given in the class works ideal when the active devices are selected.  When I wrote the Tape Op DIY JFET article, I built at least 6 different designs, most of which were far more complicated than what I ultimately published.  The funny thing was, I kept going back to the simple "tube-like" design because it seemed to just sound better when I was playing music through it, maybe because it sounded more like tubes, which I am clearly more used to listen to.  This circuit has limited headroom running off a wallwart (which isn't always a bad thing if it limits in a pleasing way, which I believe it does), has a fairly healthy 2nd and 4th harmonic (similar to a tube circuit) when driven hard, and will produce abundant distortion if overdriven into saturation.  However, even with these limitations (keep in mind the parts and pcb for this DIY workshop cost about $8/person), I still think this sounds better than an op amp based circuit when put in a recording situation, with real instruments and mics.  This is my opinion, based simply on playing my own music through different designs for years, so take it as just my opinion.  The design has very good bandwidth, and under moderate input levels (i.e. using a pad for hotter inputs), will deliver a very present, relatively uncolored signal.  When overdriven, it makes a very musical, thick distortion which I have found pleasing on vocals, bass, and guitars.  The reason I presented the circuit in the form it is in, is to show you how simple you can actually make something, and still have it sound good pretty good.  In many ways this circuit is a bit more of a musical instrument, than it is a clinical design, so enjoy it in that way.  Cheers.  SH

Back to Micpre kit page