Are all ESS dac ICs voltage out? (i.e., no need for i/v)

Status
Not open for further replies.
In short, yes. ES9018 and its derivatives can support voltage mode or current mode. It depends on the input impedance where the next stage has. When it sees a low impedance, it's current mode. ES9023 has integrated amp for voltage out. I don't know if it can support current output.
 
Last edited:
In short, yes. ES9018 and its derivatives can support voltage mode or current mode. It depends on the input impedance where the next stage has. When it sees a low impedance, it's current mode. ES9023 has integrated amp for voltage out. I don't know if it can support current output.
Yeah, I noticed that and some discussion about using ESS DACs with an I/V. Then, I recall ESS being used in Dragonfly and other ultracompact designs, and those use no i/v.
Too bad ESS discloses no detailed info on its designs like other DAC manuf. do in their datasheets--then we (the DIY community) would better know how to implement these DACs.
If anyone has access to ESS application datasheets (or more-detailed info), please post a link (or PM me).

Thx!
 
This is from the TwistedPear guide for their "BUFFALO-IIISE (STEREO EDITION) 2-CHANNEL DAC":


Analog characteristics:

The Buffalo III-series boards do not include an output stage but are designed to mate with a Twisted Pear
Audio current-to-voltage stage (I/V stage) like the IVY III or Legato 3.1. For this reason, the mounting
holes have the exact same lay-out as the Buffalo II version of the DAC module.
It is not recommended to use the Buffalo III/IIISE output directly, but it is possible. The output can be used
as a voltage source into a high impedance but THD+N will suffer. THD+N as a voltage source is about
-108db whereas when used with a very low input impedance current to voltage stage such as the IVY III it
can achieve -120db THD+N in mono mode.

So, is TP trying to sell I/V boards, or does the ESS9018 DAC (on its own) really sound worse as voltage out (with just a buffer/output stage)?
 
Last edited:
ES9018 has some DC offset voltage on the output, so you need something to get rid of the DC before sending it to your amp. It makes about 3Vpp with a few mA of current if you parallel 4+4 channels for stereo. It sounds pretty good with just a transformer, but bass is lacking and mids not full bodied compared to Twisted Pear Buffalo32 with IC i/v stage. I sometimes wished for more than 1.5 V output, but it was usually enough.

You can get the 9018 technical specsheet if you sign nondisclosure form.

Twisted Pear seems to take a few production versions to work out all the bugs in its designs. But the latest (3rd or 4th) version of Legato sounds really excellent and seems reliable. No coupling caps. I'm not sure it's still worth the money anymore, but it does sound great top to bottom if you give 9018 very low jitter I2S data.

Audio application doesn't really need the higher dynamic range (134db iirc?) that current mode provides. Voltage mode is only a few dB less, but still excellent, and way more than redbook at 96dB, vinyl is 40dB on a good day. The noise in any music recording, or from any amp, or the ambient noise in your listening room makes the dynamic range spec silly. The real factor is whether there is a timbral difference between the two modes when well-designed output stages with 1ohm or 1Mohm load impedance are compared. I don't know the definitive answer to that. I've heard it sound good and bad in both ways. I think experienced professional designers will choose I/V and this is what ESS recommends, but it is easy to make either sound bad. Many ESS bashers eager to move onto the next fad DAC chip have not heard it sound its best, or it would leave a mark not easy to forget.

I'm building a simple SE, cap coupled, voltage mode output stage for my Buffalo2 using single STMicro MC33079 dual opamp with a little gain to make 2V. This opamp sounds open, balanced and very musical even in unity gain compared to most opamps. I like it better than TI 49xxx with low gain.
 
Agreed on "fadism"!

I think experienced professional designers will choose I/V and this is what ESS recommends, but it is easy to make either sound bad. Many ESS bashers eager to move onto the next fad DAC chip have not heard it sound its best, or it would leave a mark not easy to forget.
I'm not sure what output topology (VO or i/v) the AQ Dragonfly uses (??), but it employs an ESS and has achieved a decent reputation -- all in a very minimal package. This makes me question TwistedPear, et. al.

I agree with remark about fads. It's also about boredom (not jut DIYers!) and lack of hindsight. Heck, there is even some now claiming first gen. Wadia's and Theta's are superior.
 
9018 (9012 in this case), a good 18 ohm resistor (Z naked foil) and a Lundahl MC trafo isn't bad 😉

//
 

Attachments

  • fotor_(59).jpg
    fotor_(59).jpg
    161.4 KB · Views: 550
LL1941 (as 1:32)

It is a moving coil trafo. You must load down the output in order not to saturate it. But this is what the 18 ohm does so it works well - maybe not full 2V out... I haven't measured it. It works OK straight into my Hypex NC400 with normal (88dB) sensitive speakers. 3-6 dB more wouldn't hurt 🙂

If 9018K2M supports I-out it's OK.

//
 
Status
Not open for further replies.