Now with the aid of the scope and the +3db I should be able to see it at the different points, after the D/A then after the I/V and finaly the output to see which one goes into cliping.
My aim is that (hopefully it's not the D/A) is to get the +3db as clean as the 0db at least that should stop the slight over tone at 0db.
Any one got any other ideas?
Cheers George
My aim is that (hopefully it's not the D/A) is to get the +3db as clean as the 0db at least that should stop the slight over tone at 0db.
Any one got any other ideas?
Cheers George
....These last scope shots show ADC overload. ...and the performance is good. No wrap-around etc. The DAC get data only. The only valid data is 1 or 0. 16-bit (or 18-bit, 20-bit, 24-bit, etc) is serially fed to the DAC (most cases) and latched in. Think of the DAC as a fast, programmable power supply. Based on the input word value, a programmed output current is generated. You can only feed a min value 0x000000...... to 0xFFFFFFFF..... The DAC cannot latch in a larger word size, nor can it accept other than binary values. You can overload Digital Acquisition. Hopefully it will have smooth overload performance. You cannot overload a DAC.
mrshow4u, you kind of lost me a bit there, what your saying I believe is that the clipping is the D/A converter itself and that it looks good to you?
Then why if you can explain simply to a digital pleb like me, can I hear a slight distortion on the esl's speakers at 0db digital but not at -3db digital or -10db digital all being played back at the same analog volume?
It seems to me that the scope cannot pick up the slight distortion that I hear, but my ears can at 0db digital.
Cheers George
Here are the notes of those three tracks
Then why if you can explain simply to a digital pleb like me, can I hear a slight distortion on the esl's speakers at 0db digital but not at -3db digital or -10db digital all being played back at the same analog volume?
It seems to me that the scope cannot pick up the slight distortion that I hear, but my ears can at 0db digital.
Cheers George
Here are the notes of those three tracks
Attachments
Sort of. The current that the DAC is putting out is programmed, but not distorted. Yes, the waveform is a distorted audio waveform, but the DAC is not clipping. In digital, you can program a word value between all bits off to all bits on. Say you have a scale from zero to one. The range is divided up this evenly by the word length (16-bit, etc) for linear PCM. So that flat-top overload that you see will happen on every DAC. That's what the data is "telling" the DAC to output. That flat-topped waveform that you see was either: generated at the conversion Analog overloading a ADC, or the overload occured in DSP through addition or multiplication of non-overloading data. If your finding overload occuring and dependent on the DAC circuit, it could be from the I/V converter with too much gain or too small rail voltage.
......I just saw the added info about 0dB FS with audible distortion but not -3dB FS.. Make sure when making this comparison, that the volume control is knocked down 3 dB when playing the -0dB FS track. Maybe the panel is distorting a little bit with the increased level? Kepping the same acoustic level will determine that.
Yes the same acoustic level was carfully matched.
I posted the circuit back one page of the DAC.
mrshow4u, I'm using these AD825 http://www.analog.com/UploadedFiles/Data_Sheets/218761852AD825_f.pdf
for the I/V with 15+ and 15- they are fet input, the originals were NE5532's and they are bi-polar input, do you think the problem lays there? As the Ad825's sound so much better than the NE5532's and the specs kill them also.
Cheers George
I posted the circuit back one page of the DAC.
mrshow4u, I'm using these AD825 http://www.analog.com/UploadedFiles/Data_Sheets/218761852AD825_f.pdf
for the I/V with 15+ and 15- they are fet input, the originals were NE5532's and they are bi-polar input, do you think the problem lays there? As the Ad825's sound so much better than the NE5532's and the specs kill them also.
Cheers George
Hi!
Why not settle the controversy by sticking a distortion analyzer at the output of the CD player? It's extermely difficult to see "slight" distortion of sine waves.
You may be better off digitally synthesizing a triangle wave that peaks at the maximum possible digital values and looking at that on the tek scope - that "may" be more visible. I can easily syth that with a program and send it to you as a wav file if you wish - then just burn it into a cd-r.
Could the 0dBFS be overloading something else in the playback chain?
Cheers!
Clem
Why not settle the controversy by sticking a distortion analyzer at the output of the CD player? It's extermely difficult to see "slight" distortion of sine waves.
You may be better off digitally synthesizing a triangle wave that peaks at the maximum possible digital values and looking at that on the tek scope - that "may" be more visible. I can easily syth that with a program and send it to you as a wav file if you wish - then just burn it into a cd-r.
Could the 0dBFS be overloading something else in the playback chain?
Cheers!
Clem
Hi Georgehifi, Does the problem lie in the NE5532??? Hmmm. Well I can't say. I looked at the -0 dB FS (full-scale) and did not see overload. It "looks" good. If you play the -0 dB FS 1 kHz tone and turn down the volume a little, does it still sound distorted? From the scope shot, it looks all good. Are you scoping at the line out? I think the distortion you might be hearing may not be coming from the DAC. ...just a guess. You can check the P-to-P voltage (at 0 dB FS) and see if the voltage swing is a couple of volts away from the rail voltage. The AD devices should saturate pretty close to the rail, but safety margin is always good. You might want totry with good headphones (with a volume control!!) and if you hear the distortion at all play level, reverse the headphones (L>R, R>L) and see if the distortion is really in the content.
Really, All DACs and I/V converters will/must show clipping above -0 dB FS. The distortion is in the data.
Really, All DACs and I/V converters will/must show clipping above -0 dB FS. The distortion is in the data.
mrshow4u said:
The slight distortion is there at all levels, even a whisper when playing the 0db digital, and not there at all levels at -3db digital.
Yes scoping at the line out with 22k load, also listening from the line out with headphones, and it is equally there in both channels so reversing does nix
Cheers George
Hi Clem,
You suggested a very good idea with the triangular waveform. One we normally reach for on the bench for checking linearity. Good application. An FFT would show the slightest issue with a sine wave. An alternative would be to run the DAC output into a distortion analyzer. Connect your scope on the residual output and look at that.
Hi mrshow4u,
I agree with your points. The D / A can not clip, but it can be overloaded on it's output. Shouldn't be too much of a problem on a current output D/ A convertor. Is that ref pin accessible? What happens if it's loaded down?
Hi George,
I normally stay with the same family when changing op amps. So to replace a NE5532, stay with another bipolar type. Op amps are optimized for specific applications. There is no one "best" device for that reason. The NE5532 isn't that bad a chip either.
-Chris
You suggested a very good idea with the triangular waveform. One we normally reach for on the bench for checking linearity. Good application. An FFT would show the slightest issue with a sine wave. An alternative would be to run the DAC output into a distortion analyzer. Connect your scope on the residual output and look at that.
Hi mrshow4u,
I agree with your points. The D / A can not clip, but it can be overloaded on it's output. Shouldn't be too much of a problem on a current output D/ A convertor. Is that ref pin accessible? What happens if it's loaded down?
Hi George,
I normally stay with the same family when changing op amps. So to replace a NE5532, stay with another bipolar type. Op amps are optimized for specific applications. There is no one "best" device for that reason. The NE5532 isn't that bad a chip either.
-Chris
Hi Chris,
Thanks - I remember my brother telling me about the triangle wave and it makes sense...
Wouldn't a computer with FFT be able to pick up the distortion georgehifi is hearing? To be able to hear overtones would mean significant harmonics well visible above even a mediocre ADC in a sound card I think - specially at 1KHz..
But can a DAC run out of current? In what would appear to be the most 'logical' topology, where various resistors are used to produce the current, no, but that's not how DACs are built anyway, given it's not easy to balance currents over that wide a span.
Cheers!
Clem
Thanks - I remember my brother telling me about the triangle wave and it makes sense...
Wouldn't a computer with FFT be able to pick up the distortion georgehifi is hearing? To be able to hear overtones would mean significant harmonics well visible above even a mediocre ADC in a sound card I think - specially at 1KHz..
But can a DAC run out of current? In what would appear to be the most 'logical' topology, where various resistors are used to produce the current, no, but that's not how DACs are built anyway, given it's not easy to balance currents over that wide a span.
Cheers!
Clem
Hi Clem,
George, it may be labeled as a mic jack. Please have another look. Soundcards are not expensive either. The Soundblaster Live! 24 bit is very good sounding but limited to about 19 KHz. Enough for you. They are around $35 CDN around here.
-Chris
Yes. It would stand out like a sore thumb. Unfortunately his computer has no line in jack.
George, it may be labeled as a mic jack. Please have another look. Soundcards are not expensive either. The Soundblaster Live! 24 bit is very good sounding but limited to about 19 KHz. Enough for you. They are around $35 CDN around here.
D/A chips are amazing these days. Such high accuracy. Current or voltage out is about the same work. Most people go with current out so they can use whatever op amp they want, or go with a resistor and tube. Whatever they want. On board op amps have not received good reviews.
-Chris
Ok, I have an extra tek distortion analyzer module - designed to plug into one of those tek housings with a programmable power supply. Problem with it the pcb that extends into the housing cracked off, which is why I got it cheap. It still works, I ran it off an external, fixed voltage(s) supply. Anyone want it? You pay for S/H...
Getting to swap it for a dead TDS-210 would make me even happier... (I have a tds210 with a bad LCD)...
Cheers!
ps: It requires something like +5, -12 and +12 if I recall.. Easily achieved from a PC power supply...
Getting to swap it for a dead TDS-210 would make me even happier... (I have a tds210 with a bad LCD)...
Cheers!
ps: It requires something like +5, -12 and +12 if I recall.. Easily achieved from a PC power supply...
clem_o said:
I'm back, soooo much info, i'm getting cranial overload.
Clem_o that thought crossed my mind also, fets are harder to drive because of their input capacitance, I wonder what the output stage of the D/A is like, if it's weedy it may have trouble, in the next couple of days I'll try the NE5532's back in and see what happens, if it works (bummer) there goes my direct coupling as they have 20 x the offset dc of the AD825's. not easy to trim out.
Time to go for a surf and forget about 1's and 0's for a while, and think about 6ft glassy barrels.
Cheers George
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.