Hello everyone,
I'm almost ready to put my ChipAmp in his (or her?) new house.
The designs of the house are almost done can be seen here: http://www.diyaudio.com/forums/chip-amps/257001-introduction-project-info.html
But i'm wondering which wires are most suitable for both Power, signal and speakers.
"Normal" powercord wiring for the power lines in the amp?
Thick copper wire to the speaker terminals?
And perhaps some silver wire as signal wiring?
If possible can you please post links to best places to buy them and or photo's of how you've did your wiring?
I'd really love some neat wiring in my case, i'm open for all suggestions!
Regards,
Peter
I'm almost ready to put my ChipAmp in his (or her?) new house.
The designs of the house are almost done can be seen here: http://www.diyaudio.com/forums/chip-amps/257001-introduction-project-info.html
But i'm wondering which wires are most suitable for both Power, signal and speakers.
"Normal" powercord wiring for the power lines in the amp?
Thick copper wire to the speaker terminals?
And perhaps some silver wire as signal wiring?
If possible can you please post links to best places to buy them and or photo's of how you've did your wiring?
I'd really love some neat wiring in my case, i'm open for all suggestions!
Regards,
Peter
Don't fall for any audiophool fluff. Good old copper is all you need. Use shielded or twisted pair wire for the input signals. Twist the speaker leads. Also twisting the power leads inside the cabinet won't hurt. Route the input leads away from the speaker leads. If you have a volume control, I like to ground the casing of the pot so you don't get any hum when you touch it.
.
Above all, use flux. The flux in the solder core is not enough by a mile. The general purpose flux sold by Radio Shack will do the job. Cut a pointy kind of flux applicator out of something like an empty plastic bottle, a milk carton or such.
Otherwise, tossing in 2 cents, mostly just agreeing with Pjotter. For the circuitry I seldom need wire because the component leads are long enough. When I do need jumpers I use AWG (American Wire Gauge) 22 or 24, commonly called just #22 or #24. Very best of all is bell wire, which is silver colored, not copper colored, because bell wire is, by definition, pre-tinned.
The only heavy current flow is in the power supply wires to the amp chip, and to the speakers, but "heavy" is relative, there's still only a couple of amps or so, meaning #16 or #18 wire is fine. Stranded is flexible, single-conductor is not.
Twist all pairs so the magnetic field of current going one way semi-cancels the magnetic field of current current going the other way. There's one ground point on the whole assembly, and one only. Keep power lines away from children, pets, and input lines.
I know this is disappointing, and I'm sorry, but nobody has even shown that wire costing $10/ft sounds better. Making matters worse, neatness has nothing to do with the materials used, it's all in the hands of the builder.
As a matter of fact, striving for neatness has been the downfall of many. It can be difficult to come up to your own standard until you have some familiarity with how this kind of thing goes. It might be wise to concentrate on the practical side, and let the art develop on its own terms.
.
Above all, use flux. The flux in the solder core is not enough by a mile. The general purpose flux sold by Radio Shack will do the job. Cut a pointy kind of flux applicator out of something like an empty plastic bottle, a milk carton or such.
Otherwise, tossing in 2 cents, mostly just agreeing with Pjotter. For the circuitry I seldom need wire because the component leads are long enough. When I do need jumpers I use AWG (American Wire Gauge) 22 or 24, commonly called just #22 or #24. Very best of all is bell wire, which is silver colored, not copper colored, because bell wire is, by definition, pre-tinned.
The only heavy current flow is in the power supply wires to the amp chip, and to the speakers, but "heavy" is relative, there's still only a couple of amps or so, meaning #16 or #18 wire is fine. Stranded is flexible, single-conductor is not.
Twist all pairs so the magnetic field of current going one way semi-cancels the magnetic field of current current going the other way. There's one ground point on the whole assembly, and one only. Keep power lines away from children, pets, and input lines.
I know this is disappointing, and I'm sorry, but nobody has even shown that wire costing $10/ft sounds better. Making matters worse, neatness has nothing to do with the materials used, it's all in the hands of the builder.
As a matter of fact, striving for neatness has been the downfall of many. It can be difficult to come up to your own standard until you have some familiarity with how this kind of thing goes. It might be wise to concentrate on the practical side, and let the art develop on its own terms.
.
Hi John and Bentsnake,
Thanks for the replies!
I'm sorry for my late reply, there where some other things i had to deal with.
It's a good tip to use extra flux, will definitively do that!
Using just copper is fine with me, i was more wondering whether there's a huge difference in using different wires.
If it's a very obvious difference or not.
I'm not disappointed at all, rather relieved that it's o.k. the way i'm doing it right now.
I am wondering however how i can twist certain wires - signal or speaker output - when im using a true star ground.
Every single ground wire in my amp goes straight to the star. you can see that very obviously in the thread i was referring to in my first post here.
The speaker out is straight from the LM3886 to the binding post + and the - comes from the star.
It's not like there's really a + and - pair coming from my chip.
Thanks again, both of you. It's much appreciated!
Thanks for the replies!
I'm sorry for my late reply, there where some other things i had to deal with.
It's a good tip to use extra flux, will definitively do that!
Using just copper is fine with me, i was more wondering whether there's a huge difference in using different wires.
If it's a very obvious difference or not.
I'm not disappointed at all, rather relieved that it's o.k. the way i'm doing it right now.
I am wondering however how i can twist certain wires - signal or speaker output - when im using a true star ground.
Every single ground wire in my amp goes straight to the star. you can see that very obviously in the thread i was referring to in my first post here.
The speaker out is straight from the LM3886 to the binding post + and the - comes from the star.
It's not like there's really a + and - pair coming from my chip.
Thanks again, both of you. It's much appreciated!
You think i'm spoiling peoples time here by asking questions that i'm not serious about?
No, i think we've all seen amp's with very exotic wires whether it's silvered copper or pure silver with a Teflon sleeve.
I've seen several people building gainclones with those conductors.
That's basically why i was wondering whether gainclones where sensitive for the type of wire they're connected to.
I didn't mean to waste anyone's time. I'm really glad that they took the effort to answer my questions. I've always been told that stupid questions don't exist. Stupid answers on the other side.... do.
In case you think i'm just trolling here, i'm not.
What rubbish. If the joints are clean the flux in the solder is more than adequate. Extra flux only adds corrosion problems later in life.
Copper is absolutely fine but you do need heavy gauge wire especially for the 0V (GND) connections, and the power rails. The output wiring should also be substantial.
Last edited:
'
<< I am wondering however how i can twist certain wires >>
Easy to show, difficult to explain.
Hold one end of two lengths of wire with pliers. Individual wires, not combined into a cable. Separate the wires as much as the plier jaws allow. Hand this to your gf and tell her to hold tight.
A second pair of pliers, which you hold. Same arrangement of the wires. Put as much tension on the wires as the two of you can manage.
Your gf holds her pliers steady, you twist yours. Rather incredibly, the wires will twist into a neat spiral. The engineering term is "tightly twisted," which means tight, but don't overdo, you can break the wires this way.
More incredibly, do this with three wires and the result will be the same. But the more separation the better.
I can't imagine who thinks this stuff up.
.
<< I am wondering however how i can twist certain wires >>
Easy to show, difficult to explain.
Hold one end of two lengths of wire with pliers. Individual wires, not combined into a cable. Separate the wires as much as the plier jaws allow. Hand this to your gf and tell her to hold tight.
A second pair of pliers, which you hold. Same arrangement of the wires. Put as much tension on the wires as the two of you can manage.
Your gf holds her pliers steady, you twist yours. Rather incredibly, the wires will twist into a neat spiral. The engineering term is "tightly twisted," which means tight, but don't overdo, you can break the wires this way.
More incredibly, do this with three wires and the result will be the same. But the more separation the better.
I can't imagine who thinks this stuff up.
.
Last edited:
'
<< Every single ground wire in my amp goes straight to the star. >>
This is good, very good. But c'mon, did you really think it would be that simple?
Before anything else, leave what you've got alone, it's fine. But for future reference I'll give you my understanding of present-day thinking, which is to use a "binary star" ground.
The idea here is that very large current pulses flow in the power supply, and very dirty hash is injected into ground by the bypass/decoupling capacitors. The plan is to have these distortions swamped by the transformer and smoothing capacitors, so they don't distort the much smaller input signal. By the way, "swamped" really is an engineering term.
To accomplish this, the center tap of the transformer, the ground side of the smoothing capacitors, and the ground side of the bypass/decoupling capacitors are all joined together. These form the main ground point, and join to the chassis ground.
Coming together at a separate point are the grounds for the input circuit, including the feedback circuit, and also including the output (speaker) ground. This is the signal ground, the second star of the binary star ground. The signal ground is then jumpered (jumper = a short length of wire) to the main ground point.
Note: Do not ground the input with a connection to the chassis at the input jack. Rather, use an insulated input jack, and run a separate conductor to the signal ground point. Use the input cable shield if present, or else that tightly-twisted pair.
I should add that there appears to be disagreement among the gurus. Douglas Self seems to sometimes recommend a third, separate ground point for the bypass/decoupling capacitors, which is also jumpered to the main ground point ("Audio Amplifier Design Handbook").
The engineers at Texas Instruments do go along with the binary star ground idea, but they don't seem to go as far as a separate bypass/decoupling ground point. Here's the TI reference, mainly of interest is Figure 2. Grounding Principles - The Signal - Archives - TI E2E Community
On the other hand, looking at all of this from a different angle, it seems the grounding circuit is rapidly becoming more complicated than the amplifier circuit. I still think the star ground you've got now will be fine.
.
<< Every single ground wire in my amp goes straight to the star. >>
This is good, very good. But c'mon, did you really think it would be that simple?
Before anything else, leave what you've got alone, it's fine. But for future reference I'll give you my understanding of present-day thinking, which is to use a "binary star" ground.
The idea here is that very large current pulses flow in the power supply, and very dirty hash is injected into ground by the bypass/decoupling capacitors. The plan is to have these distortions swamped by the transformer and smoothing capacitors, so they don't distort the much smaller input signal. By the way, "swamped" really is an engineering term.
To accomplish this, the center tap of the transformer, the ground side of the smoothing capacitors, and the ground side of the bypass/decoupling capacitors are all joined together. These form the main ground point, and join to the chassis ground.
Coming together at a separate point are the grounds for the input circuit, including the feedback circuit, and also including the output (speaker) ground. This is the signal ground, the second star of the binary star ground. The signal ground is then jumpered (jumper = a short length of wire) to the main ground point.
Note: Do not ground the input with a connection to the chassis at the input jack. Rather, use an insulated input jack, and run a separate conductor to the signal ground point. Use the input cable shield if present, or else that tightly-twisted pair.
I should add that there appears to be disagreement among the gurus. Douglas Self seems to sometimes recommend a third, separate ground point for the bypass/decoupling capacitors, which is also jumpered to the main ground point ("Audio Amplifier Design Handbook").
The engineers at Texas Instruments do go along with the binary star ground idea, but they don't seem to go as far as a separate bypass/decoupling ground point. Here's the TI reference, mainly of interest is Figure 2. Grounding Principles - The Signal - Archives - TI E2E Community
On the other hand, looking at all of this from a different angle, it seems the grounding circuit is rapidly becoming more complicated than the amplifier circuit. I still think the star ground you've got now will be fine.
.
Hi Bentsnake
Thanks for your extensive answers!
I'm in the process of making a new case around my amplifier and will redo pretty much all wiring.
So thanks for tip, i'll use it in the near future ;-)
What i meant to ask about twisting the wires is how i'm going to twist them if they're not a pair all the way.
The speaker out wire from the LM3886 to the + binding post is just a few centimeters long.
Then i have one wire from the - binding post to the star ground which is roughly 15 cm long and a few other ground wires to the chip.
Is it wise to twist them anyway? I can of course twist them from the binding posts towards the chip, from there the + wire will go into the chip and the - will go to the star.
There seems to be a lot of confusion about grounding, there are so many ways to go and even more opinions on how to do it properly.
I'm not an electrical engineer but just doing the best i can.
My gain clone is right now - in test build - quite quiet. Nothing's shielded or separated.
I've basically tied it all together to see if it works, it works quite well.
There's a very little bit of distortion left at this point. I think that's because of
1. There's just one star ground, no separated or binary (?) stars. Definatily reading upon this.
2. It's all open, all my signal and power wires running trough and next to each other.
I've been reading a lot on this subject, i think this can make or break a project like mine.
I didn't saw the link you shared before, thanks for that!
The following link will also cover a part of this subject http://www.diyaudio.com/forums/diya...udio-component-grounding-interconnection.html
Thanks again!
It's getting late here now, i'll dive into it asap.
- Peter
Thanks for your extensive answers!
I'm in the process of making a new case around my amplifier and will redo pretty much all wiring.
So thanks for tip, i'll use it in the near future ;-)
What i meant to ask about twisting the wires is how i'm going to twist them if they're not a pair all the way.
The speaker out wire from the LM3886 to the + binding post is just a few centimeters long.
Then i have one wire from the - binding post to the star ground which is roughly 15 cm long and a few other ground wires to the chip.
Is it wise to twist them anyway? I can of course twist them from the binding posts towards the chip, from there the + wire will go into the chip and the - will go to the star.
There seems to be a lot of confusion about grounding, there are so many ways to go and even more opinions on how to do it properly.
I'm not an electrical engineer but just doing the best i can.
My gain clone is right now - in test build - quite quiet. Nothing's shielded or separated.
I've basically tied it all together to see if it works, it works quite well.
There's a very little bit of distortion left at this point. I think that's because of
1. There's just one star ground, no separated or binary (?) stars. Definatily reading upon this.
2. It's all open, all my signal and power wires running trough and next to each other.
I've been reading a lot on this subject, i think this can make or break a project like mine.
I didn't saw the link you shared before, thanks for that!
The following link will also cover a part of this subject http://www.diyaudio.com/forums/diya...udio-component-grounding-interconnection.html
Thanks again!
It's getting late here now, i'll dive into it asap.
- Peter
with that flux thing i simply do not agree with none of you.
thing is, some say what the solder stuff has is enough.
others say, its not enough.
well, i'm certain both are wrong.
really depends on:
-> the quality of the solder
-> the wire gauge
for a thin wire any decent solder has enough flux.
for thick wires it does not.
thing is, some say what the solder stuff has is enough.
others say, its not enough.
well, i'm certain both are wrong.
really depends on:
-> the quality of the solder
-> the wire gauge
for a thin wire any decent solder has enough flux.
for thick wires it does not.
.
<< My gain clone is right now - in test build - quite quiet. Nothing's shielded or separated...I've basically tied it all together to see if it works, it works quite well. >>
You've stated the heart of the matter right there. Contrary to what some say, you don't need gold wire, or indicator lights made from real emeralds. The fact of the matter is, you can get away with murder with these things.
This is not by accident. Remember that the market for audio chips is worldwide, and not every supervising engineer has a doctorate from MIT. So to serve their markets the factory makes these chips as foolproof as they possibly can. Little black chips are sold not by the thousands, but by the millions, and you don't have sales figures like that unless your product works, and works every time.
<< ask about twisting the wires is how i'm going to twist them if they're not a pair all the way >>
Why twist wires at all? Because every current-carrying wire has a magnetic field around it. This magnetic field is what you're interested in, you want to keep it localized unto itself, not turning every conductor in your circuit into a transformer.
So you twist a pair of wires that carries both sides of the same signal, such as an input pair, or a speaker pair. Especially good is to twist a power supply pair (+ and -). All of which is done because twisting partially cancels (i.e. localizes) the magnetic field around the wires.
BUT routing is much more important than twisting. The consideration is that power supply wires, and speaker wires, carry heavy currents and high voltages (relatively), while input wires carry small currents and small voltages. So worry less about twisting, more about routing.
This means don't run wires parallel to each other unless they are a pair. It also means that distance is your friend, try not to have non-pair wires close to each other at all, and especially keep inputs away from power and speakers. Try to have exterior wires approach your circuit board vertically, as this minimizes magnetic interaction.
Keep components as close together as works for you. Simply because shorter conductors have less opportunity to pick up exterior noise.
But notice all of this "try." It's not an ideal world, you do what you have to.
<< There's a very little bit of distortion left at this point. >>
Chances are slim to none that a factory circuit (as found in the data sheet) produces any distortion you can hear. Foolproof, remember. Suspect your speakers or your player before you suspect your factory circuit. Otherwise, you're probably right that distortion is coming in from the outside.
But what kind of distortion? Hiss? Hash? Hum? Buzz? Just an overall impression of distortion? If distortion creeps in as volume is increased, then a too-small power supply is a primary suspect.
Grounding: I looked at the article you referenced ( http://www.diyaudio.com/forums/diya...udio-component-grounding-interconnection.html ). Be aware that "balanced lines," as in XLR connectors, have nothing whatsoever to do with the audio signal directly. The purpose of a balanced line is to minimize nose pickup by connecting cables. It usually has no application outside of a studio, or a performing environment.
Grounding: You certain are right about confusion on the subject of grounding. Possibly because a lot of self-appointed experts--such as myself--set themselves up as gurus. Which is why I point to the grounding method discussed in Grounding Principles - The Signal - Archives - TI E2E Community . At least you know the engineers at Texas International have degrees.
However, I still think your present star ground will work fine. All by itself it's a lot better than the ground buss that used to be used.
.
<< My gain clone is right now - in test build - quite quiet. Nothing's shielded or separated...I've basically tied it all together to see if it works, it works quite well. >>
You've stated the heart of the matter right there. Contrary to what some say, you don't need gold wire, or indicator lights made from real emeralds. The fact of the matter is, you can get away with murder with these things.
This is not by accident. Remember that the market for audio chips is worldwide, and not every supervising engineer has a doctorate from MIT. So to serve their markets the factory makes these chips as foolproof as they possibly can. Little black chips are sold not by the thousands, but by the millions, and you don't have sales figures like that unless your product works, and works every time.
<< ask about twisting the wires is how i'm going to twist them if they're not a pair all the way >>
Why twist wires at all? Because every current-carrying wire has a magnetic field around it. This magnetic field is what you're interested in, you want to keep it localized unto itself, not turning every conductor in your circuit into a transformer.
So you twist a pair of wires that carries both sides of the same signal, such as an input pair, or a speaker pair. Especially good is to twist a power supply pair (+ and -). All of which is done because twisting partially cancels (i.e. localizes) the magnetic field around the wires.
BUT routing is much more important than twisting. The consideration is that power supply wires, and speaker wires, carry heavy currents and high voltages (relatively), while input wires carry small currents and small voltages. So worry less about twisting, more about routing.
This means don't run wires parallel to each other unless they are a pair. It also means that distance is your friend, try not to have non-pair wires close to each other at all, and especially keep inputs away from power and speakers. Try to have exterior wires approach your circuit board vertically, as this minimizes magnetic interaction.
Keep components as close together as works for you. Simply because shorter conductors have less opportunity to pick up exterior noise.
But notice all of this "try." It's not an ideal world, you do what you have to.
<< There's a very little bit of distortion left at this point. >>
Chances are slim to none that a factory circuit (as found in the data sheet) produces any distortion you can hear. Foolproof, remember. Suspect your speakers or your player before you suspect your factory circuit. Otherwise, you're probably right that distortion is coming in from the outside.
But what kind of distortion? Hiss? Hash? Hum? Buzz? Just an overall impression of distortion? If distortion creeps in as volume is increased, then a too-small power supply is a primary suspect.
Grounding: I looked at the article you referenced ( http://www.diyaudio.com/forums/diya...udio-component-grounding-interconnection.html ). Be aware that "balanced lines," as in XLR connectors, have nothing whatsoever to do with the audio signal directly. The purpose of a balanced line is to minimize nose pickup by connecting cables. It usually has no application outside of a studio, or a performing environment.
Grounding: You certain are right about confusion on the subject of grounding. Possibly because a lot of self-appointed experts--such as myself--set themselves up as gurus. Which is why I point to the grounding method discussed in Grounding Principles - The Signal - Archives - TI E2E Community . At least you know the engineers at Texas International have degrees.
However, I still think your present star ground will work fine. All by itself it's a lot better than the ground buss that used to be used.
.
.
Arguments about use-flux-don't are meaningless. The only guy who knows anything is the guy holding the soldering iron.
If you have soldering skills, then yay you. But to your amazement, not everybody is as good as you. Some people need a little help. In fact, judging by the "it doesn't work" questions you see on forums, a whole lot of people need help.
That help comes in the form of a little dab of flux. This little bit of extra cleans the joint, and makes all the difference when you're first starting out. The difference between success and, "Aw, I just can't do it."
So you don't need it, so you're that good, more yay you. But if you want to withhold help from somebody just starting out, then maybe you'd want to rethink exactly what you're doing, and why.
.
Arguments about use-flux-don't are meaningless. The only guy who knows anything is the guy holding the soldering iron.
If you have soldering skills, then yay you. But to your amazement, not everybody is as good as you. Some people need a little help. In fact, judging by the "it doesn't work" questions you see on forums, a whole lot of people need help.
That help comes in the form of a little dab of flux. This little bit of extra cleans the joint, and makes all the difference when you're first starting out. The difference between success and, "Aw, I just can't do it."
So you don't need it, so you're that good, more yay you. But if you want to withhold help from somebody just starting out, then maybe you'd want to rethink exactly what you're doing, and why.
.
I believe that the load returns should be connected to the bypass capacitors. But if you do connect the load returns to star ground then position the star ground between the two amps, as should the + and - rail.
Nearly every ground/return lead is returning signal/current and should be twisted together with its supply lead.
Nearly every ground/return lead is returning signal/current and should be twisted together with its supply lead.
The Speaker current comes from the PSU, it has to return to the PSU.
I wrote a long garbled reply a couple of days ago on the Speaker Return Route. I'm not sure anyone could easily follow it and may have given up trying to understand what I wrote.
I'll try again.
Trace out the ROUTE that the current takes in flowing from the PSU through the amplifier to the Speaker. It is a long route, think carefully about each part and don't miss any part out. It's a little bit easier if you consider only the +ve half of the AC waveform and follow that through the +ve half of the output stage. and DON'T miss out the local decoupling caps. They pass the fastest changing currents and create the highest interference fields. Mark refers to "bypass capacitors", I think he meant local HF & MF decoupling capacitors.
The RETURN ROUTE must exactly follow that FLOW ROUTE.
And twist wherever possible.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.