Osborne 1 - Lifeless Character Patterns

Seek advice, give advice or tell others about your repair and restoration projects

Osborne 1 - Lifeless Character Patterns

Postby BitFracture on Wed Apr 08, 2015 3:06 pm

Hello everyone,
I have a late version of the Osborne 1, it was clearly stored in a moist environment as many screws are corroded. When powered on, it beeps once and vertical bars are displayed across both the internal and external displays. Two alternating columns of characters fill the screen. Pressing reset seems to scramble which character fills the screen, and occasionally there are many characters, but always ordered in some way. These screen remains completely static until reset is pressed. Initially, after two or sometimes three presses of the reset button the floppy disk drives would rotate continuously and the beeper wouldn't stop. Now the beeper doesn't beep after the initial chirp, but in addition there is now more than one pattern of characters on the screen at once.
I have verified that all voltages are correct, and the RAM is receiving correct voltages as well. The RAM chips in bank D (closest to the front) are the warmest but nothing on the board gets hot enough for me to assume a short. Removing the double-density board doesn't seem to make a difference. Here are the pictures of the symptoms shown on the external display:
Image Image Image Image
Additionally, when I remove the processor, the vertical bars and initial characters remain on the display, but the reset button no-longer has any effect on the system. Removing the character ROM eliminates all characters from the display, but doesn't change the bars going down the display as shown in the first picture. Removing the main ROM causes a continuous war-cry from the beeper. Keyboard input never has any effect on the symptoms. Pressing enter or tab at boot doesn't cause either disk drive to respond.

So, given all of that, I am inclined to suspect the processor or RAM, but I don't know where to start. I also figure the vertical bars and the failure to boot may be entirely separate issues. I read a comment by Tezza at the following URL:
http://www.vintage-computer.com/vcforum ... 13108.html
The piggyback test seems a good 'quick-and-dirty' method to diagnose some cases of bad RAM. While obviously not definitive, it did work with my Osborne.

Does my issue also seem like RAM? Would piggy-backing be a good place to start? I don't have any spare chips at this point. I am hoping anyone here can help me identify what my symptoms mean. I don't trust myself to de-solder multiple RAM chips without ruining solder pads and/or at least one chip, so I hope to do it as little as possible.

Thanks!
BitFracture
 
Posts: 11
Joined: Mon Apr 06, 2015 9:56 am
Location: United States of America

Re: Osborne 1 - Lifeless Character Patterns

Postby RonTurner on Wed Apr 08, 2015 11:57 pm

Have you checked the DRAM refresh from the CPU ? I wonder because removing the CPU makes no difference.

Obviously the RAM isn't functional so because removing the CPU (and the refresh state it produces) makes no change I suspect the CPU, bearing in mind those old RAMs are known to fail quite a bit.

The vertical bars go hand in hand with the failure to boot, your Osborne has no functional memory producing the video output and preventing the CPU from having a stack.

You have a good char ROM and the display is getting good timing signals.

Since you want to do as little as possible why not flick the CPU out of its socket and replace it instead of starting out removing loads of soldered in IC.
RonTurner
 
Posts: 339
Joined: Tue Dec 17, 2013 9:43 pm

Re: Osborne 1 - Lifeless Character Patterns

Postby BitFracture on Thu Apr 09, 2015 1:40 am

RonTurner wrote:Have you checked the DRAM refresh from the CPU ? I wonder because removing the CPU makes no difference.

I've stuck my scope on a few CPU pins but the biggest problem is that I don't know what to look for. Removing the CPU doesn't make a difference in the initial photograph, but pressing reset doesn't do anything when the CPU is removed (so the following pictures never happen).

RonTurner wrote:Since you want to do as little as possible why not flick the CPU out of its socket and replace it instead of starting out removing loads of soldered in IC.

That's probably a good idea. They're quite inexpensive, and even though this is an NEC version, isn't it compatible with all other processors of this generation? I swear everything I have runs a Z80 or 8080 (I'm not too sure of the difference). The processor chip name is NEC D780C-1, and the RAM is NEC D416C.
BitFracture
 
Posts: 11
Joined: Mon Apr 06, 2015 9:56 am
Location: United States of America

Re: Osborne 1 - Lifeless Character Patterns

Postby RonTurner on Thu Apr 09, 2015 1:57 am

The reset button is connected to the CPU reset pin, this would be why nothing happens when you remove the CPU and press reset.

A Z80 by any other manufacturer is still a Z80.
RonTurner
 
Posts: 339
Joined: Tue Dec 17, 2013 9:43 pm

Re: Osborne 1 - Lifeless Character Patterns

Postby BitFracture on Thu Apr 09, 2015 2:46 am

Alright, good to know. I see a nice lot of these that run at 6Mhz (my original runs at 4):
https://dl.dropboxusercontent.com/u/572 ... PSC%29.pdf

And this is my original. I'll probably buy them later today unless you think there is a reason I shouldn't. Here is the datasheet of my original. Thanks!
https://dl.dropboxusercontent.com/u/572 ... eet%29.pdf

EDIT:
It is things like this that make life horrible:
http://en.wikipedia.org/wiki/Zilog_Z80
The NEC 78K series microcontrollers are based on the Z80; they share the same basic BC,DE,HL register structure, and has similar (but differently named) instructions; not binary compatible.

Is an NEC 78K not the same thing as an NEC 780C-1? Half the time I'm certain that things are compatible I find that they're not after I buy them. In the darned chip datasheet is says "Fully Z80® -compatible" so I guess it's fine, but people sure do love to make things as confusing as possible.
BitFracture
 
Posts: 11
Joined: Mon Apr 06, 2015 9:56 am
Location: United States of America

Re: Osborne 1 - Lifeless Character Patterns

Postby RonTurner on Thu Apr 09, 2015 8:46 am

All I know is that my Z80 are Z80 not microcontrollers , surely a Z80 is Z80 no matter who manufacturer's it (dips anyway. ). But it has been a while since I looked inside my Osborne.

Every site says the Osborne has a Z80 CPU.

Don't try a 24 bit eZ80 !!! Just get the same package, a 40 pin dual inline pin package (dip).

http://www.cpu-world.com/CPUs/Z80/NEC-D780C-1.html
RonTurner
 
Posts: 339
Joined: Tue Dec 17, 2013 9:43 pm

Re: Osborne 1 - Lifeless Character Patterns

Postby BitFracture on Thu Apr 09, 2015 10:51 am

Thanks, it does have a Z80, I just don't know the difference between a microcontroller and a microprocessor, I've always assumed they were the same thing. I'll post an update when I try a new processor.
BitFracture
 
Posts: 11
Joined: Mon Apr 06, 2015 9:56 am
Location: United States of America

Re: Osborne 1 - Lifeless Character Patterns

Postby tezza on Thu Apr 09, 2015 10:59 am

BitFracture,

At least you can take heart that the video, clock and the power supply seem to be working.

I'd swap the processor with a standard known-working Z80. If you don't have one, could you poach one from another machine for diagnostic purposes?

I would suspect the CPU or RAM, or (if you are really unlucky) a bad logic chip on the data or address bus. Piggybacking RAM is quite easy to do...do you have a spare socketed RAM chip in another machine you could borrow for the task? From memory they are 4164s right? Mind you not all RAM chip problems can be diagnosed by piggybacking. It depends on the nature of the fault in the IC. It's certainly worth doing though.

If you find it is a RAM chip the best way to replace it is not to desolder it, but rather clip it off leaving the pins in place. Then remove the pins one at a time by heating the base while gently pulling them. Then clean up the hole with solder wick.
Tez (Terry Stewart) (Administrator)
Collection: https://www.classic-computers.org.nz/co ... /index.htm
Projects and Articles: https://www.classic-computers.org.nz/blog/index.htm
Twitter: @classiccomputNZ | YouTube: Terry Stewart
Trade Me: tezza5
tezza
Site Admin
 
Posts: 2382
Joined: Thu Jan 01, 1970 12:00 pm
Location: Palmerston North, New Zealand

Re: Osborne 1 - Lifeless Character Patterns

Postby BitFracture on Thu Apr 09, 2015 11:11 am

tezza wrote:I'd swap the processor with a standard known-working Z80. If you don't have one, could you poach one from another machine for diagnostic purposes?

I believe my TRS80 Model 1 and Zenith Z90 both sport a Z80A, let me check that out. I REALLY don't want to harm those computers too! But I guess if I did I'd know what to replace.

tezza wrote:If you find it is a RAM chip the best way to replace it is not to desolder it, but rather clip it off leaving the pins in place. Then remove the pins one at a time by heating the base while gently pulling them.

I actually messed around with an old board today using a foil shield and heat gun. I was able to remove almost everything successfully without an iron at all. I melted one socket... but it seems rather successful. Probably not a bad idea to clip leads if the chip is certainly bad, but honestly all of the darned chips should be in sockets... another day I guess.

My North Star is rather accessible right now actually, I might see what chip is inside that. I'll be right back!
BitFracture
 
Posts: 11
Joined: Mon Apr 06, 2015 9:56 am
Location: United States of America

Re: Osborne 1 - Lifeless Character Patterns

Postby BitFracture on Thu Apr 09, 2015 11:30 am

The processor from the Osborne works in the Advantage, but not vice-versa. So, more than likely the problem here is RAM. Thanks for sparing me the pain of buying a processor.
How do I know that RAM is compatible with this machine? Is anything with the same capacity and pin-outs going to work?

EDIT:
Here is the datasheet for the RAM in the Osborne:
https://dl.dropboxusercontent.com/u/572 ... 20DRAM.pdf

And for the RAM in my Zenith Z-90:
https://dl.dropboxusercontent.com/u/572 ... 0-1%29.pdf

My senses tell me it is compatible... am I right? It's all socketed too, which makes this super convenient if it works.
BitFracture
 
Posts: 11
Joined: Mon Apr 06, 2015 9:56 am
Location: United States of America

Re: Osborne 1 - Lifeless Character Patterns

Postby tezza on Thu Apr 09, 2015 12:59 pm

Yes, they are of the same type. One thing that can make a difference in compatibility is speed. Hopefully the access speed value of the borrowed chip is the same or lower than the ones in the Osborne. If the chips from the Northstar are higher, they may not work so well. They might be ok for a diagnostic exercise like this though.

Be very careful you align the chip correctly i.e. that its the right way around when piggybacking. Otherwise you are likely to cook it.

Incidentally, sometimes TWO or more RAM chips are faulty. I had this in an Apple II once. In this case piggybacking is less useful, but if there is a significant change in the screen, it may indicate a problem with the chip being tested.

Good luck!
Tez (Terry Stewart) (Administrator)
Collection: https://www.classic-computers.org.nz/co ... /index.htm
Projects and Articles: https://www.classic-computers.org.nz/blog/index.htm
Twitter: @classiccomputNZ | YouTube: Terry Stewart
Trade Me: tezza5
tezza
Site Admin
 
Posts: 2382
Joined: Thu Jan 01, 1970 12:00 pm
Location: Palmerston North, New Zealand

Re: Osborne 1 - Lifeless Character Patterns

Postby RonTurner on Thu Apr 09, 2015 1:36 pm

Sorry I didn't know you had already tested the Z80.
RonTurner
 
Posts: 339
Joined: Tue Dec 17, 2013 9:43 pm

Re: Osborne 1 - Lifeless Character Patterns

Postby BitFracture on Thu Apr 09, 2015 2:49 pm

tezza wrote:Yes, they are of the same type. One thing that can make a difference in compatibility is speed. Hopefully the access speed value of the borrowed chip is the same or lower than the ones in the Osborne. If the chips from the Northstar are higher, they may not work so well. They might be ok for a diagnostic exercise like this though.

I tested the RAM, found three chips that modified what was on the display. Two of them far more than the other. In fact, one of those two caused keyboard input to start effecting the display and the beeper. In all cases it appears like the two chips were "fighting" each-other, which makes sense. The good thing is that the Z-90 RAM passed the RAM test even after all of that handling! I was expecting otherwise.

Should I go ahead and buy some RAM and sockets? Is there anything else I should try first?
BitFracture
 
Posts: 11
Joined: Mon Apr 06, 2015 9:56 am
Location: United States of America

Re: Osborne 1 - Lifeless Character Patterns

Postby tezza on Thu Apr 09, 2015 3:35 pm

Well it's not definitive. Did those changes happen every time the chip was piggybacked? Problem with these computers is that a lot of the I/O is memory mapped so just random garbage in RAM can cause keyboards to beep, drives to start spinning etc.

The ICs are quite cheap. It wouldn't hurt to have a few around if you are going to dabble with machines of this vintage. If it was me I'd replace them and see what happens. Be prepared for disappointment though. It is somewhat hit and miss.
Tez (Terry Stewart) (Administrator)
Collection: https://www.classic-computers.org.nz/co ... /index.htm
Projects and Articles: https://www.classic-computers.org.nz/blog/index.htm
Twitter: @classiccomputNZ | YouTube: Terry Stewart
Trade Me: tezza5
tezza
Site Admin
 
Posts: 2382
Joined: Thu Jan 01, 1970 12:00 pm
Location: Palmerston North, New Zealand

Re: Osborne 1 - Lifeless Character Patterns

Postby BitFracture on Fri Apr 10, 2015 2:08 am

tezza wrote:Well it's not definitive. Did those changes happen every time the chip was piggybacked? Problem with these computers is that a lot of the I/O is memory mapped so just random garbage in RAM can cause keyboards to beep, drives to start spinning etc.

I understand, but it was direct keyboard presses that queued the beeping and the changes on the display. I am curious if static through the keyboard input could have blown out a RAM chip or two, if they are connected.

For the sake of documentation, I want to leave this list here. These chips are all 16,384 x 1 bit DRAM rated at 150 nanoseconds (or faster). All of them should theoretically be valid replacement chips.

    NEC D416C-3 (current)
    NEC D416C-5
    Fujitsu MB8116H
    AMD AM9016FDC
    AMD AM9016FPC
    TI TMS4116-15
    Mostek MK4116P-2
BitFracture
 
Posts: 11
Joined: Mon Apr 06, 2015 9:56 am
Location: United States of America

Next

Return to Repair and Restoration

Who is online

Users browsing this forum: No registered users and 32 guests

cron