For a given circuit, let's say especially a small-signal circuit like a line stage, what is the relative impact of optimizing the power supply circuit for low noise vs. providing mechanical shielding between the power supply and the audio circuit, or isolating the power supply in a separate chassis?
Perhaps better to find similar equipment with measured noise levels that you think are acceptable to you, and see if they have mechanically isolated powering or not. If not, then you know that either path can get you where you want to be.
The problem with taking in examples from a forum is that they are highly likely to be different from what you would build, and hum/noise has many many ways to permeate in to an audio path, so the only practical way forward is likely to set up the amp yourself, and the tools to measure hum noise, and be prepared to make changes (as opposed to thinking you can design in all relevant noise mitigation measures and so will have no need to make any changes or do any faultfinding).
The problem with taking in examples from a forum is that they are highly likely to be different from what you would build, and hum/noise has many many ways to permeate in to an audio path, so the only practical way forward is likely to set up the amp yourself, and the tools to measure hum noise, and be prepared to make changes (as opposed to thinking you can design in all relevant noise mitigation measures and so will have no need to make any changes or do any faultfinding).
Are you saying here, the design and implementation of the audio circuit makes it more or less susceptible to different types of hum/noise, so installing the same power supply in three differently designed amps could have different amounts of noise/hum?
How would I determine the source of a hum? Experimentation implementing different filters or shields? Is there a way to look at a type of noise and say "Ahh yes, this sort of noise is most likely introduced because of a lack of filters" or "this sort of hum is induced by the power transformer emitting electromagnetic waves"?
Heads up that the topic of hum/noise is quite technical in terms of appreciating how it is generated and how it passes around circuitry and how it is then experienced and measured.
Perhaps it is worthwhile if you can elaborate on why this is of interest for you (eg. it seems that you have noticed a hum/noise but are not sure how to suppress it). Maybe we can focus on that aspect rather than delve head first in to a web of technical topics.
Perhaps it is worthwhile if you can elaborate on why this is of interest for you (eg. it seems that you have noticed a hum/noise but are not sure how to suppress it). Maybe we can focus on that aspect rather than delve head first in to a web of technical topics.
There are two separate but related issues that lead to my curiosity.
1. I observe 60Hz harmonic noise on my the power amp build I recently completed. Not high enough to be problematic, but enough to make me wonder if I could reduce them.
2. I once had an Audio Research preamp with an external power supply. Ever since I’ve been curious about the impacts of separating the power supply from the main chassis. I’ve seen it discussed in other contexts, especially around line stages and phono preamps. I’ve also seen discussion of separating the power supply and audio circuits with a metal bulkhead to achieve similar results.
So much in hifi audio is old wives tales, I’m trying to separate fact from fiction.
1. I observe 60Hz harmonic noise on my the power amp build I recently completed. Not high enough to be problematic, but enough to make me wonder if I could reduce them.
2. I once had an Audio Research preamp with an external power supply. Ever since I’ve been curious about the impacts of separating the power supply from the main chassis. I’ve seen it discussed in other contexts, especially around line stages and phono preamps. I’ve also seen discussion of separating the power supply and audio circuits with a metal bulkhead to achieve similar results.
So much in hifi audio is old wives tales, I’m trying to separate fact from fiction.
Do you have measuring equipment or a way to measure the hum you presently observe? Preferably any measuring technique explicitly identifies hum related frequencies, as ears are prone to misinterpreting the frequencies involved.
There are some starting points for that type sleuthing in the end appendix of the linked article: https://www.dalmura.com.au/static/Hum article.pdf
There are some starting points for that type sleuthing in the end appendix of the linked article: https://www.dalmura.com.au/static/Hum article.pdf
Yes, I'm using the FFT function of Room EQ Wizard. I've attached an example image.
I'm less interested in addressing this specific issue as I am in understanding the general/theoretical benefits of separating the power supply from the audio circuit, either by shielding, or by placement in a separate enclosure. 10,000ft stuff.
I'm less interested in addressing this specific issue as I am in understanding the general/theoretical benefits of separating the power supply from the audio circuit, either by shielding, or by placement in a separate enclosure. 10,000ft stuff.
Attachments
Are you saying that the highest hum component (mains 2H), which is about -85dB of 1W level, and other related components are what you are interested in suppressing even further? I guess for starters it would be best to characterise your measurement setup with a spectrum measurement under the same conditions but with the probe connection at the amp output moved across to the 0V connection at the amp output (ie. the noise floor of the same measurement setup, including what may be being picked up by the probing arrangement).
We're getting away from my main question, which isn't related to this amp, or any specific amp.
My question is about external power supplies. Why do it, why not do it?
My question is about external power supplies. Why do it, why not do it?
Your question is imho way too open ended to progress in anything but a very subjective comment, and any comment is most likely based on a single example where there was either no objective measurements made, or the example related to a very specific set of circumstances that few would be able to benefit from in any objective way.
Of course if you locate the power supply a mile away from the amp then any field related coupling is not a concern any more, but amps are made with the power supply in the same chassis and still there is no measurable concern. For every plausible hum coupling mechanism there is likely a reasonable technique to suppress it if the power supply is local.
Perhaps if you also consider other non-hum reasons why the power supply was sometimes made on a separate chassis - as they could well be the 100% reason for that decision - eg. a single chassis would be far too heavy, or too large a footprint.
Of course if you locate the power supply a mile away from the amp then any field related coupling is not a concern any more, but amps are made with the power supply in the same chassis and still there is no measurable concern. For every plausible hum coupling mechanism there is likely a reasonable technique to suppress it if the power supply is local.
Perhaps if you also consider other non-hum reasons why the power supply was sometimes made on a separate chassis - as they could well be the 100% reason for that decision - eg. a single chassis would be far too heavy, or too large a footprint.
It's a common "idea" that I've seen around the audiophile world that having a preamp/phono stage/line stage in a separate box from it's power supply will reduce noise. I'm trying to sus out the fact from fiction, as we know that makers of audiophile equipment do not need a "real" reason to do something and advertise it as beneficial.
What I gather from your comments is there is almost never a need to physically separate a power supply from an audio circuit, even if the circuit is dealing with small signals like a phono preamp, because the various mechanisms by which the power supply circuit would create noise is the audio circuit can all be addressed in other ways, circuit designs, etc. And, if I can extrapolate, if you were designing some arbitrary piece of audio gear, your first thought would not be "well, I'd better put the power supply in a separate box to reduce noise issues".
What I gather from your comments is there is almost never a need to physically separate a power supply from an audio circuit, even if the circuit is dealing with small signals like a phono preamp, because the various mechanisms by which the power supply circuit would create noise is the audio circuit can all be addressed in other ways, circuit designs, etc. And, if I can extrapolate, if you were designing some arbitrary piece of audio gear, your first thought would not be "well, I'd better put the power supply in a separate box to reduce noise issues".
Yup. At some point engineering considerations can and do come in to play. For example if the chassis was on the small size, and the PT was quite close to sensitive circuitry, and an easy to get PT was used but caused too high a level of coupling then a better PT (eg. better screening or orientation) may or may not be available, and may or may not be too expensive for the product, or fitting in shields to the chassis may be too expensive. Another example is that guitar amp chassis are often quite rectangular, with one benefit being that the OT is at one end and the PT is at the other - yes there are other benefits of such a layout such as the number for control knobs that have to be fitted to the front panel, but it does mean that inter-transformer coupling is rarely an issue.
Marketing will always spin any technical aspect into a big plus.
Marketing will always spin any technical aspect into a big plus.
So, externalizing/shielding/distancing the power transformer from the audio circuitry can be a perfectly valid and effective method for reducing noise. But, there are many other ways to achieve the same thing. So like all engineering decisions, it comes down to what makes the most sense based on costs, available materials, practicality, usability, etc, etc.
Is the key issue at play here transformer coupling? IE, if I moved just the transformer far away from the rest of the circuit, but left all of the other power regulators, filters, etc in close proximity to the main circuit, would I have accomplished all that can be accomplished by having a "separate power supply"?
(I appreciate you bearing with me on this 🙂)
Is the key issue at play here transformer coupling? IE, if I moved just the transformer far away from the rest of the circuit, but left all of the other power regulators, filters, etc in close proximity to the main circuit, would I have accomplished all that can be accomplished by having a "separate power supply"?
(I appreciate you bearing with me on this 🙂)
There are many ways for hum/noise to couple - and every situation may accentuate a particular mechanism or not - there are no rules that dictate a particular mechanism is worse than another. Many people experience a few forms and so alleviate them in their next build, and happily suggest they me be the cause of other peoples issues - which may or may not end up being true, so should be seen as good examples of where hum has caused problems. DIY is very prone to a wide range of mechanisms as not everyone is schooled in particular practices that alleviate hum - and the diy chase for hum/noise elimination can tend to be all engrossing for some and go far beyond any tangible benefit, not that that is a wasted effort (far from it). Just the range of ways a tube itself can exacerbate a particular hum mechanism is enlightening. Google, time and effort are your friends on this topic.
Thank you very much. This has been helpful.
I resisted my power supply noise. I don't remember exactly, but I think the above graph was taken with spring clips on the output posts. The input to my sound card might have also been loose previously. The below graph was taken using a load box connected with proper spade lugs.
As you can see, I still have a similar noise distribution, but it's 10-15 dB down from previous. (Ignore the signal harmonic differences, I've been making adjustments).
If you have any suggestions from reducing this sort of noise further, I'm all ears. However, I'm much less worried about it than I was before. For reference, I've built a Burning Amp 3 with the diyA Universal Power Supply.
I resisted my power supply noise. I don't remember exactly, but I think the above graph was taken with spring clips on the output posts. The input to my sound card might have also been loose previously. The below graph was taken using a load box connected with proper spade lugs.
As you can see, I still have a similar noise distribution, but it's 10-15 dB down from previous. (Ignore the signal harmonic differences, I've been making adjustments).
If you have any suggestions from reducing this sort of noise further, I'm all ears. However, I'm much less worried about it than I was before. For reference, I've built a Burning Amp 3 with the diyA Universal Power Supply.
Attachments
As per post #8, a benchmark measurement is the best way to confirm that a measurement plot is not just showing measurement set-up foibles.
It’s somewhat common to use a steel partition to help with radiated stuff from a power transformer in line level gear, and amps often use steel covers for their transformer
As for junk from the mains, rectifier diodes, there are simple ways to help that too, just have to search.
If there’s a pc involved, the monitors can be very noisy if plugged in near certain line level equipment.
As for junk from the mains, rectifier diodes, there are simple ways to help that too, just have to search.
If there’s a pc involved, the monitors can be very noisy if plugged in near certain line level equipment.
Tape head demagnetizer is an excellent tool to chase down magnetic coupling. It tells you immediately if you've got your current loops right (minimized loop area and preferably makeing the remaining loop self-cancelling like in twisted pair wiring). Good luck with any audio transformers, though.
To isolate the results from the normal mains pickup you already have in the circuit, use a different frequency and time-domain averaging on the scope. Or, alternatively, at least move the tranny out of the enclosure and place it far away from the circuit.
To isolate the results from the normal mains pickup you already have in the circuit, use a different frequency and time-domain averaging on the scope. Or, alternatively, at least move the tranny out of the enclosure and place it far away from the circuit.