Lampy cd player
#46
It may also help matters if you pict a test frequency thats a whole ratio of the sample freq, it will make what the output stage is doing easyier to see.
Whenever an honest man discovers that he's mistaken, he will either cease to be mistaken or he will cease to be honest.
#47
Thanks Nick.
I reinstated 4x OS and it scopes a lot better but starts to show that similar waveform after 5KHz.
By 20KHz it was looking like this again,
Now this has got me wondering what's going on as the only mods are larger caps on the 1541 decoupling pins and the analogue output straight into the valve stage.
The standard machine appears completely clean.
Also checked both chanels and both are identical with and without NOS.
If it was interference I'd expect it to be different across the channels. So it must be more than just NOS I suppose.
How happy is a 1541 working into 82R? Would it be better with a lower IV R do you think?
I reinstated 4x OS and it scopes a lot better but starts to show that similar waveform after 5KHz.
By 20KHz it was looking like this again,
Now this has got me wondering what's going on as the only mods are larger caps on the 1541 decoupling pins and the analogue output straight into the valve stage.
The standard machine appears completely clean.
Also checked both chanels and both are identical with and without NOS.
If it was interference I'd expect it to be different across the channels. So it must be more than just NOS I suppose.
How happy is a 1541 working into 82R? Would it be better with a lower IV R do you think?
#48
Have you seen these similar traces before then Nick?Nick wrote:As long as you don't look at them on a scope, NOS 1541 without a reconstruction filter should sound good. You may prefer others, or with the OS and or reconstriction filtering. But IMHO the sound you are describing is not down to the DAC.
I played other games with it (driving a pair of 1541 differentially via a CPLD) but if you have them on hand, try cap coupling the I/V resistor via a set of 1:10 MC step up transformers to give you voltage gain. You may be supprised how it sounds. (just to get a sense of what the DAC is doing).
#49
HI Nick, the frequencies are just those on the test cd;Nick wrote:It may also help matters if you pict a test frequency thats a whole ratio of the sample freq, it will make what the output stage is doing easyier to see.
2, 4, 8, 17, 20, 31.5, 50, 100, 200, 500, 1k, 2k, 5k, 10k, 15k, 20k,
I can show any ofr those if it will help.
- Dave the bass
- Amstrad Tower of Power
- Posts: 12276
- Joined: Tue May 22, 2007 4:36 pm
- Location: NW Kent, Darn Sarf innit.
#50
I was going to ask that question too!Richard wrote:
Have you seen these similar traces before then Nick?
It's very pretty but not right natch. It looks like a repetition of the original of the signal but delayed a tiny bit x a dozen times.
DTB
"The fat bourgeois and his doppelganger"
#51
Its what I would expect with no reconstruction filter. If the frequency was a sub multiple of 44.1khz, then you would see the clear steps, one at each sample period. But because the scope is locking on the 1khz (or whatever) then the 44100 steps a second happen at different parts of the 1kHz sine wave, so you see the blur of each pass of the dot on the scope traces a different path. If you had a storage scope and just took one sample you would see the staircase that is the different value every 1/44100th of a second.
Whenever an honest man discovers that he's mistaken, he will either cease to be mistaken or he will cease to be honest.
-
- Shed dweller
- Posts: 2300
- Joined: Fri Aug 03, 2007 10:57 am
- Location: From the land of the Bodgers
#52
Yes, agree with that, the traces are due to the lack of filtering and similar to the ones I saw on my DAC. Even though NOS can sound good I found that the timbre of instruments could be affected, listening to a test tone NOS vs OS with filter you could hear how the NOS tone was different. It's still a good DAC though and I can see why people like it....oops thread drift, I'll stop there!
#53
Get a copy of Audacity and create a 0dB 882Hz test tone, burn to CD and use that for the tests, you will see what I mean.
Whenever an honest man discovers that he's mistaken, he will either cease to be mistaken or he will cease to be honest.
#54
Hi Nick, so does your dac trace look like that, or are there ways around it, or is it a different chip/filter?
Hi Neal, yes, when I did the NOS mod it was attractive in some ways but I didn't think it sounded "right". What I have here is a musical instrument How did you end up, was it a worthwhile trip or just an experience? Trying to decide what to do next.
Hi Neal, yes, when I did the NOS mod it was attractive in some ways but I didn't think it sounded "right". What I have here is a musical instrument How did you end up, was it a worthwhile trip or just an experience? Trying to decide what to do next.
#55
If you remove the reconstruction filter, all DAC will look like that. If the sample rate is higher, you have more steps in a given time, but thats what DAC's do.Hi Nick, so does your dac trace look like that, or are there ways around it, or is it a different chip/filter?
A 1kHz tone and a 44.1k samples/sec sample rate, means you have 44.1 samples in the space of a single 1kHz cycle.
But remember, we rarely build kit to look at the output on a scope. Unless your amp is flat to 44k, it won;t look like that by the time it gets to the speaker. And unless your speaker is flat to 44k it wont be like that by the time it hits your ears. And I know my ears are not flat at 44kHz.
Whenever an honest man discovers that he's mistaken, he will either cease to be mistaken or he will cease to be honest.
- Mike H
- Amstrad Tower of Power
- Posts: 20189
- Joined: Sat Oct 04, 2008 5:38 pm
- Location: The Fens
- Contact:
#56
Throwing my oar in only to mention ECC83 wouldn't have been my first choice either. I would've thought its gain was too high for that. Not that I know a lot about it
"No matter how fast light travels it finds that the darkness has always got there first, and is waiting for it."
-
- Shed dweller
- Posts: 2300
- Joined: Fri Aug 03, 2007 10:57 am
- Location: From the land of the Bodgers
#57
I ended up buying a Beresford Caimen DAC! The best I got from my Arcam TDA board was with a 3-pole filter before the output opamp, can't recall what filter shape I ended up using (will check tomorrow and find the cct) but it did help clean the output up a bit. It's catch 22 though as Nick pointed out, you have to move the sampling freq out from the audio band for any filter to work with minimum impact, simple passive filters have too many issues and are not steep enough in profile for NOS so you have to OS.Richard wrote:
Hi Neal, yes, when I did the NOS mod it was attractive in some ways but I didn't think it sounded "right". What I have here is a musical instrument How did you end up, was it a worthwhile trip or just an experience? Trying to decide what to do next.
Having said that the Arcam DAC sounded dam fine all the same but I always thought I could do better and the little Beresford DAC proved that.
#59
Hi Neal, thanks, very kind, it would be good to meet up again one day and catch up
Hi Nick and all, thanks for the explanations so far. I think I can see what's happening now, does this sound correct;
the first 4 traces I posted were NOS with no reconstruction filter. I'd bypassed the 7220 4x chip before the dac and the valve stage has no deliberate low pass. Reinstating 4x allowed what "natural" filtering exists in the circuit/load to effectively be (4x?) more effective and so the second 20KHz trace looks (4x?) better than the first 20KHz one but is still not clean.
There is no "proper" reconstruction low pass as that was lost with the opamp stage. From putting the values into a calc it was a sallen key 2 pole at 44.1KHz on the second opamp stage. The irritating noise I gather may be aliasing. Certainly some highs seemed to take on a mind of their own.
Tomorrow I'll try Thorsten's for 100R IV with the low pass (-0.7dB 20KHz) on the input,
http://www.fortunecity.com/rivendell/xe ... ltibit.gif
Hi Nick and all, thanks for the explanations so far. I think I can see what's happening now, does this sound correct;
the first 4 traces I posted were NOS with no reconstruction filter. I'd bypassed the 7220 4x chip before the dac and the valve stage has no deliberate low pass. Reinstating 4x allowed what "natural" filtering exists in the circuit/load to effectively be (4x?) more effective and so the second 20KHz trace looks (4x?) better than the first 20KHz one but is still not clean.
There is no "proper" reconstruction low pass as that was lost with the opamp stage. From putting the values into a calc it was a sallen key 2 pole at 44.1KHz on the second opamp stage. The irritating noise I gather may be aliasing. Certainly some highs seemed to take on a mind of their own.
Tomorrow I'll try Thorsten's for 100R IV with the low pass (-0.7dB 20KHz) on the input,
http://www.fortunecity.com/rivendell/xe ... ltibit.gif
#60
Yes. Though I will repeat again, what you see on the scope is not real, whats real looks a lot less like noise.
Whenever an honest man discovers that he's mistaken, he will either cease to be mistaken or he will cease to be honest.