Hi people !
I've tested my GainClone with SMPS, using four ordinary PC power supplyes, and I found that it sounds much better indeed ! I'm impressed with the result.
But, these pc power supplyes can't deliver a good amout of current, they're very cheap, those chinese ones...
I found this SMPS at elliot's site, and I think that is very well designed:
http://sound.westhost.com/project89.htm
But, I need this to work on 127VAC. YES, I KNOW that I can't use it directly on the mains, because it's dangerous, so I'll have a transformer 127 --> 127Vac (I already have it), just for isolation. But I need a SMPS that work with 127VAC.
I was looking the datasheets of the mosfets from a computer supply, and they work till 400V !
What modifications I must do, to get this SMPS from elliots working on 127VAC?
Thank you very much !
I've tested my GainClone with SMPS, using four ordinary PC power supplyes, and I found that it sounds much better indeed ! I'm impressed with the result.
But, these pc power supplyes can't deliver a good amout of current, they're very cheap, those chinese ones...
I found this SMPS at elliot's site, and I think that is very well designed:
http://sound.westhost.com/project89.htm
But, I need this to work on 127VAC. YES, I KNOW that I can't use it directly on the mains, because it's dangerous, so I'll have a transformer 127 --> 127Vac (I already have it), just for isolation. But I need a SMPS that work with 127VAC.
I was looking the datasheets of the mosfets from a computer supply, and they work till 400V !
What modifications I must do, to get this SMPS from elliots working on 127VAC?
Thank you very much !
Please don't consider this... Elliots design is meant to operate from 12 Volts. 127 VAC will require a total redesign of the entire circuit.
For that much effort, you would be better to build a a simple linear supply... transformer-rectifier-cap.
😉
For that much effort, you would be better to build a a simple linear supply... transformer-rectifier-cap.
😉
Nando-
Since Rod's SMPS already uses a transformer, you really don't need an isolation transformer for use, but you might want to invest in one for the designing and testing stages. You could get away with just changing the MOSFETs in Rod Elliot's power supply with 500V MOSFETs (I would be cautious using the 400V ones), changing the winding ratios on the transformer, using hi-voltage insulation tape separating primary and secondaries, and providing a source to power the SG3525 chip itself. This would be a marginal approach, at best.
Pros and Cons.
The only Pro I can think of for using this approach is: each MOSFET is driven from a ground-referenced supply, eliminating the need for a floating-supply for the upper MOSFET.
The Cons include requirement of 800V MOSFETs with low Rds(on) (which can be a bit expensive) for the circuit to be able to operate off 240VAC (not necessary for your application, though), and very high voltage spikes (~700V) associated with the push-pull topology at these input levels.
Because of this, the center-tap Push-pull topology is very rarely used for line-voltage applications. It is used almost always for low-voltage inputs like 12- or 24VDC. BTW, is this for abord ship or on an aircraft? 127VAC sounds a little unusual for a ground source.
A much better approach would be to modify one of those "cheap Chinese" SMPSs. This is not as complicated as it might sound, and if done properly, can supply more than enough power for your gainclones. Any supply over, say, 230W would be a good candidate for conversion.
There a number of threads in the Power Supply Design Forum on this very subject, so hit the "search" button and I would recommend looking at each of these before proceeding. Each will give good insight and different results obtained. Let us know what rout you want to go, there are a number os PS gurus that can hekp (EVA, Jackinnj, etc.) Hope this helps.
Cheers,
Steve
Since Rod's SMPS already uses a transformer, you really don't need an isolation transformer for use, but you might want to invest in one for the designing and testing stages. You could get away with just changing the MOSFETs in Rod Elliot's power supply with 500V MOSFETs (I would be cautious using the 400V ones), changing the winding ratios on the transformer, using hi-voltage insulation tape separating primary and secondaries, and providing a source to power the SG3525 chip itself. This would be a marginal approach, at best.
Pros and Cons.
The only Pro I can think of for using this approach is: each MOSFET is driven from a ground-referenced supply, eliminating the need for a floating-supply for the upper MOSFET.
The Cons include requirement of 800V MOSFETs with low Rds(on) (which can be a bit expensive) for the circuit to be able to operate off 240VAC (not necessary for your application, though), and very high voltage spikes (~700V) associated with the push-pull topology at these input levels.
Because of this, the center-tap Push-pull topology is very rarely used for line-voltage applications. It is used almost always for low-voltage inputs like 12- or 24VDC. BTW, is this for abord ship or on an aircraft? 127VAC sounds a little unusual for a ground source.
A much better approach would be to modify one of those "cheap Chinese" SMPSs. This is not as complicated as it might sound, and if done properly, can supply more than enough power for your gainclones. Any supply over, say, 230W would be a good candidate for conversion.
There a number of threads in the Power Supply Design Forum on this very subject, so hit the "search" button and I would recommend looking at each of these before proceeding. Each will give good insight and different results obtained. Let us know what rout you want to go, there are a number os PS gurus that can hekp (EVA, Jackinnj, etc.) Hope this helps.
Cheers,
Steve
Poobah, why? The implementation of the PWM control will be exactly the same ! I don't think so... 😉
Steve, thank you for your interest !
As you said, the elliots SMPS already have a transformer, so if I convert it in a 127VAC SMPS, I'll can connect it to the mains directly. Anyway, I have a isolator transformer, and I can use it if it's necessary.
I don't know yet how many of these SMPS I'll have to build, all them will have to be 127VAC, and I'm sure that isn't few supplyes that I'm going to build. If I modify the Chineese SMPS, I can't be sure about how reliable that thing will be, they aren't well built at all. Another con, is that if I modify the chineese, I'll can't intregrate it on the amplifier PCB
So I think that convert the elliot's SMPS to 127VAC is really the best way!
As I saw in the schematic, appears to be like you said, just change the primary windings of the transformer and the mosfets !
So, how can I calculate how many turns of copper wire and what copper wire will I need ? It's related to the mosfets right? Any sugestion for mosfets?! 🙂
Thank you !
Steve, thank you for your interest !
As you said, the elliots SMPS already have a transformer, so if I convert it in a 127VAC SMPS, I'll can connect it to the mains directly. Anyway, I have a isolator transformer, and I can use it if it's necessary.
I don't know yet how many of these SMPS I'll have to build, all them will have to be 127VAC, and I'm sure that isn't few supplyes that I'm going to build. If I modify the Chineese SMPS, I can't be sure about how reliable that thing will be, they aren't well built at all. Another con, is that if I modify the chineese, I'll can't intregrate it on the amplifier PCB
So I think that convert the elliot's SMPS to 127VAC is really the best way!
As I saw in the schematic, appears to be like you said, just change the primary windings of the transformer and the mosfets !
So, how can I calculate how many turns of copper wire and what copper wire will I need ? It's related to the mosfets right? Any sugestion for mosfets?! 🙂
Thank you !
Why?
1) Transformer design for an SMPS is not a light subject.
2) You will need a bootstrap supply for the 12 Volts to run the chips.
3) If you have to ask how to change the transformer; you probably should not be working with an offline SMPS.
4) By the time you put in an isolating transformer, why not a rectifier and some caps?
😉
1) Transformer design for an SMPS is not a light subject.
2) You will need a bootstrap supply for the 12 Volts to run the chips.
3) If you have to ask how to change the transformer; you probably should not be working with an offline SMPS.
4) By the time you put in an isolating transformer, why not a rectifier and some caps?
😉
Nando-
The suggestion of modding the center-tap push-pull of Rod's design was only a suggestion. As I said in my previous post, this center-tap push-pull topology is almost exclusively used for low-voltage inputs, like 12-24VDC. You could do it this way, but I remind you of the cautions I stated earlier.
Fos a line voltage application (like 120-240VAC), you are better off going with the half-bridge, like those found in the chinese cheapies.
BYW, what makes the unreliable cheapies unreliable are bad parts- and build-quality, NOT design quality.
As poobah suggests, designing or modifying high-frequency SMPS power transformers is not a light subject, and much reading should take place before you attempt anything at line voltages. Also, you will need a small "housekeeping" supply to power the PWM chip. This can be either a small 50-60Hz transformer rated at 12V, or this can be a starved-zener regulator to power the chip at start-up. After that, an auxiliary winding on the main power transformer will provide power to the PWM chip.
Alomst forgot, it goes without saying that in order for either of these suggestions to work off 127VAC, you will need a common-mode line filter, followed by a diode bridge of suitable voltage and current ratings, ending with a capacitor filter of sufficient value and voltage rating. Roughly 220mF to 470mF rated at 200V will do.
As for MOSFET suggestions, any n-channel unit rated at 500V should do, as long as it meets your power level requirements. A good candidate would ne an IRF840 (500V, 8A), or bigger. Since I don't know the power levels (output voltages and currents), I can't recommend a specific MOSFET.
For a better guide, go to ONSemi's website and look around in their app notes and datasheets sections for the Switchmode Design guide. This .PDF doc is downloadable, and is an indispensable aid for choosing components for your SMPS.
Again, PLEASE do much research before you attempt anything at line-voltage levels.
Steve
The suggestion of modding the center-tap push-pull of Rod's design was only a suggestion. As I said in my previous post, this center-tap push-pull topology is almost exclusively used for low-voltage inputs, like 12-24VDC. You could do it this way, but I remind you of the cautions I stated earlier.
Fos a line voltage application (like 120-240VAC), you are better off going with the half-bridge, like those found in the chinese cheapies.
BYW, what makes the unreliable cheapies unreliable are bad parts- and build-quality, NOT design quality.
As poobah suggests, designing or modifying high-frequency SMPS power transformers is not a light subject, and much reading should take place before you attempt anything at line voltages. Also, you will need a small "housekeeping" supply to power the PWM chip. This can be either a small 50-60Hz transformer rated at 12V, or this can be a starved-zener regulator to power the chip at start-up. After that, an auxiliary winding on the main power transformer will provide power to the PWM chip.
Alomst forgot, it goes without saying that in order for either of these suggestions to work off 127VAC, you will need a common-mode line filter, followed by a diode bridge of suitable voltage and current ratings, ending with a capacitor filter of sufficient value and voltage rating. Roughly 220mF to 470mF rated at 200V will do.
As for MOSFET suggestions, any n-channel unit rated at 500V should do, as long as it meets your power level requirements. A good candidate would ne an IRF840 (500V, 8A), or bigger. Since I don't know the power levels (output voltages and currents), I can't recommend a specific MOSFET.
For a better guide, go to ONSemi's website and look around in their app notes and datasheets sections for the Switchmode Design guide. This .PDF doc is downloadable, and is an indispensable aid for choosing components for your SMPS.
Again, PLEASE do much research before you attempt anything at line-voltage levels.
Steve
Nando & N-Channel,
Good points N.
The motivators for a switching supply are:
1) High efficiency... less energy... less waste heat.
2) Reduced weight size; unless an isolation trans is required.
If you're dying to try your hand at an SMPS design, play with a simple buck convertor first, then a boost convertor, then a flyback. Most other designs derive from these basics... It's fun to learn; but it is complicated.
"Linear Technologies" application note #19 is a great place to start (free, on-line).
🙂
Good points N.
The motivators for a switching supply are:
1) High efficiency... less energy... less waste heat.
2) Reduced weight size; unless an isolation trans is required.
If you're dying to try your hand at an SMPS design, play with a simple buck convertor first, then a boost convertor, then a flyback. Most other designs derive from these basics... It's fun to learn; but it is complicated.
"Linear Technologies" application note #19 is a great place to start (free, on-line).
🙂
- Status
- Not open for further replies.