• Please review our updated Terms and Rules here

Northstar

Gary C

Gary C

Veteran Member
Joined
May 26, 2018
Messages
2,741
Location
Lancashire, UK
Bought a N* last year and finally getting round to it.

It was for sale at a good price but collection only right down in the southeast. Price too god to ignore so I bought it with the understanding they would hold it until I could get over that way. Then as luck would have it, the owner was passing my house and dropped it off, result.

Powered up with boards disconnected, and everything nominal, onto board and it seems to boot but nothing on my terminal.

Looking at the N* manual, the jumper to set the serial port seems to expect a modem and a null modem cable doesn't work which I suppose is expected but the standard connections of DTR/DSR/CTS/RTS less so.

1734354225445.png

For some reason the standard header seems to wire DTR into pin 5 which would normally be CTS and DSR is connected to RTS, RTS connected to DSR and CTS to DTR ! Bit spaghetti like. So to remain compatible with the plug and lead I use on my EPROM programmer I have rewired the header thus

1734362095782.png
So CTR/RTS connected together, Txd and Rxd not swapped so my null modem cable works and flow control via DTR/DSR.
 

Attachments

  • 1734353416500.png
    1734353416500.png
    54.9 KB · Views: 2
  • 1734354035141.png
    1734354035141.png
    4.6 KB · Views: 2
Last edited:
Still no joy so I have installed the PROM option detailed here https://deramp.com/downloads/north_star/hardware/ZPB-A PROM Mod.pdf, EPROM with the Diag rom from here https://github.com/codesleuth7/northstar_horizon_z80_rom and now it boots to

E000
>

but only for about 5 minutes, after that, corrupt data then nothing. Seems to be temperature related. Swapped the USARTS around but no change (and I did note that the end pins on one of the USARTS were bent and it wasn't me !, almost looks as if someone has been investigating the same issue before).

Seems to be a problem addressing the USART so first port of call, check the buffers on the PCU card to the bus. Pulled the first 74LS241 out to find pin 1 missing and I don't mean it was left in the socket but actually missing. Seems the pin has been snapped before and just pushed back harder into its socket. The pins are also very black due to the silver oxidisation. New leg soldered on but no better.
Pulled all the other 74LS241's and on every single one, pin 1 is cracked and most are near dropping off. A known issue with some Ti chips apparently and certainly every 241 needs replacing on this board. However with a bit of a clean and careful reinsertion, the machine now resets to the diag rom prompt and pressing B starts a boot and gives me

NDOS5.1DQ
+

Booted !

the machine also has a virtual sector generator already fitted which seems to work well (and shows this machine has had some 'recent' attention)
 
Anyone got an image of the N* floating point card with no IC's installed ?

Looks like an easy one to replicate if I could get the image.
 
Memory cards are a mix of two types of Industrial Micro Systems 16K static RAM cards. The first 32K are type 282 cards which seem very simple, the last two are type 501 cards which have various options including setting up a memory hole for any boot ROM's.
Oddly though, the first card on has the first 8K mapped in, then the second card does the next 16K so the machine is only using 48K at the moment even though originally it didn't have a boot ROM to work around.

So need to get my head around where the built in device ports are to see if they will conflict with a Dazzler, but its working well so far :)
 
Fitted a Gotek today.

Latest version of flashfloppy 4.7a and converted the .NSI images to HFE V3 images and they boot fine. Thought the virtual sector generator would conflict with it but everything works fine with a real drive with soft sector disks and the Gotek.

Now booting to NDOS and CPM which will probably be more useful when I complete the Dazzler.
 
Damn

My N* has been running fine while I have been testing my Dazzler, but today its started to crash when running the Macro-80 assembler and now its throwing errors on trying to run simple commands

Mon> BOOT DISK
48K CP/M vers 2.2 Horizon rev 1.2.0 QD
Product of North Star Computers,Inc.


A>stat

Bdos Err On P: Select

The vector monitor includes a memory check that throws random bytes in and tests and this is not showing any errors which seems to confirm the memory cards and the bus is working ok and I am beginning to suspect the disk interface.

Time to borrow the one from the museum and see if it makes a difference.
 
Spare disk interface performs the same so its not the interface

However, noted something strange with the Gotek.

MBASIC won't load and run from the Gotek, but if I us copy to duplicate the CP/M disk from the Gotek to the real floppy disk, it runs fine.

Swapped out the USB stick for another one and noted the display went upside down. I had but a FF.CFG file onto the stick to correct the display and make stepping instant some time ago, but it stopped the drive working, rather than mess about I deleted the file but I didn't notice that the display still remained the right way around almost as if it still had the FF.CFG on the drive somewhere. It did work after that so might just be a coincidence.
Put a new stick in with CP/M and it now works, strange.

Tried to format the dodgy stick, but its refusing to let me format it FAT32 (probably because its too big so will have to use powershell I believe) but I might just park this and go back to trying to assemble GDEMO.
 
Last edited:
With some decent USB sticks the machine has been very reliable with the Gotek. The original TM100-2A has been cleaned and installed along side a temporary 5.25 mount for the Gotek. The dazzler is working well, but I am still not happy with the video connection to the outside world. I have used some jumper pins to connect a lead to a Phono socket on the back panel, but I think I will solder a cable to the card and cable tie it using the big hole then run to an SMA connector and then to a bulkhead BNC. I don't want to alter the chassis at all but do want to be able to remove cards easily.
 
In other news, I received a DA7+IO card last month.

It was missing all the buffer IC's and the A/D converter plus a lot of the ceramic capacitors plus a couple of IC's were squashed and displaced.

Fitted the missing IC's today and straightened out the components, checked from Vcc to earth and no shorts so bunged it into the machine.

Only done a quick check of D to A and that works fine, A to D has mixed results, more work required but its basically working.

Only got three slots of the twelve left to fill now and hopefully that will be reduced to two in the coming months if the floating point card recreation is successful.
 
In other news, I received a DA7+IO card last month

1000003145.jpg

It was missing all the buffer IC's and the A/D converter plus a lot of the ceramic capacitors plus a couple of IC's were squashed and displaced.

Also found a backshell for the connector

1000003146.jpg


Fitted the missing IC's today and straightened out the components, checked from Vcc to earth and no shorts so bunged it into the machine.

Only done a quick check of D to A and that works fine, A to D has mixed results, more work required but its basically working.

Only got three slots of the twelve left to fill now and hopefully that will be reduced to two in the coming months if the floating point card recreation is successful.
 
And looking forward, how am I going to reduce the +8v rail supply. No idea why but for some reason they chose a transformer that results in the +8V rail running at 12V lightly loaded and even my machine with 9 cards runs at 9.7V

I don't really want to replace all the 7805's with modern regulators but if this machine is to be left on display for hours a day with the lid on, Im not really happy with the heat it generates.

Not quite sure what to do, ideas anyone ? (I have read someone got a custom transformer made but Im not keen on that route). Maybe it will just have to sit with a Variac.
 
@Hugo Holden came up with something in another thread about adding an additional small transformer in series with the primary of the existing transformer.

I didn't take too much notice at the time. I should have!

See what Hugo has posted recently.

This was related to a high mains voltage required to be reduced. The same should apply to your situation.

Dave
 
Gary C said:
And looking forward, how am I going to reduce the +8v rail supply.
Click to expand...
What current do you need? If your ok with modern there are lots of adjustable buck DC/DC converters though most don't say what the minimum dropout voltage they need to regulate. They are cheap enough to try one. A buck-boost converter will work though not a lot of high current ones in a quick search. Derate the current on the cheap ones.

I remember someone making a servo controller for a variac to keep the 8V at his desired voltage.
 
Found what I was looking for:

Bucking transformer...

H

Post in thread 'Screen Wobble on Commodore PET 4032 12" CRT'

jtjacques said:
As I mentioned, it is a plastic unit without a plate. The case doesn't have any mounting holes for a plate, either. I suppose I could reuse the monitor/pcb mounting bolts, but they are recessed/setback in this model which makes this more of a pain. I'm open to suggestions on materials, thickness, etc. It does feel a little like treating the symptom, rather than cause, at this point though.


Unfortunately, as mentioned, I don't have a variac to hand at the moment. More importantly, the machine will be deployed variously by others. Something that can be misconfigured or requires tuning...
Click to expand...

Dave
 
@Gary C Any chance I could get you to make an image of a blank, formatted disk for me? I have some Northstar disks I've been meaning to image but I wanted to check something about the raw formatting.
 
daver2 said:
Found what I was looking for:

Bucking transformer...

H

Post in thread 'Screen Wobble on Commodore PET 4032 12" CRT'

jtjacques said:
As I mentioned, it is a plastic unit without a plate. The case doesn't have any mounting holes for a plate, either. I suppose I could reuse the monitor/pcb mounting bolts, but they are recessed/setback in this model which makes this more of a pain. I'm open to suggestions on materials, thickness, etc. It does feel a little like treating the symptom, rather than cause, at this point though.


Unfortunately, as mentioned, I don't have a variac to hand at the moment. More importantly, the machine will be deployed variously by others. Something that can be misconfigured or requires tuning...
Click to expand...

Dave
Click to expand...
So a 240 to x voltage wired anti phase on the LV
 
The example I found online was used to reduce the primary voltage of the mains to the transformer within the power supply, but I can't see why your suggestion wouldn't work either.

Perhaps Hugo can weigh in here.

Dave
 
You must log in or register to reply here.
Back
Top