I've been reading a lot about grounding lately, but none of the sources I've read seem to be able to agree on anything. I've got a couple of small projects involving high-gain circuits in the works (mic preamp, phono preamp, etc...) and I'd like a little advice on the best way to minimize hum.
Usually when I do a project like this I use a star type ground, where each section has a thick 14 gauge solid copper ground bus connected to a single copper lug in the power supply. This seems to work exceptionally well for most circuits, but even though I'm using DC on the filaments and the B+ supply is rock solid I'm still hearing some hum in very high gain circuits.
Then I got to thinking about how the grounds for the inputs and outputs are routed. We'll use the mic pre as an example. Is it better to connect the incoming ground from the mic to the preamp section's local bus, close to where the signal enters, or should it be connected to the power supply's star ground? Would it be better to AC couple the incoming ground through a capacitor? Should right and left have separate paths to ground or be connected directly together? What about volume potentiometers? Should they be connected to a local bus close to where the signal enters the circuit, or directly to the star ground? I've also noticed that for some reason, coupling the filament supply's ground to the star ground with a 0.1μF capacitor reduces hum, why is this?
Usually when I do a project like this I use a star type ground, where each section has a thick 14 gauge solid copper ground bus connected to a single copper lug in the power supply. This seems to work exceptionally well for most circuits, but even though I'm using DC on the filaments and the B+ supply is rock solid I'm still hearing some hum in very high gain circuits.
Then I got to thinking about how the grounds for the inputs and outputs are routed. We'll use the mic pre as an example. Is it better to connect the incoming ground from the mic to the preamp section's local bus, close to where the signal enters, or should it be connected to the power supply's star ground? Would it be better to AC couple the incoming ground through a capacitor? Should right and left have separate paths to ground or be connected directly together? What about volume potentiometers? Should they be connected to a local bus close to where the signal enters the circuit, or directly to the star ground? I've also noticed that for some reason, coupling the filament supply's ground to the star ground with a 0.1μF capacitor reduces hum, why is this?
People often give grounding 'recipes', which have worked for them. They will work for you too, provided your system and environment are sufficiently similar to theirs. If not, they could make things worse.
The best option is to ignore recipes, and instead ask yourself where the currents go. Treat every conductor, no matter how thick and short, as a resistor with some inductance (so you will get some voltage drop). Treat every current loop as the primary of a transformer, with every other current loop as a secondary (so you will get some induction).
The best option is to ignore recipes, and instead ask yourself where the currents go. Treat every conductor, no matter how thick and short, as a resistor with some inductance (so you will get some voltage drop). Treat every current loop as the primary of a transformer, with every other current loop as a secondary (so you will get some induction).
Here's my suggested reading list:
Morrison: "Grounding and Shielding: Circuits and Interference". I have the third edition, but he's up to the fifth now...
Ott: "Noise Reduction Techniques in Electronic Systems", 2nd ed.
Walt Jung: "OpAmp Applications Handbook". I think you can actually find the full book on-line from Analog Devices.
I think my main point is: Don't rely on Internet "sources" (unless you find some actual university research). Hit the books instead. I agree with DF96's statement about "following recipes". Think through the problem and understand how the various noise and coupling paths work it your circuit and go from there. Grounding can be a bit of a black art. But you know... It's only black art until someone turns on the light. 🙂
~Tom
Morrison: "Grounding and Shielding: Circuits and Interference". I have the third edition, but he's up to the fifth now...
Ott: "Noise Reduction Techniques in Electronic Systems", 2nd ed.
Walt Jung: "OpAmp Applications Handbook". I think you can actually find the full book on-line from Analog Devices.
I think my main point is: Don't rely on Internet "sources" (unless you find some actual university research). Hit the books instead. I agree with DF96's statement about "following recipes". Think through the problem and understand how the various noise and coupling paths work it your circuit and go from there. Grounding can be a bit of a black art. But you know... It's only black art until someone turns on the light. 🙂
~Tom
tomchr has it right. Those are reliable sources of valid information he listed; be cautious about random internet sources, at least take them with a large grain of salt. I have Morrison's book "Grounding and Shielding Techniques in Instrumentation"; it's quite good.
Last edited:
,
Hi, thank you for suggestions.
For a total newbie which one should be more understandable?
Regards
the "articles section" has the readable version. I have not seen any adverse comment on the accuracy of the content of DD's article.
Audio Component Grounding and Interconnection - diyAudio
They're both textbooks that would fit in a college curriculum, so they're not the easiest to understand. That said, I think Morrison does a good job of introducing the fundamental physics concepts and explain how they apply to grounding and shielding. If you have some recall of your high school physics you should be able to follow Morrison.
It's been a while since I've had Ott's book cracked open. I recall it being pretty good, but I have no recollection of the prerequisites needed to read it.
~Tom
Of course, there is the classic Rane text on this subject:
Grounding and Shielding Audio Devices
and my own modest contribution:
http://www.ianbell.ukfsn.org/EzTubeMixer/docs/EzTubeMixer/SimpleMixer/grounding101v2.pdf
Cheers
Ian
Grounding and Shielding Audio Devices
and my own modest contribution:
http://www.ianbell.ukfsn.org/EzTubeMixer/docs/EzTubeMixer/SimpleMixer/grounding101v2.pdf
Cheers
Ian
Tony Waldrons input:
Tony Waldron's EMC ranting and ravings
And Henry Otts view of single point at the power supply.
Also Henry Otts 'electromagnetic compatibility engineering"
This recently referenced on the Blowtoch thread by JN:
http://www.ewh.ieee.org/r5/denver/rockymountainemc/archive/2004/October/Experimental_Demo.pdf
There is no standard solution and staring is only one way, and not always the best. or having long tails on screens back to a single star point. Your shielding and grounding will be affected by your environment, the amount of noise, the makeup of your system, how much digital, how much analogue how they are interconnected .
Wrapping your interconnects round your mains leads can help reduce noise...
Tony Waldron's EMC ranting and ravings
And Henry Otts view of single point at the power supply.
Also Henry Otts 'electromagnetic compatibility engineering"
This recently referenced on the Blowtoch thread by JN:
http://www.ewh.ieee.org/r5/denver/rockymountainemc/archive/2004/October/Experimental_Demo.pdf
There is no standard solution and staring is only one way, and not always the best. or having long tails on screens back to a single star point. Your shielding and grounding will be affected by your environment, the amount of noise, the makeup of your system, how much digital, how much analogue how they are interconnected .
Wrapping your interconnects round your mains leads can help reduce noise...
All excellent replies and references above. IMO, Ott is great, but a bit advanced for many. Grounding and shielding is approached differently at low frequencies vs. high frequencies. Preventing audio hum can require a different "recipe" than solving a HF EMC problem. You need to identify the enemy before designing a grounding stratagy, and the answer may be a combination of star grounding for the LF issues, and distributed local grounding for the HF problems.
Hello Audiofox -
Ian has provided an excellent primer on hum and noise, IMO. Others here have presented excellent suggestions, if there is no 'canned' solution seen it is because each design is different and each build will have its own unique combination of problems.
Hum is a complicated subject, no two designs will have the same cause or solution. There are some common problems that are commonly seen (input connector grounding, letting the input-stage ground ride on the power-stage ground, etc.) but each build and design will have its own gremlins...
The best way to approach hum is to think in terms of currents and loops. Where are the current paths? What kind of loop (or, more likely, loops) does the grounding scheme have? How can I route grounds inside my amplifier to break these loops? (Here I am talking only about conducted noise - radiated noise and safety grounding are two other very important aspects of the generic 'grounding' problem.)
For instance, 'star' grounding is an excellent philosophy to follow, but if you have your input connector ground (shield) connected to your input amplifier ground, and that ground in series with your PA ground and then take the PA ground to your 'star' ground then you will almost certainly hear hum when you connect your source to your input connector.
This is because the topology I describe above sets up a loop by stringing all of the grounds together. Even though you are using copper wire to connect the grounds, that copper wire has some resistance and when ground currents run through that wire Mr. Ohm says you will develop a voltage and viola' - HUM.
The above is grossly oversimplified, but gives you an idea of why recommended schemes sometimes fail to solve a problem. Each amplifier will have unique design problems that need to be solved - unless you are building a Heathkit and follow the instructions to the letter, you must turn into a technician and spend hours and hours to track down each problem.
The simplistic scenario above completely neglects other sources of hum - radiation from transformers, radiation from the filament wires (if you use AC for filaments, which most Power Amps using tubes with a Cathode do), coupling from external equipment or lights near your amp, computers or other noisy equipment plugged into the same power strip as your amp - the list is long, and all of these sources of noise interact and layer upon each other to really make your hair turn grey trying to track them all down and eliminate them...
Long story short - anybody building an audio amplifier from scratch has a LOT more work to do than just follow the schematic correctly. The build part of an amplifier is generally the easiest part - it's the 'getting it to work properly' part that consumes all of the time.
So learn the basics of what causes hum (and other noise), learn to think in terms of currents-and-loops, and be prepared for some really frustrating evenings when nothing seems to work.
It's like giving birth - once you make that last 'fix' and turn on your system and it is noise-free, all the pain of getting it to that point is forgotten as you bask in the warm sound of something you worked hard to create!
(Then, of course, it's on to your next project - and more hair-pulling... 🙂 )
Sam
Ian has provided an excellent primer on hum and noise, IMO. Others here have presented excellent suggestions, if there is no 'canned' solution seen it is because each design is different and each build will have its own unique combination of problems.
Hum is a complicated subject, no two designs will have the same cause or solution. There are some common problems that are commonly seen (input connector grounding, letting the input-stage ground ride on the power-stage ground, etc.) but each build and design will have its own gremlins...
The best way to approach hum is to think in terms of currents and loops. Where are the current paths? What kind of loop (or, more likely, loops) does the grounding scheme have? How can I route grounds inside my amplifier to break these loops? (Here I am talking only about conducted noise - radiated noise and safety grounding are two other very important aspects of the generic 'grounding' problem.)
For instance, 'star' grounding is an excellent philosophy to follow, but if you have your input connector ground (shield) connected to your input amplifier ground, and that ground in series with your PA ground and then take the PA ground to your 'star' ground then you will almost certainly hear hum when you connect your source to your input connector.
This is because the topology I describe above sets up a loop by stringing all of the grounds together. Even though you are using copper wire to connect the grounds, that copper wire has some resistance and when ground currents run through that wire Mr. Ohm says you will develop a voltage and viola' - HUM.
The above is grossly oversimplified, but gives you an idea of why recommended schemes sometimes fail to solve a problem. Each amplifier will have unique design problems that need to be solved - unless you are building a Heathkit and follow the instructions to the letter, you must turn into a technician and spend hours and hours to track down each problem.
The simplistic scenario above completely neglects other sources of hum - radiation from transformers, radiation from the filament wires (if you use AC for filaments, which most Power Amps using tubes with a Cathode do), coupling from external equipment or lights near your amp, computers or other noisy equipment plugged into the same power strip as your amp - the list is long, and all of these sources of noise interact and layer upon each other to really make your hair turn grey trying to track them all down and eliminate them...
Long story short - anybody building an audio amplifier from scratch has a LOT more work to do than just follow the schematic correctly. The build part of an amplifier is generally the easiest part - it's the 'getting it to work properly' part that consumes all of the time.
So learn the basics of what causes hum (and other noise), learn to think in terms of currents-and-loops, and be prepared for some really frustrating evenings when nothing seems to work.
It's like giving birth - once you make that last 'fix' and turn on your system and it is noise-free, all the pain of getting it to that point is forgotten as you bask in the warm sound of something you worked hard to create!
(Then, of course, it's on to your next project - and more hair-pulling... 🙂 )
Sam
One other aspect of grounding I would like to emphasize is safety. Any exposed metal (this includes painted metal or metal control shafts with even a plastic knob) must be securely bonded together and to the 3rd-wire 'earth' ground on your 3-wire plug.
Commercial products are required to have a dedicated BOLT with a star washer (to bite well into the metal chassis) connected to the 3rd-wire on your power cord (doesn't matter if you use an IEC or a dedicated power-cord - securely attach that 3rd wire to chassis with a bolt and nut that you tighten until it screams!).
Further, no other connections to this safety-ground are allowed (to preclude loosening the nut on the safety ground).
I generally put two studs on my chassis in close proximity - one for the dedicated safety ground, the other for my 'star' ground.
There is no reason to compromise safety when it comes to anything that plugs into a wall outlet - it is very cheap insurance, only takes a few extra minutes, and will not (by itself) cause noise or hum problems.
(BTW - 'exposed metal' includes your speaker outputs! Connect the 0V winding of the output side of your output transformer to chassis at your star-ground along with all of your signal and power grounds. You never know when a primary-to-secondary short might develop on your output tranny, and you certainly don't want B+ exposed on the speaker wires!)
Commercial products are required to have a dedicated BOLT with a star washer (to bite well into the metal chassis) connected to the 3rd-wire on your power cord (doesn't matter if you use an IEC or a dedicated power-cord - securely attach that 3rd wire to chassis with a bolt and nut that you tighten until it screams!).
Further, no other connections to this safety-ground are allowed (to preclude loosening the nut on the safety ground).
I generally put two studs on my chassis in close proximity - one for the dedicated safety ground, the other for my 'star' ground.
There is no reason to compromise safety when it comes to anything that plugs into a wall outlet - it is very cheap insurance, only takes a few extra minutes, and will not (by itself) cause noise or hum problems.
(BTW - 'exposed metal' includes your speaker outputs! Connect the 0V winding of the output side of your output transformer to chassis at your star-ground along with all of your signal and power grounds. You never know when a primary-to-secondary short might develop on your output tranny, and you certainly don't want B+ exposed on the speaker wires!)
Lots of excellent stuff already, but I'd just add that the important thing about "ground" is that there ain't no such critter. There's only wires and such. For me, a good exercize is to erase all the "ground" symbols on a schematic and to think about how things are really connected.
All good fortune,
Chris
All good fortune,
Chris
I don't see that what you've described is star grounding. It sounds more like bus grounding.
Without a diagram showing how and where the wires are connected, it's hard to say what it sounds like. Topologies are never "pure" - the verbiage presents a simple example of how things can go wrong with a rationale of why, in that particular case.
Well done and well timed reminder.
I have been banging this drum for years and many refuse to listen.
One Member even accused me of exaggerating.
two plug vs polarized three plug grounding
Old thread I know but I did a search looking for a grounding answer.
I'm sort of converting a pair of Heathkit EA-2s into a pair of UA-2's. (this only additions I"m making are the sheldon stokes PS boards) I already have the front plate tops back. But before I dive into this thing, I'd like to hear from a wiser person about the grounding scheme. Are there any practical safety reasons to change it to a modern three plug design, or are the two plug ground layout safe enough?
Old thread I know but I did a search looking for a grounding answer.
I'm sort of converting a pair of Heathkit EA-2s into a pair of UA-2's. (this only additions I"m making are the sheldon stokes PS boards) I already have the front plate tops back. But before I dive into this thing, I'd like to hear from a wiser person about the grounding scheme. Are there any practical safety reasons to change it to a modern three plug design, or are the two plug ground layout safe enough?
From the safety pint of view you really should use a modern 3 prong mains connector. Just make sure the safety earth is connected to the chassis where the cable enters it.
Cheers
Ian
Cheers
Ian
- Status
- Not open for further replies.