3577A有ROM错误

Gerald，检查您的数字5V电源。
这可能会导致失败。
戴夫



     以下为原文

  Gerald,

Check your digital 5V supply.   That can cause a failure.

Dave

谢谢Dave，我在开始时检查过它可能会有多余的波纹。
但唉，它在100mVdc和Gerald之内



     以下为原文

  Thanks Dave,

I checked that at the start thinking that it might have excess ripple.  But alas, it was within 100mVdc and <80mVp-p ripple measured with a 200MHz DSO.  All other supplies were also good.

Gerald

我忘了，3577岁30岁，还是35岁？
什么是序列号？
将1960添加到前缀的前两位数字，您将知道它的制造年份。



     以下为原文

  I forget, is the 3577 30 years old, or 35 years old?  What's the serial number? Add 1960 to the first two digits of the prefix and you will know the year it was manufactured.

这个特定的单元已有29年历史，足以容纳EPROM数据保留问题。
许多集成电路上的日期代码显示1984年末的工作周，因此我认为它是在1985年制作的。所述的S / N从23开始，指向1983年，即模型引入的那一年。
CRT没有任何烧伤证据，既明亮又清晰。
这表明该单位29年来只有轻微使用。
我的实验室里有3570个，在39年后仍在运行！
当它构建正确时它不会因为支持消失而过期;-)。
当启动设备启动时，它的ROM测试报告两个块错误，它的RAM测试（通过）然后停止ROM测试灯仍然打开和蜂鸣器现在响亮。
我已经检查了卡上裸露的顶部的所有MC68000引脚，它们看起来都是强大的逻辑电平。
我也接触到ROM模块上的芯片选择线，它们看起来也像TTL一样合理。
我还没有尝试过更乏味的探测EPROM上所有缓冲地址和数据线的工作。
以同样的方式，我还没有使用在网络上找到版本C代码的转储来更新所有新的EPROM。
所有八个都焊接在:-(。我想有两个同时出去很可能不是真正的EPROM。谢谢，杰拉尔德



     以下为原文

  This particular unit is 29 years old, old enough to have EPROM data retention issues.  The date codes on many of the ICs show work weeks late in 1984, therefore I believe it to have been made in 1985.  The S/N as stated begins with 23 which points to 1983, the year the model was introduced.

The CRT has no evidence of any burns and is both bright and sharp.  That would suggest that the unit had only light use for all its 29 years.  I have a number of 3570As in my lab that are still running after 39 years!  When it's built right it doesn't expire just because support vanishes ;-).

When booting the unit comes up, does its ROM tests where it reports the two block errors, does its RAM test (which passes) and then stalls with ROM test light still on and beeper now blaring.  I've examined all of the MC68000 pins that come to J3 on the exposed top of the card and they all appear to be robust logic levels.  I've also reached in to the chip select lines on the ROM blocks and they also look reasonable as TTL.  I have not yet tried the more tedious effort of probing all of the buffered address and data lines on the EPROMs.  In like manner I've not yet changed out all the EPROMs with new ones using a dump found on the web for version C code.  All eight are soldered in :-(.  I would like to think that having two go out at the same time is most likely not the actual EPROMs.

Thanks,
Gerald

通过更换所有ROM来修复单元。
实际上只发现U19有缺陷。
据报道，U19和U21都被POST报告错误，但两个显示的“坏”校验和都是错误的？
当在U20-U26单元外测试时，所有校验和都读得很好..即匹配网上找到的二进制文件。
U19-U26似乎是掩模ROM而不是OTP EPROM。
他们都有Signetics门牌号，我没有Signetics制作OTP版本的记录。
ROM部件号与HP IC的服务手册或主列表不匹配。
现在，该单元的工作效果非常好，因为每件HP都应该;-)。
如果有人能指出3577A服务笔记，我仍然会很感激。
谢谢！
杰拉尔德



     以下为原文

  Unit fixed by replacing all the ROMs.  Only U19 was actually found defective.  Both U19 and U21 had been reported bad by POST but both of the displayed "bad" checksums were in error??  When tested outside the unit U20-U26 checksums all read as good.. i.e. matched the binaries found on the web.

U19-U26 appeared to be masked ROMs and not OTP EPROMs.  They all had Signetics house numbers and I have no record of Signetics making OTP versions.  The ROM part numbers did not match either the service manual or master listings of HP ICs.

The unit now works beautifully as every HP piece should ;-).

If anyone can point me to 3577A service notes, I would still appreciate that.

Thanks!
Gerald

错误的校验和报告的谜团现在已经解决了。
固件版本C为每个ROM提供两个校验和，并使用与原理图和PCB丝网印刷不同的编号块编号方式。
U19是块1和3，U20是块2和4，... U25是块13和15，U26是块14和16.两个块的mod 16bit总和等于通常报告的校验和。
只读存储器。
我刚刚修复了第二个3577A报告的ROM错误报告了13和15的坏块，这两个块都不存在原理图和丝网印刷。
U25（示意图块7）很糟糕。
第二个单元在电源端也有一个松散的5V电源连接器。
人们想知道这种情况是否会导致5V电源瞬变，这可能对某些IC造成危害。
杰拉尔德



     以下为原文

  The riddle of the erroneous checksum reports has now been solved.  Version C of the firmware gives two checksums per ROM and uses a different way of numbering blocks than is on the schematic and PCB silkscreen.  U19 is block 1 and 3, U20 is block 2 and 4,.. U25 is block 13 and 15 and U26 is block 14 and 16.  The mod 16bit sum of the two blocks is equal to the checksum that would normally be reported for the ROM.

I just repaired a second 3577A with reported ROM errors that reported bad blocks of 13 and 15, both of which do not exist per the schematic and silkscreen.  U25 (schematic block 7) was bad.  This second unit also had a loose 5V supply connector at the power supply end.  One wonders if this condition can cause 5V supply transients which might be hazardous to some ICs.

Gerald

在查看一些旧的Bench Brief问题时，我看到了t的服务说明



     以下为原文

  In looking over some old Bench Brief issues I saw mention of service notes for t