Hello all,
I have this LM3886 amplifier with integrated NE5534 preamp and power supply. This is what it looks like:
I have been using this for a while with an old non-toroidal transformer, without a case. I did not take care to ground it, as it was without a case (I should have). I never had problems with noise.
Since I like it a lot, I decided to purchase a nice toroidal transformer and an aluminum case to put everything in. However, I am faced with trouble when trying to connect it to my sound source. The problem seems to be related to grounding. I hear the term ground loop noise but I am not sure if this is the problem I am dealing with.
The toroidal transformer is 2x30V, 160VA power. I have joined the primary coils and insulated the connection. I also connected the secondary coils and got a center tap to connect to the middle socket of the power supply input. Obviously, this is the three prong terminal at the lower left of the board. Speakers are connected to the right side of the board.
Now when it comes to choosing a ground point to connect to the case, the manufacturer recommended the bottom left hole. Indeed there is a visible trace there that makes that point electrically very close to the grounded sides of the two big capacitors. I had soldered a lead there before I asked him the question, so I left it there considering it ok:
Now, as to how I am connecting to source, I use these screw type RCA terminals with the 3-pin connector supplied with the board. You will notice it is currently outside the aluminum case as the noise problem I am about to describe is so severe and not affected by having the board inside or outside the case.
So, what I am dealing with, is noise. Again, I mention that I never had this problem before, because it was outside a case and things did not lead me to consider grounding it with the IEC inlet ground. The thing is, if I leave it ungrounded, there is just a slight noise from EMI because it's not in a case. The moment I connect the IEC ground lead to the ground point show above, I get a lot of noise.
I am not sure it is EMI noise. Because the amp board is connected to an audio interface and then electrically with a computer, I 'm dealing with noise that seems to originate from the computer. Essentially I am hearing all the electrical activity of the computer as quite loud noise. You can imagine the volume levels of listening to some relatively low to mid volume calm music, only it isn't calm. By electrical activity of the computer I mean for example, I can hear the wireless receiver of my mouse when I move it, or a difference in noise when I turn on the computer keyboard LEDs on/off (on means more noise). Again, this is much louder than say, some residual EMI of a fairly well put together setup.
The moment I disconnect the IEC inlet ground from the board, the noise goes away, and there's only very low residual noise, I presume EMI. Also, to be noted, if I connect my mobile phone at the input, I can ground the board via the IEC inlet ground with no probems whatsoever. Obviously the phone operates from a battery, and is floating without a ground.
Can you firstly help me identify the type of problem here? Is it a so called ground loop noise issue? Both the computer and the amp are on the same multi socket inlet connected to mains.
Obviously, I am still expecting various parts to complete my amp, most importantly proper RCA terminals to connect to the back panel. But this thing is so severe that I think it requires immediate investigation since it happens both inside and also outside the case.
Are there things I should make sure before starting to troubleshoot this? And how would I go about doing that and hopefully fixing this? It is so prominent an issue that it must be solvable.
Many thanks for reading, any pointers/help are more than welcome.
I have this LM3886 amplifier with integrated NE5534 preamp and power supply. This is what it looks like:
I have been using this for a while with an old non-toroidal transformer, without a case. I did not take care to ground it, as it was without a case (I should have). I never had problems with noise.
Since I like it a lot, I decided to purchase a nice toroidal transformer and an aluminum case to put everything in. However, I am faced with trouble when trying to connect it to my sound source. The problem seems to be related to grounding. I hear the term ground loop noise but I am not sure if this is the problem I am dealing with.
The toroidal transformer is 2x30V, 160VA power. I have joined the primary coils and insulated the connection. I also connected the secondary coils and got a center tap to connect to the middle socket of the power supply input. Obviously, this is the three prong terminal at the lower left of the board. Speakers are connected to the right side of the board.
Now when it comes to choosing a ground point to connect to the case, the manufacturer recommended the bottom left hole. Indeed there is a visible trace there that makes that point electrically very close to the grounded sides of the two big capacitors. I had soldered a lead there before I asked him the question, so I left it there considering it ok:
Now, as to how I am connecting to source, I use these screw type RCA terminals with the 3-pin connector supplied with the board. You will notice it is currently outside the aluminum case as the noise problem I am about to describe is so severe and not affected by having the board inside or outside the case.
So, what I am dealing with, is noise. Again, I mention that I never had this problem before, because it was outside a case and things did not lead me to consider grounding it with the IEC inlet ground. The thing is, if I leave it ungrounded, there is just a slight noise from EMI because it's not in a case. The moment I connect the IEC ground lead to the ground point show above, I get a lot of noise.
I am not sure it is EMI noise. Because the amp board is connected to an audio interface and then electrically with a computer, I 'm dealing with noise that seems to originate from the computer. Essentially I am hearing all the electrical activity of the computer as quite loud noise. You can imagine the volume levels of listening to some relatively low to mid volume calm music, only it isn't calm. By electrical activity of the computer I mean for example, I can hear the wireless receiver of my mouse when I move it, or a difference in noise when I turn on the computer keyboard LEDs on/off (on means more noise). Again, this is much louder than say, some residual EMI of a fairly well put together setup.
The moment I disconnect the IEC inlet ground from the board, the noise goes away, and there's only very low residual noise, I presume EMI. Also, to be noted, if I connect my mobile phone at the input, I can ground the board via the IEC inlet ground with no probems whatsoever. Obviously the phone operates from a battery, and is floating without a ground.
Can you firstly help me identify the type of problem here? Is it a so called ground loop noise issue? Both the computer and the amp are on the same multi socket inlet connected to mains.
Obviously, I am still expecting various parts to complete my amp, most importantly proper RCA terminals to connect to the back panel. But this thing is so severe that I think it requires immediate investigation since it happens both inside and also outside the case.
Are there things I should make sure before starting to troubleshoot this? And how would I go about doing that and hopefully fixing this? It is so prominent an issue that it must be solvable.
Many thanks for reading, any pointers/help are more than welcome.
Perhaps you will find some useful tip here, its from a member here.
https://neurochrome.com/pages/grounding
https://neurochrome.com/pages/grounding
Your choice of the place for attaching the ground wire to chassis looks to be the cause of your troubles. I'd try attaching the brown wire (second pic) at the input ground terminal - that's the one going to the red wire in your last pic.
2x30 VAC will result in about ±40 V under load at nominal mains voltage. But this assumes that the transformer actually outputs 2x30 V under those conditions. Most will output more as they're designed to provide the specified voltage under full load. So you could have more like ±45 V on the supply. That's beyond the ±42 V spec of the LM3886 so you're definitely playing with fire there.
I generally recommend 2x22 VAC, 160-200 VA for a stereo LM3886 amp. That'll result in ±30 V rectified which is just right for the LM3886. You can read more about that here: https://neurochrome.com/pages/output-power
This spot would be better for the ground reference because the charging currents that flow through the power supply capacitors should ideally cancel there. I doubt that's your issue, but it wouldn't hurt to move the wire over.
The manufacturer should have soldered the mounting pin for the heat sinks. The rectifiers (or regulators) will develop bad solder joints pretty quickly if you don't. I'd give the mounting pins a dab of epoxy.
You will always have ground loops. If nothing else, you'll have a loop through the RCA cable that connects the input of the amp to the source. The trick is to 'encourage' the ground current to flow in the power cord instead of the RCA cable. If you haven't already, ensure that the RCA connectors are isolated from the chassis.
Do you have pictures of the amp as it is installed in the chassis?
Tom
I generally recommend 2x22 VAC, 160-200 VA for a stereo LM3886 amp. That'll result in ±30 V rectified which is just right for the LM3886. You can read more about that here: https://neurochrome.com/pages/output-power
This spot would be better for the ground reference because the charging currents that flow through the power supply capacitors should ideally cancel there. I doubt that's your issue, but it wouldn't hurt to move the wire over.
The manufacturer should have soldered the mounting pin for the heat sinks. The rectifiers (or regulators) will develop bad solder joints pretty quickly if you don't. I'd give the mounting pins a dab of epoxy.
You will always have ground loops. If nothing else, you'll have a loop through the RCA cable that connects the input of the amp to the source. The trick is to 'encourage' the ground current to flow in the power cord instead of the RCA cable. If you haven't already, ensure that the RCA connectors are isolated from the chassis.
Do you have pictures of the amp as it is installed in the chassis?
Tom
I presume "playing with fire" was just a random choice of words, right? 😉
Thank you guys for the warning. Indeed my post rectification voltage is +-45VDC so definitely in the danger zone, especially if we're talking long term use.
Is there a practical way I can lower the voltage at the toroidal or just before reaching the LM3886? Although I suspect the answer would be something like "just get the correct toroidal'.
Thank you guys for the warning. Indeed my post rectification voltage is +-45VDC so definitely in the danger zone, especially if we're talking long term use.
Is there a practical way I can lower the voltage at the toroidal or just before reaching the LM3886? Although I suspect the answer would be something like "just get the correct toroidal'.
@NanoFarad thanks for your suggestion. Funny thing is, my speakers specifications keep this a mystery: "4-8Ω". Should you care to have a look what I want to drive with my amp:
https://www.magnat.de/en/detail/index/sArticle/705
I would have to manually measure using some technique. Could you briefly elaborate on why things are more or less ok with an 8Ω load?
https://www.magnat.de/en/detail/index/sArticle/705
I would have to manually measure using some technique. Could you briefly elaborate on why things are more or less ok with an 8Ω load?
For higher impedance load you need higher voltage swing for maximum output power but +/-42VDC is the absolute maximum operating voltage for LM3886 (+/-47VDC with no signal condition). With such high operating voltage you are already exceeding power dissipation limit of the chip which isn't GOOD for long term use. But under load condition & with a underpowered transformer the power supply voltage will sag to a more reasonable level which maybe ok for higher impedance loads.
Anyway for such spkrs i suggest reducing the supply voltage to +/-30VDC which is safe & healthy for 4-8R loads.
Anyway for such spkrs i suggest reducing the supply voltage to +/-30VDC which is safe & healthy for 4-8R loads.
Since I only need a little voltage drop to fall into a safer range of say +-40V, what would be the recommended way to do that? I'm thinking about resistance in series on the transformer secondary or a voltage divider but I do realize these pose problems when done and I am not sure which one to choose. An additional voltage regulator is out of the question and I won't even dare open the toroidal to remove windings from the secondary, it's more than certain I will just ruin it.
H
HAYK
Wind 3 turns of any wire through the donut and measure the ac voltage to determine the number of turns you need to decrease. Put in series of opposite phase to reduce the total voltage. The required thickness of the wire you find on AWG table.
For the noise, think of linking electrically the chassis to the planet.
For the noise, think of linking electrically the chassis to the planet.
I have not read the whole thread, but I have no idea why anybody wants to connect an amplifier board ground to a mains safety ground. The safety ground to the chassis is, so you don't get electrocuted under a faulty condition. The ground on an amplifier board should be treated like a signal ground.
I think I understand what the principle of what you suggest is and it's a good idea. I'd much prefer removing windings however. Not sure how successful a novice can be at this though, what do you think?
I think I will skip all existing transformer modifications. I just found a nice 2x25VAC, 225VA toroidal for a very good price. I expect this to deliver +- 33-34VDC post rectification.
LM3886 spec sheet mentions a constant output power of 50W at 8Ω using +-35VDC. Since my speakers ambiguously mention 4-8Ω impedance, I wanted your advice on what gauge zip cord to purchase for connecting them for optimal operation. Also, should I use the same gauge to connect the speaker outputs of the power amp board to the speaker terminals on the back panel?
So, given +-35VDC and kind of unknown speakers, what gauge/gauges would you guys pick from practical experience?
LM3886 spec sheet mentions a constant output power of 50W at 8Ω using +-35VDC. Since my speakers ambiguously mention 4-8Ω impedance, I wanted your advice on what gauge zip cord to purchase for connecting them for optimal operation. Also, should I use the same gauge to connect the speaker outputs of the power amp board to the speaker terminals on the back panel?
So, given +-35VDC and kind of unknown speakers, what gauge/gauges would you guys pick from practical experience?
I would not. The absolute maximum voltage allowed for an LM3886 with signal applied is ±42 V. See page 3 of the LM3886 data sheet.
HAYK's suggestion is good. You can also use a bucking transformer. You may also be able to unwind the secondary a bit. You probably don't need to take that many turns off.
tom
Buy a 22-0-22 VAC transformer instead which is perfect for variety of loads. I'm using dual 22VAC 220VA transformer for my setup.
For spkr wire 14G pure copper wire is all you need.
You said the opposite in Post #7. Maybe you like setting people up for failure. I prefer to set people up for success.I already explained that in post #9, but if you want to play with fire...he he he
Tom