The commercial market doesn't build what I want at an agreeable price.
I just want five channels of amplification, nothing else. I have no need for a tuner, no need for a clock, no need for a source switch. The plan is a HTPC (probably MythTV), though this is a couple years off. It's never to early to start planning.
I want to build a five channel chipamp. Wikipedia sent me here.
I know I can use a quintet of LM3886 chips to achieve my goal, as I need 4ohm stability. (DIY speakers.)
The pin layout is straightforward on the chips.
What makes little sense to me is the power supply.
I searched, but there's tons of posts in here that confuse me further, and nothing that says "hey newbie: do this, this is why!"
So, tell me what to do and tell me why.
I just want five channels of amplification, nothing else. I have no need for a tuner, no need for a clock, no need for a source switch. The plan is a HTPC (probably MythTV), though this is a couple years off. It's never to early to start planning.
I want to build a five channel chipamp. Wikipedia sent me here.
I know I can use a quintet of LM3886 chips to achieve my goal, as I need 4ohm stability. (DIY speakers.)
The pin layout is straightforward on the chips.
What makes little sense to me is the power supply.
I searched, but there's tons of posts in here that confuse me further, and nothing that says "hey newbie: do this, this is why!"
So, tell me what to do and tell me why.
Is there something in perticular about the power supply that is confusing you? Something more specific?
Otherwise... make an unregulated power supply with a couple of big electrolytic capacitors. Why? because it is easy and efficient (doesn't make a lot of heat).
Otherwise... make an unregulated power supply with a couple of big electrolytic capacitors. Why? because it is easy and efficient (doesn't make a lot of heat).
Not really particular, more general. I'm not an engineer.
For starters, I ran across this when looking for power supply schematics and formulas:
http://my.integritynet.com.au/purdic/power1.html
There's a number of grounds in the schematic, both on the primary and secondary windings. Is that ground at the ac power (far left), or is that "earth"?
If that's earth, what about the ground on the secondary side of the transformer, by the "load" (far right)?
Do those symbols mean the same thing?
For starters, I ran across this when looking for power supply schematics and formulas:
http://my.integritynet.com.au/purdic/power1.html

There's a number of grounds in the schematic, both on the primary and secondary windings. Is that ground at the ac power (far left), or is that "earth"?
If that's earth, what about the ground on the secondary side of the transformer, by the "load" (far right)?
Do those symbols mean the same thing?
Hi,
that V2 site is good if you want to get deeper into understanding electronics.
The schematic you posted is a single polarity supply.
Most solid state, including chipamps, use a dual polarity supply.
The unregulated supply is easy, cheap and most of all it works.
Don't go regulated until YOU can design a discrete power amplifier. The complexities of designing a good regulated PSU that sounds good with your chosen amplifier are just a difficult as designing the amplifier and some designers say it is more complex to get them to work together.
Have a look at ESP site for LOTS and LOTS of practical designing and kit building and even some good theory.
that V2 site is good if you want to get deeper into understanding electronics.
The schematic you posted is a single polarity supply.
Most solid state, including chipamps, use a dual polarity supply.
The unregulated supply is easy, cheap and most of all it works.
Don't go regulated until YOU can design a discrete power amplifier. The complexities of designing a good regulated PSU that sounds good with your chosen amplifier are just a difficult as designing the amplifier and some designers say it is more complex to get them to work together.
Have a look at ESP site for LOTS and LOTS of practical designing and kit building and even some good theory.
AndrewT said:Don't go regulated until YOU can design a discrete power amplifier. The complexities of designing a good regulated PSU that sounds good with your chosen amplifier are just a difficult as designing the amplifier and some designers say it is more complex to get them to work together.
Have a look at ESP site for LOTS and LOTS of practical designing and kit building and even some good theory.
Ah, thanks for the tip. A bit of searching there gives me this:
http://sound.westhost.com/power-supplies.htm
This makes things a whole lot clearer:
http://www.sound.westhost.com/psu-wiring.htm
Now for specifics:
Mains capacitors (between the fuse and primary windings on the transformer): yes/no and why? If yes, how big?
Can I connect multiple LM3886s to one power supply if I have a big enough transformer/fuse/switch? Should I fuse the connections?
What kind of effect does using bigger/smaller supply caps have on the power supply? (Instead of four 4,700uF caps, how about four 10,000uF or four 1,000uF?)
http://www.sound.westhost.com/psu-wiring.htm
Now for specifics:
Mains capacitors (between the fuse and primary windings on the transformer): yes/no and why? If yes, how big?
Can I connect multiple LM3886s to one power supply if I have a big enough transformer/fuse/switch? Should I fuse the connections?
What kind of effect does using bigger/smaller supply caps have on the power supply? (Instead of four 4,700uF caps, how about four 10,000uF or four 1,000uF?)
beerman said:
What kind of effect does using bigger/smaller supply caps have on the power supply? (Instead of four 4,700uF caps, how about four 10,000uF or four 1,000uF?)
As you know already, a diode turns AC into a DC current with a ton of ripple, the bigger the cap the less ripple there will be, this is very important when large DC currents are quickly drawn from the supply. The bigger the cap the less ripple will be on the supply line and less voltage sag when power is drawn from the supply.
So for example, if your using 4,700uF caps on the supply and you want 1mV of ripple MAX then you are able to draw say 2.5Amps.
But if you use 1,000uF caps and still want 1mV of ripple it will be able to draw .5A MAX from the same supply. And say you use 10,000uF caps then you can draw 5A MAX at 1mV of ripple.
P.S those numbers are only examples, there not real.
So to put it easy, the bigger the caps the stronger the supply is up to a point, the transformer is another big issuse for stronger supplys.
I hope this helps and isn't too confusing.
ifrythings said:
As you know already, a diode turns AC into a DC current with a ton of ripple, the bigger the cap the less ripple there will be, this is very important when large DC currents are quickly drawn from the supply. The bigger the cap the less ripple will be on the supply line and less voltage sag when power is drawn from the supply.
So for example, if your using 4,700uF caps on the supply and you want 1mV of ripple MAX then you are able to draw say 2.5Amps.
But if you use 1,000uF caps and still want 1mV of ripple it will be able to draw .5A MAX from the same supply. And say you use 10,000uF caps then you can draw 5A MAX at 1mV of ripple.
P.S those numbers are only examples, there not real.
So to put it easy, the bigger the caps the stronger the supply is up to a point, the transformer is another big issuse for stronger supplys.
I hope this helps and isn't too confusing.
This place and the ESP site pages are clearing things up quite a bit. Thanks to all. I'll make a drawing this weekend and submit for criticism.
From what I understand, I'll need at least a 500VA transformer, 35V (+35?) secondary, and tons of capacitance from the current draw.
Two more newb questions:
How do the big boys do a master volume control on multichannel amps?
Should I fuse between the power supply and the chipamps?
Yes you should, it's a RFI filter but it isn't mandatory. You might like to just buy an IEC mains socket with this built into it, for simplicity.beerman said:Mains capacitors (between the fuse and primary windings on the transformer): yes/no and why?
Yes. There is a little to be gained from using separate supplies but you certainly don't have to.Can I connect multiple LM3886s to one power supply if I have a big enough transformer/fuse/switch?
Yes, bad things may happen if you don't. 🙂Should I fuse the connections?
Edit: BTW, you can't increase a power supplies capability by increasing the rating of the fuse. The fuse must blow first.
Hi,
500VA is at the top end for supplying 5channels each about 50 to 60W into 4ohms.
35-0-35 is far too high. This will result in >+-50Vdc and fry your chipamps.
18Vac to 25Vac is chipamp suitable but the range 21Vac to 25Vac should only be used for 8ohm speakers. For your 4ohms speakers use 20Vac or less.
Download the 3875 datasheet and go through it very carefully.
Ask about the parts you don't understand (if any).
500VA is at the top end for supplying 5channels each about 50 to 60W into 4ohms.
35-0-35 is far too high. This will result in >+-50Vdc and fry your chipamps.
18Vac to 25Vac is chipamp suitable but the range 21Vac to 25Vac should only be used for 8ohm speakers. For your 4ohms speakers use 20Vac or less.
Download the 3875 datasheet and go through it very carefully.
Ask about the parts you don't understand (if any).
This much I do understand. I've been playing with car audio for a while, just never jumped this far into how things work.lndm said:Edit: BTW, you can't increase a power supplies capability by increasing the rating of the fuse. The fuse must blow first.
How about 70?AndrewT said:Hi,
500VA is at the top end for supplying 5channels each about 50 to 60W into 4ohms.
AndrewT said:35-0-35 is far too high. This will result in >+-50Vdc and fry your chipamps.
18Vac to 25Vac is chipamp suitable but the range 21Vac to 25Vac should only be used for 8ohm speakers. For your 4ohms speakers use 20Vac or less.
Download the 3875 datasheet and go through it very carefully.
Ask about the parts you don't understand (if any).
I was planning on the 3886, but I'll still look at the datasheet.
Just a formality 😉 A person has to be careful on this forum when talking about the dangerous end of a piece of equipment.beerman said:This much I do understand.
Hi,
I don't know why I looked up the wrong datasheet, sorry.
have you read note 11?
It states a typical 68W when tested at 25degC. That is effectively telling the user that one cannot get 68W out of the amp in practice.
My rule of thumb for multichannel amp VA requirement is
VA~=1.5times highest power channel + sum of all other channels. For 60W *5 you need about 330VA for a good transformer. What happened to the 0.1? Is that a sixth channel or are you doing something else for the bass?
I don't know why I looked up the wrong datasheet, sorry.
have you read note 11?
It states a typical 68W when tested at 25degC. That is effectively telling the user that one cannot get 68W out of the amp in practice.
My rule of thumb for multichannel amp VA requirement is
VA~=1.5times highest power channel + sum of all other channels. For 60W *5 you need about 330VA for a good transformer. What happened to the 0.1? Is that a sixth channel or are you doing something else for the bass?
Hmm, add "compressed nitrogen coolant" to the list.... Just kidding. 😀
The 0.1 is a plate amp. I'm looking at a partsexpress.com 500w. I want big bass, wife wants a smallish box. Big power means best of both worlds.
I don't feel at all comfortable building an amp with that kind of power and my lack of experience, someone's gonna get hurt, probably me.
The chip amps are simple enough from pics I've seen. I feel comfortable enough with those.
The 0.1 is a plate amp. I'm looking at a partsexpress.com 500w. I want big bass, wife wants a smallish box. Big power means best of both worlds.
I don't feel at all comfortable building an amp with that kind of power and my lack of experience, someone's gonna get hurt, probably me.
The chip amps are simple enough from pics I've seen. I feel comfortable enough with those.
I need some reassurance I understand "ground" correctly:
The PSU's center tap is "0v" thanks to the bridge rectifier, correct?
Assuming I construct the PSU correctly,
If I probe ground and one of my rails with a DMM set to Vdc, I should see XXv, where XX = my desired rail voltage?
If I probe both rails and not ground, I my DMM would read YYv, where YY = XX * 2?
After reading the data sheet and seeing "typical application" circuits, I see that the input and output are both connected to ground (0Vdc at the PSU) at given points. Is this correct?
And this "ground" is not the same as "earth", the third prong on my North American 120vac 60hz plug, right?
The PSU's center tap is "0v" thanks to the bridge rectifier, correct?
Assuming I construct the PSU correctly,
If I probe ground and one of my rails with a DMM set to Vdc, I should see XXv, where XX = my desired rail voltage?
If I probe both rails and not ground, I my DMM would read YYv, where YY = XX * 2?
After reading the data sheet and seeing "typical application" circuits, I see that the input and output are both connected to ground (0Vdc at the PSU) at given points. Is this correct?
And this "ground" is not the same as "earth", the third prong on my North American 120vac 60hz plug, right?
One more question:
Are these "miniature" capacitors okay to use on the power supply?
http://www.partsexpress.com/pe/showdetl.cfm?&Partnumber=020-1268
Are these "miniature" capacitors okay to use on the power supply?
http://www.partsexpress.com/pe/showdetl.cfm?&Partnumber=020-1268
- Status
- Not open for further replies.
- Home
- Amplifiers
- Power Supplies
- Help?