• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

High sensitivity horns versus signal-to-noise

Head_Unit

Major Contributor
Forum Donor
Joined
Aug 27, 2018
Messages
1,341
Likes
688
I am getting closer to my dream of custom speakers with large horns-some of which can hit 110 dB/1W/1m. Such horns are known to be sensitive for revealing hiss in amplifiers and upstream components. And I am really REALLY REALLY sensitive to hiss whether on cassettes or computer fans or whatever, I could not stand it if I could hear hiss out of the speakers

I was just reading the Denon AVR-X3700H, where Amir purports the signal-to-noise is 90dB at 5W, call it 84 dB re 1 watt. Or 104 dB at full power, which is around the same 84 dB. 110 dB from a horn - 84 dB leaves the noise at 26 dB which should be inaudible.

Hmmm, really? Am I missing something? My current AVR-2312Ci and previous AVR-1612 seemed quiet, but the slim Infinity towers they powered are low in sensitivity so that proves nothing.
 

richard12511

Major Contributor
Forum Donor
Joined
Jan 23, 2020
Messages
4,335
Likes
6,700
I am getting closer to my dream of custom speakers with large horns-some of which can hit 110 dB/1W/1m. Such horns are known to be sensitive for revealing hiss in amplifiers and upstream components. And I am really REALLY REALLY sensitive to hiss whether on cassettes or computer fans or whatever, I could not stand it if I could hear hiss out of the speakers

I was just reading the Denon AVR-X3700H, where Amir purports the signal-to-noise is 90dB at 5W, call it 84 dB re 1 watt. Or 104 dB at full power, which is around the same 84 dB. 110 dB from a horn - 84 dB leaves the noise at 26 dB which should be inaudible.

Hmmm, really? Am I missing something? My current AVR-2312Ci and previous AVR-1612 seemed quiet, but the slim Infinity towers they powered are low in sensitivity so that proves nothing.
My horns definitely have audible hiss, and they’re quite a bit less efficient than 110dB, though it doesn’t bother me.
 

hardisj

Major Contributor
Reviewer
Joined
Jul 18, 2019
Messages
2,907
Likes
13,908
Location
North Alabama
My horns definitely have audible hiss, and they’re quite a bit less efficient than 110dB, though it doesn’t bother me.

Ditto. But I don’t use my horn setup for music. If I did then the hiss would bother me.

I did spend a lot of time trying to come up with a solution early on. It resulted in me admitting defeat because the only options were (5 years ago) to spend way more than I could afford on a very *high* SNR amplifier.
 
Last edited:

Blumlein 88

Grand Contributor
Forum Donor
Joined
Feb 23, 2016
Messages
20,522
Likes
37,050
Seems one of the Purifi amps like this one:
https://www.audiosciencereview.com/...easurements-of-purifi-1et400a-amplifier.7984/

Or one of the better Hypex amps is what you need. Way too much power, but I don't think you'll find much that is quieter. Maybe a Benchmark AHB2.

You could help the situation somewhat by going through some voltage dropping resistors at the output of the amp. Lets you run the amp at higher power further above the noise floor, and the voltage division will reduce the noise further still. You could easily gain 12 db less noise that way.
 

Blumlein 88

Grand Contributor
Forum Donor
Joined
Feb 23, 2016
Messages
20,522
Likes
37,050
And thinking on this some more, you might do fine with one of the super performing headphone amps. Some of those will put out 1 or 2 watts into relatively low impedance loads while having SOTA noise levels.

What are the impedance characteristics of your horn speaker?
 

dc655321

Major Contributor
Joined
Mar 4, 2018
Messages
1,597
Likes
2,235
And thinking on this some more, you might do fine with one of the super performing headphone amps. Some of those will put out 1 or 2 watts into relatively low impedance loads while having SOTA noise levels.

I think @pos did exactly that for his DIY speakers, for exactly the reasons discussed here.
 

restorer-john

Grand Contributor
Joined
Mar 1, 2018
Messages
12,579
Likes
38,278
Location
Gold Coast, Queensland, Australia
Go vintage.

The residual noise is the key spec, not the signal to noise ratio itself as SNR is "gamed" as it is not referenced to 1W, it's referenced to full power. So, the more power you have on tap, the better the SNR number. Trouble is, more powerful amplifiers have more paralleled output devices and more standing current and hence more residual noise.

The sweet spot in my experience is about 100-150 watts per channel at 8ohms. You will find plenty of top models from the 1980s with truly excellent residual noise specs, better than any of the class Ds for sure.

You can quantify the residual noise spec yourself with a sensitive audio-millivoltmeter and a few weighting networks. You can amplify the residual noise and look at its spectrum- some hiss is audibly worse than others. Amplifier residual is pretty much white noise, so A-weighting it and comparing, you will easily find the quiet ones. Use a shorting RCA/XLR for the input, open the level pots (if any) and measure. Also with integrated amps, you can determine where on the volume/gain the amplifier is the noisiest- often it is right in the typical listening position!
 
Last edited:

Blumlein 88

Grand Contributor
Forum Donor
Joined
Feb 23, 2016
Messages
20,522
Likes
37,050
The Benchmark power amp lists 7.1 microvolts a wtd at idle. The Purify is claiming 11.7 microvolts. Some of the Hypex units are below 20 microvolts. These are the lowest I'm aware of in general. I do seem to recall some of the old analog amps got to around 20 micro-volts maybe some did better. Several of the hypex units are around 20-35 uvolts idle noise. So that gives you an idea of what is available. Depending upon the impedance of your horns this will all translate into not much better than 90 db below 1 watt. Which might be enough, and the idea of wasting a little power thru dropping resistors could let you lower that noise some more.

Of course the JDS Labs Atom headphone amp might well power your speakers just fine. What is that like $200? If it won't, the AAA 789 Massdrop headphone amp might be enough at $349. Both put out over 1 watt into 33 ohms. Just a matter of whether or not they get current starved at a lower impedance.
 
Last edited:

restorer-john

Grand Contributor
Joined
Mar 1, 2018
Messages
12,579
Likes
38,278
Location
Gold Coast, Queensland, Australia
The Benchmark power amp lists 7.1 microvolts a wtd at idle. The Purify is claiming 11.7 microvolts. Some of the Hypex units are below 20 microvolts. These are the lowest I'm aware of in general. I do seem to recall some of the old analog amps got to around 20 micro-volts maybe some did better.

Here's a Yamaha M-80. The big one from the mid/late 80s at 250W/ch @8R 0.003%. 16uV. 127dB A-WTD S/N.

1606896655721.png


Here's another, the MX-630 from 1990/1. This was a relatively inexpensive power amp in their separates range. Sold for AUD$999 here, so likely half that in the US. The last one I picked up for $130 second hand. The MX-600 is another, with nice red power meters. They are extremely quiet. You have to put your ear on the tweeter to hear anything.

1606898162200.png
 
Last edited:

Absolute

Major Contributor
Forum Donor
Joined
Feb 5, 2017
Messages
1,084
Likes
2,125
I experienced the terror of hiss on my M2 clones, which have about 100-105 dB sensitivity on the tweeter at most after the passive filter.
Using the Hypex DLCP as pre/dsp and Hypex Nc400 there was noise that I just couldn't live with. One thing is the amount of noise, another is the noise spectrum.

I removed some resistor to reduce the amplifier gain from 26 to about 13 dB and away went the hiss. And all power as well because I'm stupid and didn't factor in the gain chain and the DLCPs max volt out. (2,56v).

Changed the amplifier to the far more noisy Crown cdi 4/1200 and won back all the noise and then some, plus all the power and far more.
Although the hiss is quite a bit louder than the Nc400, the spectrum itself is far more comfortable and doesn't annoy me nearly as much.

Someone once told me it's best to have a digital dsp, extremely low-noise dac with high signal level and then a ultra-silent volume control after both.
That will ensure best SNR and protect your ears/speakers if you for some reason mess the dsp up. I can attest to the fact that both DLCP and MiniDSP have given me near-death experiences where something goes haywire during uploading of a new filter and sending out full-scale signal to the amplifiers, so I'm convinced a separate volume control after the dsp/dac could be a life-saver on high sensitivity devices.

A headphone amp with spectacular SNR to the tweeter could be gold, but so could any low-gain low-noise amp.
 

Blumlein 88

Grand Contributor
Forum Donor
Joined
Feb 23, 2016
Messages
20,522
Likes
37,050
Well keep in mind if you had a source component with 2 volt max output and a -120 db noise floor it will put out 2 microvolts itself. Send that to an amp with 26 db gain and you'll have 40 microvolts at the amp output with no signal. You'll have eaten into your noise budget quite a bit right there. A passive preamp would work here, but the output impedance is going to be high enough thermal noise of the resistors will be eating into the noise budget. A no gain buffer would do the trick. A really good buffer with some current capability is really all you'd want to drive a speaker like that.

So in terms of practical gear, I think a digital source with digital volume control directly feeding the amp, a quiet amp with the gain changed to nothing or nearly nothing, and some voltage divider power resistors at the output.

Maybe a 90 db rated speaker isn't so bad.
 

tuga

Major Contributor
Joined
Feb 5, 2020
Messages
3,984
Likes
4,281
Location
Oxford, England
I am getting closer to my dream of custom speakers with large horns-some of which can hit 110 dB/1W/1m. Such horns are known to be sensitive for revealing hiss in amplifiers and upstream components. And I am really REALLY REALLY sensitive to hiss whether on cassettes or computer fans or whatever, I could not stand it if I could hear hiss out of the speakers

I was just reading the Denon AVR-X3700H, where Amir purports the signal-to-noise is 90dB at 5W, call it 84 dB re 1 watt. Or 104 dB at full power, which is around the same 84 dB. 110 dB from a horn - 84 dB leaves the noise at 26 dB which should be inaudible.

Hmmm, really? Am I missing something? My current AVR-2312Ci and previous AVR-1612 seemed quiet, but the slim Infinity towers they powered are low in sensitivity so that proves nothing.

Do you think it's wise to use an AVR for powering110dB horns?

I could be mistaken but perhaps a < 10 w (class A?) amplifier might be a better option considering that you are very sensitive to noise.
Also distortion levels with transistor amplification tend to drop at high output and it'll be much easier to adjust the playback volume without having to use analogue or digital attenuation.
 
OP
H

Head_Unit

Major Contributor
Forum Donor
Joined
Aug 27, 2018
Messages
1,341
Likes
688
Do you think it's wise to use an AVR for powering110dB horns?
I am not aware of any other practical solutions to run multichannel with bass management and room correction and most of all decoding 5.1:
- A pre/pro and power amps, but $$$ and this forum doesn't indicate to me that performance is any better
- miniDSP et al have all kinds of stuff, but not decoding that I'm aware of. I guess maybe some kind of home theater PC could, but that also seems like a mess.
I used to run horns off a little NAD 3020. Now maybe those weren't 110 but they were up towards 100. That NAD had a switch to change the volume knob gain range, and was quiet in the first place, so not directly comparable.
 
OP
H

Head_Unit

Major Contributor
Forum Donor
Joined
Aug 27, 2018
Messages
1,341
Likes
688
microvolts
SNR is "gamed" as it is not referenced to 1W,
Well, but in @Amir's case, he does reference it, to 5W (90 dB) and to full power. (104 dB). Those translate to around -84 dB from 1 watt. Using an online converter, that is 0.000004 mW = 0.000 000 004 watts. At 4 ohms, P=V^2/R --> 0.000000004 = V^2/4 --> V^2 =0.000000016 --> V=0.00012649 or 126 mV...did I do that correctly?

Assuming so, back to the original question, if 84 dB down from 110 is 26 dB, how can the hiss from a horn be audible? (Cuz I know sometimes it IS, not even just horns)
 
Top Bottom