|
|
Post by honk on Dec 5, 2010 5:43:39 GMT -5
Sorry to say it again, the gain of the XDA1 is simply a joke! I don't care if it's linear instead log., or if it's better working with AVs in between, or with some additional resistors to the XLR inputs.
NO! The XDA-1 should work excellent and practically with the XPA-1, if it doesn't, the target CLEARLY failed.
|
|
|
|
Post by jimangie1973 on Dec 5, 2010 6:05:09 GMT -5
An in-line attenuator (or potentiometer) will have a series resistance followed by a shunt resistance to ground. The relative values of these two resistances determines the attenuation. It is critical that output impedance of this attenuator does not exceed a couple of KOhms. So for example, a 10KOhm pot will have the highest output impedance at 6dB attenuation when the series and shunt resistances are each 5KOhms. The output impedance will be 5KOHm in parallel with 5KOhm which is 2.5 KOhm. A 20KOhm pot will have a worst case output impedance of 5KOhm, a 50KOhm pot will have a worst of 12.5KOhm.
The reason this high output impedance causes a problem is because it interacts with the capacitance of the cable between the attenuator and the amp and, usually of even bigger significance, the input capacitance of the amplifier. Most amplifiers have an input capacitance ranging from 200pF to 600pF. When the output impedance is high, it creates an RC lowpass filter with the capacitance and starts to attenuate the top octave.
The best compromise between maintaining a lowish output impedance (not too high a value) and avoiding overloading the source component (not too low a value) is 10KOhm.
The Goldpoints are really nice but I'd go with 10K instead of the 20K that is recommended on their site.
With regard to the XDA volume control, the specs say it is logarithmic, "Internal Volume Control: Digitally controlled in increments of 0.5db". It doesn't make sense to me that Analog Devices would make a linear taper on an audio DAC. If any of you have a multimeter, read the output voltage at different volume settings with a fullscale 60Hz sinewave as the input. You can create a wav using Audacity which is freeware.
|
|
|
|
Post by khonfused on Dec 5, 2010 6:53:05 GMT -5
Sorry to say it again, the gain of the XDA1 is simply a joke! I don't care if it's linear instead log., or if it's better working with AVs in between, or with some additional resistors to the XLR inputs. NO! The XDA-1 should work excellent and practically with the XPA-1, if it doesn't, the target CLEARLY failed. I do have to agree 100%, although I've found a work around for me. My XDA-1 is directly connected to the XPA-5 and 0.5 is sometimes too loud for background listening. I then use the volume control on itunes. This is okay since it's for background listening only. For serious quality listening 0.5 or 1.0, very rarely up 2.0 is okay, so I can live with the current situation and will not look into additional equipment to add to my set-up for now. I do hope though that Emotiva will address this issue in the future and supply a solution, because the XDA-1 is a fantastic piece of equipment sound wise. |
|
|
|
Post by rockguitar on Dec 5, 2010 7:55:26 GMT -5
With regard to the XDA volume control, the specs say it is logarithmic, "Internal Volume Control: Digitally controlled in increments of 0.5db". It doesn't make sense to me that Analog Devices would make a linear taper on an audio DAC. If any of you have a multimeter, read the output voltage at different volume settings with a fullscale 60Hz sinewave as the input. You can create a wav using Audacity which is freeware. It's already been done by cfelliot in this post on page 1 of this thread emotivalounge.proboards.com/index.cgi?board=preampdac&action=display&thread=15010&page=1That's how I concluded there is a linear taper vs log. in the same thread. 80 vol. setting = 2volts, 40 vol. setting = 1 volt --> linear taper. Logarithims do not exactly double. I tested the XDA-1 outputs tonight by doing the following: I created a pink noise wav file on my PC normalized to 0 dB. I burned this file to an audio CD. I played the track with my old CD player and observer 2.1V out with my multi-meter. I played the track on my PC to the XDA-1 at a volume level of 80 and also got about 2V. Note - unbalanced! I lowered the volume to 40 and got about 1V. Bottom line: nothing wrong with the XDA-1! |
|
|
|
Post by Chuck Elliot on Dec 5, 2010 10:08:53 GMT -5
The AD1955 controls volume with a 14 bit linear register.
The max value in a 14 bit register is 16383.
The dB reduction at any step is = 20*log10(x/16383) where x is the value put into the register.
This means that max ref volume @ 0 dB = 20*log10(16383/16383)
Minimum volume is -84.3 dB = 20*log10(1/16383)
Volume ranges of 80-0 or 0dB to -80 dB can be just calculated steps within the linear range to optimize to .5 dB steps.
The granularity of the volume step in the linear register will not allow .5 dB steps below a raw register value of about 18 or -60 dB.
So, clearly the XDA-1 is able to run this way.
Was it just a mistake the a linear stepping of the Ad1955 was used?
|
|
|
|
Post by jimangie1973 on Dec 5, 2010 14:13:58 GMT -5
The AD1955 controls volume with a 14 bit linear register. Was it just a mistake the a linear stepping of the Ad1955 was used? You're right. Analog Devices used enough bits in the volume taper to allow the user control the taper how he wishes. There's no reason a large range logarithmic taper cannot be implemented with the AD1955, if the right values are sent to the part. |
|
BigRed
Seeker Of Truth
Posts: 7 
|
Post by BigRed on Dec 6, 2010 0:08:57 GMT -5
Just finished hooking up the XDA-1 blue light "show". Very nice fit with the UPA-2. I was a little concerned though just prior to the install after reading this section. It was definitely not my intent to purchase a USP-1 preamp for my 2-channel Apple TV (original) system. Especially if the only reason I needed it was to expand volume control. That is precisely why the XDA-1 is so perfectly appealing to me, it does it all. I had no problems since I was able to fine tune the LEVEL output on the back of the Amp, but I suppose all the Amps don't come with that feature.
|
|
henne
Minor Hero
Posts: 12
|
Post by henne on Dec 6, 2010 3:07:04 GMT -5
I did read somewhere (a few weeks ago): "the AD1955 has an on-chip clickless analog volume control, without loss in resolution".
Hendrik Jan.
|
|
henne
Minor Hero
Posts: 12
|
Post by henne on Dec 6, 2010 3:16:00 GMT -5
Most of the discussion is too technical for me. Do I understand it correctly that there is indeed no loss in resolution??
Hendrik Jan.
|
|
|
|
Post by gman1234 on Dec 6, 2010 9:05:34 GMT -5
From what I take from all of this is the volume is an analog resister controlled volume, with no affect on the digital resolution, ie bits. Setting the unit at 80 (highest) should provide a straight connection with no degredation.
My questions is, is highest setting 2 volts? Or did Emo ramp this output up a little due to it potential use without a preamp. Could this cause an overload issue for some using it with a preamp?
|
|
|
|
Post by Chuck Elliot on Dec 6, 2010 10:50:56 GMT -5
From what I take from all of this is the volume is an analog resister controlled volume, with no affect on the digital resolution, ie bits. Setting the unit at 80 (highest) should provide a straight connection with no degradation. Nor should any value below 80 because it's analog. My questions is, is highest setting 2 volts? Or did Emo ramp this output up a little due to it potential use without a preamp. Could this cause an overload issue for some using it with a preamp? Most all CD players and DVD players output about 2V at full signal, as the XDA-1 appears to do. This should not overload an input meant for this type of signal. |
|
|
|
Post by merlinwerks on Dec 6, 2010 14:14:00 GMT -5
Do I understand it correctly that there is indeed no loss in resolution?? Hendrik Jan. Yes |
|
|
|
Post by Chuck Elliot on Dec 6, 2010 19:59:19 GMT -5
Ok, regardless of technical discussion, the XDA-1 volume control does not work as advertised.
It is not a 0.5 dB/step as specified on the product page.
Emotiva(read Lonnie): Is there a fix??? Firmware update possible through the USB? Send back for fix?
This may not be an issue for those going into a USP-1, future XSP-1 or AVR, but it is for most of us running direct.
|
|
|
|
Post by ajani on Dec 6, 2010 22:11:22 GMT -5
Ok, regardless of technical discussion, the XDA-1 volume control does not work as advertised. It is not a 0.5 dB/step as specified on the product page. Emotiva(read Lonnie): Is there a fix??? Firmware update possible through the USB? Send back for fix? This may not be an issue for those going into a USP-1, future XSP-1 or AVR, but it is for most of us running direct. Seconded... I really hope they come up with a fix (preferably one that doesn't require shipping the unit back, as I'm an overseas customer)... And since I use my XDA-1 direct to my XPA-2, I really would like to have proper range with the volume control... |
|
|
|
Post by sharkman on Dec 6, 2010 22:26:11 GMT -5
I use the XDA directly into my XPA-3 and get usable volume up to 30 depending on the cd(some are only 14). I assumed that with some devices the upper end of the volume would come into play, like with an Ipod or computer as source. Perhaps I am incorrect?
I have XLR cables coming in about a week, and I wanted to compare volume performance with these cables.
|
|
|
|
Post by Topend on Dec 6, 2010 22:54:38 GMT -5
It would be great if Lonnie could weigh in on this thread.
Dave.
|
|
|
|
Post by rockguitar on Dec 7, 2010 9:58:53 GMT -5
The AD1955 controls volume with a 14 bit linear register. Was it just a mistake the a linear stepping of the Ad1955 was used? You're right. Analog Devices used enough bits in the volume taper to allow the user control the taper how he wishes. There's no reason a large range logarithmic taper cannot be implemented with the AD1955, if the right values are sent to the part. In my investigation of the AD1955 DAC chip, you look correct. The Lavry DA10 uses the same chip and appears to use the same on-board volume control but uses a logarithmic taper instead of linear. So it is definitely up to the designer to pick what he wants to use. I posted a lot more indepth about the Lavry on page 2 of this thread. emotivalounge.proboards.com/index.cgi?board=preampdac&action=display&thread=15010&page=2Log vs. Linear volume control -- why is this so critical? Because our hearing is roughly logarithmic so a log volume control better correlates with how loud we perceive the sound to be. |
|
|
|
Post by Chuck Elliot on Dec 7, 2010 12:49:40 GMT -5
Talked to Nick at Emotiva support today.
It sounds like they are aware of the issue with the volume control, but not much else regarding a fix or firmware updatability.
I guess we will just have to wait for Lonnie.
|
|
|
|
Post by srb on Dec 7, 2010 13:48:31 GMT -5
Talked to Nick at Emotiva support today. It sounds like they are aware of the issue with the volume control, but not much else regarding a fix or firmware updatability. I guess we will just have to wait for Lonnie. It's hard to imagine how this could have been missed during the prototyping and testing stages. Hooking it up to their own amplifiers should have revealed the potentially problematic volume adjustment. At this point, they should probably at least modify their website and change Internal Volume Control: Digitally controlled in increments of 0.5db to Internal Volume Control: Digitally controlled in linear increments of 0.5 on a 0 to 80 scale. Not that prospective buyers would catch the the linear vs logarithmic distinction. I guess at least drop the "db" part or the "in increments of 0.5db" (which brings up the minor point of using the correct unit; "dB", instead of "db"). Steve |
|
|
|
Post by Chuck Elliot on Dec 7, 2010 13:59:03 GMT -5
....At this point, they should probably at least modify their website and change Internal Volume Control: Digitally controlled in increments of 0.5db to Internal Volume Control: Digitally controlled in linear increments of 0.5 on a 0 to 80 scale. ..... Steve But, even that is not correct. The way is is now there is no uniform step. The lower volume settings make a huge change in output, while the upper region makes very little. |
|
