Considering placing an Lpad on the RCA outputs coming from a Rega Dac (2.2V / 620Ω) into a Luxman L-505u line level input (.18V / 42kΩ), but not sure what resistor values / preparations might work best for this case. Often see starting with 10kΩ on the source and solving for ground--about 930Ω. Does this make sense?
10k in series, 1k to ground which will be in parallel with the 42k = 977, 977/(10k+620+977)=0.084, times 2.2V = 0.185V, close enough I guess.
Cheers,
Cabirio
Cheers,
Cabirio
Thanks cabirio. With this particular case, would there be any reason (or not) to experiment with a lower (5kΩ/470Ω) or higher (20kΩ/1875Ω) L-pad? For example, does that >1:10 source-to-amp load impedance matching ratio I see tossed around factor into this at all?
You could go lower assuming the source does not break a sweat (and measured results for the Rega DAC seem to indicate that there is little difference even down at 600 ohms), keeping down the source impedance seen by the input.
A >20 dB attenuator is a bit too much though. (That would mean maximum output would be just enough to get full power output of the amp with the volume cranked all the way up.) Assuming you're not on unusually sensitive speakers (but rather the usual 85-90 dB/W/m fare), you only need to "trim the fat" that CDs have gained in the last 25 years or so, which is 10-12 dB tops. Try anything from 3k3/1k to 6k8/2k2. How's your volume control set right now? Something like 9 o'clock?
A >20 dB attenuator is a bit too much though. (That would mean maximum output would be just enough to get full power output of the amp with the volume cranked all the way up.) Assuming you're not on unusually sensitive speakers (but rather the usual 85-90 dB/W/m fare), you only need to "trim the fat" that CDs have gained in the last 25 years or so, which is 10-12 dB tops. Try anything from 3k3/1k to 6k8/2k2. How's your volume control set right now? Something like 9 o'clock?
sgrossklass, the first attempt at this was a pair of Harrison Labs 12dB attenuators, which only moved the listening sweet spot on the volume pot from about 8:30 to 10:00. So figured should try something greater and not necessarily spend money on what seem like pretty simple devices. Speakers are Harbeth C7s rated at 86dB/1W/1m.
2.2Vac is the maximum output of the DAC @ 0dBfs
What is the 0.18Vac value?
Is it the signal required to drive the Receiver to clipping?
or, is it the signal required to give an output equivalent to 1W into 8ohms speaker/s?
0.18Vac input to give 2.83Vac output would be a reasonable sensitivity and indicates a gain of ~15.7times (+24dB)
0.18Vac input to give 20Vac (50W into 8ohms) output would be a very sensitive input and indicates a gain of ~111times (+41dB)
If you maximum output is 20Vac and your maximum input is 2.2Vac then you only need a gain of 9.1times (19.2dB)
Subtract 19.2 from your Receiver gain and you arrive at the attenuation you require to get full use of your vol pot for that DAC.
What is the 0.18Vac value?
Is it the signal required to drive the Receiver to clipping?
or, is it the signal required to give an output equivalent to 1W into 8ohms speaker/s?
0.18Vac input to give 2.83Vac output would be a reasonable sensitivity and indicates a gain of ~15.7times (+24dB)
0.18Vac input to give 20Vac (50W into 8ohms) output would be a very sensitive input and indicates a gain of ~111times (+41dB)
If you maximum output is 20Vac and your maximum input is 2.2Vac then you only need a gain of 9.1times (19.2dB)
Subtract 19.2 from your Receiver gain and you arrive at the attenuation you require to get full use of your vol pot for that DAC.
Last edited:
Andrew, I have no more information other than 180mV is the quoted input sensitivity from the manual. I've assumed that's the minimum required to drive the amp to full power--100W into 8ohms?
Which is about what I'd have expected - and usually quite enough to fix super-sensitive volume-pot syndrome and channel imbalance. If you go up higher, you'll eventually be confronted with increasing amplifier hiss levels and potentially worsening distortion as well.
What are you feeding the DAC from, a CD player? If you were using a computer with a ReplayGain capable player (like, say, Foobar2000, which also includes a RG scanner), you could use that to provide the necessary attenuation - and enjoy sensible volume levels across your entire collection to boot. Pretty big deal to me, as I have everything from early-'80s CD material to modern-day output, and being able to switch from one to the other without having to fear for your ears is pretty nifty. Plus, you can dial in extra pre-gain or pre-cut to have everything covered - it's -3.2 dB here, and Vangelis' Opera Sauvage is still giving me a hard time after deemphasis (I'd need almost -6 dB for the first track, and that's just ridiculous when the vast majority of your collection isn't anywhere close, so on very dynamic outliers like that I'll just accept the very occasional clipping). Combined attenuation on a modern-day release would be around -11..-14 dB here (as opposed to +3.69 dB for Opera Sauvage), which is just about the right ballpark for avoiding issues with classic hi-fi gear.
A PC-based music collection has a few gotchas to offer, such as proper treatment of CD material with pre-emphasis. (EAC will at least show whether tracks are noted to have PE in the TOC though it's the very last column, CUERipper will even look at subcode, and iTunes is one of the few rippers to apply deemphasis automagically. There's a gazillion ways of applying it manually, like standalone via SoX or via a Foobar plugin.) Thankfully these are not too common after about 1985, and I only have a handful myself.
That is even higher gain that I gave in post
100W into 8ohms is equivalent to 28284mVac
The gain would need to be 157times (+43.9dB) to get from 180mVac to 100W
Your 2.2Vac source only needs +22.2dB of gain to reach that 100W.
A fixed attenuator of -21.7dB would get you to using the full range of your vol pot for that 2.2Vac source.
A divide by 10, or divide by 12, attenuator is close enough.
10k series resistor with 1k shunt resistor gives a divide by 11 (=1k/{10k+1k})
- Status
- Not open for further replies.