Anything I can use for DIY amp from PC powersupplies?

Status
Not open for further replies.
Jump to Latest
#21
"I would be careful connecting DC outputs from multiple supplies in series, as each one's DC ground is also tied to earth ground, and things could get bad very quickly when you throw the switch!"

cant you just cut the connection between dc-ground and earth ground?


cant i make a simple transformer from like a piece of metal(does it have to be any special metal?) and like 6 tuns of wire(2primary 4sec) to double the voltage out of my pc-power sup?
 
#25
Olle Bard said:
hmm, seems lika a bit of trouble. but hey, im gonna try it some day, the worse thing that could happen are some broken components, rigth?

but how about cutting the connection between earth ground and dc-ground? so you can just put 2 in series to get 24v
Click to expand...

Actually.... no, because they use a single diode for the filtering... you.. hm.... I think you can't... 🙂 lol... Its ear;y.. let someone else think.. 🙄
 
#29
First of all, You may connect power supply outputs in parallel but you may have some unstability as there isn't a voltage or current feedback to them. If you connect them in series, they won't know what each other is doing and if there is any irregularities, they will blow up. They need to have voltage and current feedback routs so the PWM circuits will act accordingly and balance themselves out. With seperate supplies, with different PWM circuits, there must be some sort of controller to tell the individual PWM's what to do at whatever time, so things stay stable.

Depending on the supply topology, the voltage that the transformer uses is DC already, due to a rectifier and capacitors on the input.

Be careful, rewinding the cores. If they saturate, many amps will flow in the FET and blow it up. Changing the frequency of the PWM will affect the cores also. You need to know what core you have, to see if it will operate in that frequency range and its efficiency at whatever frequency you choose. Again, saturation of the core is a big problem. Some will get lucky but most will not be so lucky in this. Switchmode powersupplies are very fenicky about most of what is in them. This is why engineers get paid the big money to design them. It is a very precise art in which to make an efficient , stable and cheap power supply. I won't even go into EMI considerations. I spent many months in an EMI chamber trying to get a PS within specs.........

Good luck,

Chris
 
#31
EMI is Electro-magnetic interference. The power supplies I worked on were quite powerful.... 2400W ... When we turned one on, and the case wasn't bolted together, it would wipe out an FM radio for up to 100 feet, 30M. This was from the different magnetic fields within the device. The inductors aren't the only thing radiating EMI. The recitfiers and FET's have their own frequencies that they radiate. In America, the Federal Communications Commission, (FCC) has specific requirements as to the limit of EMI any electronic thing can radiate. You will notice filtering on the input and output. On the input, so you don't put these radio frequency harmonics back on to the AC power lines and on the output so you don't radiate them into your or any other equipment. You will notice the orientation of components on the board and how far they are from each other, along with they are in a metal case, etc. This is the "art" in which I refered to earlier. Not only do you want to make a robust and efficient supply, it has to be cheap to produce and very little EMI radiation.

Try to wind the transformer but beware, it may not work and it may just blow up. I hope it works for you but you need to know a little of what you are getting into. Switchmode supplies can be extremely frusterating to work with, especially when tweeking or designing them. :scratch:

Notice the apparent "double" winding. This is called bi-filar winding, 3 wires are called tri-filar, etc. When you do your winding, do the same. On car audio amps, you will see them use up to 6 wires in a bundle, wound around the core. This gives more current carrying capability along with minimal wire resistance. Lets say that each wire is 1 ohm in a single wire at a certin length, if you have 4 wires you now have 1/4 ohm and 4 times the current carrying capability. To wind with a larger gague wire doesn't give the same crossectional area (less current capability) and heat loss is greater. The wire is magnet wire, with the enamel coating and tinned and joined at the ends only.

This is probably more than you ever wanted to know and there is an ocean of other information about this.......

Like I asked before, Please let us know how it goes and tell us what you did!!!!

Good luck,

Chris
 
#32
ok so i should use six pieces of wire(an i do it with pvc-insulated wire?) and twist it like 4 turns around the core.

should i half the number of wires(6) on my secondary(8 turns around core) sonce it will only get half as many ampere?


looked in my pc-ps (rated dc output: 200w) and it has a tiny transformer(like 2" tall and 1"/1 1/2" wide).
it feels a bit small but i guess since its rated 200w it should be enough, right?


oh yea, anyone knows of an electronics project(preferably a preamp, that you can run on 5v single supply?
 
#33
I would use 2 or 3 wires and not 6 as you won't have enough room for so much wire..... PVC insulated wire takes too much room as well.... Use magnet wire. The number of turns depends on the output voltage but you will have to modify the feedback resistor network as it was designed to do the specific output voltage of its original state. This means that you will have to calculate what the voltage is at that node and then calculate what the new output voltage will be so you can get that same voltage at that point otherwise the PWM won't know what the pulse width is supposed to be and problems will occur. With switchmode supplies, it is more exact as every little aspect is carefully designed in. In analog transformers, yeah, just add or take away turns. The core is calculated on how much current it needs to handle and heat loss due to eddy currents. I suggest getting a book on switchmode power designs and topologies and design one from scratch. You'll probably get that done faster than modifying a pre-existing design, unless you really know what you're doing. I'm not trying to be negative, just realistic..... This has been my personal experience.

Good luck,

Chris
 
#34
Diode: Speaking of EMI:

Did you see the article iin the current issue of AudioXpress the "ionic tweater" -- I can just see someone firing one of these babies up when I am trying to work DX on 20 meters! Clearly the case of an individual pursuing an elusive end without a spectrum analyzer.

Diode said:


Notice the apparent "double" winding. This is called bi-filar winding, 3 wires are called tri-filar, etc. When you do your winding, do the same. On car audio amps, you will see them use up to 6 wires in a bundle, wound around the core. This gives more current carrying capability along with minimal wire resistance. Lets say that each wire is 1 ohm in a single wire at a certin length, if you have 4 wires you now have 1/4 ohm and 4 times the current carrying capability. To wind with a larger gague wire doesn't give the same crossectional area (less current capability) and heat loss is greater. The wire is magnet wire, with the enamel coating and tinned and joined at the ends only.

Chris
Click to expand...
and last, but not least, using multiple wires reduces the "skin effect" -- which becomes more of a problem as the frequency rises.

I would NOT discourage DIYrs from trying SMPS -- unlike so many other aspects of DIY you really have to pay attention to the math and the most fundamental concepts of electricity and magnetism. <p>and one more note in passing, every controller manufacturer has warnings to the effect of "don't do this on a breadboard". A standard layout PCBoard for a quiet , DIY current mode controller would probably sell a lot.
 
#35
diode>> I wasn´t really going to modify the powersupply, just build a little box that you can put at the output containing a transformer that will double the output voltage.
this should not require any changing of the components inside the pc-ps I have got now, right?

if i do this, should i have a resistor in series with the transformer primaries to prevent damage on the pc ps?
 
#36
Olle Bard said:
diode>> I wasn´t really going to modify the powersupply, just build a little box that you can put at the output containing a transformer that will double the output voltage.
this should not require any changing of the components inside the pc-ps I have got now, right?

if i do this, should i have a resistor in series with the transformer primaries to prevent damage on the pc ps?
Click to expand...

You can't just feed the tranformer with DC if that is what you meant.
 
#37
You cant just feed another transformer with the output from the supply. Transformers need AC or pulsed DC to operate so you have to do all the transforming before the output rectifier. The best way would be to remove the transformer from the supply, rewind it, change the resistor values etc so that the feedback to the PWM IC is correct and then put it all back together.
 
#38
thank you, that was what i wanted to know.


hmmm
then its not so easy.

anyone know why it doesnt work with dc?
isnt it some magnetic stuff, because i remember making electro magnets when i was young with just batteries a nail and a piece of wire. and batterys produce dc, right?

how can i get pulsed dc? cap?
 
#39
For a transformer to function correctly, the magnetic field must be continually changing. DC will produce a constant field and so no output will be formed. It's not worth the trouble trying to make a switcher, you'll lose efficiency and it'll cost more money. All the parts are already there all you have to do is rewind the transformer and fiddle with the feedback.
 
Status
Not open for further replies.
Top Bottom