[qusp]

of course there is no DC offset with the caps in , thats what they do, remove DC offset, thats the only reason they are there otherwise there would be no need for them to be there, they sure dont improve the sound by being there; not even the best ones, they can only colour it, but also degrade it. best thing they do is pass AC, but not DC, DC cannot pass through the windings of the cap; the 1mv will most likely just be your meter not measuring low enough
 
 I meant to measure the raw output of the dac, before the input to the transformers or now you have the caps in, before the caps and best if you have a test signal to measure it with, rather than music which will be fluctuating

 

[Chris Adams]

Quote:

I meant to measure the raw output of the dac, before the input to the transformers or now you have the caps in, before the caps and best if you have a test signal to measure it with, rather than music which will be fluctuating

Ah, thanks for clarifying. I have a disc with warble tones at -20db. It would be nice to have something at 0db. I can't seem to find anything on the web.
I will check the DC before the caps. I would really like the use the TXs and no caps.
The buffer is a Pass B1. Input R is set by the pot which is 20k in mine and specified at 25k...it will operate with higher or lower R to suit the need. It is class A single ended with unity gain. http://www.passdiy.com/pdf/B1%20Buffer%20Preamp.pdf
 
I can run the output of my McCormack signature cd player into the buffer with no distortion and the output of that is approximately 3.5v...quite hot, sounds great.
 
I experienced the same distortion with the TXs and caps while the dac was running into the aux of a vintage HK 380i receiver.
 
Just measured the dc before the caps...right off the dac. Right channel is 4.69v and left is 4.61v.

 

[qusp]

yeah I run the pass D1 mosfet IV (well a modified version by OPC) and in the middle of building the hotrodded DCB1 DC coupled pass buffer with salas shunts (the latest revision). knew there had to be a pot there for that reasonably low input impedance.  I wont be using a pot, as I use digital volume control in my sabre. waiting for the new ZEN IV to be realized with servo and no output caps.
 
that output is way hot, thats even hotter than a balanced signal should be. you need to attenuate that somehow, so what happens when you lower the volume on the pot in the buffer? do you still get clipping then? line level single ended audio is supposed to be 2v point to point, a bit hotter is OK depending on your pre or amp, but thats even hotter than the 4v swing of balanced. and is that with the -20db disk?? :eek: seems a bit of a silly design to try adding DNR by paralleling the dacs, but then you have to attenuate it and thus lower DNR, plus add to the output impedance. hmm, I think you need a transformer to attenuate the signal, or some sort of active stage, I would do it with some jfets or god forbid, a nice jfet input opamp, cheap fix is turning the buffer pot down (paying the penalty of lower DNR and higher noise, but that massive voltage I seriously cannot see how that wouldnt saturate the TX.

 

[Wood]

Quote:

If anything this thread may well be doing its bit to discourage someone like myself that may or may not carry out mods from actually buying this thing.  There has been little analysis of this unit, purchased as is, as compared directly to competitors at its price point.  If there is, I may have missed it with all the "fault rectifications" being carried out with this unit. 
 
The thought of an inexpensive R2R DAC design that competes well in the under 1k category - appeals to me, unfortunately for the Chameleon this category seems locked by Audio GD via plenty of users able to contrast it with competing products.

Er, yes its true, the guys here don't compare to other dacs.... they are simply trying to squeeze this dac to produce some small measure of the musicality of their analogue rigs. 

 

[Wood]

Chris, did you try adjusting the blue trimmer when you used the TXs?

 

[Chris Adams]

Quote:

Chris, did you try adjusting the blue trimmer when you used the TXs?

I did but don't know how to measure it. I only went a couple of turns in each direction, while listening, so I wouldn't go too far.

 

[Chris Adams]

Quote:

yeah I run the pass D1 mosfet IV (well a modified version by OPC) and in the middle of building the hotrodded DCB1 DC coupled pass buffer with salas shunts (the latest revision). knew there had to be a pot there for that reasonably low input impedance.  I wont be using a pot, as I use digital volume control in my sabre. waiting for the new ZEN IV to be realized with servo and no output caps.
 
that output is way hot, thats even hotter than a balanced signal should be. you need to attenuate that somehow, so what happens when you lower the volume on the pot in the buffer? do you still get clipping then? line level single ended audio is supposed to be 2v point to point, a bit hotter is OK depending on your pre or amp, but thats even hotter than the 4v swing of balanced. and is that with the -20db disk??

" class="bbcode_smiley" height="" src="http://files.head-fi.org/images/smilies//http://hfimage.head-fi.org/smilies/eek.gif" title="

" width="" /> seems a bit of a silly design to try adding DNR by paralleling the dacs, but then you have to attenuate it and thus lower DNR, plus add to the output impedance. hmm, I think you need a transformer to attenuate the signal, or some sort of active stage, I would do it with some jfets or god forbid, a nice jfet input opamp, cheap fix is turning the buffer pot down (paying the penalty of lower DNR and higher noise, but that massive voltage I seriously cannot see how that wouldnt saturate the TX.

I'm sorry I wasn't clear. The 4.69 and 4.61 is the dc offset. I can't get an accurate read for ac voltage out without a 0db tone. I did measure with the -20db and it is about .19vac at the rca jacks.
It is the dac that was clipping, not the buffer. I can run the McCormack with 3.5v into the buffer with no problems.
 

 

[skibum]

Quote:

I'm sorry I wasn't clear. The 4.69 and 4.61 is the dc offset. I can't get an accurate read for ac voltage out without a 0db tone. I did measure with the -20db and it is about .19vac at the rca jacks.
It is the dac that was clipping, not the buffer. I can run the McCormack with 3.5v into the buffer with no problems.
 

Hi Chris,  You need to measure the output with a 0dB 1 KHz sine wave. Download this free app and you can produce any frequency & level you want:  http://www.tropicalcoder.com/AudioTestFileGen.htm
 
The blue trim pot (variable resistor) should be adjusted to get an approximate 2.0 Volts output when playing a 0dB 1KHz tone.   One warning is do not turn the pot to much.  Once it goes past its end of travel it will not go back and is broken.
 
I spent yesterday afternoon trying to get a pair of output transformers working on my VALAB with no real luck.  Just couldn't get them to sound good.  No low end output at all.  But I definately had to adjust this pot when going from my caps to the transformers.

 

[Chris Adams]

skibum, thanks for that link. I'll give it a try today.

 

[Chris Adams]

skibum, thanks! I downloaded and burned a 1k 0db disc. Threw it in the transport and was able to adjust the dac to 1.9v out. Sounds very good. As soon as I have time I'll try those TXs again. That app is terrific...great tool.

 

[Chris Adams]

Well, I put the TXs in and learned something. I had set the dac output at 1.9v with the caps. With the TXs in and no other adjustments, the voltage dropped to 1.7. This may not sound like a lot but it was enough to cause some distortion on certain discs. I brought it up to 1.9 and still some distortion. Then I checked the dc in and found it had dropped slightly from 7.85 to about 7.8. Hmmm...I dropped the dac output to 1.6 and distortion went way up. After a couple of increases I ended up at 7.95dc in and 1.99ac out. Adjustments of less then a tenth of a v made a difference. I may experiment a bit more but right now it sounds wonderful. Compared to the 4.7uf Sonicaps, there have been veils lifted. The layering is now quite deep. I am hearing things I had heard but didn't recognize. There is more air and space around the instruments and singers. The dac always had good drive but now it seems to "puuump you uuup"

Rockin' good!
 
By the way, the TXs are inside the dac, yup. I had connected them outside and when things were running fine I moved them around. Side by side, above, below and next to each other as they would be in the dac. No noise, distortion or unusual sounds, just clear and clean. So in they went just like the picture a few posts back.
 
Thanks for the help. This is a great place.

 

[qusp]

Quote:

I'm sorry I wasn't clear. The 4.69 and 4.61 is the dc offset. I can't get an accurate read for ac voltage out without a 0db tone. I did measure with the -20db and it is about .19vac at the rca jacks.
It is the dac that was clipping, not the buffer. I can run the McCormack with 3.5v into the buffer with no problems.
 

 
yes I know its the dac clipping, but I was trying to rule out whether it was the dac output being too hot and clipping the buffer input, not the buffer output. for finding the right loading on the TX primaries I usually use a pair (quad in my case) of vishay foil trimmer pots across each leg to ground to adjust the resistance till it sounds right and then once thats found, measure across the trimmer pot to find where its at and replace with set value resistors.
 
glad you got it all sorted out
 

 

[Wood]

Excellent news Chris, glad it was worth it.
 
I prefer to adjust the trimmer by ear, turn it anticlockwise until the sound, just, stops distorting and then continue turning, but this time counting each turn until it begins to distort again. Finally turn it half the the number of counted turns in a clockwise direction. This will set the trimmer midway between distortion points.
 
Be great to hear more of your impressions after you have had some time to enjoy the TXs.

 

[Chris Adams]

Quote:

Excellent news Chris, glad it was worth it.
 
I prefer to adjust the trimmer by ear, turn it anticlockwise until the sound, just, stops distorting and then continue turning, but this time counting each turn until it begins to distort again. Finally turn it half the the number of counted turns in a clockwise direction. This will set the trimmer midway between distortion points.
 
Be great to hear more of your impressions after you have had some time to enjoy the TXs.

Hey, Wood, thanks for sharing that technique. I'll give it a try. Will do on the impressions.
 

 

[Wood]

Chameleon review here
 
http://www.digitalaudioreview.com.au/index.php/audio-reviews/digital-source-reviews/item/128-teradak-chameleon-16-x-tda1543-dac