Topic: A good PM exchange I had with IEMCrazy (posted with his permission) (Read 459 times)

Marvey · « **on:** October 23, 2012, 07:48:39 PM »

I love these kinds of discussions. Lots of nuanced information in a friendly exchange. Atypical of shit that goes on HF boards.

IEMCrazy
Yesterday at 9:13 am
Ok, between you and Sphinxvc, you've done it. I'm actually considering a Mjolnir.

I ignored your advice earlier and went with HE-6 after all. I don't regret it. Then I ended up upgrading my DAC to suit: Woo WDS-1. Then I realized with that I no longer needed the crummy preamp of the Marantz PM6004. I upgraded.....Marantz MM7025. Big power, low gain, balanced inputs....great for HE-6 and WDS-1. That was endgame. it was perfect.

Then sphinx dropped the price on his HD800s. I couldn't resist. So now I have HD800 too. The plan was use HD800 with the HP out on the Marantz PM6004. It's a great sounding amp with it, unfortunately there's this weird issue where when certain volume settings are combined with certain impedance loads, it produces a sharp "crackle" sound in things like soft wind-instrument passages in classical recordings below -40dB. It happens with HD800, but not D5k, HE-400, or HE-6. But with any of those if I lower the input voltage (volume down on the WDS), then the noise starts....annoying. So I'm using mostly Lyr for the moment with them.

So I'm looking for a better HD800 amp, and Mjolnir, thanks to some of your input tops the list (yeah I know folks live tubes for HD800, and I like tubes, but I don't like messing with power tubes, so I'm not sure I want my better amps to be tube...Lyr gives me enough tube-playing for now.

You commented in a thread about the balanced inputs being better, and I'm trying to plan my wiring here. I plan on putting the MJ on the OTHER side of the room for a different listening location. I intend to sometimes use with with the WDS-1, and other times use it with EITHER Bifrost or the outputs of the Behringer DEQ.

So my question for you is two-fold based on how different you found the balanced inputs to be from the unbalanced:

1) Bifrost is a fantastic value though not perfect. The Behringer DEQ is known as a so-so DAC to some, and "indistinguishable from a Bel Canto" by others. Many belive modding it significantly does make it into a great DAC either way but I'm not going that route. Is the difference in inputs such that you'd suggest using the balanced outs of the Behringer instead of the unbal outs of the Bifrost? Or you'd stick with the good-value Bifrost in that setup?

2) For the other connection from the WDS-1 to the MJ for when I want to listen that way, I'm debating going with splitting the balanced outs (one to the 7025, one to MJ) versus running unbal to the MJ from the spare unbal outs on the WDS. Since you've compared the inputs on the MJ, you're probably the best to advise on if it's worth a Y-split on the XLRs or not.

Marvey · « **Reply #1 on:** October 23, 2012, 07:48:53 PM »

purrin
Yesterday at 12:26 pm
I haven't heard a stock Behringer DEQ/DCX in a long time so it's hard to say. (My own DCX2496 which I use with the Mjolnir is significantly modded.) I do know that the stock DEQ is warmer sounding than a modified one - about the same amount as the Bifrost. So maybe this would be a better match for HE-6. It wouldn't hurt to try both your DEQ and Biforst. The HE-6 may not match well with the Mjolnir since he amp is on the unrelenting / honest side. Make sure you burn in the Mjolnir for at least 100 hours or just leave it on for a week.

Split the balanced outs. Balanced outputs have a lot of juice. They run on higher voltages (x2) compared to SE and are way more immune to noise for long cable runs. That's that they were meant more. I split my balanced outs from the PWD2 for comparing amps.

Cheers!

Marv

Marvey · « **Reply #2 on:** October 23, 2012, 07:50:34 PM »

IEMCrazy
Yesterday at 12:38 pm
I'd be using the WDS-1 with the HE-6 (and sometimes HD800.) The DEQ/SRC vs. Bifrost equation would be for sometimes use with HD800 only. And given your statement about warmer, that would definitely be a good paring for HD800.

Given the "unrelenting/honest" comments it seems in that sense to be a bad match for HD800, but since I've heard some good feedback from a number of people about MJ with HD800, I'm not too worried about that. I'm just not planning to go for tubes for the moment. If I weren't EQ'ing it would be a no-no, but I'm EQing with EQ, not with cables & tubes, so if there's harshness in the treble, I can certainly pare the treble down more. (The DEQ is just a handy little machine...)

Given the temps MJ runs at, I don't think I'd be tempted to leave it (or any class-A amp or tube amp) on unattended, but I'm familiar with Schiit burn-in from Bifrost....it would only take 2 or 3 weeks of normal usage for burn-in for me. I did read your comments on the 100 hour burn-in, but is the effect as severe as the burning with Bifrost was?

Got it, XLRs to be split....I'm loading up my cable orders now

I'll order the MJ probably next week...this week is cable week. Some XLR splitters from BJC, long XLR runs from Audio Technica...($17 vs. BJC $50....kind of a no brainer there...)

Out of curiosity, from long-long-ago memory, what kind of improvements did you get with the modding of the Behringers and/or how do the stock Berrys stack up to something like Bifrost? I know BF is an excellent value, but Berry can be. I actually will be using the SRC2496 if I go that route, not the DEQ for the analog outs. Same DAC as the DEQ, but much more like the DCX in terms of the analog section of the output. I suspect its generally better than the afterthought analog on the DEQ.

One big thing the SRC has over the Bifrost as a DAC is reclocking. That's one key thing Bifrost lacks. On the other hand, I have no idea how the analog stages would compare. Yeah, I may have to A/B it

Marvey · « **Reply #3 on:** October 23, 2012, 07:52:18 PM »

purrin
Yesterday at 2:23 pm
Yeah DEQ/DCX are nice in that regard. I love the EQ on them. [I secretly use Parametric EQ]

I felt the effects of burn-in [on the Mjolnir] have a huge [OK maybe not so big] effect. A lot smoother and much less aggressive. You turn it on for a few hours before you use it for the first several weeks if you don't leave it on all the time.

The modded DEQ/DCXs are every bit as good and possibly better than the stock Gungnir. It depends upon how you mod. I've experimented and tried all sorts of stuff for a few years now on my DCX, so I've basically made it how I wanted it to sound.

Marvey · « **Reply #4 on:** October 23, 2012, 07:53:15 PM »

IEMCrazy
Yesterday at 7:45 pm
I haven't tried the DCX, but yeah the DEQ is amazing. I felt like the odd one out on an audiophie forum where EQ is a forbidden evil talking about a "cheap" pro audio DAC, despite a smattering of folks that had used them as well. I was pretty shocked to find out you were using Behringer stuff too, of all people

I wonder what the magic burn-in is with Schiit gear....seems the same exact burn-in results as for Bifrost. Could be heat related...takes that long to melt the thermal compounds and at the heat they throw, probably reflow the solder

Interesting to hear you so high on the moddeed DEQ. Not that I have any interest in resoldering the insides of a Behringer, nor messing with the dainty traces on the board....but out of curiosity do you have a link to anywhere you've talked about the mods you've done? I found a few pages you mentioned modes, but have yet to stumble into descriptions, either here or changstar Mostly for curiosity, II'll almost certainly be running it unmodded. And I'd only even consider modding an SRC....my precious DEQ must remaine unharmed....that's the cheapest and arguably moest important backbone of my rig

(BTW, I still love that avatar! )

Marvey · « **Reply #5 on:** October 23, 2012, 07:54:12 PM »

purrin
Yesterday at 8:05 pm
The DCX / DEQ are very similar internally. Obviously DEQ has more EQ functionality (more than one would ever need) and the DCX has crossover functionality. The AKM DACs used are voltage output DACs with clean output - meaning it's absolutely not necessary [it really depends what's downstream - caution still recommended] to filter the DAC oversampling ultrasonics. The internals of the chip actually control these ultrasonics very well - at least according to spec sheets. The AKM are dirt cheap, but sound great (at least I love how they sound).

Being output voltage DACs, it means the chips already have a small built-in op-amp. If there's any possible issue, these opamps are weak and may need some help. The stock DCX/DEQ implementation takes the output from the DAC and feeds them into another set of two or thee opamps for output. Since we are using the DAC in a home environment where we don't use 50-75ft cable runs, we can get away with bypassing this last analog stage.

The DAC board carries the signal to an analog output board via ribbon cable. We snip the +/-/gnd wires on the ribbon cables, attach the +/- wires to capacitors (since there is +/ 2.5V offset from the DAC output) and voila. A balanced output suitable for home use which bypasses all that junk in the stock analog board.

Now what I did next was a little twist. Since the DCX has 6 channels (it is a crossover after all), I paralleled the the three balanced DAC outputs per channel into one. This is exactly the same technique used by those DACs which use multiple DAC chips to increase resolution and lower noise floor. It also increases current capability (especially necessary for bass and to lower output impedance.) So essentially each channel is the sum of information and current capability from 6 individual (3 + and 3 -) DAC chip channels.

I've also done a few other power supply and PS capacitor bypass tweaks. I may swap out the caps with transformers next. [good quality line stage transformers seem to sound more transparent than caps]

Marvey · « **Reply #6 on:** October 23, 2012, 07:57:12 PM »

IEMCrazy
Today at 7:02 am
Interesting about that AKM chip, I was unaware of those details of it beyond that it's documented that they prefer to be resampled to 24/96 when possible. The DCX/DEQ doesn't do that, but the SRC does and it's in the manual that it's ideal for how the D/A stage is set up to oversample first. I'd balk at it, but the oh-so-vaunted Sabre does that anyway, so...who cares.

Any idea which opamps are in the output stage that you bypass? From the description you gave I actually don't see much, on paper, wrong with the output stage unless they're using some really cruddy opamps, or if the PSU is pure garbage (which I wouldn't doubt either.) Even the WDS-1 uses opamps on the XLR outs....but they're good opamps. I actually pictured a muchworse scenario for the inside of the Behringer. The Bifrost uses JFETs as line drivers (ewww) so the opamps almost seem ideal in contrast

My main complaints with the Bifrost in contrast to the WDS-1 are the slight harshness, or perhaps ringing in the low treble, a comparative lack of clarity (noise floor?), and a non-linear bass (too little sub-bass, too much mid-bass.) That's the warmth you described, but when comparing A/B to a higher end DAC, it becomes disturbing. Minor squibbles though, in general it compares very well.

Interesting mod...also less ugly than I imagined. I wouldn't dice up my precious DEQs but for $170 ea/ the SRC would be a more interesting unit to dissect. Also has a better power supply from an analog perspective (DAC is one of it's main features, not procesing.) If I trusted myself with a soldering iron, I'd get an extra and try it out I'd be interested in seeing pictures if you ever opened it up again for any reason.

Also, VERY interesting with the DCX as well!

Marvey · « **Reply #7 on:** October 23, 2012, 07:57:28 PM »

purrin
Today at 8:37 am
The DCX/DEQ actually do resample to 24/96 first. The DSP EQ/crossover functions work in 24/96. Overall the stock SRC is slightly better sounding - better implementation - mostly because it is single purpose.

The opamps in the output stage I believe are NJM4556 (or something along similar lines). They are dirt cheap and actually used in the supremely overpriced Grado RA-1 amp (which is really a CMOY) and output stage of the nwavguy's O2 amp. They output a ton of juice, but they certainly aren't what I consider "good" opamps. It's actually more the implementation than use of opamps. I think the DAC output signal actually goes through two or even three opamps (I don't recall the schematic off the top of my head.) Just totally unnecessary.

LOL on the FET thing. I see you are not a fan of FET. FETs can sound fuzzy or have a "mist" or in general just plain sound flat and un-involving. I would prefer critical circuits like line or voltage multiplier stages not to be FET. It all depends upon implementation. The Mjolnir uses all FETs throughout, and you would be hard-pressed it hear it as such. The Asgard is a "bad FET sound" poster boy! But in defense of Schiit, the sound was intentional and I totally understand why the Asgard sounds that way. [it works well with Grados and the other entry level cans with treble ringing of yesteryear]

Indeed the Bifrost has that ever slight bit of digititus in the treble, lacks bass extension, and has a little mid-bass kick. [But what else are you going to get at that price]

The SRC is a little bit more painful to dissect. There is no handy ribbon cable as on the DCX/DEC to take the DAC chip outputs. You have to solder wires on to these tiny pads on the board.

Marvey · « **Reply #8 on:** October 23, 2012, 08:00:40 PM »

IEMCrazy
Today at 9:16 am
I thought the DSP's actually worked in 32/96, it just can't output at 32, it's internal only. Either way they may resample to 24/96 internally, but they generally sample back to the original before output. Though I think you can set it to output at an upsampled output. I forgot about that. Interesting that I've heard hooplah about the SRC's resampling in conjunction with the DEQ's. But I don't think the DEQ does VCXO reclock, SRC does.

It'll be very interesting to compare the SRC and the Bifrost. I'll be very amazed if the cheap little Behringer manages to best the Bifrost even in stock form! Obviously for most folks the balanced outs would be too messy as a drop-in replacement, even though it can be run unbalanced with an adapter cable.

Interesting about the opamps. If it's the output stage of of the O2, that's actually not bad. My gripe about O2 is that it's too dry and analytical and simply flat out painful to listen to as a result, so for a line driver that's probably a good thing. And that's simply hilarious about the RA-1! But you're right an implementation with 3 opamps can't be a good thing! Why three?

Yeah, fuzzy, misty...that describes what I found when comparing Bifrost to WDS-1. But I'm glad to hear Mjolnir doesn't have that effect. Lyr is FET as well, but over all the noise of the tubes, one doesn't notice it Asgard I have managed to not hear, amazingly. From the sound of it, I'm glad. I suspect my Headroom Micro (the original!) is another "bad FET sound poster boy" based on some comparative descriptions of tdockweilers dislike of Asgard with my dislike of the original Micro.

Yep, digititis in the treble, lack of bass extension, and that mid-bass kick. All the things that annoyed me in the contrast of the WDS, but I hadn't realized annoyed me previously except the mid-bass kick. I'd ask how you find the unmoddedSRC/DEQ/DCX in contrast (any such faults?) but it's been a while since you've heard it unmodded, so it's probably not a fair question.

And yuck about the SRC not being easy to mod. I think that eliminates the idea for me rather than risking the pricier units

Marvey · « **Reply #9 on:** October 23, 2012, 08:01:51 PM »

purrin
Today at 10:41 am
Yeah, the DEQ/DCX can benefit from a reclocker chip. [I mean super-precise clock mod] I am of the opinion that jitter, once past a certain point, is OK. Super low levels of jitter has never been a high priority for me. (I've actually got a few jitter measurements here and there.)

Let me know about the Bifrost comparison with stock SRC.

The analog / filter section of the DCX was a textbook balanced out design from a spec-sheet. I don't think much creativity was involved. Just one Lego set after another. This is understandable given its low cost.

Mjolnir still has a little bit of that effect. [low level FET fuzz] BTW, tubes done right are quieter [that is blacker] and cleaner sounding than FETs. I think FETs are best used on the power stage to provide current. The Schiit statement should really kick ass because the input FETs will be replaced by tubes.

From memory, the unmodded DCX was warmer, lusher and not sufficiently resolving for the HD800s. For speakers, they were good. I do think the stock DCX would be a good DAC for the HE-6.

Do you mind if I post this conversation in the secret members section of my website? This exchange can benefit a lot of people. A lot of good stuff here and I think you are spot-on with your impressions.

From memory, the unmodded DCX was warmer, lusher and not sufficiently resolving for the HD800s. For speakers, they were good. I do think the stock DCX would be a good DAC for the HE-6.

Author Topic: A good PM exchange I had with IEMCrazy (posted with his permission) (Read 459 times)

Marvey

A good PM exchange I had with IEMCrazy (posted with his permission)

Marvey

Re: A good PM exchange I had with IEMCrazy.

Marvey

Re: A good PM exchange I had with IEMCrazy (posted with his permission)

Marvey

Re: A good PM exchange I had with IEMCrazy (posted with his permission)

Marvey

Re: A good PM exchange I had with IEMCrazy (posted with his permission)

Marvey

Re: A good PM exchange I had with IEMCrazy (posted with his permission)

Marvey

Re: A good PM exchange I had with IEMCrazy (posted with his permission)

Marvey

Re: A good PM exchange I had with IEMCrazy (posted with his permission)

Marvey

Re: A good PM exchange I had with IEMCrazy (posted with his permission)

Marvey

Re: A good PM exchange I had with IEMCrazy (posted with his permission)

(adsbygoogle = window.adsbygoogle || []).push({});

Author Topic: A good PM exchange I had with IEMCrazy (posted with his permission) (Read 459 times)