Hi all!
Can anybody give me some advice on
SMPS transformer core/winding calculations.
Things are known:
1)Shematics/Topology = Half-Bridge/Unregulated;
2)SG3525 PWM IC frequency = 85kHz;
3)Transformer Core = ETD59;
I suppose that there are two separate SMPS for
each Amplifier channel.
I was using those Amps in reality. One Amp had
sufficient power to drive Celestion CXi1812 1200w
Subwoofer connected to each channel.
Because it takes a long time to design a transformer
from scratch, I am ready to listen to all ideas of
yours.
Maube some of you already has built a 500-1000W
Suply using ETD59 or even TOROID Core.It would be
great to take a look at this.
Can anybody give me some advice on
SMPS transformer core/winding calculations.
Things are known:
1)Shematics/Topology = Half-Bridge/Unregulated;
2)SG3525 PWM IC frequency = 85kHz;
3)Transformer Core = ETD59;
I suppose that there are two separate SMPS for
each Amplifier channel.
I was using those Amps in reality. One Amp had
sufficient power to drive Celestion CXi1812 1200w
Subwoofer connected to each channel.
Because it takes a long time to design a transformer
from scratch, I am ready to listen to all ideas of
yours.
Maube some of you already has built a 500-1000W
Suply using ETD59 or even TOROID Core.It would be
great to take a look at this.
Digital Boy,
Couple o' questions:
1) What is your source? +12V (car) or 220/240VAC mains?
2) Have you considered full-bridge? This topology is called for in circuits over 1kW.
3) Have you considered sybchronizing the two SG3525s?
If you are powering this from a 12V source, I hope you have some MOSFETs with a very low Rds(on), like 1mW.
If you are powering them from 220/240VAC mains, you will still need some very robust MOSFETs.
I would recommend synchronizing one SG3525's clock to the other's. If both SMPSs are located in the same enclosure, then this is critical, as beat frequencies will be generated from two close, but not exact, matching frequencies. Reference datasheets at OnSemi, SGS Thompson or Texas Instrumentsto see how to do this. The '3525 is a very good chip. If each SMPS is located in separate enclosures, then synchronization need not apply.
For your transformer, you could use a toroid, but make sure the size and core material are right for 85kHz. Speaking of which, is 85kHz your clock frequency or your switching frequency? If it is your clock freq., then your switching freq. will be half that (42.5kHz), as it taked two clock pulses to complete one PWM cycle. I have used Ferrite toroids from Amidon Corp. with good results. For a power level of 1.1kW, I would recommend the following core, FT-240-77 (Ferrite Toroid, 2.40" outer diameter, #77 material). per Amidon's own charts, this core is good for in excess of 1.7kW @ 20kHz. At higher frequencies, you can squeeze more power through it, but I would be wary of increased core losses.
Hope this helps. BTW, Eva might chime in here shortly, as she has done a full-birdge up to 1.8kW, complete with pictures.
Steve
Couple o' questions:
1) What is your source? +12V (car) or 220/240VAC mains?
2) Have you considered full-bridge? This topology is called for in circuits over 1kW.
3) Have you considered sybchronizing the two SG3525s?
If you are powering this from a 12V source, I hope you have some MOSFETs with a very low Rds(on), like 1mW.
If you are powering them from 220/240VAC mains, you will still need some very robust MOSFETs.I would recommend synchronizing one SG3525's clock to the other's. If both SMPSs are located in the same enclosure, then this is critical, as beat frequencies will be generated from two close, but not exact, matching frequencies. Reference datasheets at OnSemi, SGS Thompson or Texas Instrumentsto see how to do this. The '3525 is a very good chip. If each SMPS is located in separate enclosures, then synchronization need not apply.
For your transformer, you could use a toroid, but make sure the size and core material are right for 85kHz. Speaking of which, is 85kHz your clock frequency or your switching frequency? If it is your clock freq., then your switching freq. will be half that (42.5kHz), as it taked two clock pulses to complete one PWM cycle. I have used Ferrite toroids from Amidon Corp. with good results. For a power level of 1.1kW, I would recommend the following core, FT-240-77 (Ferrite Toroid, 2.40" outer diameter, #77 material). per Amidon's own charts, this core is good for in excess of 1.7kW @ 20kHz. At higher frequencies, you can squeeze more power through it, but I would be wary of increased core losses.
Hope this helps. BTW, Eva might chime in here shortly, as she has done a full-birdge up to 1.8kW, complete with pictures.
Steve
SMPS for AF Amp
Hi N-Channel. Can you maybe get hold on that design please. I'm interested. I want to use it with a 500W X 2 amp. Powered from 230V mains supply. I want to build an amplifier with high power, but isn't much effort to travel with. (Weight wise). I have only a design for a 300W - half bridge. Thanks.
Hi N-Channel. Can you maybe get hold on that design please. I'm interested. I want to use it with a 500W X 2 amp. Powered from 230V mains supply. I want to build an amplifier with high power, but isn't much effort to travel with. (Weight wise). I have only a design for a 300W - half bridge. Thanks.
This thread shows my 1800W full bridge (and also previous development stages where output was only 1KW). The low-voltage output section for 0-15V with synchronos rectification may be just replaced by some higher voltage windings, coupled buck inductors and conventional diodes.
http://www.diyaudio.com/forums/showthread.php?s=&threadid=60969
Be warned that this is not a conventional design, as I'm employing just four $1 bipolar transistors as switches, the thing works at 32Khz, and I'm making extensive use of current transformers, as opposed to nowadays trends towards expensive floating gate drive ICs, frequencies above 100Khz, dozens of paralelled MOSFETs, and plenty of air cooling (required due to the poor resulting efficiencies in the range of 80%).
http://www.diyaudio.com/forums/showthread.php?s=&threadid=60969
Be warned that this is not a conventional design, as I'm employing just four $1 bipolar transistors as switches, the thing works at 32Khz, and I'm making extensive use of current transformers, as opposed to nowadays trends towards expensive floating gate drive ICs, frequencies above 100Khz, dozens of paralelled MOSFETs, and plenty of air cooling (required due to the poor resulting efficiencies in the range of 80%
).- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.