Hi,
4mVpk from a first test selection using a 0.25mV cartrdige.
Now substitute a 0.5mV cart and we are up to 8mVpk. That's 30dB above the starting reference for 5cm/sec.
Now search through all the recordings for worse.
Add in some pops and crackles.
Aim for 35dB headroom.
Let the electronics follow the (bad) signal accurately and return from the peak as quickly as possible before feeding the next stage.
Gerhard,
listen at a maximum average level at least 20dB below the power amp's maximum capability to allow average to peak overhead.
Some recommend -30dB ratio for average to maximum levels on very dynamic material.
The cleaner the signal when it gets to the speaker the less noticeable are the clicks and pops that are inevitable.
It's very similar with the cart tracing the Lump stuck in the groove. If the cart can follow the lump and not bounce (mistrack) the cleaner sound of the better tracking cartridge seems to make the clicks and pops less noticeable.
4mVpk from a first test selection using a 0.25mV cartrdige.
Now substitute a 0.5mV cart and we are up to 8mVpk. That's 30dB above the starting reference for 5cm/sec.
Now search through all the recordings for worse.
Add in some pops and crackles.
Aim for 35dB headroom.
Let the electronics follow the (bad) signal accurately and return from the peak as quickly as possible before feeding the next stage.
Gerhard,
listen at a maximum average level at least 20dB below the power amp's maximum capability to allow average to peak overhead.
Some recommend -30dB ratio for average to maximum levels on very dynamic material.
The cleaner the signal when it gets to the speaker the less noticeable are the clicks and pops that are inevitable.
It's very similar with the cart tracing the Lump stuck in the groove. If the cart can follow the lump and not bounce (mistrack) the cleaner sound of the better tracking cartridge seems to make the clicks and pops less noticeable.
AndrewT said:
Rather than designing for insane headrooms (again, I have not seen any spec that high, Pass Xono (SS), Manley Steelhead (tubes) etc... are all well under this) there is a much simpler solution: lower the gain. If the purpose is to keep the pops and clicks in the linear region of the amp, then simply reduce the gain; you'll overload less the whole chain and the sound will improve.
Really, while I understand the need for a decent headroom, when it comes to extremes, I'm with Gerhard on this one. Assume the example above, a 0.25mV cartridge, 60dB gain and 36dB overhead. Under these circumstances, the MC preamp has to support an output of 15.8V before clipping, current limitations, etc...
Ok, we got this great MC preamp, now what? Should we assume that clipping/overloading the line preamp and/or the power amp is anywhere better than clipping/overloading in the MC preamp? Ot should we design the entire chain to support 36dB headroom (which is the equivalent of building a 1000W power amp and listening at 250mW level), just for avoiding the clicks and pops to overload the chain?
syn08 said:
Indeed, keep the perspective. Or switch to CDs... (janneman can be seen ducking and donning a bullet proof vest) 😀
Jan Didden
You have to take into account very, very different output voltage levels (per cm/s) of the MC cartridges.
janneman said:
Not CDs, of course, but the best digital recordings I've ever seen (available online from Linn) have almost 90dB dynamic. Designing a system (including room and speakers) to support such a dynamic is as well not easy.
I'd be looking for 30 to 35dB overhead throughout the electronics chain, until the power amps. There I reduce the overhead to around 20 to 26dB.
170W+170W into a pair of 92dB/W/m speakers gives a peak ability of around 110dB at an 8foot (2.4m) listening distance.
Playing very loud (average level 20dB below peak) results in 90dB average SPL.
Playing at something sensible, say -30dB below maximum peak allows upto 80db SPL of average level.
That's all easily achievable (and affordable) and allows for 30dB transients.
The system never sounds strained and the vinyl replay can be exemplary with regard to tracking the non audio artifacts.
100W into 90dB speakers reduces the transient to average overhead by ~5dB. The attenuator is there for a reason, learn to use it.
edit:
90dB of dynamic range with peaks at 110dB leaves the quiet end @ ~20dB SPL. You need a quiet listening room to gain full advantage.
170W+170W into a pair of 92dB/W/m speakers gives a peak ability of around 110dB at an 8foot (2.4m) listening distance.
Playing very loud (average level 20dB below peak) results in 90dB average SPL.
Playing at something sensible, say -30dB below maximum peak allows upto 80db SPL of average level.
That's all easily achievable (and affordable) and allows for 30dB transients.
The system never sounds strained and the vinyl replay can be exemplary with regard to tracking the non audio artifacts.
100W into 90dB speakers reduces the transient to average overhead by ~5dB. The attenuator is there for a reason, learn to use it.
edit:
90dB of dynamic range with peaks at 110dB leaves the quiet end @ ~20dB SPL. You need a quiet listening room to gain full advantage.
AndrewT said:
Agreed, 20dB headroom in the chain makes sense, but why doing this in the power amp?
Perhaps a few more details will keep this in better perspective-
1) The measured output of the cartridge is essentially integrated so the first issue is input overload of the first stage.
2) I know there are people who live life at 11, but most of us have a volume control in the chain- and for the control to have utility it has attenuation at normal operating level. Lately, because of highly compressed levels on CD's, I have seen a trend to having the volume control set at 20% of its full range (possibly 30 dB of attenuation). This is why, I think, the gain requirements for MC inputs have increased over the years even though the cartridges have more output.
The output of a CD playback chain is listed as 2V, really meaning that FS= 2.0 Vrms. That would be 20 dB of headroom from a 200 mV reference level. But CD player output tends to be higher in audiophile land (easy way to be perceived as better) and DAW's have better compression algorithms so the average to peak on CD's has been going down.
Records, for the most part, are relics of the past, with all analog chains in front of them and much higher average to peak ratios (crest factor) so the headroom required for a record at the same perceived loudness can be much higher.
1) The measured output of the cartridge is essentially integrated so the first issue is input overload of the first stage.
2) I know there are people who live life at 11, but most of us have a volume control in the chain- and for the control to have utility it has attenuation at normal operating level. Lately, because of highly compressed levels on CD's, I have seen a trend to having the volume control set at 20% of its full range (possibly 30 dB of attenuation). This is why, I think, the gain requirements for MC inputs have increased over the years even though the cartridges have more output.
The output of a CD playback chain is listed as 2V, really meaning that FS= 2.0 Vrms. That would be 20 dB of headroom from a 200 mV reference level. But CD player output tends to be higher in audiophile land (easy way to be perceived as better) and DAW's have better compression algorithms so the average to peak on CD's has been going down.
Records, for the most part, are relics of the past, with all analog chains in front of them and much higher average to peak ratios (crest factor) so the headroom required for a record at the same perceived loudness can be much higher.
Syn_08
Just a suggestion for your phono amp. This has nothing to do with the argument of whether or not the AD797 can drive the FB network.
Use a buffer driven from pin 5 of the 797. You could even put a small constant current load on the normal o/p (pin 7) of the 797. The potential advantages are that you remove any thermal changes on the 797 chip from the o/p as it drives low Z loads and you can now more easily use very low noise power supplies for the FETs and 797, with higher current and voltage supplies for the buffer and subsequent stages.
And an off-the-wall thought since you evidently have had good success with the Hawksford/Cordell EC output stages: adapt something along those lines with discretes for the buffer. 🙂 Not what I would do but...................
Phil
Just a suggestion for your phono amp. This has nothing to do with the argument of whether or not the AD797 can drive the FB network.
Use a buffer driven from pin 5 of the 797. You could even put a small constant current load on the normal o/p (pin 7) of the 797. The potential advantages are that you remove any thermal changes on the 797 chip from the o/p as it drives low Z loads and you can now more easily use very low noise power supplies for the FETs and 797, with higher current and voltage supplies for the buffer and subsequent stages.
And an off-the-wall thought since you evidently have had good success with the Hawksford/Cordell EC output stages: adapt something along those lines with discretes for the buffer. 🙂 Not what I would do but...................
Phil
john curl said:
Sure. Try it from a different point of view.
5Vp, 20kHz
SR = 2 x pi x f x Vp = 0.628V/us
This voltage trying to drive 1.8uF cap.
PH104 said:
Thanks.
I have though another option on the drawing board, much better I guess... Buffers are usually driving 100mA, exactly as much as AD797 is able to. There are a few high current buffers (like OPA634) but I'm not sure if it would be any much better.
- Status
- Not open for further replies.