Gootee and Tauro both make good points on the grounding scheme.
To help understand what's going on, here are some simple sketches. I have not shown the external equipment that is the source of the ground noise, but it doesn't matter. Any equipment connected to the RCA jack that causes hum is simply supplying a small AC current that is using the amplifier to complete a circuit. That circuit consists of 1) the RCA input jack, 2) the amplifier circuitry, and 3) the earth-ground (safety ground) on the AC plug.
Here is what I think is going on.
The original amplifier design had floating RCA input jacks and a connection to chassis at the star-ground. When you connected the external source, noise currents on the RCA jack shield go through the amplifier PCB, through the star-ground and out the power supply plug safety ground.
The hot-pink line in Figure 1 shows the current-path through the amplifier.
The external AC noise disturbs the amplifier PCB ground and is picked up by the high-gain 12AX7 stage, producing hum on the audio output.
Breaking the chassis connection at the star-ground and routing the RCA jack shield to earth-ground supplies a ground path for the external AC noise that bypasses the amplifier.
There is still a ground-loop, and AC currents from the RCA jack shield still flow to the mains-plug earth ground BUT they bypass the amplifier PCB. None of the ground loop current disturbs the amplifier PCB, and the 12AX7 does not pick them up.
The downside of this approach is that while it provides an alternate path for the external ground loop, it may not provide a safe path to earth ground. This is because now the amplifier PCB and power supply rely on the thin twisted-pair wire for safety ground for the amplifier PCB and power supply circuit.
One might think that a good way to remove the ground-loop is just remove the safety ground from chassis. This would be VERY dangerous.
If some kind of fault developed in the amplifier circuit, there would be a risk of exposing a person to dangerous voltages on either the metal chassis or the RCA jack shield.
NOT a good idea! (nor, I might point out, has anybody suggested this!!)
But there IS a way to break the ground-loop path yet maintain safety.
The method is the one pointed out by Tauro in his diagram. I have re-drawn it to show the AC ground loop current path:
You can see, in this diagram there is no ground-loop current path. Under normal operation, the amplifier floats and the currents see no path to ground through the AC mains plug.
No path = no ground loop.
The anti-parallel (back-to-back) diodes provide a path to earth-ground in the case of a catastrophic fault in the amplifier. If the amplifier is at a voltage higher than about 0.7V (either polarity), the diodes conduct and connect the amplifier to earth ground.
The resistor provides an added safety margin (in the case of a diode failure), and the capacitor provides an AC ground to very high (RF) frequencies at all times but does not conduct the lower (50-60 Hz) frequencies.
This is some of the theory behind what we are doing here, at least as I understand it.
As an experiment, and an experiment only, you can float the amplifier from earth ground by removing the earth-ground connection to your star ground.
Be very careful doing this! With the earth ground removed, the whole amplifier is "floating" with respect to earth. While this normally will not be a problem, if there IS a problem (for instance, an improperly insulated transformer) then the RCA jack shield COULD be at elevated voltage. Anybody who touches the RCA jack will then become the path for current to earth-ground - NOT a good thing!
With the star-ground still removed from chassis ground, connect the wires from the RCA jack shield to the earth-ground on the AC mains plug. This was my first proposal, and should provide a path for the ground-loop current that does not introduce hum into the amplifier.
Finally, to insure adequate safety, install the diode/resistor/capacitor network that Tauro proposed. This is a heavy-duty safety circuit, and as Gootee points out be sure to use good, solid connections in this circuit.
Hopefully, all of this will improve your noise situation and in the end give you a good, safe amplifier.
As a side note...
There might be a good argument for using only the diode/resistor/capacitor circuit and not grounding the RCA jacks. "Why have a ground loop within the amplifier at all?"
It is a good point.
I will say that I like to keep any externally exposed metal on anyelectrical equipment tied directly to earth-ground through a good, solid wire connection. This means the RCA jack shields as well as the metal chassis.
True, having ground currents flowing inside the amplifier could radiate hum into components and not solve the hum problem.
However, I believe your ground-loop currents at this time are a conducted problem, not a radiated problem. The radiated problem should be MUCH lower, and may not be noticeable at all. Having the RCA shields tied to earth-ground is the ultimate safety.
Waiting to see the results of this latest batch of experiments
~ Sam
Last edited:
Hi,
rfengineer2013 you are right on the ground loop explanation. I had a hum problem when I built my LM3886 amplifier. I did every suggestion from the net and still I was not able to fix the hum problem until I found the back to back diode. I installed it on my amplifier and the hum immediately when away.
rfengineer2013 you are right on the ground loop explanation. I had a hum problem when I built my LM3886 amplifier. I did every suggestion from the net and still I was not able to fix the hum problem until I found the back to back diode. I installed it on my amplifier and the hum immediately when away.
The only problem with connecting the RCA jack shield to anything other than the ground end of the input resistor is: It makes a lot of "enclosed loop area" between the signal and signal ground pair's paths, a big loop when connected to a source or shorted, which will have currents induced by any nearby time-varying magnetic field, which will induce a voltage across the input resistor, which will go into a high-gain amplifier...
Hi,
Why the use of the shielding cable? Normally what I do it is take two wire and twisted them with a drill that the threads are close each other and never had a problem. Now it is my understanding that the shield should be grounded at the source and leave the other end open. If we apply this ruler then the shield should be ground at the RCA connectors and leave the other end open. NO?
Why the use of the shielding cable? Normally what I do it is take two wire and twisted them with a drill that the threads are close each other and never had a problem. Now it is my understanding that the shield should be grounded at the source and leave the other end open. If we apply this ruler then the shield should be ground at the RCA connectors and leave the other end open. NO?
Finished! It was a long jurney but it seems I'm at the end now and I can enjoy wonderful music coming out of my tube SE amp.
I really appreciate support from this forum with great ideas and recommendation how to reduce hum and buzz. Special thanks to Thomas and Sam who had great ideas and guided me through this journey. It was a great learning experience.
A brief summary - there was a clearly audible 100 Hz hum in my amp.
This was completely reduced by moving grounded CT resistors network in AC heater circuit to uplifted resistor divider network between B+ and GND where I uplifted AC heater resistors 47R from buth sides of 6.3 AC voltage. This voltage divider is 100 k + 20 k while there are 220 n foil and 100 u electrolytic capacitors in parallel with 20 k resistor.
The 100 Hz hum was gone by doing this modification but there appeared a new buzz 50 Hz + some high frequency. What helped to reduce this buzz was:
- replace shielded (coaxial) cable between input RCA jack on the back panel and input terminal on the PCB by twisted pair
- shielding these twisted pair by chassis grounded copper shielding
- connecting 22 n / 630 V capacitors between input RCA socket minus pin to chassis. RCA sockets are isolated from the back panel
- connecting 100 p capacitors between plus and minus of input RCA socket on the back panel
- moving power transformer a bit farer from output RCA sockets on the back panel - it is ~ 2 inches far now
- carefully feeding twisted pairs of input wires between output panel RCA sockets and input temrin on PCB - far from the power transformer, as close to chassis corners as possible
Above mentioned modifications significantly reduced buzz but it still wasn't completely silent.
So I did an experiment with replacing signal source - preamplifier with 3 pole power socket with CD player with just 2 pole power connector - no yellow/green PE wire. And my tube amp was completely silent then.
At the same time Sam has proposed to install "ground breaker" what I did and it is completely silent with both 2 pole as well as with 3 pole power lines sources connected to my tube amp.
This ground breaker is realized using rectifier bridge (AC pins connected together and +/- pins connected together) with parallel 10 R / 5W and 10 n / 1000 V. It is connected between star ground and chassis (before there was a single wire between star ground and chassis).
As I mentioned in the beginning it was a long journey but I learnt a lot of new information that I will use in my future projects.
Once more thanks a lot for all your comments and advices and special thanks to Thomas and Sam.
vcelkamaja
It was a long jurney but it seems I'm at the end now and I can enjoy wonderful music coming out of my tube SE amp.I really appreciate support from this forum with great ideas and recommendation how to reduce hum and buzz. Special thanks to Thomas and Sam who had great ideas and guided me through this journey. It was a great learning experience.
A brief summary - there was a clearly audible 100 Hz hum in my amp.
This was completely reduced by moving grounded CT resistors network in AC heater circuit to uplifted resistor divider network between B+ and GND where I uplifted AC heater resistors 47R from buth sides of 6.3 AC voltage. This voltage divider is 100 k + 20 k while there are 220 n foil and 100 u electrolytic capacitors in parallel with 20 k resistor.
The 100 Hz hum was gone by doing this modification but there appeared a new buzz 50 Hz + some high frequency. What helped to reduce this buzz was:
- replace shielded (coaxial) cable between input RCA jack on the back panel and input terminal on the PCB by twisted pair
- shielding these twisted pair by chassis grounded copper shielding
- connecting 22 n / 630 V capacitors between input RCA socket minus pin to chassis. RCA sockets are isolated from the back panel
- connecting 100 p capacitors between plus and minus of input RCA socket on the back panel
- moving power transformer a bit farer from output RCA sockets on the back panel - it is ~ 2 inches far now
- carefully feeding twisted pairs of input wires between output panel RCA sockets and input temrin on PCB - far from the power transformer, as close to chassis corners as possible
Above mentioned modifications significantly reduced buzz but it still wasn't completely silent.
So I did an experiment with replacing signal source - preamplifier with 3 pole power socket with CD player with just 2 pole power connector - no yellow/green PE wire. And my tube amp was completely silent then.
At the same time Sam has proposed to install "ground breaker" what I did and it is completely silent with both 2 pole as well as with 3 pole power lines sources connected to my tube amp.
This ground breaker is realized using rectifier bridge (AC pins connected together and +/- pins connected together) with parallel 10 R / 5W and 10 n / 1000 V. It is connected between star ground and chassis (before there was a single wire between star ground and chassis).
As I mentioned in the beginning it was a long journey but I learnt a lot of new information that I will use in my future projects.
Once more thanks a lot for all your comments and advices and special thanks to Thomas and Sam.
vcelkamaja
Last edited:
Excellent news, Vcelkamaja! You stuck with it and now you will really enjoy that amplifier!
I have also learned a lot - your hum troubles made me think and go back to basics, and that bit about biasing the heaters was not at all obvious to me when I started following the thread, but made perfect sense when I stopped and thought it through.
Glad to hear you are done with this project - so, how long before you get the itch to build another one??
Congratulations, and happy listening to you and yours
~ Sam
I have also learned a lot - your hum troubles made me think and go back to basics, and that bit about biasing the heaters was not at all obvious to me when I started following the thread, but made perfect sense when I stopped and thought it through.
Glad to hear you are done with this project - so, how long before you get the itch to build another one??
Congratulations, and happy listening to you and yours
~ Sam
This may sound bizarre, but if you were ever inclined, you could short the lower resistor in the heater elevation divider (ie. effectively revert the heater humdinger to going to 0V) and check how influential that particular change is in isolation of all the other changes that have now been made.
Asking for help from the forum members on same issue.
Well, the major difference to the OP's case is that if I shorten the direct inputs - the amp goes almost dead silent, otherwise. ..
The info:
- it is a Japan-made piece (ran through a 120->100V 500VA tranny), the ground pin on mains connector is floating, the ground topology seems a proper star;
- the grounds from fillament windings and bridges come to the 1st filter caps' negative terminal;
- it has switched inputs - direct and variable, if I switch to variable and turn the knob leftmost the hum goes to pretty much low level;
- all the input wires seem shielded cables, layed around the far side edge + in the corners;
- the 6.3 and 12.6 V filaments are twisted everywhere, neither CT on each so they are simply grounded on one end (that end seems same phased between two filament winding);
- the B+ filter is a proper CLCRCRCR;
- all transformers are EI and locatead as far from each other;
- there is no shield layer inside of mains transformer.
The inputs are isolated from the chassis, there are 100K + 100p in parallel to the input, and the 0.27uF soldered between each RCA's groung to the separate thick ground wire which is attached to chassis nearby the inputs. The jumpwire from star-center on 1st cap also goes here. The variable input has the 100K pot instead. Then the wires go to switch -> 2.2uF electrolytic -> the 1 MOhm in parallel -> 5" & 10" shielded wires to the fist stage's grid and the shield is soldered to the stage's local ground.
As I mentioned:
- if I shorten the direct input the amp goes most silent, the huge portion of audible hum is 60 goes away. At this point the residual 120 hum is very little, but if I then attach the additional jumpwire from RCA's ground to 1st (2nd) stage's ground - the 120 goes to almost inaudible;
- it does not make much of difference if the signal source is connected to the direct input or not. When it is - slightly higher level of 60 is present, but again, if I shorten the input - the hum goes to zero;
- the variable input works a bit better: if I have the pot at e.g. 12 the level of 60 is half as much;
- if I turn the power off -the hum stays on same level and decays as the filter capacitors drain.
Please advise where to look at.
I was thinking of the decoupling input transformer, like 600:600 so its secondary would short the input, but. .. would it make any difference? if almost same level of 60 is present with no signal source.
Well, the major difference to the OP's case is that if I shorten the direct inputs - the amp goes almost dead silent, otherwise. ..
The info:
- it is a Japan-made piece (ran through a 120->100V 500VA tranny), the ground pin on mains connector is floating, the ground topology seems a proper star;
- the grounds from fillament windings and bridges come to the 1st filter caps' negative terminal;
- it has switched inputs - direct and variable, if I switch to variable and turn the knob leftmost the hum goes to pretty much low level;
- all the input wires seem shielded cables, layed around the far side edge + in the corners;
- the 6.3 and 12.6 V filaments are twisted everywhere, neither CT on each so they are simply grounded on one end (that end seems same phased between two filament winding);
- the B+ filter is a proper CLCRCRCR;
- all transformers are EI and locatead as far from each other;
- there is no shield layer inside of mains transformer.
The inputs are isolated from the chassis, there are 100K + 100p in parallel to the input, and the 0.27uF soldered between each RCA's groung to the separate thick ground wire which is attached to chassis nearby the inputs. The jumpwire from star-center on 1st cap also goes here. The variable input has the 100K pot instead. Then the wires go to switch -> 2.2uF electrolytic -> the 1 MOhm in parallel -> 5" & 10" shielded wires to the fist stage's grid and the shield is soldered to the stage's local ground.
As I mentioned:
- if I shorten the direct input the amp goes most silent, the huge portion of audible hum is 60 goes away. At this point the residual 120 hum is very little, but if I then attach the additional jumpwire from RCA's ground to 1st (2nd) stage's ground - the 120 goes to almost inaudible;
- it does not make much of difference if the signal source is connected to the direct input or not. When it is - slightly higher level of 60 is present, but again, if I shorten the input - the hum goes to zero;
- the variable input works a bit better: if I have the pot at e.g. 12 the level of 60 is half as much;
- if I turn the power off -the hum stays on same level and decays as the filter capacitors drain.
Please advise where to look at.
I was thinking of the decoupling input transformer, like 600:600 so its secondary would short the input, but. .. would it make any difference? if almost same level of 60 is present with no signal source.
Last edited:
I know this is an old thread but just a quick mention that I had a similar problem and it turned out to be mea culpa! I had modded the amp by adding a headphone jack, and the jack ground was not isolated from the chassis, forming a ground loop. Switching to a mostly-plastic headphone jack eliminated the hum in my case.
Long time ago one happy diyer finished his tube amp and with a great support from this forum resolved the issue with hum noise
Well, I changed my speakers recently to a high sensitive horns and what seemed to be almost silent before is more audible now. With my old speaker having around 90 dB sensitivity I can hear the hum when I put my ear as close as ~ 10 centimeters to the speaker. With the new horn speaker with 96-98 dB sensitivity I can hear it also in my listening position which is ~ 1.5 m in front of the speakers. It seems that it is 100 Hz (checked with an iPhone microphone / spectrum analyser app). Is it something to accept with high sensitivity speakers or should I think about some additional fine tuning or optimization?
Actually what I recognized is that sometimes during nigh listening I can also hear the sound -hum coming from the amplifier itself - I think it must be a “mechanical” noise generated by power toroidal transformer inside the chassis. Not sure if mybe “dampening” of the trafo or additioanal shielding of the trafo grounded to the chassis’s earth would help here...
Well, I changed my speakers recently to a high sensitive horns and what seemed to be almost silent before is more audible now. With my old speaker having around 90 dB sensitivity I can hear the hum when I put my ear as close as ~ 10 centimeters to the speaker. With the new horn speaker with 96-98 dB sensitivity I can hear it also in my listening position which is ~ 1.5 m in front of the speakers. It seems that it is 100 Hz (checked with an iPhone microphone / spectrum analyser app). Is it something to accept with high sensitivity speakers or should I think about some additional fine tuning or optimization?
Actually what I recognized is that sometimes during nigh listening I can also hear the sound -hum coming from the amplifier itself - I think it must be a “mechanical” noise generated by power toroidal transformer inside the chassis. Not sure if mybe “dampening” of the trafo or additioanal shielding of the trafo grounded to the chassis’s earth would help here...
Finished!
I really appreciate support from this forum with great ideas and recommendation how to reduce hum and buzz. Special thanks to Thomas and Sam who had great ideas and guided me through this journey. It was a great learning experience.
A brief summary - there was a clearly audible 100 Hz hum in my amp.
This was completely reduced by moving grounded CT resistors network in AC heater circuit to uplifted resistor divider network between B+ and GND where I uplifted AC heater resistors 47R from buth sides of 6.3 AC voltage. This voltage divider is 100 k + 20 k while there are 220 n foil and 100 u electrolytic capacitors in parallel with 20 k resistor.
The 100 Hz hum was gone by doing this modification but there appeared a new buzz 50 Hz + some high frequency. What helped to reduce this buzz was:
- replace shielded (coaxial) cable between input RCA jack on the back panel and input terminal on the PCB by twisted pair
- shielding these twisted pair by chassis grounded copper shielding
- connecting 22 n / 630 V capacitors between input RCA socket minus pin to chassis. RCA sockets are isolated from the back panel
- connecting 100 p capacitors between plus and minus of input RCA socket on the back panel
- moving power transformer a bit farer from output RCA sockets on the back panel - it is ~ 2 inches far now
- carefully feeding twisted pairs of input wires between output panel RCA sockets and input temrin on PCB - far from the power transformer, as close to chassis corners as possible
Above mentioned modifications significantly reduced buzz but it still wasn't completely silent.
So I did an experiment with replacing signal source - preamplifier with 3 pole power socket with CD player with just 2 pole power connector - no yellow/green PE wire. And my tube amp was completely silent then.
At the same time Sam has proposed to install "ground breaker" what I did and it is completely silent with both 2 pole as well as with 3 pole power lines sources connected to my tube amp.
This ground breaker is realized using rectifier bridge (AC pins connected together and +/- pins connected together) with parallel 10 R / 5W and 10 n / 1000 V. It is connected between star ground and chassis (before there was a single wire between star ground and chassis).
As I mentioned in the beginning it was a long journey but I learnt a lot of new information that I will use in my future projects.
Once more thanks a lot for all your comments and advices and special thanks to Thomas and Sam.
vcelkamaja
Last edited:
Actually While trying to identify the source of hum I tried:
- shorten input - no change, audible hum
- remove input 12ax7 tube - no change, audible hum
- remove output 6l6gc tube - no hum, silent
- switch speaker out to headphone output (with both tubes installed) - no hum, silent
So it seems there is no issues with input circuit or output wiring - rather the output 6l6gc tube?
One more question - is it realostoc that DC heater voltage (regulated) will reduce hum, compared with AC heating?
- shorten input - no change, audible hum
- remove input 12ax7 tube - no change, audible hum
- remove output 6l6gc tube - no hum, silent
- switch speaker out to headphone output (with both tubes installed) - no hum, silent
So it seems there is no issues with input circuit or output wiring - rather the output 6l6gc tube?
One more question - is it realostoc that DC heater voltage (regulated) will reduce hum, compared with AC heating?
… What he said … if you have access to an oscilloscope, it'd be good to take a look at that B+ line, while the 6L6 tubes are in place. If the B+ is 'hummy', it'll be immediately obvious.
-= GoatGuy ✓ =-
Consider dropping $65 into this ... Hantek 6022BE PC-Based USB Digital Oscilloscope 48MSa/s 20MHz 2 Channels US 612046488263 | eBay
Last edited:
Most cases of hum or power line related buzz comes from poor wiring practice or layout. These are hard to diagnose over the internet, but you could get better advice if you were to include some detail about signal grounding.
In fact, I'd venture to guess that most noise issues could be solved by the builder if a truer schematic were drawn, one that has NO ground symbols, only actual wiring. It should really be done, mentally after some experience, or literally when learning, before construction.
All good fortune,
Chris
In fact, I'd venture to guess that most noise issues could be solved by the builder if a truer schematic were drawn, one that has NO ground symbols, only actual wiring. It should really be done, mentally after some experience, or literally when learning, before construction.
All good fortune,
Chris
Thanks for all your notes and comments. Unfortunately the design is realized on PCB which I can’t change (commercial kit). I twisted heating AC wires as much as possible but no change. Therefore I thought about trying to use DC heating voltage.
B+ hum / ripple AC voltage -how do I check if the hum is coming from B+ or from heating AC? I’m a bit afraid to connect oscilloscope to B+ high voltage
What options I have if it is B+? Better filtering - bigger caps, add inductor (CLC)? Is twisting of high voltage as critical as it is with heating wires? (There are both heating as well as high voltage anode AC wires connected between the transformer and the main PCB)
B+ hum / ripple AC voltage -how do I check if the hum is coming from B+ or from heating AC? I’m a bit afraid to connect oscilloscope to B+ high voltage
What options I have if it is B+? Better filtering - bigger caps, add inductor (CLC)? Is twisting of high voltage as critical as it is with heating wires? (There are both heating as well as high voltage anode AC wires connected between the transformer and the main PCB)
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.