Looking for Better Jitter Reduction in a DAC

Here I am again asking for help to troubleshoot the "hard edge" produced by my DAC. Here's an update...

I haven’t tried the iFi spdif ipurifier yet since I’m not convinced the problem is due to poor jitter control. I’m really sold on the HDMI (I2S) interface from PWT to DAC, it presents a realistic holographic image. Music sounds wonderful thru tubes, but there’s the damn harshness in the highs, same as with 75 ohm coax.

I tried an experiment, taking all digital and analogue components and running them thru a power strip on the same dedicated AC line. The result was a very offensive harshness in the top-end. Even though all PC’s are shielded, it’s possible there is backflow from the digital devices into the mains.
Normal setup is a separate line for analogue and digital wired from a subpanel. Low noise-floor, excellent imaging, but music has that digital edge to it. Tried it with and w/o power conditioning, I prefer it without. Also tried different AC receptacles around the house, cheater-plugs and a Jensen Iso-max.

Now I’m wondering if the DAC output of 2.8V (unbalanced) is too high and what I’m hearing is distortion thru the speakers, since it is noticeable on peaks; e.g., massed strings sound fine, but brass and solo violin will sound harsh (can sound like distortion).

My computer is still kaput, so I haven't been able to try CD rips or USB. Anybody know where I can get an inexpensive DAC for a trial or loan? The Cable Company has DAC's in the $2K to 4K range, so the trial would be very costly.

I do indeed think you may well have a gain mismatch between the DAC and the amplifier. What you described is precisely how this would sound like: clipping of louder passages.
Part of the traditional criticism of digital comes from precisely this, as the 2V output for cd players as specified in the Red Book was too high for many amplifiers of the day. Hence Quad after the introduction of CD produced a special input board for their existing amplifiers. These days, there are still many amplifiers that cannot quite handle this signal level. In your case, the situation is aggravated because your DAC has a 2.8V output (what is your amplifier’s input sensitivity?). A telltale sign would be that you do not have to open the volume control very much to get a loud signal. Now you may ask why manufacturers would specify a DAC or CD player with an output level that is too high, or amplifiers with a high input sensitivity. The answer is that the human brain interprets louder as better. So in a demo room, the louder player/DAC or the more sensitive amplifier will always be interpreted as the better.
So how do you find out if this is the case (after all, the problem may be different from what I suggest, mine is only an hypothesis)? The first test is the volume control setting. If this suggests that you may have a gain mismatch issue, there are two options. The technical one is to have a look with a scope. Such clipping is easily visible. The second is to use a clean test tone and raise the volume until you hear trouble. The third one is to insert inline attenuators, and see if that improves the situation. If it does, you also have your (fortunately very cheap) solution.
See here for some demo material and discussion: http://www.harbeth.co.uk/usergroup/forum/the-science-of-audio/amplifier-matching-mismatching-and-cli...

Harshness could also be your power amp or your speakers.

@willemj 

Higher voltages on line level devices has nothing to do with making louder volumes to sound nice. (although loud does sound nice as you hear more low level detail as it lifts the low level sounds above the ambient noise floor)

Generally you only see high voltage levels on Class A+ and pro audio equipment. 

The use of higher voltage more expensive components in a line level device is to achieve better THD+N and better SNR. Better performance.

Consumer RCA = cheap crap that is popular because it is cheap manufacture and can often sound OK to non-discerning listeners

Pro Audio balanced is the gold standard = higher signal level and shielded conne tions with signal wires that arent grounded. Expensive and always equal or better than RCA. 

My Benchmark DAC 3 was recently testsd by Stereophile - it delivers 18 volts on XLR at max volume. This high cost high spec analog output circuitry is how Benchmark achieve market leading specifications.

I am not talking about absolute levels, but relative ones, i.e. a mismatch between output level of the source and input sensitivity of the amp. Pro audio is very good with a properly matching gain structure, but home hifi is often not. I agree about XLR.

It is entirely possible that the harshness could be from the speakers. Tuner and TT sound fine, but are of course lower than 2.8V output gain. I'm using Gallo floorstanders; 88dB, 8 ohms, a benign impedance curve between 4 and 8 ohms. Gallo uses a very revealing 180 degree radiating tweeter, piezoelectric, no crossover.
I have also tried some soft dome bookshelves; 88dB, 8 ohms, and they presented the same harshness or distortion in the highs from digital.

Preamp is Atma-Shere UV-1; Line stage output...12V, 17V RMS at clipping. Output impedance... 400 Ohms.
No specs for input sensitivity. It must be low since VC is at about 8 to 9 o'clock, 11 o'clock for TT. Ralph installed a 6 dB pad in the AUX input for a digital device.

I am now bypassing the preamp and DAC is going direct to amp. Amp is a TAD-60 (Tube Audio Design); Input Sensitivity is 1.0V with variable gain control, Input Impedance…100K ohms.
TAD recommends that the gain control be turned to max when using a preamp. I need to leave it at 12:00 on the dial.
The DAC has a fixed output.

@willemj , thanks for that very interesting Harbeth link.