I see that on Rod's ESP site, he measures his MOSFET amp as having noise of <2mV RMS (unweighted -54dBV). A voltage seems to be a sensible way of measuring noise, but I'm not informed enough to know how this value compares to the SNR quoted on most commercial amp specifications (e.g. 116dB), so i'm struggling to put it into context. I realise this will be an incredibly basic question for many of you, but for thicko here is it comparable or convertible in some way? Or is it good/bad/indifferent?
For those of you who have built your own version of these, I'd also be interested to hear if you felt it was quiet or noisy compared to other amps, and if you made any mistakes or improvements etc related to noise/hiss/hum and so on of the completed amplifier. Obviously there'll be other sources of noise too (interconnects, upstream equipment etc) but this question is particularly about the amp (and PSU) built to Rod's official guides.
I'm interested in this particular amp because I already have the official PCBs and most components, bought some years ago but then never built. I'm trying to decide if I'd want to build and use it for my next application: an active configuration involving potentially sensitive drivers rather close to the listening position. In this case it seems very likely that noise of lower levels than normal might become quite audible, and I really dislike hearing it.
Thanks,
Kev
For those of you who have built your own version of these, I'd also be interested to hear if you felt it was quiet or noisy compared to other amps, and if you made any mistakes or improvements etc related to noise/hiss/hum and so on of the completed amplifier. Obviously there'll be other sources of noise too (interconnects, upstream equipment etc) but this question is particularly about the amp (and PSU) built to Rod's official guides.
I'm interested in this particular amp because I already have the official PCBs and most components, bought some years ago but then never built. I'm trying to decide if I'd want to build and use it for my next application: an active configuration involving potentially sensitive drivers rather close to the listening position. In this case it seems very likely that noise of lower levels than normal might become quite audible, and I really dislike hearing it.
Thanks,
Kev
Well, to (sort of) answer my question, calculations show that a 2mV rms signal of noise would produce about 25dB at 1m (very roughly speaking) from an 8ohm driver of middling sensitivity at 1m. Double that (28dB) for a 4ohm driver.
So that is very decently quiet for most situations, and the spec says "less than" 2mv so I would assume that would be the max one could expect, and probably much less.. But for my situation.. with closer listening distance, potentially more sensitive 4ohm drivers, and between 4 to 6 of them, it isn't quite the silence I'm wanting. There is also the strong possibility that, in my hands, the finished product may not measure so well as Rod's.
Though now the OP is kind of reversed, and probably what I should have asked to begin with.. How does one quantify noise/hiss of commercial amps - is it possible to calculate the SPL of noise from any typical specs they give, such as SNR?
So that is very decently quiet for most situations, and the spec says "less than" 2mv so I would assume that would be the max one could expect, and probably much less.. But for my situation.. with closer listening distance, potentially more sensitive 4ohm drivers, and between 4 to 6 of them, it isn't quite the silence I'm wanting. There is also the strong possibility that, in my hands, the finished product may not measure so well as Rod's.
Though now the OP is kind of reversed, and probably what I should have asked to begin with.. How does one quantify noise/hiss of commercial amps - is it possible to calculate the SPL of noise from any typical specs they give, such as SNR?
As a comment on noise levels, measurements are usually expressed in terms of power rather than voltage, which is partly explained here, for example: https://resources.pcb.cadence.com/blog/2020-what-is-signal-to-noise-ratio-and-how-to-calculate-it
Rod's P101 amp is certainly quiet enough in a typical room with full-sized speakers but perhaps you are planning to use that large-ish power amp for a desktop system or maybe tweeter drivers for an active crossover? You probably won't need anything like the P101's output power there but generally and because of multiple amplification factors, any power amplifier could appear to be noisy but that noise is more likely to originate in the lower signal levels of the preamp, controls and grounding arrangements than the power amplifier itself.
A straightforward solution for a suitable power amp, would be to scale it appropriately to the load. You could use the P101 but it's rather a waste and may well be a tad noisy due to level mismatches too. If you want a simple, low cost amp with good performance, a smaller LM1875 chipamp kit is cheap, can be easy to implement and will perform reliably to its specified, high quality standard, providing more than enough power for say, your desktop system, where 5W is usually plenty. The highest cost of amplifiers is in the case and transformer so the savings there will be the most significant and the effort needed to mount and connect the optional, pre-assembled kit, is minimal.
The following example is close to what I have in use. It proves to be much more than enough for desktop use. I've also used similar, mono versions for tweeters in an active crossover arrangement. Mine were pre-assembled boards (OK, I'm lazy, they were cheap and ready to go!) which I used to replace some tiny, DIP-8 chipamps in a cheapo, desktop 2.1 system that had been fried. It's now transformed into something much better with its improved chipamps and capability of up to 20W or so. https://www.aliexpress.com/item/100...00017160885392!sea&curPageLogUid=6irlCvvKu89c
Now you don't need to finish that large amp - a 40-60VA toroid and almost any aluminium case big enough for the assembled board and power supply could act as a sufficient heat sink and you don't need to do so much wiring either.
Rod's P101 amp is certainly quiet enough in a typical room with full-sized speakers but perhaps you are planning to use that large-ish power amp for a desktop system or maybe tweeter drivers for an active crossover? You probably won't need anything like the P101's output power there but generally and because of multiple amplification factors, any power amplifier could appear to be noisy but that noise is more likely to originate in the lower signal levels of the preamp, controls and grounding arrangements than the power amplifier itself.
A straightforward solution for a suitable power amp, would be to scale it appropriately to the load. You could use the P101 but it's rather a waste and may well be a tad noisy due to level mismatches too. If you want a simple, low cost amp with good performance, a smaller LM1875 chipamp kit is cheap, can be easy to implement and will perform reliably to its specified, high quality standard, providing more than enough power for say, your desktop system, where 5W is usually plenty. The highest cost of amplifiers is in the case and transformer so the savings there will be the most significant and the effort needed to mount and connect the optional, pre-assembled kit, is minimal.
The following example is close to what I have in use. It proves to be much more than enough for desktop use. I've also used similar, mono versions for tweeters in an active crossover arrangement. Mine were pre-assembled boards (OK, I'm lazy, they were cheap and ready to go!) which I used to replace some tiny, DIP-8 chipamps in a cheapo, desktop 2.1 system that had been fried. It's now transformed into something much better with its improved chipamps and capability of up to 20W or so. https://www.aliexpress.com/item/100...00017160885392!sea&curPageLogUid=6irlCvvKu89c
Now you don't need to finish that large amp - a 40-60VA toroid and almost any aluminium case big enough for the assembled board and power supply could act as a sufficient heat sink and you don't need to do so much wiring either.
To compare to Stereophile (1W, 8R 2.83V SNR): 20log(2.83/.002)
To compare to ASR (5W, 4R 4.47V SNR): 20log(4.47/.002)
To go the other way, if you want to figure the RMS noise of an amp with 63dB SNR ref 1W 8R: 2.83*10^(-63/20)
If SNR is given for rated output power, you can normalize to 1W. So if you have a 100W 8R amp with 83dB SNR at full power, to convert that to 1W SNR: 83+10log(1/100)
Alternatively, you can calculate directly from 100W by converting output power into a voltage via ohm's law. V=sqrt(P*R); sqrt(100*8)*10(-83/20)
To compare to ASR (5W, 4R 4.47V SNR): 20log(4.47/.002)
To go the other way, if you want to figure the RMS noise of an amp with 63dB SNR ref 1W 8R: 2.83*10^(-63/20)
If SNR is given for rated output power, you can normalize to 1W. So if you have a 100W 8R amp with 83dB SNR at full power, to convert that to 1W SNR: 83+10log(1/100)
Alternatively, you can calculate directly from 100W by converting output power into a voltage via ohm's law. V=sqrt(P*R); sqrt(100*8)*10(-83/20)
Superb, thank you! I felt it should be possible but wouldn't have managed to get there without your help. This is great, and means that I can now compare the different options sensibly.
Thanks again, Kev.
Thank you very much for these thoughts, too; they make a lot of sense to me.
The P101 boards were worth considering since I have them, and there is some possibility of a linkwitz transform on the (quite small) bass drivers so power might be useful for that. But yes I agree; they are not looking to be quite what I want for this particular application.
Yep, these are small nearfield/desktop speakers and could indeed run off modestly proportioned, and more power-efficient, amps. Especially if I approach the speaker design a little differently, reducing the power requirement for low-end electronic correction. As it gets higher the extra power brings increasingly little difference in SPL anyway, so probably not worth the extra challenges for noise.
Not sure if I will choose to go with chip amps or used desktop amps (or even new ones) since the cost of all those options are more modest for lower power options. They all have their attractions, so I'll have to ponder what suits me best.
Thanks again,
Kev