SNR, THD+N, Dynamic Range and Impedance

[Johnseye]

I'm in the market for a new preamp and I'm trying to get an understanding of the following specs and how to compare them between devices. Signal to noise ratio or SNR, Total harmonic distortion plus noise or THD+N, Dynamic Range and impedance. I think I get impedance, both input and output but not sure if I'm understanding how to compare. Here are specs provided in different formats which makes it even more difficult to compare. The first two are preamps and the third is an amp. In addition to understanding how to interpret the specs, any opinions of them as they relate to the device is appreciated.

I've listed the specs as they're stated by the vendor.

Preamp #1

Noise: less than -100dBv at max volume
THD+Noise: Less than .007% at 2V RMS BAL output
Dynamic Range: 122dB
Input Impedance: 100k ohms balanced, 50k ohms single ended
Output Impedance: 600 ohms balanced, 300 ohms single ended, 20k ohms minimum load and 2000pF maximum capacitance

Preamp #2

The noise floor with a 0.5 mV cartridge playing a reference level is about 70db. It, and the THD is a bit variable due to the tubes themselves. THD is usually no more than about 0.05% at full output which is about 20 volts.
Line stage input impedance: 200 Kohm balanced, 100Kohm single ended
Output imprdance: 250 Ohms from 2Hz to 400KHz
Phono stage input impedance: 47 K ohms adjustable

My amplifier - Benchmark AHB2

1kHz <0.0003% THD+N
THD+N: 1kHz, 80kHz LPF, <-118dB (<0.00013%)
THD: 1kHz, 20kHz LPF, <-119dB (<0.00011%)
SNR & Dynamic Range: 132 dB A-weighted, 130dB unweighted, 20 Hz to 20kHz
Input Impedance: 1M Ohm common mode, 50 k Ohms normal mode

 

[Mivera]

There's no way that Benchmark amp could measure that good. Look how close the analog outs are to the mains input!

 

[Blumlein 88]

Well let's start with SNR and dynamic range. They seem almost like the same thing, but aren't quite.

Let us suppose you have a device with a 1 volt maximum output. Let us say you are using it to put out 1 khz at 1 volt. Now remove the signal and measure the level of noise across that band. In this case it is the noise across 0hz to 20 khz or the audible band. If that were a flat white noise floor at 31.6 microvolts that would be -90 db below 1 volt or -90 db V (relative to 1 volt) That would be signal to noise ratio of 90 db.

Now let us suppose the same situation, but the device has the same white noise floor at -90 db plus some 60 hz hum at -87 db. The noise level would be raised to roughly -85 db when the hum and white noise is combined. The SNR would therefore be about 85 db. However the dynamic range is the highest signal level minus the highest single component in the noise floor. In this case the -87 db hum at 60 hz. So the SNR is 85 db, but the dynamic range is 87 db.

 

[Blumlein 88]

Now THD+N. Let us say you have same device with a -90 db white noise floor. So you put a 1 volt 1khz tone thru it and measure the result. If you notched out the 1 khz tone what is left is noise and distortion. If your device has THD of -90 db then it would combine with the noise and a meter reading that 20 khz band would read -87 db relative to the 1 volt input. In this case we know the noise is -90 dbV. If we didn't know that we could only say it has THD+N of -87 dbV. Some of it could be noise, or hum or distortion.

So what happens if your device is very low distortion like -120 db. We use a meter to read the result with the 1 khz fundamental notched out and read about -90 dbV. Because the distortion contribution is so much below the noise floor is makes little difference in what the meter reads. So THD+N would be -90 dbV even though the THD part is much, much lower than this. That is why on amplifier THD vs power level you sometimes see the THD+N rise at very low power levels below 1 watt. The THD is likely way below the noise floor, but you are mostly reading noise.

 

[Blumlein 88]

What if we have hum from the mains going into the analog outs? What if that Bent faston connection snaps off and shorts out somewhere? Look at those live terminals 1/8" away from the right analog outs on the PCB. What if there's a voltage surge and it arcs to the analog out? How about that crooked chip with the silicone blobbed on it? Is that how they get the dynamic range?

You should at a minimum start your own thread about the pictured amp. Clearly you aren't trying to contribute usefully to the topic of the thread. As this is recurring behaviour were I the moderator you would have just earned a minimum 30 day ban. Similar activity in the future would make it permanent as it indicates you aren't here to be helpful.

 

[Blumlein 88]

Input and output impedances. These are roughly like resistance only with AC signals. So other factors like capacitance and inductance might mean a different resistance to current flow at different frequencies. If something says input impedance of 100 k ohms you would expect it to have the resistance to current flow across the 20 khz bandwidth or other bandwidth specified.

In general lower output impedances are better and higher input impedances are better. Though that is simplifying it some. Usually with audio you want a 20 to 1 ratio between output impedance and input impedance of the following device.

 

[amirm]

I think I get impedance, both input and output but not sure if I'm understanding how to compare.

A good rule of thumb is to look at the input impedance of your power amplifier, and make sure the output impedance of your preamp is not more than 10% of it. This way the impact of this parameter essentially goes away.

In your case, the Benchmark amp has a 50K input impedance, which means that you want the pre-amp output impedance to be less than 5000 ohm. Both of your pre-amp qualify in this regard so you are good there.

 

[amirm]

In general lower output impedances are better and higher input impedances are better. Though that is simplifying it some. Usually with audio you want a 20 to 1 ratio between output impedance and input impedance of the following device.

Oops, you had beat me to it.

 

[amirm]

THD+Noise: Less than .007% at 2V RMS BAL output

Unfortunately THD+N figures are not useful. They give you a sum of distortion products, not what they are. You want to see the spectrum of the distortion in order to understand audibility. This is a quick example from stereophile review:

We see that the spikes are all at -100 db so clearly inaudible.

Had we had higher first harmonic than others we could have dismissed that fact because the early harmonics get masked more than the later one.

That aside, numbers such as you list are very low and very doubtful that the distortion is audible.

 

[Johnseye]

Looks like some of Minerva's posts were moderated. Much appreciated. He still has one of the Benchmark amp. I don't mind the picture, but his comment isn't helpful. Also, Ray had a post that was deleted. He asked what my sources are.

Sources:

DAC: Audio Alchemy DDP-1
Nominal THD+N (1 volt) <0.001% at 10 volts output from analog input, 0.009% at 7 volts from XLR output, 3.5 volts from RCA output
SNR: (1kHz) 103 dB unweighted
Input impedance: 50 k ohms
Output impedance: 75 ohms

VPI Prime with Ortofon Quintet Black cart
Output voltage at 1000 Hz, 5cm/sec -0.3 mV
Internal impedance, DC resistance -5 Ohm
Recommended load impedance - > 20 Ohm


Thank you for the helpful info @Blumlein 88 . I'll be re-reading it a few times to process.

 

[Johnseye]

Unfortunately THD+N figures are not useful. They give you a sum of distortion products, not what they are. You want to see the spectrum of the distortion in order to understand audibility. This is a quick example from stereophile review:

We see that the spikes are all at -100 db so clearly inaudible.

Had we had higher first harmonic than others we could have dismissed that fact because the early harmonics get masked more than the later one.

That aside, numbers such as you list are very low and very doubtful that the distortion is audible.

Based on the specs of the two pre's how can I discern which one is better, or is it more a question of which one matches my amp better? Maybe the answer is in Blumlein's responses and I'll try to take that info to answer the question, but this is new to me.


Stereophile measured the Benchmark. That's what convinced me to buy the amp.

https://www.stereophile.com/content/benchmark-media-systems-ahb2-power-amplifier-measurements

Fig.9 Benchmark AHB2, left channel, spectrum of 50Hz sinewave at 50W into 8 ohms (blue) and spectrum of 50Hz sinewave with Audio Precision SYS2722 analyzer looped back to generator at 20V (red) (linear frequency scale, DC–1kHz).
Read more at https://www.stereophile.com/content...er-amplifier-measurements#YE2khmbMJ7akGRl1.99

 

[amirm]

Based on the specs of the two pre's how can I discern which one is better, or is it more a question of which one matches my amp better? Maybe the answer is in Blumlein's responses and I'll try to take that info to answer the question, but this is new to me.

Matching is trivial today if you are dealing with solid state devices. Only tube amps can present issues. So in general, this is not an important consideration.

 

[amirm]

Stereophile measured the Benchmark. That's what convinced me to buy the amp.

https://www.stereophile.com/content/benchmark-media-systems-ahb2-power-amplifier-measurements

Great. Now if you have that for the two preamps you are considering, then we can analyze them and figure out which one is better. This is considerably more useful than just having the THD number.

 

[RayDunzl]

Also, Ray had a post that was deleted. He asked what my sources are.

I deleted that myself after seeing your interest in a phono input. I didn't read closely enough before opening my mouth.

That (along with, assumedly, digits) pretty much covers your sources for me.

 

[Blumlein 88]

Based on the specs of the two pre's how can I discern which one is better, or is it more a question of which one matches my amp better? Maybe the answer is in Blumlein's responses and I'll try to take that info to answer the question, but this is new to me.


Stereophile measured the Benchmark. That's what convinced me to buy the amp.

https://www.stereophile.com/content/benchmark-media-systems-ahb2-power-amplifier-measurements

Fig.9 Benchmark AHB2, left channel, spectrum of 50Hz sinewave at 50W into 8 ohms (blue) and spectrum of 50Hz sinewave with Audio Precision SYS2722 analyzer looped back to generator at 20V (red) (linear frequency scale, DC–1kHz).
Read more at https://www.stereophile.com/content...er-amplifier-measurements#YE2khmbMJ7akGRl1.99

"Conventionally it is assumed that the measurement system used to test a product is much better than the device under test. However, the Benchmark amplifier's specifications are close to those of my Audio Precision system, both regarding the purity of its signal generator and the dynamic range of its analyzer."

Yes, the Benchmark is pushing the limits of what is possible in a few areas.

 

[DonH56]

I'm in the market for a new preamp and I'm trying to get an understanding of the following specs and how to compare them between devices. Signal to noise ratio or SNR, Total harmonic distortion plus noise or THD+N, Dynamic Range and impedance. I think I get impedance, both input and output but not sure if I'm understanding how to compare. Here are specs provided in different formats which makes it even more difficult to compare. The first two are preamps and the third is an amp. In addition to understanding how to interpret the specs, any opinions of them as they relate to the device is appreciated.

I've listed the specs as they're stated by the vendor.

Preamp #1

Noise: less than -100dBv at max volume
THD+Noise: Less than .007% at 2V RMS BAL output
Dynamic Range: 122dB
Input Impedance: 100k ohms balanced, 50k ohms single ended
Output Impedance: 600 ohms balanced, 300 ohms single ended, 20k ohms minimum load and 2000pF maximum capacitance

Preamp #2

The noise floor with a 0.5 mV cartridge playing a reference level is about 70db. It, and the THD is a bit variable due to the tubes themselves. THD is usually no more than about 0.05% at full output which is about 20 volts.
Line stage input impedance: 200 Kohm balanced, 100Kohm single ended
Output imprdance: 250 Ohms from 2Hz to 400KHz
Phono stage input impedance: 47 K ohms adjustable

My amplifier - Benchmark AHB2

1kHz <0.0003% THD+N
THD+N: 1kHz, 80kHz LPF, <-118dB (<0.00013%)
THD: 1kHz, 20kHz LPF, <-119dB (<0.00011%)
SNR & Dynamic Range: 132 dB A-weighted, 130dB unweighted, 20 Hz to 20kHz
Input Impedance: 1M Ohm common mode, 50 k Ohms normal mode

You've already potten good advice. I have some info but it's on another site and no better than you've already received.

A few comments on the preamp specs:

• -100 dBV is not all that great for a preamp these days but good enough that as long as you do not have highly sensitive speakers it won't be audible. The missing information is if this is unweighted or weighted, and if the latter what weighting is applied. Weighting rolls off high and low frequencies and is a big player in the final noise (and SNR) numbers.
• Unless both are for the phono stage, the noise numbers are not comparable. Phono stages add a lot of gain and may not have great noise specs because, well, neither do records!
• 122 dB dynamic range is good, but there is no comparable spec for the second preamp.
• Input and output impedance are fine for both, a wash. With 1 M-ohm input impedance neither preamp will have any problem driving you amplifier.
• Preamp 2's 20 V output is very high, about 10x what most amps need for full-scale output. Thus, while likely meaning low distortion at more typical 1-2 V output, it also means you are giving up some in noise and dynamic since you more than likely can't take advantage of a 20 V output.

Bottom line is just the specs provided are not enough for me to say which is better.

The specs on the Benchmark amp are great, BTW.

HTH - Don

 

[Johnseye]

Matching is trivial today if you are dealing with solid state devices. Only tube amps can present issues. So in general, this is not an important consideration.

The pre's are both tube.

Pre 1 is an Audio Research LS28. I haven't found any spectrums on it, but it's better brother is the Ref 6 and there are spectrums for it.

https://www.stereophile.com/content/audio-research-reference-6-line-preamplifier-measurements

Fig.7 Audio Research Reference 6, balanced spectrum of 50Hz sinewave, DC–1kHz, at 2V into 10k ohms (left channel red) (linear frequency scale).
Read more at https://www.stereophile.com/content...preamplifier-measurements#xmHCPaCCJChCKDez.99



Pre 2 is an Atma-Sphere MP-3. I haven't found any spectrum for it.

 

[Johnseye]

You've already potten good advice. I have some info but it's on another site and no better than you've already received.

A few comments on the preamp specs:

• -100 dBV is not all that great for a preamp these days but good enough that as long as you do not have highly sensitive speakers it won't be audible. The missing information is if this is unweighted or weighted, and if the latter what weighting is applied. Weighting rolls off high and low frequencies and is a big player in the final noise (and SNR) numbers.
• Unless both are for the phono stage, the noise numbers are not comparable. Phono stages add a lot of gain and may not have great noise specs because, well, neither do records!
• 122 dB dynamic range is good, but there is no comparable spec for the second preamp.
• Input and output impedance are fine for both, a wash. With 1 M-ohm input impedance neither preamp will have any problem driving you amplifier.
• Preamp 2's 20 V output is very high, about 10x what most amps need for full-scale output. Thus, while likely meaning low distortion at more typical 1-2 V output, it also means you are giving up some in noise and dynamic since you more than likely can't take advantage of a 20 V output.

Bottom line is just the specs provided are not enough for me to say which is better.

The specs on the Benchmark amp are great, BTW.

HTH - Don

Thanks Don. Other than stating the pre's are the Audio Research LS28 and the Atma-Sphere MP-3, what information can I provide that can help determine which is better?

It's also worth noting that the LS28 is without a phono stage. We can leave that piece out of it and evaluate the line stages of both only.

 

[DonH56]

Uh, how they sound? My guess is either would be a good choice. One advantage most tube pres have over most SS preamps for phono inputs is the tubes have very high voltage overload so better handle the dynamic range of ticks and pops of records without clipping and splattering harmonics.

Tube components are notoriously under-specified though these days tend to do better. Let me get back to you, or more likely someone else will pick this up. I'm just fly-by posting awaiting a conference call. A quick look does not show SNR specs so hard to compare noise floors.

Do you need a phono stage? AFAIK the LS28 is a line stage only...

I like the balanced design of the MP-3. There are no SNR specs that I saw but everything else looks really good. If you need a phono stage that is the obvious choice, and it appears to have good reviews as well. I've owned ARC gear in the past (still have an SP3A1a in storage) but have a soft spot for Atma-Sphere because they seem more down-to-earth with a little less marketing gloss than some other companies. Note tube phono stages tend to be a bit noisier than SS, probably not a big deal with MM cartridges but in the past I used a transformer with a MC cartridge into a tube preamp.

Probably not much in the way of concrete help, sorry...

 

[amirm]

Pre 1 is an Audio Research LS28. I haven't found any spectrums on it, but it's better brother is the Ref 6 and there are spectrums for it.

https://www.stereophile.com/content/audio-research-reference-6-line-preamplifier-measurements

Ah, now you tell us . There are much more useful measurements there than the spec sheet. Let's review them:

This says there is a very mild 0.5 db roll off at 20 Khz. I am not going to hear that but if you are much younger or have better hearing, you may hear a slight roll off.

That is fairly noisy at low volumes.

Other than that, this is a type of tube amp that tries hard to not be a tube amp with all the problems that brings .