KJF Audio Ltd
- Jan 21, 2014
- HiFi Trade?
That is a biggie, I was asked about input sensitivity yesterday elsewhere which is a related question so I will paste that below, any more questions just ask.Super review, thank you!
I have more than a passing interest, as my Primare I32 is just coming up to its 4th birthday, and earlier this year I bought a Nord from a fellow Wammer. I agree with the comments that these designs get surprisingly close to ‘perfect’ and definitely sound nothing like some attribute to class D.
Perhaps Stefan @orangeart might be able to kindly explain the power ratings - Hypex emphasise the peak outputs, which is understandable because (a) the numbers are bigger, and (b) peak power is arguably more relevant when playing music. However, what I don’t understand is how does a nominally 50 watt amplifier yield the 136/256 into 8/4ohms shown in the Hifi World review? That’s double the rated continuous power, but not as high as the peak output claimed by Hypex.
I was asked about amplifier input sensitivity in another thread and as it isn't as simple as it first looks I thought I would add my answer here for anyone else who was interested.
OK here is amp gain in a little more detail. As I said earlier the input sensitivity is a function of the rated power available and the load (speakers) on the amp. Some manufacturers will massage the rated load upward and some will be more conservative. I'll explain the maths and then give an answer to your last question about how the measurement by HiFi world can be different.
Input sensitivity of am amp is given by the following formula. √(Rated power Wrms* Load Ω)/10^(25.6/20)= sensitivity V RMS. To plug in some numbers for real life situation, the NC500MP has a rated power of 500Wrms into a 4 ohm load, therefore we get √(500* 4)/𝟏𝟎^(𝟐𝟓.𝟔/𝟐𝟎)=2.35V RMS. That is to drive the system to full power into 4 ohm load we need to input 2.35 Vrms. Of course real world loads are not a flat impedance, they generally rise as frequency rises and have a spike at resonance and maybe some dips here and there as well so the above only gives a rough outline into an average impedance load.
Now we can see that if we change some of those numbers, the input sensitivity changes, so for the same amp module into an 8 ohm load it is rated at 270Wrms, this same formula gives us a value of 2.43 Vrms to drive to a full signal.
Into a 2 ohm load it is only rated at 400Wrms which gives us only 1.48 Vrms before we hit the limit.
Now manufacturers will take different things into consideration when providing a max rated power. It might be based on the thermal properties of the output devices, get to close to thier thermal breaking point and bang. It might be based on the maximum voltage available in the PSU for the output stage. The first part of the formula - √(Rated power Wrms * Load Ω) gives us the voltage needed to drive a load to the power we want for any amplifier, 500Wrms into a 4 ohm load requires 44.72Vrms available from the PSU. Let us assume there is some headroom and the PSU can produce 45V from the 230V supply available from your house wall socket. Now what happens if that supply is 210V or 254V? If that means we get less voltage from the PSU we may not be able to drive the speakers to the same level.
The second part of the formula is the amplifier gain part of the formula where 25.6 is the gain in dB of the amp, this figure is made up of 2 sections of gain in most amps, the input circuit (ususally some sort of op-amp) which provides initial gain and the output devices which provide the rest. Ideally they would be completely linear, load agnostic, heat agnostic, manufaturing varience agnostic, etc. These modules have a really tight tolerance, however minimum gain would be expected to be 25dB and max gain 26dB for these particual modules, changing that number in the formula gives you a different voltage that you need to supply to get the full rated output power.
Made more difficult is that some pre amp manufatures provide Vrms capabilities for thier output and others provide peak voltage. The formula above is for Vrms which is not the peak V. Peak volage is the voltage at the top of the sine wave, the RMS voltage is a mathematically derived average voltage along the length of a whole sine cycle. To get to peak voltage from the RMS voltage we need to multiply by √2 or 1.414, to go the other way you would multiply by .707. Mathmaticians amongst you might recognise those numbers, they crop up a lot in maths.
Now when an amp is tested by the magazines, they are pretty ruthless. They use a dummy load which is a bank of high power resistors and apply a signal at the input. These are provided by signal generators that can produce a high magnitude sine wave without clipping it's own signal. A scope is attached to the speaker output/load and the input sine is increased until just before the output looks like it is clipping (flattening on the top of the waveform). In other words the maximum they can manage, pretty brutal. Remember this 500Wrms into 4 Ohms is not the continuous expected operation, average level is expected to be down at 100Wrms, Music is dymanic so amp continuous operation is usally given at about a fifth of full signal, and also usually where stuff like distortion is measured. The recent HiFi world measurment show the amp exceed the rated continuous power by a large margin as these tests represent a continuos load, the distortion figures are also given at full load, showing just how impressive these Hypex modules are. As Noel put it 'almost startingly good measured performance'
To answer more directly though, assuming a 4 Ohm load the 125W modules require 1.17Vrms to manage full power, the 250W modules need 1.66Vrms and the 500W modules require 2.35Vrms The amp will display clipping on the button ring if you are pushing to hard