Burson Conductor ESS9018 vs PCM1793

burson1_94Q1377

Recently Burson updated their Conductor line up to include 2WPC variants similar to the concept of the Soloist SL to the Soloist. Less power output minus pre-amp functions to provide a cheaper option for mainly headphone users.

Along with this change Burson also re-introduced the PCM1793 based DAC first seen in the Burson HA-160D yet with a new FET input stage featuring fewer components for better transparency. Therefore potential customers now have the choice of either the Conductor SL PCM1793 and the Conductor SL ESS9018, while the original Burson Conductor (non-SL) will still stick with ESS9018.

I received a request to review the new PCM1793 based DAC but since I already have a Conductor in my possessions, I thought it’d be more economical to do the review on the 4WPC Conductor (instead of the 2WPC Conductor that the PCM1793 is intended to be paired with) by swapping the DAC module inside the Conductor. The question here is if the PCM1793 can compete with the supposedly higher end ESS9018 DAC. According to Burson it can as the design surrounding the DAC chip plays a big role in the final sound.

 

SOUND IMPRESSIONS

I’ve compared the two DACs and found that the PCM1793 to have a bigger soundstage, more air,  more relaxed and more flow. The ES9018 is stiffer, more forward, more energy and engagement. The result was not that surprising as previous experiences with ES9018-based and PCM179X-based tend to give similar adjectives. While the ES9018 is more in line with the Burson sound signature with the forward and energetic sound, I find the PCM1793 to help relax the Burson sound a little. In general I enjoy the PCM1793 more on the Conductor, not only due to the more spacious sound but also to the more relaxed sound. Pace and PRaT is definitely better on the ESS and for fast paced songs needing PRaT the ESS should be the way to go while the PCM1793 is the better option for instrumentals and Jazz requiring good ambiance.

In the big view picture, the final sound output of both DAC modules are still influenced a big deal by the design of Burson’s output stage, and so both DACs do inherit a little “Burson DNA” that is fast paced, forward, well-articulated sound.

 

Special Note:

If you wish to obtain the PCM1793 DAC module or the ES9018 separately (for existing Burson amps with the other DAC module), they can be purchased direct from Burson for $250 AUD for the PCM1793 and $350 for the ES9018.

 

 

 

share us on your favourite network
  • Dave Ulrich

    Since the SL is a bit more laid back than the Soloist, do you think that the SL with the 1793 would be too laid back, or would it still bear the hallmark of Burson?

    • http://www.headfonia.com Mike

      Good question. Still overall Burson is not laid back. Violectric is laid back

  • Rūdolfs Putniņš

    Oh, my… If Burson hopes to at least scratch the surface of what these DAC chips have to offer they must learn to do better PCB layouts. To at least shoot for true 24bit resolution (120dB SNR) you must use smd parts and drastically reduce trace lengths. Even advanced DIY designs are vastly better than this, Acko DAC for example…

    They are starting to resemble Audio GD and not in a good way.

    • Julio César Bedoya

      Heck, even creative pulls out better (measured) sound cards….

    • http://www.headfonia.com Mike

      SMD parts are known to produce inferior sound than through the hole. Especially when you start using boutique TTH parts.

      • Rūdolfs Putniņš

        Known to whom?

        It is true that larger parts like higher capacity electrolytic and film caps aren’t feasible to produce in smd, however ceramic caps, resistors and all small signal active elements are as good or better in smd packages. For high current and sometimes high voltage applications smd may be unfeasible due to physical constrains that limit heat dissipation. For high voltage it is better to have well separated leads because kilovolt currents know how to fly pretty well.

        However there is one application where only smd will do – high frequency circuits. I can’t make out the lettering on the crystal on the Sabre board but stable operation can be only attained by using >85Mhz crystals with 100Mhz being the optimum. At these frequencies it is paramount to have extremely short traces, plenty of capacitive decoupling, impedance optimized tracing (to tame wave reflectance) and separate planes for power/power ground and signal/signal ground so you don’t have any vias that increase trace inductance.

        Smd also helps in high quality low noise analog circuits where the designer hopes to attain >120dB SNR (which is true 24bit resolution). Same goes for high bandwidth analog with high slew rate devices – screw up there and you get all kinds of HF crazy. I mean if you use a 24bit DAC you might as well design rest of circuit to benefit from that extra resolution.

        • Opinionated but experienced

          All your points are valid, apart from high end audio. It takes years of experience to find out that in many cases the opposite of your statement can be true.

          Known to many who design high end:

          SMT capacitors are pizzo electric and hence microphonic. Anything less than NP0 sound terrible as they have varying dC/dV. Many SMT resistors are not linear with voltage or power, so unless melf are used you will not get the best out of the 134dB capable ESS. The large xtal is a good idea, not some tiny SMT device working in overtone adding jitter.

          This guy is using Dale resistors, Wima capacitors and Elna Silmic II electrolytics, which means he listens, and listens well.

          24bit is 144dB.

          • Rūdolfs Putniņš

            You are talking about parts, yet there is no mention of the circuit they are used in.

            Also when talking about the 134dB and 144dB SNR please do check on the Johnson noise figures for resistors used in room temperatures.

            • Opinionated but experienced

              You were talking about parts too…

              With capacitors I specifically mean filters: High pass and low pass. All parts deviate from ideal when a voltage or current passes through them, an so need careful selection. SMT parts tend to be worse (smaller size gives higher ppm/C, dC/dT on X7R and below, and non linear resistors)

              Agreed, it is not practical to have liquid nitrogen cooling on audio products. I was just making clear that these DACs have huge potential dynamic ranges. I would rather the Johnson noise of a good linear resistor be the limitaion than the quantisation of a DAC though.

              • Rūdolfs Putniņš

                Yes, however I was mentioning the application of these parts as well. For high performance low bandwidth/high voltage/current analog circuits TTH is perfectly acceptable and even necessary due to physical limitations of tiny parts. However when we are talking about small signal high precision high frequency digital circuits SMT is a must. Sometimes even the smaller the better due to shorter trace length and lower parasitic capacitance/inductance.

                Problems start with high performance analog solid state circuitry that approach digital circuits in speed. Very easy to get these to ring with too long leads. Also for getting that >120dB SNR SMT is very much necessary.

                The problem is that I’ve heard and built both high quality TTH and SMT circuits. My current headamp the QRV08 uses more than 300 SMT parts and achieves what many would call high end performance. Also companies like Resonessence and others don’t shy away from using these parts where appropriate.

                • Opinionated but experienced

                  SMT is not “a must” for 120dB SN. Matching of traces can be done at these frequencies with PTH on the analogue side.

                  But the right part for the right job is. I think your original criticism of Burson is not appropriate here. Sure some SMT for decoupling and where steady state signals exist is advisable. However where there are audio signals use MELF, PTH resistors, or NPO SMT, polypropelene, or polystyrene if you care about the sound. The same applies to semiconductors: if the signal is dynamic, cascode that little SMT transistor, or use a bigger one.

                  • Rūdolfs Putniņš

                    As for Burson – there is one way to know for sure – scope it out.

                    In any case many thanks for your input – I will try to use the parts you mentioned in one my future projects. My belief is that a good circuit should function well without matched or premium parts. To get those extra zeroes is where the really good parts come in (good industrial grade, not “magic” grade).