[Jon L]

Big thanks to our member Skullguise, I was able to compare CS4398 vs. AD1852 DAC board in my Oritek Zhaolu 2.0. I had been listening to the AD1852 board for a long time, and without changing anything else in my system, I popped in the CS4398 board and listened. The differences are definitely very audible..

AD1852 board is on top, CS4398 board on bottom. Virtually identical, except for the couple of electrolytics around the DAC chip itself.

Rear view with the Oritek discrete output stage board and DAC board visible. Notice the 75 Ohm BNC jack and the nicer RCA spdif jack installed as well as the better RCA analogue output jacks. So far, comparisons have been via my speaker rig only, which is always more magnifying of differences compared to my headphones.

The biggest difference I noticed right away was the upper-midrange area. Basically, the CS board is quite bit more pronounced and highlighted in this area, which has the effect of making vocals, especially female vocals, a lot more forward and closer to the listener.

This gives you a feeling of more 'apparent' detail and speech intelligibility, leading to more "texture" in voices and instruments. I don't mind this usually, and lots of audiophiles love this kind of sound, which usually makes you go, "Wow, listen to how detailed, immediate, and textured the midrange is!"

Of course, as with all things audio, this has its downsides. This effect of CS4398 DAC board is akin to turning up the contrast knob a bit on your TV; there's a feeling of more apparent detail, but it also introduces a bit more grain and harshness. This is exactly the downside I hear with many recordings that are already recorded "hot" in this area to begin with.

I'm not really sure if the top treble is a bit rolled off compared to AD1852 or not, but this is because the emphasis in upper-midrange tends to draw your attention away from the high treble. Also, the midbass of CS DAC seems just a bit less impactful, though this is nothing like the difference noted above.

Due to these differences, the soundstaging of CS DAC tends to put performers at or in front of my speakers more compared to AD DAC, which stages at or behind my speakers. CS soundstaging is more forward, intimate, and grabs you more, but the AD gives a more naturally deep, separated stage.

The other odd thing I noticed was the CS DAC seems to have a louder output than AD DAC, probably in the 1.5-2.0 dB range, but some of this may be due to the more highlighted upper-mids, which can seem subjectively louder. In my setup, the CS DAC is giving me a higher physical noisefloor, which manifests as hisses via my 95 dB sensitive speakers. I can turn down the volume knob on my amps, which decreases the background noise to the level with my AD DAC, but to do this, I'm turning the knob down probably 1/8th of a turn. This shouldn't be a problem for more normal speakers in the usual 86-88 dB sensitivity rating.

At any rate, so far, I prefer the AD DAC version for its more "analogue" and fluid rendering of upper-midrange and vocals, as well as deeper, more naturally resolved soundstage and better bass impact.

However, this is somewhat unfair to the CS DAC b/c my system has been tweaked with AD DAc in place. So I have some work to do with tube, cable rolling, etc. I will report back after some more experimentation. Stayed tuned.

 

[JahJahBinks]

nice review, can you say the CS is more analytical?

 

[tkam]

The CS4398 does in fact have higher output than the AD1852 and many other dacs.

 

[daggerlee]

Thanks for the writeup Jon. I had both sets as well. When I had the AD1852 I was using it with a very neutral sounding amp (the original master, comparable to the gilmore lite imo), that plus my K701s combined for a very laid back, dry sound, too much so. I bought a CS4398 board from Eddie, and it returned some of the forwardness/warmth missing from the AD1852. Now that I have a Heed canamp, the sound has become very rich/full - a good combination with the somewhat lean K701s IMO. Now I'm just waiting on a Zapfilter 2 to complete my Zhaolu.

I'd be interested in hearing how you will go about tweaking your setup, I'd love to get some tips on bringing the CS4398 to its full potential.

 

[audioengr]

Good one Jon. I have done a lot of comparing of the CS4396/7 and the AD1853, both current-output DAC's. I found similar results. The AD1853 D/A is a bit less noisy and more detailed, with deeper sound stage. It is more accurate I believe. It's a newer design, so it should be. The CS4396/7 D/A is always a joy to listen-to though, with a bit warmer sound. Have not found anyone that does not like the sound yet. I personally prefer the AD chips. The only ones that I think are a bit better are the BB PCM1704's.

Keep in mind too that each of these chips has different power delivery requirements, so similar decoupling caps on these modules may do a disservice to one of the chips. One chip may work great with very little decoupling cap and the other may need expensive caps and several values strategically located. Op-amps are like this too. Some require special power delivery techniques to sound really good.

Steve N.
Empirical Audio

 

[Ori]

Jon, what a clear and very nice review!
I'm pretty much in agreement with you on all accounts and heard just about the same differences on two systems very different than yours.
I also like to add a word of caution: the differences were not the same on a stock D2 (caps bypassed, OPA2604/DY2000) and in fact I prefered

the CS on that setup.
A method to this madness?! Well, there might well be... My hypothesis is that the CS board has a slightly higher RF noise on the output and the audibility of it greatly depends on the analog section that follows and the rest of the amplification chain. The opamps have enough "smoothing" effect to take away the negatives and leave the CS sounding so much more relaxed in the stock Zhaolu.
Bottom line: don't hurry up and toss those CS boards out the window...

 

[Sean H]

Very interesting, Jon, thanks. I tend to agree for the most part but there are times I actually feel the CS4398 has a bit more depth and has equal and at times a more hefty midbass. I do use a Mapltree tube amp and Grado HF-1's though which both reputedly have excellent midbass punch and heft. I have a Zhaolu 2.5C and also a 2.0A. In the 2.0 I definitely preferred the A model pre shorting the DC caps, and like Ori said I found the 4398 chips sounded much better after shorting them and started to really like the 4398 so I got a 4398 model 2.5. The 4398 to me is a bit more dense, warmer and slightly darker sounding at times which are things I like. I do think the 1852 chip is crisper though and bit brighter but has that excellent definition and clarity. For some reason I have felt like the depth is more flat but wider but it's been some time since I have had the 2.0A in my main rig to remember right. I think both chips sound quite nice and is why I have both.

 

[Jon L]

Quote:

Originally Posted by Sean H /img/forum/go_quote.gif Very interesting, Jon, thanks. I tend to agree for the most part but there are times I actually feel the CS4398 has a bit more depth and has equal and at times a more hefty midbass. I do use a Mapltree tube amp and Grado HF-1's though which both reputedly have excellent midbass punch and heft. I have a Zhaolu 2.5C and also a 2.0A. In the 2.0 I definitely preferred the A model pre shorting the DC caps, and like Ori said I found the 4398 chips sounded much better after shorting them and started to really like the 4398 so I got a 4398 model 2.5. The 4398 to me is a bit more dense, warmer and slightly darker sounding at times which are things I like. I do think the 1852 chip is crisper though and bit brighter but has that excellent definition and clarity. For some reason I have felt like the depth is more flat but wider but it's been some time since I have had the 2.0A in my main rig to remember right. I think both chips sound quite nice and is why I have both.

Yeah, even Ori says above that in non-Oritek Zhaolu, the CS4398 sounded better.

Why don't you send one of your Zhaolu's to Ori for fun? You can spare one to experiment with, right?

I'm kind of wondering if Ori has modded any of the Zhaolu 2.5 and whether there's any improvement from modded 2.0 models myself..

 

[skullguise]

Nice write-up Jon, glad the module got there OK. I never tried it in my Ori-mod'd D2A, just too lazy

I agree with all that's been written here, the Ori-D2A is amazing (even before his latest mods, which I hear add some missing bass depth - I'll be in contact soon Ori!).

The D2.5C is great as well, and I appreciate Sean H's comments on opamp rolling; it definitely helps tailor the sound a bit. It's VERY good, and I often use the discrete headphone amp on it; but the Ori D2A still blows it away IMO.

 

[Jon L]

Part Deux

After I switched to the CS4398 DAC, I spent quite a bit of time in order to tweak my setup to optimize this DAC chip. I decided to leave everything else alone but to only use tube rolling in order to reduce variables. I was able to achieve some great results by rolling in some nice NOS tubes, especially with Telefunken EL84 and Amperex EL84, in combination with driver tubes such as Amperex BugleBoy 12AX7, RCA triple mica 5751, etc.

When I thought I achieved a great sound, I then switched again to the AD1852 board. This time, I noticed even bigger difference in sound compared to first time around. I have come to these conclusions for myself in my setup.

AD board has significantly more treble extension, sparkle, and resolution. This is what leads to better imaging focus and soundstaging accuracy over CS DAC. Tubes can be rolled to change tonal balance, smooth out grain, etc, but you can't really create what's not there to begin with, i.e. resolution.

CS board has significantly fuller midrange/lower-midrange presentation. Usually I would dig this kind of sound, but this presentation also had a puffier, artificially bloated feel to it, reducing sense of speed and control over the notes in this range. In fact, through the treble and midrange (even through bass), AD board had a tighter grip on the music's leading edge, note, harmonics, and decay. This is SPEED without sounding artificially speedy.

CS board also had a bit more mid-bass energy; however, it was not the super-tight variety that carries the PRAT along. There was a puffy halo around the bass notes that dragged a bit. When a vicious bass note hits, CS board sounded like being hit with a rubber-capped hammer whereas the AD board was like a steel hammer.

So far, it sounds like AD board is killing the CS board, but that's not the whole story at all. For example, if I were involved in a shoot-out between these two boards with music I'm not intimately familiar with, I might declare the CS board the winner! The richer midranges and somewhat softer-edged, gentler sound CAN sound more attractive and "musical" to many folks, especially with thin, sterile dgital recordings out there. If the system was all solid-state, it would be even more likely that more people would prefer the CS board.

However, I am a tube lover, and there will always be some tubes in my system, both speaker and headphone systems. In this environment, it's much easier to choose some yummy tubes that will bring out that last iota of extra "magic" out of the AD board while preserving the scary resolution, PRAT, and purity.

So, as usual, there will not be one-fits-all answer in this matter of CS4398 vs. AD1852 preference, but that's my story and I'm sticking with it

 

[cotdt]

Amazing and very interesting review Jon!

 

[daggerlee]

Quote:

Originally Posted by Jon L /img/forum/go_quote.gif If the system was all solid-state, it would be even more likely that more people would prefer the CS board.

I read your review in trepidation until I got to that line. Perhaps I had been a little preemptive in selling my AD1852? But I've only added more SS components to my setup, and I'm quite happy with the CS4398 sound through them. Upgraditis averted

 

[Jon L]

An excerpt from the new 6Moons Review out:


"After discussing my findings with Steve Nugent, he felt that the highs might be smoother in the Northstar because it has a discrete output stage versus the op amp output stage of the Benchmark. The greater clarity of the Benchmark is likely due to the newer generation, better spec, AD1853 DAC chip. The Northstar uses the older CS4396/7 D/A chip, which is the same as used in the Perpetual Technologies P-3A DAC that Steve used to modify. Perhaps a DAC with the AD1853 DAC chip and a discrete output stage (along with Steve's other goodies) would be the best of both worlds."

I count myself fortunate to be already enjoying the combination of AD1852/3 chip with discrete output stage

 

[ginetto61]

Hi !
very interesting thread.

Big thanks to our member Skullguise, I was able to compare CS4398 vs. AD1852 DAC board in my Oritek Zhaolu 2.0.....

 
Hi and thanks for the very interesting information
If i understand well this dac allows for a dac boards swap ? very interesting.
I understand right that you prefer in the end the AD1852 DAC board ?
I have an Apogee Rosetta 200 that should have this same dac chip inside.
I understand that the implementation also counts a lot ... but nevertheless a very comforting opinion.   I mean it seems that the AD1852  is one of the good of the delta-sigma kind.
Good !  and thanks a lot again !
Regards,  gino