Easyamp,
Hi,I didn't really understand what you explained, but If it was refering to using the AT transformer then thats not what I meant. meaning if your calculations where based on using the 494 whit a rewinded AT transformer.
jackinnj,
I know that i cant use an AT transformer, I meant using a bigger transformer, maby a toriod for a 500 watt SMPS.
But I had an idea for an off line SMPS, why cant we wind a bigger transformer, and replace it with the small one on the AT/ATX board(we'll conect the transformer with wires because it will be to big for the board, and also other moddes to make it work) as well as higher output voltage on the seconderys so you can get more then 5V+and 12v+/.
will it work? because my mom and dad want me to make a 5 channel amp(based on OPA549) for our DVD player. and a 500 watt off line SMPS will be great for it. (a 500VA toriod in Isaerl will cost between $100-$150 out of my budget)
Hi,I didn't really understand what you explained, but If it was refering to using the AT transformer then thats not what I meant. meaning if your calculations where based on using the 494 whit a rewinded AT transformer.
jackinnj,
I know that i cant use an AT transformer, I meant using a bigger transformer, maby a toriod for a 500 watt SMPS.
But I had an idea for an off line SMPS, why cant we wind a bigger transformer, and replace it with the small one on the AT/ATX board(we'll conect the transformer with wires because it will be to big for the board, and also other moddes to make it work) as well as higher output voltage on the seconderys so you can get more then 5V+and 12v+/.
will it work? because my mom and dad want me to make a 5 channel amp(based on OPA549) for our DVD player. and a 500 watt off line SMPS will be great for it. (a 500VA toriod in Isaerl will cost between $100-$150 out of my budget)
Kinser --
the torroid you are talking about is for operating at line frequencies, a torroid for an off-line switcher is much smaller and considerably less expensive --
if they have it in the budget to spend $100 on a transformer great -- you can build a linear supply which is much quieter, and infinitely simpler to implement.
the torroid you are talking about is for operating at line frequencies, a torroid for an off-line switcher is much smaller and considerably less expensive --
if they have it in the budget to spend $100 on a transformer great -- you can build a linear supply which is much quieter, and infinitely simpler to implement.
jackinnj,
I was not talking about the big 240Vac powerd toriodes, i was talking about a suituble toriod CORE for an SMPS. I would rather try make an off line SMPS, because it will be cheaper for me, a price of an ATX PSU will be about $35 and an SMPS xformer for about 10$ in total.
But is it possible to do this with an ATX PSU?
Thanks,
Kinser
I was not talking about the big 240Vac powerd toriodes, i was talking about a suituble toriod CORE for an SMPS. I would rather try make an off line SMPS, because it will be cheaper for me, a price of an ATX PSU will be about $35 and an SMPS xformer for about 10$ in total.
But is it possible to do this with an ATX PSU?
Thanks,
Kinser
TOROIDS
Kins-
Amidon Corporation of California sells toroids of all sizes. For a 500W supply, either +12VDC or off-line, try their FT-240-77 core.
This toroid measures 2.40" (outer diameter), and made of their #77 ferrite material, and has sufficient cross-sectional area to pass ALOT of power. The nomograph for the #77 material at Amidon's website shows, that at a 40kHz switching frequency (80kHz clock frequency), the FT-240-77 can pass through 1.7kW peak, and 1.0kW easily, at 40kHz. The max power varies with frequency. For anything above 500w for an off-line unit, you will want to use a FULL-BRIDGE topology. Your driver circuit will be a little more complicated (you're now driving 4 transistors instead of two), but you will have lower primary-side currents.
BTW, what is Israel? 120V or 240V? 50Hz or 60Hz? Not that the Hz matters, I was just wondering.
Steve
Kins-
Amidon Corporation of California sells toroids of all sizes. For a 500W supply, either +12VDC or off-line, try their FT-240-77 core.
This toroid measures 2.40" (outer diameter), and made of their #77 ferrite material, and has sufficient cross-sectional area to pass ALOT of power. The nomograph for the #77 material at Amidon's website shows, that at a 40kHz switching frequency (80kHz clock frequency), the FT-240-77 can pass through 1.7kW peak, and 1.0kW easily, at 40kHz. The max power varies with frequency. For anything above 500w for an off-line unit, you will want to use a FULL-BRIDGE topology. Your driver circuit will be a little more complicated (you're now driving 4 transistors instead of two), but you will have lower primary-side currents.
BTW, what is Israel? 120V or 240V? 50Hz or 60Hz? Not that the Hz matters, I was just wondering.
Steve
N-Channel,
Do YOU thing that If I where to change the transformer of an ATX power supply (300watts) with a bigger toriod transformer(about 500+ but not peak), with higher output voltage (about 28Vdc+/-) and change the filter and caps and a bit more mods i will have a good power supply to power 5 OPA549's?
Because it will be in my budget (as i mention befor I can get a 300 watt ATX PSU for about $35)
BTW In Israel the voltage is 240V and I think 60Hz.
Thanks,
Kinser
Do YOU thing that If I where to change the transformer of an ATX power supply (300watts) with a bigger toriod transformer(about 500+ but not peak), with higher output voltage (about 28Vdc+/-) and change the filter and caps and a bit more mods i will have a good power supply to power 5 OPA549's?
Because it will be in my budget (as i mention befor I can get a 300 watt ATX PSU for about $35)
BTW In Israel the voltage is 240V and I think 60Hz.
Thanks,
Kinser
Re: TOROIDS
I used to purchase torroid cores from Amidon when their distributor was Byetmark, they sell small quantities to ham radio operators and the price for the core that Steve suggests is $9.00. It is going to a bit of a tight squeeze to fit a core that sized onto an ATX chasis --
btw, take a look at the underside of any ATX or other off line switching supply -- you will see that great care is taken to isolate the "line side" of the PS from the DC side -- the peak voltages and currents can be much greater than what one comes to suspect working on linear supplies -- I really want you to be careful if you start playing around with a switcher.
N-Channel said:
I used to purchase torroid cores from Amidon when their distributor was Byetmark, they sell small quantities to ham radio operators and the price for the core that Steve suggests is $9.00. It is going to a bit of a tight squeeze to fit a core that sized onto an ATX chasis --
btw, take a look at the underside of any ATX or other off line switching supply -- you will see that great care is taken to isolate the "line side" of the PS from the DC side -- the peak voltages and currents can be much greater than what one comes to suspect working on linear supplies -- I really want you to be careful if you start playing around with a switcher.
Hi Kinser,
On the second page second post you made the above comment, and If I understand what your saying correctly you wish to rewrap the AT core for use with a 12v primary and you wish to get 56vdc +/- at the secondaries and use a tl494 control circuit.
If that is correct then I wanted to let you know that in square wave circuits like the one you will build there is no 1.41 rms multiplier like you have shown with your "40-0-40=56Vdc+/-".
In a sine wave or a standard line driven device 50~60 cps you'll have a sine wave with a +peak a zero crossing and a -peak along a curving line . Along this curving line you'll have an infinite number of times you could compute the power delivered to your load, most of them being incorrect. This is where root mean squared (rms) comes in. It's a simple calculation to find the average power being delivered to a load in a sine wave based circuit. This however does not work so well with transformers because the transformers inductance cause's the current wave to lag behind the voltage wave, but this is not important now.
In a square wave based system as most switch mode supplies are you need not worry about rms so much becuase your dealing with a square wave that is either fully negative (or ground) or fully positive with as little time as possible anywhere else, well in an unregulated push pull forward converter anyway, there are many topollogies that zero in between switching. So if you wish to get 56Vdc+/- you will need to generate peak to peak square wave voltages of roughly 112Vpp for each secondary leg of transformer or a total of 224Vpp because a 50% duty cycle square wave when rectiifed to DC is about have its square wave peak to peak value. When you calculate your turns use 24vdc for your primary. So you need roughly 1:4 wind ratio on your transformer.
Cheers,
On the second page second post you made the above comment, and If I understand what your saying correctly you wish to rewrap the AT core for use with a 12v primary and you wish to get 56vdc +/- at the secondaries and use a tl494 control circuit.
If that is correct then I wanted to let you know that in square wave circuits like the one you will build there is no 1.41 rms multiplier like you have shown with your "40-0-40=56Vdc+/-".
In a sine wave or a standard line driven device 50~60 cps you'll have a sine wave with a +peak a zero crossing and a -peak along a curving line . Along this curving line you'll have an infinite number of times you could compute the power delivered to your load, most of them being incorrect. This is where root mean squared (rms) comes in. It's a simple calculation to find the average power being delivered to a load in a sine wave based circuit. This however does not work so well with transformers because the transformers inductance cause's the current wave to lag behind the voltage wave, but this is not important now.
In a square wave based system as most switch mode supplies are you need not worry about rms so much becuase your dealing with a square wave that is either fully negative (or ground) or fully positive with as little time as possible anywhere else, well in an unregulated push pull forward converter anyway, there are many topollogies that zero in between switching. So if you wish to get 56Vdc+/- you will need to generate peak to peak square wave voltages of roughly 112Vpp for each secondary leg of transformer or a total of 224Vpp because a 50% duty cycle square wave when rectiifed to DC is about have its square wave peak to peak value. When you calculate your turns use 24vdc for your primary. So you need roughly 1:4 wind ratio on your transformer.
Cheers,
jackinnj,
I will be careful, for you and for me!
What I was thinking of is taking the board out of the ATX box with the new transformer mounted beside it... But befor I go and buy an ATX PSU, I need someone to tell if I can do this for shure.! because I really whant to do this. I will have to change the caps, and the filter on the output (correct me if im wrong) and many more things.
Another thing is on the ATX article the output was 31-0-31 but after rectification it was 14Vdc, Is this got to do with what easyamp expaind? And how do you calculate the rea Dc output compared to what the winding output is?
Thanks,
Kinser
I will be careful, for you and for me!
What I was thinking of is taking the board out of the ATX box with the new transformer mounted beside it... But befor I go and buy an ATX PSU, I need someone to tell if I can do this for shure.! because I really whant to do this. I will have to change the caps, and the filter on the output (correct me if im wrong) and many more things.
Another thing is on the ATX article the output was 31-0-31 but after rectification it was 14Vdc, Is this got to do with what easyamp expaind? And how do you calculate the rea Dc output compared to what the winding output is?
Thanks,
Kinser
Toroid Core
Kinser-
As Jackinnj suggests, you will not be able to stay inside the AT/ATX box if you go with the FT-240-77 core, as it is too big to fit inside the case, but you can do it on your own boards. Also, he very astutely points out that there is a big separation betweeen the Hi-voltage input section and the low-voltasge DC output side of the board. This is no accident: the separation is necessary for safety.
If you want to get large amounts of power from a smaller core, try STACKING two FT-140-77 cores. Same material, but with a 1.40" o.d. and peak power capacity @ 40kHz of something like 300-350W, each core.
Also, by changing the output section to +/- 28VDC, you will have to go with a new board, one of your own design, as the existing ATX board does not have the necessary layout to accomodate a high-current negative side supply. Since you will have to go with a new board, your choices for new output caps are much bigger. You can get away with 35V-rated caps, but for safety's sake, I would go with 50V units. Some output diodes to consider are the MUR1620CT and the MUR1620CTR. The 'CT is a common-anode pair for the (+) half of the outout, and the 'CTR is a common-cathode pair for the (-) half of the supply. This setup is found in almost every commercial car amplifier. Look up MUR1620CT & CTR at the following Onsemi links:
http://www.onsemi.com/pub/Collateral/MUR1620CT-D.PDF
http://www.onsemi.com/pub/Collateral/MURB1620CTR-D.PDF
Steve
Kinser-
As Jackinnj suggests, you will not be able to stay inside the AT/ATX box if you go with the FT-240-77 core, as it is too big to fit inside the case, but you can do it on your own boards. Also, he very astutely points out that there is a big separation betweeen the Hi-voltage input section and the low-voltasge DC output side of the board. This is no accident: the separation is necessary for safety.
If you want to get large amounts of power from a smaller core, try STACKING two FT-140-77 cores. Same material, but with a 1.40" o.d. and peak power capacity @ 40kHz of something like 300-350W, each core.
Also, by changing the output section to +/- 28VDC, you will have to go with a new board, one of your own design, as the existing ATX board does not have the necessary layout to accomodate a high-current negative side supply. Since you will have to go with a new board, your choices for new output caps are much bigger. You can get away with 35V-rated caps, but for safety's sake, I would go with 50V units. Some output diodes to consider are the MUR1620CT and the MUR1620CTR. The 'CT is a common-anode pair for the (+) half of the outout, and the 'CTR is a common-cathode pair for the (-) half of the supply. This setup is found in almost every commercial car amplifier. Look up MUR1620CT & CTR at the following Onsemi links:
http://www.onsemi.com/pub/Collateral/MUR1620CT-D.PDF
http://www.onsemi.com/pub/Collateral/MURB1620CTR-D.PDF
Steve
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.