TASCAM DA-P20 DAT Deck Low noise microphone pre-amplifier modifications

Author: Graeme W. Gill
Date:   1994/4/20


    These modifications are not for the faint of heart. No responsibility is taken by the author if any of these notes are incorrect, or you stuff things up. The modifications are electrically straight forward, but mechanically demanding, as there is not much space in the recorder to add extra circuitry.


    The TASCAM DA-P20 DAT recorder is (was) an economical recorder capable of reasonable performance and unlike other low cost (consumer) DAT records, came equipped with XLR inputs. Unfortunately the Microphone input amplifiers have a less than wonderful noise performance. A look at the circuit diagram quickly shows why. The differential input amplifier is of the most simplistic design, using an ordinary op-amp in the standard differential input configurations. This means that both inputs have a considerable resistor noise contribution, a contribution equivalent to about 10k ohms. This is not good.

The modifications described here overcome this problem and give the DA-P20 a microphone noise performance that is essentially limited by the op-amp used.

Because of the construction of the input amplifier (lots of surface mount devices) it is a risky operation to cut tracks. It is also difficult to back out of the modifications if tracks have been cut. For this reason I have devised the modifications to completely avoid cutting tracks.

Components needed:

1   x   LM837    Low noise quad op-amp, in a 14 pin DIP package
1   x   NEC PGA2A-05 5 Volt dual reed relay or equivalent.
1   x   BC337 or equivalent npn transistor.
1   x   1N1418 or equivalent small signal diode
4   x   47k resistors
4   x   1k0 1% or better resistors
2   x   47R resistors
1   x   5k6 resistor
1   x   10n bypass capacitor

1/4 watt resistors are satisfactory, but 1/8 watt sized resistors will make construction easier. The matching of the 1k0 resistors will determine the Common Mode Rejection Ratio. If you are adventurous you could try using a lower noise op-amp than the LM837, but you will need to make sure they are unity gain stable, 4 of them need to fit in a space not much larger than a 14 pin DIP, and they need to be able to drive a 600 ohm load. If the op-amp draws its input current the opposite direction to the LM837 then you will need to reverse the polarity of DE3 - DE6.

Circuit description:

Here is a circuit diagram of the analog input section before modification:

Circuit Diagram - before modifications.

The changes made are:

A differential gain stage is inserted in front of the differential amplifier already in the DA-P20. The LINE/MIC/DIGITAL switch is bypassed so that the existing amplifier has a fixed unity gain. The MIC/LINE gain change is made in the new section by switching the reed relay. In order to avoid cutting tracks the new gain stage is inserted in place of the input coupling capacitors, and the gain switching is done via an inverting transistor relay driver, and spare contacts on the LINE/MIC/DIGITAL switch.

With the relay open each one of the LM837 op-amps is a unity gain follower stage. With the relay closed the differential mode gain is increased to 2000/47 = 32db, while the common mode gain remains at 1. The input noise voltage is determined by the 47 ohm resistor and the op-amp noise voltage performance.

Here is the the analog input section after modification:

Circuit Diagram - after modifications.

Removing the analog input PCB:

You need to keep track of which screw came from where, since they vary in size, length and thread pitch. (Note in particular the screw near the DC in socket. It looks like the other small black screws, but has a coarser thread pitch).

1) Eject the cassette carrier, and unclip the cover. It unclips by moving it towards the front.  Push the cassette carrier back into its closed position.

2) Remove the battery cover and battery.

3) Undo the 12 screws that hold the bottom cover on. With the front forward there are 3 along the right hand side, 5 on the bottom (including the one in the recess), 1 on the left hand side and 3 on the back. Remove the bottom plastic cover.

4) Undo the 5 screws inside that fasten the internal framework to the top cover (you have to peer past the circuit board to find them). Undo the 2 screws either side (on the outside) at the front, and the screw at the DC in socket at the back. Remove the top cover. Watch out for the eject key falling off.

5) De-solder the 3 green earth leads that attach the front to the bottom PCB shield. Remove the recording knobs. Gently pry off the front panel. This will force the volume control knob off, and disconnect two PCB right angle plugs. Undo the eject switch PCB. Unplug the two front panel looms. The front panel is now free.

6) The analog input PCB is the one with the record level pot on it. Undo the screws that hold the PCB shield on to the card. Be careful of the input selector switch falling off. Loosen the record level pot collar nut. Undo the two screws that hold the connector plastic and XLR connectors on. Undo the two loom connectors to completely free the analog input PCB.

Making the modifications:

    There is a limited amount of room and clearance for the extra components to go onto the PCB, basically the length of SW2, and not much beyond the edge of the surrounding ground plane outline. The height of the components should only stick 5mm or so above the height of SW2.

DE1 and DE2 need to be carefully folded over to lie in the opposite direction (as indicated by the dotted outline in the layout diagram). DE7 also needs to lie on its side, and some extra lead length should be created by heating its pads and drawing it out slightly.

There are two test points circled in the middle of the card, and marked +V and -V on the layout diagram. The LM837 can both be powered and supported by these points. I prepared the LM837 by cutting all the pins except 4 and 11 short, and then bending the short pins slightly upwards. Pins 4 and 11 are then bent under the IC. These two pins can be soldered into place to the +V and -V points. The 10n bypass capacitor can then be soldered between pins 4 and 11. The 1k0 and 47k resistors are shown spread out in the diagram for clarity, but were actually mounted vertically. The various interconnections were made with teflon insulated wire wrap wire.  (There is less chance of accidentally melting the insulation if it is teflon) The ground connections to the 47k resistors were made by poking the stripped ends of the wires through the appropriate via holes of R29 and R25 as indicated in the diagram, and soldering to the ground end of the resistors.

DE3, DE4, DE5 and DE6 were removed from the PCB, and the +ve lead carefully bent upwards. The holes closest to the center are used to connect the outputs of the LM837 (Pins 1, 7, 8 and 14) to the existing differential amplifier. DE3 - DE6 were put back into the PCB, and the free (+ve) lead connected to the LM837 inputs (Pins 3,5,10 and 12).

The leads of the relay were cut short and bent outwards, and it was glued to the top of DE1 and DE2 with a dob of silicon sealer to hold it in place. The near 1k0 resistors can be connected almost directly to pins 7 and 8 of the relay, while the 47R resistors with some sleeving on their leads can connect between pins 1 and 14 of the relay to the far side 1k0 resistors.

The transistor is mounted upside-down and connected with the 5k6 resistor and diode to pins 2 and 6 of the relay. The emitter of the transistor can connect to the ground side of DE7, and the +5V can be got by connecting a wire through the via hole that leads to CN13 pin 5.

On the solder side of the PCB pins 6 & 7, 14 & 15, and 19 & 20 of SW2 need to be connected together. Pins 2 & 4 also need connecting, and then a wire run from pin 2 to the component side of the card.  I did this by drilling a small hole in the PCB in an area clear of tracks, near the 5k6 resistor that it needs to connect to.

You need to make sure that there are no shorts, or any chance of components coming loose. Everything needs to
be kept compact.

The third diagram shows both sides of the circuit card and how the new circuitry is built and connected:

Modifications to circuit card.

Also here are a couple of photos (Sorry the depth of field is awful):

Photo of modified analog input card

Photo of DAT main unit and input card.


The modifications can be tested by re-connecting the analog input PCB, plugging the 2 looms back into the front panel, and plugging the back panel back in. You will need to supply power with the plug pack, and may have to cover the mechanism with something to stop outside light
upsetting operation.

If everything is working correctly there should be approximately +/- 5.2 volts on the +/- Vcc, 0V on all the outputs of the LM837 and the uPCV4572,
and if you set the recorder to Pause record and plug a microphone in, (selecting 0db and Mic of course) the amplifiers should be working correctly (with a lot less noise than before!). The line level and 20db pad should also be tested.


Because all the extra components extend the circuitry closer to the underlying PCBs, I made up a thin copper U shaped shield that sits in the frame opening in which the analog input PCB sits. Its edges are clamped by the PCB, while its insides and outsides were protected from shorts by covering it with some self-adhesive vinyl sheet.

Re-assembly is then the reverse of the disassembly procedure.


There you have it. I hope you find this information useful. Feel free to mail comments, suggestions, praise or news of success. I don't want to hear of
any failures.

    Graeme Gill