In reading up on ground loops and loop breaking schemes I find there is quite a bit of info on the net, some of it seemingly conflicting. There are recommendations to use the lowest resistance cable shield possible, loop breakers/resistors between signal ground and protective earth, isolation transformers, etc. Bill Whitlock states that ground loop hum and buzz is caused by ground currents inducing voltages across the cable shield resistance, so reducing that resistance is key.
However, I've also seen some suggest connecting a low value resistor (4R7 up to 100R) between the incoming RCA jack shield/return and its connection to ground at the star point, to reduce the ground current and thus decrease the noise. But...doesn't that just increase the effective shield resistance and if so wouldn't that increase the noise???
All POVs are welcome.
However, I've also seen some suggest connecting a low value resistor (4R7 up to 100R) between the incoming RCA jack shield/return and its connection to ground at the star point, to reduce the ground current and thus decrease the noise. But...doesn't that just increase the effective shield resistance and if so wouldn't that increase the noise???
All POVs are welcome.
simply put, some of the apparent conflicts has to do with how to mitigate transmitted noise vs radiated noise.
noise that is transmitted thru AC power supplies, componet power supplies, and power rails are frequently mitigated using virtual grounds, islotation windings etc.
noise that is radiated/recieved from adjacent sources are mitigated by ground lifting, shunt filters, etc.
ground loops in general can cuased by differences in ground references due to resistive changes (cumaltive resistance of all serial lenths voltage or signal has to transverse) or negitaive power factor while compontes are in operation. negative power factor is a reverse inductive type of wave that incoming voltage has to transverse in addition to the resistive nature of the conductive circut. the negative power factor is in pulses as inductive componets like transformers activate. The pulse is in sync to the voltage cycles (60hz, 50hz, etc). These pulses tempoarily add to the ground reference and that cyclic interuption results in the hum or buzz.
This accounts why ground loops percieved to be signal line related disappear when devices are unplugged and allowed to run off store power in their capacitors. If the ground loop was signal line related the loop would remain with the cord unplugged.
The cround loop symtoms disappear when unplugging and allowing devices to run off their capacitors becuase you eleimtated the source of the negative power factor, the cause of the reverse inductive wave, which rasies the ground reference in sync with each ac power cycle. the soruce of the hum or buzz.
noise that is transmitted thru AC power supplies, componet power supplies, and power rails are frequently mitigated using virtual grounds, islotation windings etc.
noise that is radiated/recieved from adjacent sources are mitigated by ground lifting, shunt filters, etc.
ground loops in general can cuased by differences in ground references due to resistive changes (cumaltive resistance of all serial lenths voltage or signal has to transverse) or negitaive power factor while compontes are in operation. negative power factor is a reverse inductive type of wave that incoming voltage has to transverse in addition to the resistive nature of the conductive circut. the negative power factor is in pulses as inductive componets like transformers activate. The pulse is in sync to the voltage cycles (60hz, 50hz, etc). These pulses tempoarily add to the ground reference and that cyclic interuption results in the hum or buzz.
This accounts why ground loops percieved to be signal line related disappear when devices are unplugged and allowed to run off store power in their capacitors. If the ground loop was signal line related the loop would remain with the cord unplugged.
The cround loop symtoms disappear when unplugging and allowing devices to run off their capacitors becuase you eleimtated the source of the negative power factor, the cause of the reverse inductive wave, which rasies the ground reference in sync with each ac power cycle. the soruce of the hum or buzz.
Magz your amps draw so much current they will be at different potential of whatever you attach them to. The potential difference can be reduced by shorter A/C cords of a higher gauge wire. (eer... that should be lower gauge wire)
Also watch for Satellite receivers plugged into your system as they WILL have an outside ground giving your system two grounds.
Solution:
JENSEN TRANSFORMERS, INC. - ISO-MAX® CI-2RR Stereo Audio Input Isolator No more humm.
Also watch for Satellite receivers plugged into your system as they WILL have an outside ground giving your system two grounds.
Solution:
JENSEN TRANSFORMERS, INC. - ISO-MAX® CI-2RR Stereo Audio Input Isolator No more humm.
It's an interesting topic.
I'm noticing a small hum/buzz both 60Hz and harmonics (only audible right next to the speaker) when the Midlife Crisis Amps are connected up to the rest of the system. When I run them with a shorting plug in their inputs they are completely SILENT, even with my ear to the drivers, so I know it's not an internal issue with the amp's noise floor. It's not audible from the listening position 12 feet away, but still if I can mitigate it why not do it.
Now, I have some things I'd like to try to reduce or eliminate the residual hum.
My first candidate is the RF filters on the two soft start boards in each monoblock. These filters are basically just Y-caps from hot and neutral to chassis ground and can be easily disconnected (well, after I get someone to help me lift and carry the amps!). I connected them at assembly in order to help remove RF from the incoming power but I think they are also introducing ground currents into the chassis, and there are TWO of the filters in parallel since I have two boards for the staged start-up, so the problem will be doubly bad if it exists. From my understanding these filters are really meant to prevent the amp from polluting the mains (like a ClassD amp), not for cleaning the mains entering the amp. I should probably have just left them unconnected in the first place, but it's an easy enough fix.
The second candidate is the 8R resistor/.1uF cap network I added to the ground line from the isolated RCA input jack to the star ground (following the suggestion I noted in my first post). The more I thnk about this, the more I think it's just adding to the problem by increasing the resistance of the shield current path - what do you think?
I'm noticing a small hum/buzz both 60Hz and harmonics (only audible right next to the speaker) when the Midlife Crisis Amps are connected up to the rest of the system. When I run them with a shorting plug in their inputs they are completely SILENT, even with my ear to the drivers, so I know it's not an internal issue with the amp's noise floor. It's not audible from the listening position 12 feet away, but still if I can mitigate it why not do it.
Now, I have some things I'd like to try to reduce or eliminate the residual hum.
My first candidate is the RF filters on the two soft start boards in each monoblock. These filters are basically just Y-caps from hot and neutral to chassis ground and can be easily disconnected (well, after I get someone to help me lift and carry the amps!). I connected them at assembly in order to help remove RF from the incoming power but I think they are also introducing ground currents into the chassis, and there are TWO of the filters in parallel since I have two boards for the staged start-up, so the problem will be doubly bad if it exists. From my understanding these filters are really meant to prevent the amp from polluting the mains (like a ClassD amp), not for cleaning the mains entering the amp. I should probably have just left them unconnected in the first place, but it's an easy enough fix.
The second candidate is the 8R resistor/.1uF cap network I added to the ground line from the isolated RCA input jack to the star ground (following the suggestion I noted in my first post). The more I thnk about this, the more I think it's just adding to the problem by increasing the resistance of the shield current path - what do you think?
I'd rather not add something to the signal path, but I'll keep those in mind if the other fixes don't pan out. Thanks.
I recommend the "copy what works" appraoch. When I was struggling with ground loops I studied some circuits and chassis layouts that worked well. For example, go to Audio Note Kits and down load a manual for one of their kits (the use loop breaker of cap parallel resistor). For example, check out the Interstage Monoblocks.
ANK Audiokits - The Authority in High End Audio Kits
ANK Audiokits - The Authority in High End Audio Kits
I've read that and my grounding scheme is what he recommends, except for the resistor/cap on the input jack connection to the star ground. It can jump that out of the circuit and see if it improves.
The RCA jack is connected to the chassis and the resistor is from the jack to the audio input circuit common.
So, you're saying that if I use a resistor from the shield connection of the jack to the circuit common, that the jack must not be isolated from the chassis? That makes sense, actually, since then the jack connection to chassis becomes the single point where circuit ground meets chassis Earth ground.
Therefore, since my input jacks are isolated from the chassis, I cannot add resistance between the input jack shield connection and circuit ground without raising the impedance of the shield and thus increasing hum.
Please verify the above analysis. I only want to lift these things one more time....
The shield of an input coaxial cable is the Signal Return.
It MUST be connected to the Signal Circuit at both ends, to maintain the close coupled Signal Pair.
You can use a twisted pair for this internal Signal Pair. With this scenario, I think no one would consider breaking one end of the Signal Pair.
The Signal Circuit may have an on PCB/board connection to the Power Ground.
If this is made then the reference voltage between the Signal & the Output is established.
If that signal Return to MAG link is not on board then you must make an external single wire connection. BUT !!!!! you must NOT make more than ONE wire connection.
It MUST be connected to the Signal Circuit at both ends, to maintain the close coupled Signal Pair.
You can use a twisted pair for this internal Signal Pair. With this scenario, I think no one would consider breaking one end of the Signal Pair.
The Signal Circuit may have an on PCB/board connection to the Power Ground.
If this is made then the reference voltage between the Signal & the Output is established.
If that signal Return to MAG link is not on board then you must make an external single wire connection. BUT !!!!! you must NOT make more than ONE wire connection.
You have to distinguish between the shield resistance responsible for the signal transmission and the resistance used to give protection against electrical charges etc. The first has to be kept as low as possible, i.e. by using high quality cables and good connectors. The latter is a compromise between protection and ground loop current. Consider sharing of the ground current between the cable shield (e.g. 0.1 Ohm) and the "low value resistor" (e.g. 10 Ohm). These resistors are put in series considering the loop current. In this example the ground current is reduced to 1/100th and thus the voltage "contaminating" the audio signal and noise (hum) is reduced by this factor. Of course the audio signal must be connected directly to the RCA connectors ground and all other wirings must be done properly.
Last edited:
Right. So another way of stating it is that within the input current return path, resistance must be kept low. After that point resistance may be added to decrease ground loop current.
As soon as I can get help moving the amps I'll remove the 8R resistor from the input return path, and I bet that will remove the low hum/buzz. The system sounds very good otherwise!
Thanks for the help thinking it through!
As soon as I can get help moving the amps I'll remove the 8R resistor from the input return path, and I bet that will remove the low hum/buzz. The system sounds very good otherwise!
Thanks for the help thinking it through!
Thanks, Andrew. Other than the resistor in the wrong place, my input wiring is fine (small loop areas, plenty of distance from the PS) since the amp is very quiet until it is connected to another grounded component. Removal of the R should do the trick.
I may keep the RFI filter on the power input connected for the big 2500V power transformer since it has no electrostatic shield between windings, and disconnect the filter on the input/driver stage power transformer, since that transformer has a shield between the windings. That should reduce the ground currents while still helping with RFI.
Just to tie this up.
I removed the ill-advised "loop breaker" resistor and cap from the input ground circuit and disconnected the Y-cap filters from the power inputs. Ground loop hum dropped to basically nothing...a very faint background noise with ear to the driver. No problem whatsoever.
Thanks to all for helping me think it through.
I removed the ill-advised "loop breaker" resistor and cap from the input ground circuit and disconnected the Y-cap filters from the power inputs. Ground loop hum dropped to basically nothing...a very faint background noise with ear to the driver. No problem whatsoever.
Thanks to all for helping me think it through.
- Status
- Not open for further replies.