[adriaan243]

I tested my Valab thoroughly using a professional AD converter.
Result: Frequency response flatter then I expexted, -0.2 20hz, -0.5 10khz,
-1.2 20 khz.
My Valab is definitely hard clipping at -3 db, no question about it!
I found out that the 340 ohm resistor in the output is actually a 390 ohm resistor.
I don't know how the output works, but it seems the answer lies there and it may also be a disadvantage of this "direct coupled" output.

 

[erin]

The resistor works like this:
The TDA1543 outputs current, when a resistor is put after the dac, ohms law comes into effect.
Volts = Current x Resistance. The resistor turns the current into a voltage.

The TDA1543 datasheet says that the average output current is 2.3mA so eight DAC's paralled should output 18.4mA full scale, so the max output voltage with the 390 Ohm resistor should be about 7.16V peak / 2 /1.414 = 2.53 V rms.

It seems to me that the bias adjust resitor could be the culprit for the clipping on the negative half cycle. This is the resistor that goes from pin 7 (V ref) of the DAC IC's to ground.

Also, this resistor adjusts the mid point of the sine wave.

Changing the value of this resistor is the more elegant solution to this clipping problem. It will maintain the same output voltage, but have the midpoint set just above where it is now avoiding the clipping of the Negative half cycle.

the specifications of each Valab will be slightly different so thats probably why many people are having issues with clipping.
Some units may be clipping on the positive half cycle. The only way to properly adjust these units would be to install a potentiometer instead of the bias resistor and whilst playing a sine wave through the DAC, adjust the trim pot unitl the clipping goes away

 

[sfogar]

Hi,

a friend of mine confirm that the Lite DAC AH, a similar Dac, has a trimmer on the pin 7...

If this is confirmed this is a bad unit design.

All the best

Alessandro

Quote:

Originally Posted by erin /img/forum/go_quote.gif The resistor works like this: The TDA1543 outputs current, when a resistor is put after the dac, ohms law comes into effect. Volts = Current x Resistance. The resistor turns the current into a voltage. The TDA1543 datasheet says that the average output current is 2.3mA so eight DAC's paralled should output 18.4mA full scale, so the max output voltage with the 390 Ohm resistor should be about 7.16V peak / 2 /1.414 = 2.53 V rms. It seems to me that the bias adjust resitor could be the culprit for the clipping on the negative half cycle. This is the resistor that goes from pin 7 (V ref) of the DAC IC's to ground. Also, this resistor adjusts the mid point of the sine wave. Changing the value of this resistor is the more elegant solution to this clipping problem. It will maintain the same output voltage, but have the midpoint set just above where it is now avoiding the clipping of the Negative half cycle. the specifications of each Valab will be slightly different so thats probably why many people are having issues with clipping. Some units may be clipping on the positive half cycle. The only way to properly adjust these units would be to install a potentiometer instead of the bias resistor and whilst playing a sine wave through the DAC, adjust the trim pot unitl the clipping goes away

 

[dcolak]

Quote:

Originally Posted by tubes /img/forum/go_quote.gif Just curious, have you guys with the distortion problems tested any other dacs? There are all kinds of sources of distortion with digital, you have jitter, upsampling, filters etc.. As far as I'm concerned I hear distortion with every piece of digital gear I've owned and the Valab to my ears is the least offensive yet. I have the 2009 Teradak which is basically the same as the Valab with a few different caps. I looked into the LT1084 Voltage Regulator as a possible source of distortion, if the 1084 adjust pin is bypassed you need a 150uf cap on it, I have a 220uf Rubycon ZL cap so I replaced the 47uf Sanyo Oscon and the sound fleshed out nicely and is sounding nicer in my system than other digital I've had including the Benchmark and the CA 740C CD player.

I also have M-Audio FastTrack DAC, NuForce Icon DAC, SqueezeBox DAC, and why not mention also, M-Audio Revolution 5.1 DAC and the DAC in NAD C541i.

Of course, they do not distort or clip the music.

 

[adriaan243]

I changed the 390 ohm resistor, now a Caddock 330 ohm and adusted the blue trimpot for best clipping behavior.
The DAC now clips at the maximum level, as it should.
A slightly lower level may even be safer.
For more information: http://www.dddac. DDDAC 1543

 

[tubes]

Hi Adriaan243, a full discription of how you adjust the blue trimpot would be a great help for those of us that have no fear of arming ourselves with a soldering iron but are not yet electrical engineers. Thanks.

 

[prog]

I/V and Ref resistors explained:
diyAudio Forums - Non Oversampling DAC-complementing CD-PRO - Page 6

 

[adriaan243]

Quote:

Originally Posted by tubes /img/forum/go_quote.gif Hi Adriaan243, a full discription of how you adjust the blue trimpot would be a great help for those of us that have no fear of arming ourselves with a soldering iron but are not yet electrical engineers. Thanks.

I did it by adjusting for lowest distortion near clipping using an analyzer that can measure harmonic distortion.
It should also be possible using a sine wave and listen to it or to use an oscilloscope, it has been described here: DDDAC 2000 DDAC1543

 

[Angsila]

Quote:

Originally Posted by microzyma /img/forum/go_quote.gif It's audible with music. I also thought the fault was in my speakers. Apparently a simple fix. Just ordered some 340ohm resistors. It's annoying that I had already ordered some nice i/v resistors but whatever. We bought a dac from a company we never heard of at a cheap price from a far away land known for erratic quality control. There is no point whining about it. It is our responsibility as consumers to vet our purchases. That said, I will not be buying anything else from Valab. I don't think they have the technical expertise or auditory acuity to be credible in this business. Lesson learned. New customers may decide Valab just spent a good deal of time learning more about what they are doing and decide it's a worth wagering. Both opinions are valid and caveat emptor.

microzyma,

Can you provide a list of the CD's and tracks where you say it's audible?

 

[Angsila]

Quote:

Originally Posted by dcolak /img/forum/go_quote.gif I also have M-Audio FastTrack DAC, NuForce Icon DAC, SqueezeBox DAC, and why not mention also, M-Audio Revolution 5.1 DAC and the DAC in NAD C541i. Of course, they do not distort or clip the music.

I think we all know where you are coming from now.

 

[PatOMalley]

Quote:

Originally Posted by microzyma /img/forum/go_quote.gif This is a fairly significant sonic detriment. All of the good things are still good things. But we are talking about enough distortion that it sounds like two notes are being played. I didn't notice it at all in spdif(didn't put tones across it in spdif yet, maybe it's OK?) but when I changed over to usb on Saturday the dynamics were so good I wanted to think the distortion in peaks was in my speakers(looking for an excuse to upgrade). If you were to add up all of the tweaks you could do(caps, i/v resistor, ers paper, damping, isolation, etc) it would not make as much of a difference as getting this sorted out. Makes me think the 2008 on spdif are OK because you would definitely notice it Dialing back -3db in foobar makes the distortion go away but is missing some low level details and dynamics.

I have listened hard for it and have not heard two notes at the same time or any of the description of clipping I have read here. I HAVE heard clipping on an AMP and that would be unmistakable to me. I CAN hear when the original recording is distorted much better now

but distortion qua distortion I do not hear. ["hearing" is subjective, admittedly] I do not use a PC. I use the SPDIF coming in from an NAD C541i over glass Toslink.

 

[TeraDak]

For those concering the clipping issue.
We current use a 390R resistor for the i/v conversion and it will output 2.19V at the full range audio data. It is impossible to listen the 0-db audio signal from the DAC , to pre/power amplifier and to speaker without any attenuation. It the source doesn't exceed -3db, there will be no clipping issue. However, if you would like to attenuate the signal in the preamplifier and keep the range as full as possible and without clipping. The i/v resistor in the TDA1543 output should be lowered to 340R ~ 350R to avoid clipping. The clipping issue is just a tradeoff for signal attenuation and it depends on the user listening condition. We will show the picture for analog output with i/v resistor 340R ~ 350R. For signal analysis, it should be 340R~ 350R, but it may not be the best for users in most real listening conditions.

(1) For user using the original 390R resistor, if you care about the clipping issue, you can reduce the resistor to 340R ~ 350R without loss the audio quality. If you don't have 340R ~ 350R resistor at hand ( 340R ~ 350R resistor is difficult to buy), you can parallel a 3.3k or 3.6k on the original 390R resistor. However, a 330R resistor is good aslo. The main difference is the maximum output voltage when a 0db signal is input.

(2) After that, connect the USB line to our DAC and repeat playing a 1kHz 0DB sine wave in PC. The wave could also be downloaded form our share space ( Public - Windows Live )
Use a meter and set it to measure AC voltage segment. Measure the RCA output or our DAC. The AC output is approximate 2.0V initially. Trim the VR nearby TDA1543 to find the maximum AC output. We usually trim the VR with the help of scope. But a meter is also good enough for tuning. If you have a scope, you can trim the VR such that the sine-wave is the most symmetric. The most symmetric output implies the maximum AC output in a meter.

The second step can be skipped it there is no meter or scope at hand. There is only few difference after step (2) adjustment.
/********************** Notice **********************/
The most important should be noticed is the VR trimming should be after the DAC power-on 30 mins. After that, the DAC will be in a stable state for adjustment. In the first 30 min, the TDA1543 was not in a stable stage. Trimming in the unstable state is no use.
/**********************************************/

 

[tubes]

Michael, thanks for the update.

 

[guitarplayer]

I have been working closely with Michael from Teradak on improving this DAC and solving any issues that the community has discovered on the product that he OEM's to another customers specifications.

I have found him very competent, accommodating, open and honest. I believe he will work through these problems and any future products will be rock solid stable.

Anyway, my $.02.

Peace,

Lee

Quote:

Originally Posted by TeraDak /img/forum/go_quote.gif For those concering the clipping issue. We current use a 390R resistor for the i/v conversion and it will output 2.19V at the full range audio data. It is impossible to listen the 0-db audio signal from the DAC , to pre/power amplifier and to speaker without any attenuation. It the source doesn't exceed -3db, there will be no clipping issue. However, if you would like to attenuate the signal in the preamplifier and keep the range as full as possible and without clipping. The i/v resistor in the TDA1543 output should be lowered to 340R ~ 350R to avoid clipping. The clipping issue is just a tradeoff for signal attenuation and it depends on the user listening condition. We will show the picture for analog output with i/v resistor 340R ~ 350R. For signal analysis, it should be 340R~ 350R, but it may not be the best for users in most real listening conditions. (1) For user using the original 390R resistor, if you care about the clipping issue, you can reduce the resistor to 340R ~ 350R without loss the audio quality. If you don't have 340R ~ 350R resistor at hand ( 340R ~ 350R resistor is difficult to buy), you can parallel a 3.3k or 3.6k on the original 390R resistor. However, a 330R resistor is good aslo. The main difference is the maximum output voltage when a 0db signal is input. (2) After that, connect the USB line to our DAC and repeat playing a 1kHz 0DB sine wave in PC. The wave could also be downloaded form our share space ( Public - Windows Live ) Use a meter and set it to measure AC voltage segment. Measure the RCA output or our DAC. The AC output is approximate 2.0V initially. Trim the VR nearby TDA1543 to find the maximum AC output. We usually trim the VR with the help of scope. But a meter is also good enough for tuning. If you have a scope, you can trim the VR such that the sine-wave is the most symmetric. The most symmetric output implies the maximum AC output in a meter. The second step can be skipped it there is no meter or scope at hand. There is only few difference after step (2) adjustment. /********************** Notice **********************/ The most important should be noticed is the VR trimming should be after the DAC power-on 30 mins. After that, the DAC will be in a stable state for adjustment. In the first 30 min, the TDA1543 was not in a stable stage. Trimming in the unstable state is no use. /**********************************************/

 

[tubes]

That's good news, Thanks.