I've just finished a +/- 35v smps using the tl494. It runs beautifully. right now its driving 2 lm3886 50W amps with no drop in voltage. It runs from 12vdc and sucks around 13 amps when fully driving 2 8ohm subs. I tried a toroid but due to the small physical size of it it could supply more than 20watts. i am now using a AT PC ferite transformer wich i can get around 150watts.
Super simple, cheap circuit that allmost anyone can build.
If u want more info just email me or add me to your msn contact list. Creamed_Emu@hotmail.com
Super simple, cheap circuit that allmost anyone can build.
If u want more info just email me or add me to your msn contact list. Creamed_Emu@hotmail.com
Rod Elliot has a SMPS based on the SG3525. You can get them from Oatleys or national on a sample order (I think).
I've just re-drawn the schematic that i've used in my design. The 2 ic's are a tl494 (also known as 7500 in some pc power supplies) and a 4050 hex buffer. the capacitor and resistor attached to pins 5 and 6 of the 494 are to set the push pull frequency. you may need to change these depending on the type of core you use and the number of primary turns you have wound. these seemed fine for me even using a number of different cores.
The 4050 is a buffer to give nice n quick rise and fall times for the output stage to keep the output transistors/mosfets out of their linear range to increase efficiency and reduce heat.
The transformer i used was out of a 200w at pc power supply, as was the tl494 and pretty much everything else (except the 4050) to seperate the cores without cracking the ferite i usually boil them in water for around 5 min and take them out and seperate them as quickly as possible. then you unwind the windings (save the wire, it could be usefull) and let the cores cool.
To wind the transformer you wind 4 turns of wire (because of the high frequency its best to combine a number of thinner wires to create the one wire to handle the current, i used 6 .8mm wires twisted into the one large wire using a power drill) then leave around 10cm of wire and cut them. then wind another 4 turns of wire over(or beside) the first 4 in the same direction and leave 10cm and cut them. now join the lead from the first primary (the one from the fourth turn) and join it to the start of the 1st winding of the second primary. so you have a 8 turn center tapped primary. get it?
Now for the secondary. using the same technique as the primary, wind around 18-20 turns of wire then another 18-20 turns. so u have a 40 turn center tapped primary. ok, that should be the transformer done, replace the cores and add a little glue to hold them together and your done.
The output stage can be what ever you want, i first used mosfets but they kept getting destroyed probably due to spikes and at $4 each i didnt want to waste any more money, so i used the transistors used in the pc powersupply. theyre good for around 8 amps each with good heatsinks. you cant drive them directly so u need to get another smaller npn transistor and connect the emmitor to the base of the big transistor, the collector of the small to the collector of the big transistor. the emmiter of the big transistor goes to ground, and the collector goes to the primary of the transformer. you can find these npn transistors in the pc power supply also. The center tapped winding goes to the 12v rail. same with the other side of the push pull configuration.
on the secondary side, just put a bridge rectifier with a few caps after it, center tap goes to ground. connect the positive output to the feedback terminal. with the variable resistor towards the output side, power it up (through a fuse, of course and not with a car battery just yet, use a 12v dc supply). there should be 5v on the output. now turn the variable resistor back and it should increase, not linearly though.
I think that has covered everything. if not, or if i didnt explain things as clearly as i hoped just email me and ill try some more. if theres enough requests ill create a website with detailed instrustions and a whole lot of pics for u guys, hope i helped.
-Ben
The 4050 is a buffer to give nice n quick rise and fall times for the output stage to keep the output transistors/mosfets out of their linear range to increase efficiency and reduce heat.
The transformer i used was out of a 200w at pc power supply, as was the tl494 and pretty much everything else (except the 4050) to seperate the cores without cracking the ferite i usually boil them in water for around 5 min and take them out and seperate them as quickly as possible. then you unwind the windings (save the wire, it could be usefull) and let the cores cool.
To wind the transformer you wind 4 turns of wire (because of the high frequency its best to combine a number of thinner wires to create the one wire to handle the current, i used 6 .8mm wires twisted into the one large wire using a power drill) then leave around 10cm of wire and cut them. then wind another 4 turns of wire over(or beside) the first 4 in the same direction and leave 10cm and cut them. now join the lead from the first primary (the one from the fourth turn) and join it to the start of the 1st winding of the second primary. so you have a 8 turn center tapped primary. get it?
Now for the secondary. using the same technique as the primary, wind around 18-20 turns of wire then another 18-20 turns. so u have a 40 turn center tapped primary. ok, that should be the transformer done, replace the cores and add a little glue to hold them together and your done.
The output stage can be what ever you want, i first used mosfets but they kept getting destroyed probably due to spikes and at $4 each i didnt want to waste any more money, so i used the transistors used in the pc powersupply. theyre good for around 8 amps each with good heatsinks. you cant drive them directly so u need to get another smaller npn transistor and connect the emmitor to the base of the big transistor, the collector of the small to the collector of the big transistor. the emmiter of the big transistor goes to ground, and the collector goes to the primary of the transformer. you can find these npn transistors in the pc power supply also. The center tapped winding goes to the 12v rail. same with the other side of the push pull configuration.
on the secondary side, just put a bridge rectifier with a few caps after it, center tap goes to ground. connect the positive output to the feedback terminal. with the variable resistor towards the output side, power it up (through a fuse, of course and not with a car battery just yet, use a 12v dc supply). there should be 5v on the output. now turn the variable resistor back and it should increase, not linearly though.
I think that has covered everything. if not, or if i didnt explain things as clearly as i hoped just email me and ill try some more. if theres enough requests ill create a website with detailed instrustions and a whole lot of pics for u guys, hope i helped.
-Ben
Attachments
yes I repeat palesha
But better post full schematic of power supply. We will like to see your layout of the PCB if you have. what ferrite core you have used toroid or EE and your observation & experience about calculating the winding turns. I also have trying on sg3525 but some problems with them so please put here your schematic and pcb layout with your experience
But better post full schematic of power supply. We will like to see your layout of the PCB if you have. what ferrite core you have used toroid or EE and your observation & experience about calculating the winding turns. I also have trying on sg3525 but some problems with them so please put here your schematic and pcb layout with your experience
Hi.
I have builded smps similar to http://valveaudio.tripod.com/images/schematics/switcher1.JPG but u used sg3525 and 2+2 buz11 mosfet for swtching primary.
SG3525 was driving this 4 mosfets directly. (75Khz)
I used EE core etd49 , 4+4 primary and 11+11 sec.
I menaged to run 1 channel 100w amplifer (35+35volts) at full power, with two channels I have voltage drops . That because I have wounded only 11+11 on secondary . When I found some wires Ill try wounding 19+19 on secondary to get more power.
Mosfets remainde cool without load and warm when full loaded, core is always cold. Im using very small heatsinks for now.
Also I haved bug in pcb, I have connected praimary and secondary ground and got some kinda distortion + noise when playing audio. When I have separeted primary and secondary ground everything was OK.
Concurently I'm trying preamplifiers , I want to be able to bridge to channels so I must integrate inverting signal to one channel.
Now for testing purpose Im using two old 6800uF/50v capacitors, first time I was using 1000uF/200volts capacitors, and I think thats smart because if something is wrong secondary voltage can get 120+120 volts or more (especialy without load and no feedback to sg3525 error amplifer)
Greetings
I have builded smps similar to http://valveaudio.tripod.com/images/schematics/switcher1.JPG but u used sg3525 and 2+2 buz11 mosfet for swtching primary.
SG3525 was driving this 4 mosfets directly. (75Khz)
I used EE core etd49 , 4+4 primary and 11+11 sec.
I menaged to run 1 channel 100w amplifer (35+35volts) at full power, with two channels I have voltage drops . That because I have wounded only 11+11 on secondary . When I found some wires Ill try wounding 19+19 on secondary to get more power.
Mosfets remainde cool without load and warm when full loaded, core is always cold. Im using very small heatsinks for now.
Also I haved bug in pcb, I have connected praimary and secondary ground and got some kinda distortion + noise when playing audio. When I have separeted primary and secondary ground everything was OK.
Concurently I'm trying preamplifiers , I want to be able to bridge to channels so I must integrate inverting signal to one channel.
Now for testing purpose Im using two old 6800uF/50v capacitors, first time I was using 1000uF/200volts capacitors, and I think thats smart because if something is wrong secondary voltage can get 120+120 volts or more (especialy without load and no feedback to sg3525 error amplifer)
Greetings
Another "vote" for the SG3525 chip here. It can drive FETs directly, which is a good feature - the TL494 requires an gate driver.
Mind you, if you're driving a lot of FETs with big gate capacitances (for some sort of insane high power project or something) then you'll need an external gate drive anyway, and it doesn't really matter what you use.
Throwing another chip into the mix, we use a UC3846 chip at work for doing this kind of stuff - creating high B+ voltages from batteries. The SG3525 is voltage mode and perfectly fine (especially in open loop), but if you're using feedback and don't mind spending a few extra bucks for a current transformer then it's probably the best thing out there.
now a question; one thing I've noticed (and never really understood) is that most car audio amps lack inductors after their secondary side rectifiers, which goes completely against my training and experience building this type of power supply. Apart from cost reasons, what kind of benefits / caveats arise in doing this?
Mind you, if you're driving a lot of FETs with big gate capacitances (for some sort of insane high power project or something) then you'll need an external gate drive anyway, and it doesn't really matter what you use.
Throwing another chip into the mix, we use a UC3846 chip at work for doing this kind of stuff - creating high B+ voltages from batteries. The SG3525 is voltage mode and perfectly fine (especially in open loop), but if you're using feedback and don't mind spending a few extra bucks for a current transformer then it's probably the best thing out there.
now a question; one thing I've noticed (and never really understood) is that most car audio amps lack inductors after their secondary side rectifiers, which goes completely against my training and experience building this type of power supply. Apart from cost reasons, what kind of benefits / caveats arise in doing this?
inductor issue
Good question. Here's the good answer. The caveats are numerous. Without an inductor the input current is drawn in short pulses or bursts. This makes the input filtering much more difficult. A very large input cap is needed to supply the short bursts of current. Also, the output filter caps are stressed very severely without an inductor, and the output ripple and noise is high. The power transistor(s) doing the switching must sustain a very large rms current due to the large ac ripple current present. Likewise with the transformer windings, the rectifiers, and all wiring and pcb traces. The increased ac ripple current in the transformer results in greater losses due to skin effect, and greater core losses due to the discontinuous core flux. An inductor reduces all of these stresses and provides a quieter output. With an inductor, efficiency is much higher, particularly when the inductor is operated in the continuous conduction mode. As far as benefits go (inductorless), the only one that comes to mind is the cost of the inductor, although the headaches caused by not having one just aren't worth the small savings. The increased cost of the transformer, and input and output filter caps surpasses the pocket change saved by omitting the inductor. I would never omit the inductor. Like you, doing such goes against my experience and judgement as well.
Good question. Here's the good answer. The caveats are numerous. Without an inductor the input current is drawn in short pulses or bursts. This makes the input filtering much more difficult. A very large input cap is needed to supply the short bursts of current. Also, the output filter caps are stressed very severely without an inductor, and the output ripple and noise is high. The power transistor(s) doing the switching must sustain a very large rms current due to the large ac ripple current present. Likewise with the transformer windings, the rectifiers, and all wiring and pcb traces. The increased ac ripple current in the transformer results in greater losses due to skin effect, and greater core losses due to the discontinuous core flux. An inductor reduces all of these stresses and provides a quieter output. With an inductor, efficiency is much higher, particularly when the inductor is operated in the continuous conduction mode. As far as benefits go (inductorless), the only one that comes to mind is the cost of the inductor, although the headaches caused by not having one just aren't worth the small savings. The increased cost of the transformer, and input and output filter caps surpasses the pocket change saved by omitting the inductor. I would never omit the inductor. Like you, doing such goes against my experience and judgement as well.
gmarsh said:
The only time you can get away with not using an output inductor is if your supply has no feedback and it runs at 50% duty cycle all the time. If you take a center tapped secondary with two out of phase square waves and rectify it, you get a fairly clean, constant DC signal even with only small amounts of filtering. Don't listen to anyone who tells you otherwise - they are probably blowing transistors, heating their output capacitors, and wondering why things are breaking.
I vote for the SG3525 as well for simple converters
I have a DIY car amp (200W RMS) using TL494. I don't think there is a lot of problems in using TL494 (I didn't have any, and they're rugged enough for me) than 3525 etc. SG3525's cost twice as much as a TL494. besides being cheap, I got several from old comp PSU's. I'm already in the way of making another car amp (~500W RMS) with the TL494 based PSU. SMPS is already done. just need the amp.
gud day KAbayan
could you posibbly send info/diagrams on how to utilize that TL494? i have so many links/pdfs here but so little info to gather.... yup, TL494 is way much cheaper than SG3525.... by at least 1/3 the price of 3525...
- how to make it a simple oscillator(no feedback) (as simple as it gets)
-how to operate TL494 with "softstart"
- howto on Deadtime
- how to make it a regulating(needs feedback)
hope you could help me on this....
bro, pls mail me at: beamrx5 at yahoo dot com
tnx
could you posibbly send info/diagrams on how to utilize that TL494? i have so many links/pdfs here but so little info to gather.... yup, TL494 is way much cheaper than SG3525.... by at least 1/3 the price of 3525...
- how to make it a simple oscillator(no feedback) (as simple as it gets)
-how to operate TL494 with "softstart"
- howto on Deadtime
- how to make it a regulating(needs feedback)
hope you could help me on this....
bro, pls mail me at: beamrx5 at yahoo dot com
tnx
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.