Audiophile recording and playback - Tascam DA-3000

If you want to take the DA-3000 further and can justify $900, synchronize its clock input to an external oven-controlled clock. To record at DSD128 I connect it to an Antelope LiveClock at 44.1kHz. This significantly improves both ADC and DAC sections of the Tascam. The LiveClock has a 12-18Vdc power port that can be upgraded with a good linear power supply as an alternative to the stock wall wart. The LiveClock can also be synchronized to an 10mHz master atomic clock. This may seem like overkill, but with this set-up DSD128 recordings played back through my Esoteric K-01X get very close to the LP. It’s been rather astonishing to "listen to" a clock with at least 50x more precision than the Tascam’s internal clock.

I thought about but didn’t pop for the Decware mods, as it appears that Decware does nothing to improve the ADC section.

The Tascam has a Word Clock Sync 75ohm BNC input for connection to an external master clock, and a setting on the screen menu to switch from internal to external master clock. A higher precision master clock will improve both recording and playback of SD files, and also conventional DAC operation from the SPDIF input.

It’s a significant improvement: more detail, more accurate pitch and timbre, tighter LF, bigger soundstage, livelier dynamics, calmer background. Of course the old recordings that you made with the internal clock won’t sound as good as new ones.

I’ll sum up my experience after nine months with a DA-3000. The stock unit is capable of very good DSD128 recordings. Quite convincing, if recordings are played back through a better DAC than the one that’s incorporated in the unit.

With lite internal modifications to the ADC and power supply sections, the recordings become nearly indistinguishable from the LP source.

Externally, the unit responds to a better power cord, footers(Stillpoints Minis), and above all, synchronization to a good Word Clock(Antelope LiveClock powered by a Hynes LPS).

Tascam is unresponsive to requests for schematics, so I limited the scope of internal modifications to what I could understand from inspection and a manual trace of the ADC section. The simplest upgrade is to replace the three-pin regulators on the power board with Belleson +/-12V discrete regulators. This is an improvement, but nothing like digging into the ADC section.

The ADC section is a simple circuit, comprised of two stages of electrolytic coupling caps and NE5532 SOIC op amps, a balanced JRC NJW1195A volume control chip, and a Burr Brown PCM4202 ADC chip.

The NE5532 is a generally well-respected op amp. The twenty or so coupling and power supply filtering caps in the ADC section are mediocre Suncon/Sanyo parts. I replaced them all with Panasonic FM and raised the value of the op amp filtering caps from 22uf to 100uf. I bypassed eight 47uf electrolytic coupling caps with a combination of small film caps (.01uf MIT RTX polystyrene and .01uf Russian FT-1 teflon.) This was a bit of work, but nothing beyond what a basic technician could do in 2-3 billable hours.

The modified ADC and power supply section takes the unit up several notches. Now it’s close enough to the analog source to confuse me as to which is which. You sit on the couch, stroking your beard, musing that what looks like and prices out for the Pro Audio market, is really anything but.

There is little on the DIY or Pro forums about modifying the unit. The DECware mod addresses just the DAC section. Lampizator tried converting the DA-3000 ADC section to tubes, but doesn’t appear to have productized the modifications. In any case, the above mods are within the reach of anyone with good soldering skills and a spare afternoon.

@mre28m5... The Bellesons are drop-in replacements for the original regulators. I reused the original thermal pads and screws.

BTW, desoldering the stock Suncon electrolytics from the audio board requires a little technique.  The OEM solder is so dry and desiccated
that  it won't melt for extraction without first flowing on a bit of fresh solder to reactivate it. Thereafter extraction goes smoothly with removal braid or a vacuum bulb. The multilayer PCB is of high quality and there is no tendency for pads or traces to detach.

Good luck if you try it! Now I just need to find the motivation to re-record everything I did prior to the mods. 

@mre28m5

Glad that worked out for you. I had forgotten that I used insulating shoulder washer Digikey p/n HS418-ND to float the Belleson metal tab from ground.

Now that the audio board mods are breaking in, I’ll take a shot at tracing the circuit on the power board. Tascam uses four 2200uf and 6800uf Panasonic FK filtering caps that are keepers, but there are two Suncon 100uf filtering caps that appear to be at output of the Belleson regs and three small Jamicon electrolytics that may be worth replacing. IIRC, Belleson recommends a higher capacitance than 100uf at output of the SPZ-- which is why I increased filtering capacitance at the downstream op amps from 22uf to 100uf.

If you try the audio board mods, be gentle with the wide, flat ribbon cable. It has many tiny foil traces, several of which began to fray and peel back a mm or so during reinsertion. The solution was to use long sharp scissors to trim back the ribbon a tad and restore clean traces. As this type of cable is not designed to survive many reinsertions, it’s best to get all ADC mods done in one pass.

The Panasonic FM electrolytic coupling caps for the ADC should have their .01uf film bypass caps placed on top and soldered point-to-point to the legs of the FM cap before mounting the FM to the PCB. There not enough room to do the bypassing as an afterthought.

Just an update that I replaced the remaining stock Jamicon and Suncon electrolytic caps on the power board with Panasonic FC. This produced a modest improvement in SQ in DSD128 recordings, but worthwhile given the small effort of removing the power board once more. I think this is the finish line for the Tascam.

This little project has progressed in parallel with SACD rips via PlayStation 3. With a decent cartridge, tonearm, turntable, and phono stage as analog source for DSD128 recordings(in my case Stanton 981LZS>SME 3012R/Kuzman 4P>Luxman PD-444>modified Pass XP-25), vinyl recordings on the modified Tascam DA-3000 are on par with DSD64 SACD files. Playback for all files is QNAP TVS-882>SOtM modified D-Link EN switch>SOtM SMS-200 Ultra>SOtM USBultra>Esoteric K-01X, with all components synchronized to a Stanford Research Systems Perfection 10mHz rubidium clock and powered by a four-rail Paul Hynes SR7 linear power supply.

In short, the Tascam is a sleeper and a killer.