JLH1969M SHO （已翻译为中文）

millwood

大家好，

Millwood兄的这篇高输出功率的１９６９Ｍ，　实在是一篇，　很精彩的文章。　请饶恕我未曾得到　Millwood兄的允许前，　擅自给他翻译，　但，　为了让更多坛友能分享这篇精彩的ＤＩＹ文章，　我给他做了翻译，　我小心的将文里的精义做翻译，　任何不恰当的处理，　请　Millwood兄原谅。。。。红山老猫敬启

ＪLH1969（场管版本或晶体版本）公认被评断输出功率偏低， 原版的JLH1969有10W的输出能力，而在较高电供动作的JLH1969M则有输出20至25W到8欧姆阻抗的能力， 在大部分的情况下， 这是充足的。然而当您希望求取更大的输出功率，本篇这就是我与大家讨论更大输出功率的 JLH1969.


高功率输出的甲类动作扩音机， 可行性是较不实际的。所以我们在此讨论能在甲类或乙类状况下工作的1969M。

为了让1969M输出更多的功率， 有两件重要的事要解决。


1. 增阻电路 （the bootstrap network） ： 原版的增阻电路无法提供足够的摆幅及在高输出状况下，驱动输出级需索的电流量， 改用恒流源电路取代增阻电路，是合理的安排。然而，这会产生两个新的问题 a）：稳定性 - 原版线路利用增阻电路，大于100Khz的自然衰减特性 ，求得更佳的稳定性。使用恒流电路就必须使用米勒电容令机器的高频段稳定。这会影响了机器的上升速率 b):输出摆幅，原本的增阻电路有输出大于电供电压10V的摆动能力， 这对场管版本是非常有利的。使用恒流源电路会牺牲掉这个特性。解决办法是采用更高的电供电压。

2. 场管的跨导特性： 场管在高输出时， 也许提供更多的电流输出， 我们必须在场管的闸极施加更大的电压输入摆幅， 然而，有时这还是不足够的， 我们必须 a): 使用纵向场管 （又称竖直结构场管）以求去更高的跨导特性， 或 b): 并联更多的输出场管

目的就是求取超强输出能力 （SHO）的1969M在8欧姆负荷输出100W功率， 在4欧姆的状况下倍增

以下就是 ＳＨＯ1969Ｍ的电路

以上蓝色字体是我做的翻译　


one valid critism of JLH1969 (BJT or MOSFET) is that it has limited power output. The original JLH1969 is good for 10w rms, and the JLH1969M, working at higher voltage, can deliver 20 - 25w rms into an 8ohm load. while that's mostly adequate, there are cases where you would want more. that's why I started this thread, to discuss higher output versions of the JLH1969.

Given that Class A operation is unrealistic at high output power, we will instead focus on JLH1969M that works in both Class A or Class B.

two issues need to be resolved for the JLH1969M to deliver more power:

1) the bootstrap network: the original bootstrap network wouldn't be able to deliver the needed voltage swing, nor current capabilities, required by the output stage for high power output. an obvious solution is to use a true constant current source. This, however, create two problems: a) stability - the original designs rely on the roll-off of gains from the bootstrap network at > 100khz for better stability. With a true CCS, you will have to use miller cap to ensure stability at high frequencies. This may have slew rate implications. b) voltage swing: the original bootstrap network can swing above the supply rail for as much as 10v, and that's very helpful for the mosfet version. with a true CCS, you can swing at most to 1 - 2v to the rail voltage. the way to solve that is to use a higher supply rail.

2) the transconductance of MOSFETs: to deliver higher output power, we have to deliver higher output current. that requires more voltage swing on the gates of the mosfet. Sometimes, even that is not sufficient so we have to a) use vertical mosfets for their considreably higher transconductance, or b) parallel more mosfets to enhance the gain.

the design goal is to produce a SUPER HIGH OUTPUT ("SHO") version of the 1969M that is capable of more than 100w rms into a 8ohm load and double that into a 4 ohm load.

here is the schematic for the JLH1969M SHO.

millwood

这个线路看起来， 与标准线路相近， 除了两个例外：

1. 使用了真正的恒流源 （R87, V27 (一个红色LED）， R85（微调电阻），与Q16）： V27是个红色LED， 其顺偏电压约为1.9V R87控制着流过LED的电流（对大部分的红色LED， 大约为12ma), R85 则控制恒流源的输出电流量 = （1.9-0.7）/49 大约是 25ma。

2. 增设了R90 ： R90有两个功能 a)它给，Q18 切断时， 提供了电流通路 b)它把讯号电流转换成开/关 M7/M8 的电压讯号。 恒流源输出的 25ma 则平均的分配到 R90 及 R103， R90的电压降 是 25ma/2 X 220欧姆 = 3V -恰好足够启动场管

您会发现， 更动 R85会改变场管的静态电流

R89/R100 是个微调电阻， C39这个米勒电容应当越小越好， 前提是机器不会产生应不稳定的自激

R88/R86/R101/R99 要求不严格， 在 11欧姆 - 470欧姆间都行

Q17 是个小讯号PNP 晶体管, Q18 是个中功率NPN 晶体管, and M7/M8/M9/M10 则是纵向功率场管 (TO264/TO3P).

每个输出场管在静态时的损耗约是9W， 所以需要安装在一个大小合理的散热器上

此机的输出功率？ 此机能输出44v 峰值讯号到负载. 就是 44v*44v/2/4=~240w rms ，4欧姆负载 或 120w rms 8欧姆负载.

蓝色字体由红山老猫翻译

it looks quite similar to the regular JLH1969M, with two exceptions:

1) the use of a true CCS (R87, V27 (a red  LED), R85 (a pot) and Q16): V27 is a red LED with a voltage drop of 1.9v, R87 controls the current going through the LED (about 12ma - works for most of them). R85 sets of the output current for the CCS: = (1.9-0.7)/49=~25ma.
2) the addition of R90: R90 does two things - a) it allows a current path when Q18 shuts off; b) it converts the current signal to a voltage signal to turn on / off M7/M8. the 25ma current from the CCS is evenly split between R90 and R103, so the voltage drop on R90 is 25ma/2*220=3v, just enough to turn on the mosfets.

As you can see, changing R85 will change the the idle current going through the mosfets.

R89/R100 is a trimmer. C39 is the miller cap and it should be as small as possible without causing the amp to oscillate. R88/R86/R101/R99 aren't critical - anything in the 11ohm - 470ohm will do.

Q17 is a small signal PNP transistor, Q18 is a medium power NPN, and M7/M8/M9/M10 just a power mosfet (TO264/TO3P).

At idle, the mosfets dissipate about 9w each so a reasonably sized heatsink is required.

output power? the amp outputs about 44v peak into a load. That means 44v*44v/2/4=~240w rms into a 4ohm load and 120w rms into a 8ohm load.

jt1000027dsl

高手翻译一下，谢谢先，占位学习
4个场效应管IRFP140

xiaolu36

希望有人翻译

millwood

频率响应 : 平直的伸展超过 100Khz ， 下潜低于10hz

总谐波失真 ： 相当小，同时迅速递减

好好享受音乐

蓝色字体由红山老猫翻译

FR: flat to over 100khz and down to below 10hz.

THD: quite small and diminishes fast.

enjoy.

6p6pj

放个制作图出来好让人信服，不要老是只有电路图，这里是中文论坛，別E文好不好！

香港穷烧友

Would you mark the value of V28,V29 & polarity of the C36? thanks.

millwood

我明白， 有些朋友对 OCL 线路喜爱， 这是您可以尝试的理念. 输出摆幅达 +/- 50v. 在两对输出场管 可提供 300w rms 到 4欧姆负载 或 150w rms 到 8欧姆负荷。

它没有良好的电源纹波抑制比， 这是您需要关注的。

请注意： 这已不是 JLH1969 的出发理念， 而是个电压回馈放大器 ， 不过的确可以去除了输出交联电容

蓝色字体由红山老猫翻译

I know some of you love OCL so here is a concept that you can try. it can swing about +/- 50v. and with two pairs of output device it can deliver 300w rms into a 4ohm load and 150w rms into a 8ohm load.

it does NOT have good PSRR so you will have to address that first.

PS: it really isn't a JLH anymore as it is a voltage feedback power amp. But it does get rid of the output capacitor.

millwood

原帖由 香港穷烧友 于 2009-3-17 21:54 发表
Would you mark the value of V28,V29 & polarity of the C36? thanks.

V28是供给做电源纹波抑制比分析的， 不必要的

V29 是我较早的分析用的， 不必要

C36的 + 脚在左边， -脚在右边

蓝色字体由红山老猫翻译


V28 is for psrr analysis and you don't need it.

V29 was from my earlier analysis. you don't need it.

C36's positive on the left and negative on the right.

locky_z

这个电路考虑到了温度补偿

可以在网上找到"Construction Notes for the N-Channel Amplifier Module"，就是多这个电路的说明。

millwood

温度补偿有几个办法

1) 把LED安装在散热器上
2) 把一串二极管 (1n4148) 安装到散热器上
3) 用个晶体管， 替代LED

蓝色字体由红山老猫翻译

there are a few way to implement temp compensation.

1) mount the LED on the heatsink.
2) use a string of diodes (1n4148) and mount them on the heatsink.
3) use a transistor in place of the LED.

lxhj

LZ4个场效应管IRFP140换成450行吗？

LAOQIANG

等待高手翻译，这位的e文看着太困难！

hy751102

就是就是，我去香港网站就尽量用繁体字回帖，去外国网站我是想用英文回帖（不过我也看不懂外文帖子）
希望“贵宾”用翻译软件后发帖的好！

millwood

原帖由 lxhj 于 2009-3-17 23:35 发表
LZ4个场效应管IRFP140换成450行吗？

任何纵向场管(IRF 型号) TO3P, TO264及 TO247 都行

不建议使用TO220的， 风险较大

蓝色字体由红山老猫翻译


any vertical mosfet (IRF types) of to3p, to264 and to247 should be OK.

it would be quite risky to use to220 mosfets.

xiaolu36

我想拿2SK1529作末级输出，不知道其他数值怎么取。。

rifjft

  哇，不错不错。E文看不懂就看图算了

waltz

英文有点看不清楚，可否翻译解说啊

j10

原帖由 locky_z 于 2009-3-17 23:05 发表
这个电路考虑到了温度补偿
726890
可以在网上找到"Construction Notes for the N-Channel Amplifier Module"，就是多这个电路的说明。

老大，你这个正负70V比较厉害！

DarkHorse

bang ding!
qi shi zhe xie dian lu jie gou bi jiao shi he xiao gong lv de ji zi.
da gong lv de dong xi you bi zhe geng hao geng wan shan de OCL...
哈哈,看我的E文不错吧,只要是学过汉语的都读得懂...