• 回复
  • 收藏
  • 点赞
  • 分享
  • 发新帖

500W隔离式DC/DC模块电源

39-72V宽范围输入电压,输出24V/500W,隔离式通信模块电源,开关频率200KHz,使用什么拓扑效率才能最高呢?
1.PWM全桥+同步整流.较低的输入电压下移相全桥可能没有什么优势,甚至环流的存在可能使效率降低,LLC半桥可能也是如此;
2.推挽正激+同步整流.这是一种新的电路拓扑,但未见实际的产品设计;
3.两级结构Buck+开环全桥+同步整流.效率上不知道有无优势,不过采用更多器件,成本自然就上去了;

有没有人实际做过类似的设计?讨论一下各种拓扑的优、缺点.
全部回复(54)
正序查看
倒序查看
正弦芯
LV.8
2
2008-06-21 15:16
看看电流型推挽+同步整流是否合适?
0
回复
wjjzqy
LV.6
3
2008-06-21 16:42
@正弦芯
看看电流型推挽+同步整流是否合适?
电流型推挽+同步整流或3.
电流型推挽+同步整流为优选
0
回复
2008-06-21 22:00
这是“Temp1234”的回复:

Hi, Becareful patents issue:
2.推挽正激+同步整流.这是一种新的电路拓扑,但未见实际的产品设计;
Please check this is a patent of Bronad TeleCom Power.
3.两级结构Buck+开环全桥+同步整流. This one should belong to Synqor!
Good Luck!
0
回复
2008-06-21 22:01
@wjjzqy
电流型推挽+同步整流或3.电流型推挽+同步整流为优选
啥叫"电流型"推挽?指的是电流模式控制还是...?
0
回复
2008-06-22 08:30
3525加全桥比较好
0
回复
2008-06-22 13:51
@江湖电源
3525加全桥比较好
3525?太老的东西,又是电压模式控制的,如果是PWM全桥还是使用uc38025/uc3808吧
0
回复
呆头鹅
LV.7
8
2008-06-22 21:17
为何不考虑用有源钳位?
0
回复
themarine
LV.5
9
2008-06-22 21:59
有源箝位正激,肖特基整流,我们这700W都可以做.
或则有源箝位+正反激做.24V用同步整流可能没什么优势吧,如果实在想用觉得做两个同步整流的12v,叠加到一起效果可能好点吧.
楼主没说关键的一点,模块的尺寸大小如何!?是全砖,半砖,还是1/4砖,还是定制模块!?
0
回复
wjjzqy
LV.6
10
2008-06-23 08:19
@斜阳古道
啥叫"电流型"推挽?指的是电流模式控制还是...?
就是电流模式控制
0
回复
2008-06-23 09:44
这么大的功率,并且是一个全砖,用有源钳位的话,磁芯利用率不太高、原边电流太大,使用两象限拓扑可能更合适一点...关注的是效率与成本.

楼上的,你这700W在48V输入时能做到多大的效率?另外,你的输出也是24V?
0
回复
temp1234
LV.5
12
2008-06-23 10:46
@斜阳古道
这么大的功率,并且是一个全砖,用有源钳位的话,磁芯利用率不太高、原边电流太大,使用两象限拓扑可能更合适一点...关注的是效率与成本.楼上的,你这700W在48V输入时能做到多大的效率?另外,你的输出也是24V?
Hi, A FB/500W should be not too hard, you had better forcus on cost! The most curent products are 700W/SR, and they are all focus on the efficiency. You can refer to ERICSON, SYNQOR, DELTA and POWER ONE.
0
回复
2008-06-23 15:40
@斜阳古道
这么大的功率,并且是一个全砖,用有源钳位的话,磁芯利用率不太高、原边电流太大,使用两象限拓扑可能更合适一点...关注的是效率与成本.楼上的,你这700W在48V输入时能做到多大的效率?另外,你的输出也是24V?
我觉得直接用硬开关全桥来做就可以了,推荐你intersil的ISL6742,有同步驱动信号端.
0
回复
themarine
LV.5
14
2008-06-24 11:20
@斜阳古道
这么大的功率,并且是一个全砖,用有源钳位的话,磁芯利用率不太高、原边电流太大,使用两象限拓扑可能更合适一点...关注的是效率与成本.楼上的,你这700W在48V输入时能做到多大的效率?另外,你的输出也是24V?
是28V输出的,效率就90%多点.用有源钳位做的话效率相对低点,但成本也较低.用硬开关全桥做也是一个好方案.
另外楼主你这东东有没有EMC要求的,这个得事先考虑进去,不然会搞死人的.
0
回复
2008-06-24 11:24
关注,支持!
0
回复
2008-06-24 15:39
各位朋友,其实我并不是要做产品设计,只是关注模块电源这个方面的设计,怎样才是效率较高的电路拓扑,同时成本又不太高?
还是觉得“两级结构Buck+开环全桥+同步整流”的效率应该最好,Temp1234讲这是Synqor的专利?可以贴出这个专利或给出专利号吗?谢谢
0
回复
temp1234
LV.5
17
2008-06-24 16:27
@斜阳古道
各位朋友,其实我并不是要做产品设计,只是关注模块电源这个方面的设计,怎样才是效率较高的电路拓扑,同时成本又不太高?还是觉得“两级结构Buck+开环全桥+同步整流”的效率应该最好,Temp1234讲这是Synqor的专利?可以贴出这个专利或给出专利号吗?谢谢
Synqor has around 20 patents for these types, cover Interleave Forward, Push-Pull, HB and FB topoligrs!

For your reference:11/16/2007

SynQor Announces Patent Infringement Lawsuit

Boxborough, MA - November 16th 2007 - SynQor announced today that it filed a lawsuit against several of its competitors for infringement of three patents relating to bus converters and/or non-isolated point of load converters used in intermediate bus architectures. The patents at issue are U.S. patents 7,072,190, 7,272,021, and 7,269,034. The suit was filed in Federal Court in the Eastern District of Texas.

The defendants in the action are Artesyn Technologies, Inc., Astec America, Inc., Emerson Electric, Co., Bel Fuse Inc., Cherokee International Corp., Delta Electronics, Inc., Delta Products Corp., Murata Manufacturing Co., Ltd., Murata Electronics North America, Inc., Power-One, Inc., Tyco Electronics Corp., and Tyco Electronics Ltd.

Sidley Austin LLP will be representing SynQor in this litigation.
0
回复
2008-06-24 17:54
@temp1234
Synqorhasaround20patentsforthesetypes,coverInterleaveForward,Push-Pull,HBandFBtopoligrs!Foryourreference:11/16/2007SynQorAnnouncesPatentInfringementLawsuitBoxborough,MA-November16th2007-SynQorannouncedtodaythatitfiledalawsuitagainstseveralofitscompetitorsforinfringementofthreepatentsrelatingtobusconvertersand/ornon-isolatedpointofloadconvertersusedinintermediatebusarchitectures.ThepatentsatissueareU.S.patents7,072,190,7,272,021,and7,269,034.ThesuitwasfiledinFederalCourtintheEasternDistrictofTexas.ThedefendantsintheactionareArtesynTechnologies,Inc.,AstecAmerica,Inc.,EmersonElectric,Co.,BelFuseInc.,CherokeeInternationalCorp.,DeltaElectronics,Inc.,DeltaProductsCorp.,MurataManufacturingCo.,Ltd.,MurataElectronicsNorthAmerica,Inc.,Power-One,Inc.,TycoElectronicsCorp.,andTycoElectronicsLtd.SidleyAustinLLPwillberepresentingSynQorinthislitigation.
"Artesyn Technologies, Inc., Astec America, Inc., Emerson Electric, Co., ..."

Yeah, I knowed the patent fight. But I believe there some differences between the competitors and Synqor. For example, what SynQor announced were mainly the bus converters. But we can only do very small change. It should be a appropriate strategy.

Thank for your information.
0
回复
temp1234
LV.5
19
2008-06-24 18:10
@斜阳古道
"ArtesynTechnologies,Inc.,AstecAmerica,Inc.,EmersonElectric,Co.,..."Yeah,Iknowedthepatentfight.ButIbelievetheresomedifferencesbetweenthecompetitorsandSynqor.Forexample,whatSynQorannouncedweremainlythebusconverters.Butwecanonlydoverysmallchange.Itshouldbeaappropriatestrategy.Thankforyourinformation.
OK! Synqor's Ideas are using a step-down feed-back form secondary to regualtion owrking-voltage(or bus-voltage) and then using 50/50 ratio DC transformer to direct divide target output voltage to secondary! This stage has many versions, which cover Interleave Forward, Push-Pull, HB and FB. Using "电流型" so the smooth chock is in primary side. due to the Dc transformer stage, synqor suit his compitors. It's conceptial Infringement, so it will take long time to fighting!

I hate typing, so every time I try fewest word to explain it!
0
回复
2008-06-24 19:54
@temp1234
OK!Synqor'sIdeasareusingastep-downfeed-backformsecondarytoregualtionowrking-voltage(orbus-voltage)andthenusing50/50ratioDCtransformertodirectdividetargetoutputvoltagetosecondary!Thisstagehasmanyversions,whichcoverInterleaveForward,Push-Pull,HBandFB.Using"电流型"sothesmoothchockisinprimaryside.duetotheDctransformerstage,synqorsuithiscompitors.It'sconceptialInfringement,soitwilltakelongtimetofighting!Ihatetyping,soeverytimeItryfewestwordtoexplainit!
No, no..., you should do it to share others your ideas and thoughts. Thanks a lot.
0
回复
ic494
LV.2
21
2008-06-24 21:47
@斜阳古道
这么大的功率,并且是一个全砖,用有源钳位的话,磁芯利用率不太高、原边电流太大,使用两象限拓扑可能更合适一点...关注的是效率与成本.楼上的,你这700W在48V输入时能做到多大的效率?另外,你的输出也是24V?
全砖如果效率不做到92%现在没人敢用.
0
回复
temp1234
LV.5
22
2008-06-25 10:07
@斜阳古道
No,no...,youshoulddoittoshareothersyourideasandthoughts.Thanksalot.
OK! Now you know, Why Synqor suits everyone have bus-converter without secondary output chock! And after this matter, you see the competitors propose their new 5% regulation type of bus-converter with output-chock.
0
回复
temp1234
LV.5
23
2008-06-25 10:09
@ic494
全砖如果效率不做到92%现在没人敢用.
I don't think so, you can check all vicor's FB!!
0
回复
michael_xu
LV.2
24
2008-06-25 13:42
在39-72V低压输出的情况下,并且开关频率只有200kHz,LLC半桥(全桥)和移相全桥均没有优势.反而有可能造成通态损耗增加.在正激,半桥,全桥的拓扑中,选择,全桥做500W这个功率等级最优优势.在同样的磁芯利用率的情况下,全桥原边电流比半桥小,有效占空比比正激大,因此输出电感比正激小,Vin=39-72V,Vo=24V/21A,如果用同步整流,用全桥硬开关在半砖上做,48V满载效率应该能做到95%.
0
回复
2008-06-25 20:14
@temp1234
OK!Nowyouknow,WhySynqorsuitseveryonehavebus-converterwithoutsecondaryoutputchock!Andafterthismatter,youseethecompetitorsproposetheirnew5%regulationtypeofbus-converterwithoutput-chock.
Great, bus converter with output filtering inductor is a very clever approach not to violate Synqor's patents.
0
回复
2008-06-25 20:17
@michael_xu
在39-72V低压输出的情况下,并且开关频率只有200kHz,LLC半桥(全桥)和移相全桥均没有优势.反而有可能造成通态损耗增加.在正激,半桥,全桥的拓扑中,选择,全桥做500W这个功率等级最优优势.在同样的磁芯利用率的情况下,全桥原边电流比半桥小,有效占空比比正激大,因此输出电感比正激小,Vin=39-72V,Vo=24V/21A,如果用同步整流,用全桥硬开关在半砖上做,48V满载效率应该能做到95%.
同意楼上的说法,或许还可以尝试一下交错并联的PWM硬开关全桥,同时也可以考虑副边绕组串联加同步整流的方式.
0
回复
2008-06-25 21:15
@斜阳古道
同意楼上的说法,或许还可以尝试一下交错并联的PWM硬开关全桥,同时也可以考虑副边绕组串联加同步整流的方式.
其实最高效率的拓扑没有出现在你们的讨论中,它是两级结构的Delta Boost+交错LLC,估计见到过的也没几个.以色列出品.同志们还需努力啊,优秀的拓扑设计是一门艺术,而艺术家很少,所以并不是每个人都适合做这一行
1
回复
michael_xu
LV.2
28
2008-06-25 22:14
@buckbuckbuck
其实最高效率的拓扑没有出现在你们的讨论中,它是两级结构的DeltaBoost+交错LLC,估计见到过的也没几个.以色列出品.同志们还需努力啊,优秀的拓扑设计是一门艺术,而艺术家很少,所以并不是每个人都适合做这一行
boost+交错LLC,做到48V满载效率到95%以上应该也有难度.我没有看到过你说的产品.首先boost从48V到75V以上这一级效率就不容易做到98.5%以上.当然两级方案可以考虑,比如说前级用单电感的buck-boost,从48V输入到48输出左右,效率应该能优化到98.5%左右.第二级采用,LLC全桥的2:1转换到24V左右,效率能尽量做到97以上,总之整体效率要能超过96%很难.
目前看到的产品,可以参照synqor的1/4brick,300W产品,第一级用buck电路,将Vin=36-75转换到Vbus=36-48V,第二级用boost隔离型全桥,将Vbus36--48V调整到输出Vo=12/25A.两级都是在Vin=48点取得最高效率.因此整体效率做到最高.但两级都调整带来动态性能不好,等等问题.
0
回复
michael_xu
LV.2
29
2008-06-26 11:13
@temp1234
TwostageTopologiesarelessefficiencythanthesinglestageTopologies!Nomatterithasboost,orbuck!Youcanseebread-newEricsson's400WQBcanreach95.5%on12V/33A!(it'sahard-switchFB+SR).Actuallytwostage'sadvantageisonmore-widerinputrangenotforbetterefficiency!OK!hereis:http://www.ericsson.com/campaign/powermodules/hp/BMR_453.htm
why? Ericsson'400W QB reach 95.5%, I cannot see the module from the website. I guest, it's a bus converter or Vin=36-60V range module.

I've not see any module which can reach 96% at Vin=36-75V,Vo=12/25A in QBrick except SynQor's two stage module.

actually I ever tested SynQor's module, it has the best efficiency at Vin=48V.
0
回复
michael_xu
LV.2
30
2008-06-26 11:22
@斜阳古道
同意楼上的说法,或许还可以尝试一下交错并联的PWM硬开关全桥,同时也可以考虑副边绕组串联加同步整流的方式.
交错并联的全桥我们也用过,效率来讲比单个全桥硬开关要高,成本也差不多.重要的是pcb的布线要注意.并且一二相移相90度就行了,你可以试试看,应该能取得好的效果.
另外,我从来没用过倍流输出,与全波整流相比在12V电压输出时有优势吗?
0
回复
temp1234
LV.5
31
2008-06-26 11:31
@michael_xu
boost+交错LLC,做到48V满载效率到95%以上应该也有难度.我没有看到过你说的产品.首先boost从48V到75V以上这一级效率就不容易做到98.5%以上.当然两级方案可以考虑,比如说前级用单电感的buck-boost,从48V输入到48输出左右,效率应该能优化到98.5%左右.第二级采用,LLC全桥的2:1转换到24V左右,效率能尽量做到97以上,总之整体效率要能超过96%很难.目前看到的产品,可以参照synqor的1/4brick,300W产品,第一级用buck电路,将Vin=36-75转换到Vbus=36-48V,第二级用boost隔离型全桥,将Vbus36--48V调整到输出Vo=12/25A.两级都是在Vin=48点取得最高效率.因此整体效率做到最高.但两级都调整带来动态性能不好,等等问题.
Two stage Topologies are less efficiency than the single stage Topologies!No matter it has boost, or buck! You can see bread-new Ericsson's 400W QB can reach 95.5% on 12V/33A!(it's a hard-switch FB+SR). Actually two stage's advantage is on more-wider input range not for better efficiency!

OK! here is :http://www.ericsson.com/campaign/powermodules/hp/BMR_453.htm
0
回复