| ianpengelly |
This applies to people using the Chipamp.com Dual Mono PCB boards and possibly others using a similar arrangement.
I have been pulling my hair out trying to solve the hum problem from my amp, using large csa copper wires to connect the chassis ground points and trying input capacitors... not to mention watching every thread on this website hoping an answer would become clear. I had my amp set up as shown below. |
|
|
| ianpengelly |
Listening to the hum the amp produced is seemed all too similar to that of the transformer which got me thinking. I had all the same issues that other people had, connect one channel to the pre-amp and all was fine, connect the second and it all got noisy. This indicated that there was some interaction between the channels and the only two points of connection are the chassis ground wires, which when disconnected from each other worsen the hum and the connection through the power supply.
So with the amp unplugged I disconnected the chassis grounds and measured the resistance between the two RCAs to find that there was approximately 5M Ohms between them, confirming my suspision that the transformer noise was what I was hearing.
To validate this theory I connected both amps to a single rectfier board and voila, hum gone! :smash:
So there are two solutions to this problem, one is to use a single rectifier board and the other is to use the amps in true dual mono and use two transformers, one for each channel! |
|
|
| ianpengelly |
If you have one of these kits or you are running two rectifier boards from a single transformer one of these approaches should solve your hum problem.
Vikash, if you could get hold of some nice potted 160VA transformers the would sell like hotcakes with the dual mono chipamp boards you sell (I'll be buying another 300VA one from you in the very near future.... nothing like overkill! :devilr: )
I hope this helps a few of you that have been stuggling with this problem for a while.
I also look forward to hearing other people's experience of trying this solution and others. |
|
|
| Nordic |
It looks like a groundloop... I seee no ground disconnect network between ground and mains earth....
Do you only get the hum with the preamp connected?
Can you confirm your RCA sockets are isolated from the case.... i.e. no continuity... |
|
|
| ianpengelly |
The problem exists whether the chassis grounds are connected to earth or not.
There is some noise when the preamp is completely disconnected, although this goes if the rca is shorted.
The RCAs are indeed isolated from the chassis. Though they are effectively connected when the chassis grounds are connected to earth, as indicated in the diagram. |
|
|
| raromachine |
| Are you using sheilded cable from the RCA's to the amp? |
|
|
| ianpengelly |
| No, just a pair of twisted silver coated copper cable. |
|
|
| ianpengelly |
Just to clarify I am posting a solution, I have spent a lot of time reading through threads on this site which have suggested some extreme measures to eliminate the hum problem, including suggesting that the boards are badly designed and that you need to cut some of the traces. Now I am not being funny, but if you had just shelled out a load of cash on some PCBs it would be nice to have a solution that didn't involve hacking them to death in a potentially vain search for a solution that is relatively easy.
I think what is perhaps missing from the chipamp.com installation manual for these kits is to point out a suggested implementation to ensure interaction free operation, i.e. just use one rectifier for a pair of amp boards or use a transformer for each rectifier board, i.e. one per channel. |
|
|
| Stuey |
Ian,
I bought the stereo boards from Chipamp (single rectifier board) and paid no special attention to earthing, and have no noise problems at all.
In fact, contrary to most advice on here, my chassis grounds are joined together and earthed to the main chassis case, about 15cm from the mains (safety) earth. The chassis is a simple sheet aluminium affair I knocked up in a couple of hours. Nothing else is earthed to the chassis except the body of the pot, via its attachment nut.
In fact, there IS noise if I leave the chassis grounds joined, but remove them from the connector on the chassis.
My implementation is most basic, with simply a pot on the inputs. If I turn the pot up full, there is the faintest hiss.
FYI, anyway...
Stuey |
|
|
| ianpengelly |
Stuey,
Thanks for the post, you have confirmed one of my findings that the use of a single rectifier board works just fine and that the problem has nothing to do with the chassis ground and its connection to the safety earth.
Ian |
|
|
| CarlosT |
The main noise plaguing my IGC comes when I connect the RCA interconnect to my laptop.
My earth ground is connected to the chassis for safety but the rest of the signal and amp grounds are isolated from the chassis and earth. |
|
|
| ianpengelly |
| I used to have problems with my computer and my amp, possibly caused by another component in the system, but I could measure 100V between the 3.5" jack and the computer chassis. You could literally see a spark and feel it if you were touching both of them! :hot: |
|
|
| okapi |
this thread is timely for me as i am in the process of rebuilding my first chip amp and was considering using a second rectifier board.
i am motivated to try a second rectifier as i have read at least a few times on this forum that it can improve channel separation. Limited channel separation is definitely an issue with a single board (at least compared to my monoblocks).
prior to the rebuild the amp has been dead quiet with absolutely no attention paid to wire routing or grounding.
i will try using two rectifier boards and will post my results. |
|
|
| ianpengelly |
Sounds good, I'll look forward to the result.
I don't suppose you have a couple of transformers so you can test if it's still noise free in this arrangement? |
|
|
| impsick |
| i too have a steareo amp, 1 toroidal, 1 power supply. As for me all my grounds meet at one point on the chassis. with the pot cranked to max there is hardly a hiss. |
|
|
| quickshift |
I hope you don't mind me drawing over your diagram :)
After a few experiments, this is the grounding scheme I use for my 4ch GC - it runs off a single 500VA and uses 4 rectifier boards. Only 2 chanels shown but you get the idea.
No hum, no hiss, no turn on/off thump either :) |
|
|
| AndrewT |
Hi Quick,
I hope you are joking.
The RCA grounds sharing the speaker grounds seems, just wrong.
Is that what you found worked? |
|
|
| quickshift |
Hi AndrewT,
Yup, that's what has worked. I've used the same scheme on both my amps (I took the top off one to double check before posting the diagram)
It's the only scheme I tried that gave zero hum. I disconnected one 0V or ground at a time - wired them to the chassis ground, tested for hum and so on until I ended up with everything connected to the main chassis ground and no hum.
The circuit is roughly as shown on the attachment - I've used the optional cap Ci on both amps.
|
|
|
| AndrewT |
Hi Quick,| quote: | | It's the only scheme I tried that gave zero hum. I disconnected one 0V or ground at a time - wired them to the chassis ground, tested for hum and so on until I ended up with everything connected to the main chassis ground and no hum. |
this is different from your diagram.
Your diagram shows the RCAs returns connected to the speaker returns and not to the chassis ground. |
|
|
| quickshift |
Hi AndrewT,
The NS circuit diagram is just what the amps are based on component wise, both the speaker returns and RCA barrels are connected to the main chassis earth in the actual amps as in the diagram I drew over. |
|
|
| ianpengelly |
| It maybe that doing part of your suggestion Quickshift may work, i.e. binging all the 0V connections together at a single point then feed out to the boards, but keep the RCA and speaker boards connecting to their respective positions on the board. I might have to experiment, though I suspect I am still going to with a dual transformer option in the long run. |
|
|
| halo0925 |
Most people make more out of the hum then is really needed, if it is 60hz hum and I think most of us are familiar with this sound it can only be one thing A GROUND LOOP.
Most people connect there amp board ground to the mains ground :dead: this will almost always provide a path for variable resistance in the system even if no other component is earthed.
The boards are well designed and simple both Brians and Peters you dont have to hack traces, the idea behind dual secoundarys is to issolate against 60HZ feed back resistance paths.
1- break the earth ground via 2 diodes a resistor and cap Rod Elliotts site has a simple circuit for this.
2- dont inter connect the center points of both supplys, this will negate all benifits that might be derived from the dual secoudary/ bridge supply. Unless useing only one supply board for both channels.
3- issolate the input grounds from the system grounds, this is the culprit 90% of the time, this is very true if another component is earth grounded, most earth grounded equipment will have a ground lift. lift via a 1 to 3 ohm resistor on the RCA grounds and do not combine channels.
4- Make sure all wireing is neet and away from the trans. and AC power in.:D |
|
|
| AndrewT |
Hi Ian,
going to two transformer, or quad secondary windings on a single transformer, will allow the two channels to have an independant floating audio ground. Both these floating grounds must be kept separate if you are to avoid an earth loop with connected mains ancilliaries.
However, the floating audio grounds must be made safe and this demands that they are connected to safety earth.
You will need a disconnecting network for each channel to make those safety connections.
An alternative may be Bob Cordell's version using a loop breaking resistor in the RCA return link. |
|
|
| ianpengelly |
Halo, we have the joys of 50Hz hum here! And it is a big deal when people have spent a lot of time and money trying to solve the problem by all sorts of means and there are relatively simple answers.
I agree with you that the boards are well designed.
There is a loop issue, but it has nothing to do with the safety earth connection, I have tried removing the chassis ground connections to the safety ground for testing purposes and it made no difference (other than stopping a few pops being caused by other appliances in the house, such as the fridge). Also none of my other components have safety earths, so there can't be a loop through the safety earth wiring.
The problem lies with using two rectifier boards with a single transformer that only has 2 sets of secondary windings. As AndrewT has pointed out a quad secondary transformer would be one answer, the second is to use two transformers and the final is to use a single rectifier board. I have yet to try binding the power grounds from the two rectifier boards yet, so can't comment on whether this is a valid solution. It may slightly negate the cross talk benefit of them being separate for the two channels, but you have got twice the current handling and twice the reservoir size by this method.
Hopefully this week or next weekend I can try a couple more configurations.
Andrew, thanks for the tips I'll look up those disconnecting networks and get them installed asap, so that any possible issues with this can be eliminated. |
|
|
| jimbo1968 |
I've had the same problems but haven't got as far as tying discinnecting networks. I'm using dual trafos for the moment, but it would be handy if someone managed to solve the problem for a single trafo, dual PS configuration.
When I finish my main project amps soon, they will be dual PS single trafo but only 1 input channel (active crossover), which I'm hoping won't have the same problem? But maybe I'm being too hopeful.
Stereo or proper dual mono seems to be the safe route to go but 300VA per LM3886 seems extravagant. |
|
|
| ianpengelly |
Klass,
Good find! Its basically agreeing in part with quickshifts suggestion and one of the ideas I wanted to try.
Jimbo, it would be interesting to know if you suffer from the same problem if you are using an active crossover in the amp case, hopefully if you can keep the paths from the input to the active crossover short you might not have the same problem that we (or is it just me?) are suffering from. |
|
|
| CarlosT |
Yes...but in that solution, Peter is basically recommending keeping it to one recitifier board.
What happens when you have two trannies and two rectifier boards? Why would there be a need for an interconnect between the two amp stargrounds? |
|
|
| AndrewT |
| quote: | Originally posted by CarlosT
......What happens when you have two trannies and two rectifier boards? Why would there be a need for an interconnect between the two amp stargrounds? |
| quote: | | However, the floating audio grounds must be made safe and this demands that they are connected to safety earth. |
two floating audio grounds each with their own input RCA and output speaker return terminal are a health hazard.
Both must be made safe by connecting them to safety earth in event of an internal mains fault making these (and other) components "LIVE". |
|
|
| CarlosT |
AndrewT:
I know well how you feel about ground earth safety...even if I don't agree 100%.
I thought the focus of this discussion was noise though... |
|
|
| kvholio |
| quote: | | I thought the focus of this discussion was noise though... |
Yes, but only if the solutions proposed meet basic safety-requirements.
Please try to remember there are a lot of readers you are unaware of... we want them all to eventually grow old listening to music ;)
Klaas |
|
|
| halo0925 |
| I am trying to think when the last time I bought a amplifier except for commercial, that had the center amp ground connected to the earth, hell they dont even have a 3 prong plug just live and nuetral. |
|
|
| kvholio |
Is it safe to assume a potential reader is aware of the safety-standards that have to be met to safely build a piece of equipment without a connection to safety-earth ?
I dont think so.
Klaas |
|
|
| AndrewT |
Hi,
I have said it before and I repeat,
I don't know how to design double insulated (Class11) equipment.
I do know, at least in part, the risks involved in assembling conventional mains operated (Class1) equipment that is necessarily protected by the third earthing pin.
This third pin must effectively keep all exposed conductive parts at a low enough (safe) voltage until the fuse breaks the supply. |
|
|
| ianpengelly |
| quote: | Originally posted by CarlosT
Yes...but in that solution, Peter is basically recommending keeping it to one recitifier board.
What happens when you have two trannies and two rectifier boards? Why would there be a need for an interconnect between the two amp stargrounds? |
I think what would happen is that you would connect the chassis ground from each board to the safety earth (as Andrew previously suggested) either directly or via 2 disconnecting networks, one for each channel.
I agree that we need to ensure that the safety of the user is maintained, but I think what we are also saying is that the connection to the safety earth does not form part of the problem, so however it is connected, as long as it IS connected its fine. |
|
|
| CarlosT |
| quote: | Originally posted by AndrewT
Hi,
I have said it before and I repeat,
I don't know how to design double insulated (Class11) equipment... |
AndrewT:
As I stare at the innards of my Adcom GFA535II amp which just has a two-prong plug, I don't see or understand how any of it is "double-insulated."
What is "double-insulated" anyway? Most commercial amp chassis I know of are metal with a metal backplane where all the metal connectors are just mounted to...
When I go to Sears and pick up a "double-insulated" power tool, I see that all I'm touching is made of plastic...no extraneous exposed metal. How could you achieve this in audio amp design? What's the big deal anyway?
If we tried to reduce all exposure to hazzards, we'd just walk around in giant body-sized condoms...that'd take care of disease and electrocution :D |
|
|
| AndrewT |
Hi Carlos,
to reduce our exposure to avoidable hazard, all one has to do is follow the regulations set out to protect us.
Ignore them at your peril.
Why ask me about Class11 devices when I have told you I have no expertise in that field? |
|
|
| ianpengelly |
Carlos, I found this which might help:
Class 11, double insulated, is simply the use of two layers of insulation for protection. Internal insulation separates mains conductors from one another, an outer layer of insulation then acts as a barrier to outside contact. For a shock hazard to occur, both layers must be breached. |
|
|
| sek |
Sorry to chime in, but I was directed here by chance.
Isn't it Roman numbers?! Like Class I and Class II as in Class 1 and Class 2? Note that this question is rhetorical! :D
I didn't follow the thread but the difference between those two classes is actually straight forward.
In Class 1 equipment, every conductive (i.e. metal) part that could possibly maybe be touched by a user (i.e. enclosure and connectors) has to be connected to mains protective earth (PE), the green-yellow wire that someplace has a different color but the same purpose. :D
In Class 2 equipment, all mains related circuitry has to have double insulation between life or neutral and anything else in the unit, even internal circuitry right next to the mains wiring. Double insulation is achieved both through creepage (physical distance) and coverage (appropriate materials wrapped around).
Class 1 is achievable for the DIY enthusiast. That's why every project site recommends it. If your unit fails, it fails safe - unless you skimped on the PE wiring.
Class 2 requires every single component that has to do with the mains (primary) side to be certified for class 2 usage by regulation and safety authorities. Compliant wire, compliant board connectors, compliant board layout techniques, compliant transformer, etc. That's why no DIY project site recommends it. It's only reasonable for someone who has training and/or experience in that field.
Rest assured that a DIY product of the average hobbyist would fail the first couple of times of UL, IEC or VDE testing in a jiffy. That's why nobody tries: You are responsible for what you switch on and leave running. And when your house burns down or someone dies, you're still responsible.
In short: we want our amplifiers and sources to be class I (one) devices, because we can. If it gives us headaches because we mess up the ground layout, we improve upon it. If we get into the realm of multiple homegrown class I devices humming together to no avail, we convert to balannced interconnects! :D
Should we at some point feel the need to mess with class II (two) equipment, we first make shure we know exactly what serves which purpose - and we replace broken parts with certified and approved parts only.
| quote: | | What's the big deal anyway? |
You're dead if you do it wrong. If you do it wrong and survive out of luck, it'll be your kids who die. Or your kitty... You get my point, Carlos! It's not funny.
| quote: | | I don't see or understand how any of it is "double-insulated. |
Read up on it , it's not too difficult to learn.
Cheers,
Sebastian. |
|
|
| ianpengelly |
Andrew,
Just looking through at your suggestions for a disconnecting network, they are one or more of the following:
10r power resistor.
power bridge diode wired in inverse parallel.
high frequency capacitor (10nF to 100nF ceramic).
Can you just give me a little clarification, would you class a power resistor as 2W or more?
A power bridge diode as handling how many amps?
I would assume the arrangement is as shown in the attached diagram. |
|
|
| AndrewT |
Hi Ian,
the power resistor theoretically should never carry more than about a volt.
But ,during a failure mode the diodes may survive long enough with kA through them to develop a lot more than 1V, could we be approaching 10Vpk? that would be 10Wpk.
A wire wound can survive a much higher short term peak than any of the film types. It is survivabilty in the very short term that matters and even then only once for that fuse blowing incident.
I suppose you could deliberately set up a test with mains direct to audio ground (no PSU or amplifier inside) and close the case up. Stand back and switch on.
Tell us what happens and which parts survive the test and which are burnt out/test abnormally. |
|
|
| Nordic |
| I have seen it burn out even 1W carbon types... I only use flameproof wirewounds for that now... |
|
|
| ianpengelly |
Some progress!
First up I have now sucessfully moved over to a two transformer set up, so I have a transformer per channel now, i.e. true dual mono. This setup has resulted in no hum, as did the setup using one rectifier board to supply the two channels when using a single trasnformer. I feel pleased that I have proved a theory and hope that others can use this advice when they design their amps or find it useful in solving an exisiting issue.
I have also had a chance to install a disconnecting network as discussed above and, whilst this had no bearing on the hum issue, what it has done is surpressed noise (such as pops and cracks) from other electrical equipment in my household, i.e. the fridge, boiler and light switches.
All in all it has been an interesting learning experience and no doubt I'll get some more knowledge when I upgrade my pre-amp to include a buffer stage.
Thanks to all who have contributed to this thread. :D |
|
|
| mourip |
| quote: | Originally posted by ianpengelly
Some progress!
First up I have now sucessfully moved over to a two transformer set up, so I have a transformer per channel now, i.e. true dual mono. This setup has resulted in no hum, as did the setup using one rectifier board to supply the two channels when using a single trasnformer. |
Did you notice any audible difference between the single tranny/single rectifier and dual tranny/dual rectifier setups? |
|
|
|
