• Please review our updated Terms and Rules here

What did I do to my PDP-8 today.

I managed to build a RX01 floopy that can handle RL02 (R20A:-F:) with handlers from OS78V4 and os8diskserver devices SDA0:- SDB1:. The OS is OS/8-KBM V3Q and I use BOOT.SV V5A to boot RL.
That makes it possible to save the Data from RL02 media to os8diskserver images. Together with saving the SystemHead (PIP /Y) I'm able to reconstruct the RL02 Images in simh.
With this method I avoid any problems with the BadBlock handling of the RL02.

Did anybody integrate the os8bootloader into BOOT.SV?

Best wishes,
Volker
 
Roland Huisman said:
> If someone pays $999 for the 8e front panel then they are nuts.

I agree on that. But recently a 11/70 front without bezel was sold for $1800. To me that is nuts too... And if someone in Europe want to buy it then you end up on about $2270 for this panel with shipping and customs.

So it became a hobby for the elites I guess... That was the reason for me to start to make clone boards. I know, a front panel is harder to build with custom switches. But is the demand is on such a high level, then maybe front panel replica's would be a good idea as well...
Click to expand...
A hobby for the elite? That's almost a given.

A similar situation developed in the vintage car market. Enough people were priced out of the market for originals yet still wanted something "vintage looking" to create and sustain a thriving market for "repops." A reproduction Shelby Cobra, not the warmed over Mustangs Carol Shelby let Ford put Cobra badges on, but the modified AC Bristol with a small block Ford in it will still set you back a bit. But if you have to ask what an original 289 Cobra goes for you probably can't afford it. Forget trying to find the 427 version. Jay Leno probably owns one of each.

And then there's this:

www.manoftheworld.com

Lost and found. The peculiar case of the missing Shelby Cobra Daytona Coupe — Man of the World

written by ADAM K. RAYMOND photography by CHRISTOPHER GRIFFITH
www.manoftheworld.com www.manoftheworld.com

So essentially, the lucky few will still come across a barn find, or a physics department cleaning its basement, but if you want an 11/70 and don't already have one, be prepared to come up with some serious cash or put one together from non working parts that will also get harder and harder to find.

en.wikipedia.org

Shelby Daytona - Wikipedia

en.wikipedia.org en.wikipedia.org
 
Roland Huisman said:
I know, a front panel is harder to build with custom switches. But is the demand is on such a high level, then maybe front panel replica's would be a good idea as well...
Click to expand...
I've looked into that, and done PCBs and such for the 8/i. I estimate that the replacement cost for a front panel is expensive -- likely $500-$600. That's still 1/3 what was paid for the 11/70 panel mentioned. The 11/70 is a very nice panel, but still that is well over twice what it "was worth" in terms of replacement cost. (The PCBs are the cheap bit. Just the silk-screen panel will cost much more than the PCBs.)

Then again, a collector typically doesn't want a replica, and is willing to pay a premium.

Vince
 
gnupublic said:
I'm able to reconstruct the RL02 Images in simh.
With this method I avoid any problems with the BadBlock handling of the RL02.

Did anybody integrate the os8bootloader into BOOT.SV?
Click to expand...
A quick check shows my pull request is still pending with Kyle.

If you've got a bootable RL02 image with a working BUILD.SV, it should be possible to install SerialDisk handlers to it, assuming there are free device slots, etc. The pull request is for updates to those scripts and their documentation. You could also just do all the same steps with BUILD by hand, though that's a bit tedious and error prone.

Serving an RL02-style 10Mw SerialDisk image would be more challenging.

Your mention of BOOT.SV suggests you want to boot the RL02, then reboot to SerialDisk? I have no real experience with OS/8 and RL02, so I'm not sure whether that would be part of the workflow to back up or archive RL02 media, or perhaps to copy stuff onto the RL02?

Vince
 
Making the panels wouldn't be the deal breaker. Making the tooling to make the panels will be the bottleneck. But as we've already seen with various reproduction board projects, the cost of "small batch" pro-quality bare boards has dropped enough to make them something a garage shop hobbyist can tackle. Similar things are happening in other fabrication areas. I think we're going to see near perfect copies of vintage computer hardware available sooner than you might think.

www.instructables.com

DIY Metal Stamping

DIY Metal Stamping: So I'm all about pushing the limits of DIY fabrication, and the thought crossed my mind to try out one of the more intensive manufacturing techniques: metal stamping! I've had quite a bit of experience with 3-axis CNC Milling, so I decided to try my…
www.instructables.com www.instructables.com

www.instructables.com

DIY Sheet Metal Bender Brake

DIY Sheet Metal Bender Brake : Sheet Metal Bender Brake 805mm Mounted on a Hydraulic Portable Table I love metalworking, is my favorite hobby. So for my need to build a stainless steel barbecue I decided to make a sheet metal bender brake. It isn't a so complicated construction…
www.instructables.com www.instructables.com

makerindustry.com

CNC Mill for a Small Shop | Top 10 Budget-Friendly Options - Maker Industry

Looking for a Small Desktop CNC Mill for your shop? But need it to be cheap? Well, we've done the research and the reviews are in. CHECK THEM OUT.
makerindustry.com makerindustry.com


I'm not suggesting you buy any of the above to make one front panel. What I'm saying is that as the price of fabrication machinery comes down, the size of the shop that will have them will come down as well. That will bring down the cost of making small batches of parts to order. At some point we'll probably see someone start making reproduction card cages, chassis, backplanes, etc. We've already seen it with circuit boards and small plastic parts. The rest of the pieces will fall into place sooner or later.
 
gnupublic said:
Did anybody integrate the os8bootloader into BOOT.SV?
Click to expand...
If I understand correctly you are asking if someone has added the Serial Disk bootloader into the OS/8 BOOT program. I don't think so but I will be happy to learn otherwise.

I just looked at boot.pa and see how it works. I was never able to use it but it was clear that the way I thought it works is not how it works. I thought you typed boot and it would reboot. You need to know the 2 letter code found in the table in the OS/8 handbook. In the past when I tried to use it I didn't have a copy of the OS/8 handbook so I didn't have the magic table. I think it would be very easy to add a serial disk boot loader to BOOT.PA with SD as the magic 2 letter code. I will add this to my list of things to do with Serial Disk. In a nutshell, there is a table which has these entries:
  1. Six bit 2 character name. Examples: PT for the binary loader, RF for the RF08.
  2. Address in the BOOT.SV memory of the boot code for this device.
  3. Address in memory where the code needs to be moved to.
  4. Length of the bootstrap in words.
  5. Starting address of the bootstrap.
At the end you put in the serial disk boot code right after the RX boot.

The BOOT program would be used like this:

. BOOT /SD

At this point it halts to give you a chance to change the media. When you are ready you press continue and it will jump to the boot code.

It would be easier to just write SDBOOT which would do a specific boot to Serial DIsk.
 
My pdp8/a boots the floppy. (There should be the ROM for RL02 also, but at the moment I like booting RX.) From there I can switch to RL02 with BOOT/RL. I would also like to be able to boot the os8diskserver.
I'm not very used to write code or patch files, but I will give it a try.

I tryed to install the serial disk handlers in the images, but I noticed some problems. They do not install into OS/8 V40 or OS78.
At least the combination of RL drivers from OS/78v4 into OS/8v3Q also accepted the handlers from os8diskserver.

I would like to know the complete version tree of OS/8. In the headers of some source files of OS78 there are some hints for version numbering, but this does not seem to be the complete story.
 
I forgot to mention that BOOT has a feature I didn't know about. The command

.BOOT /ZE

will zero field zero. It is a clever piece of code and it looks like when it gets done it sits there executing zeros which is an AND 0 instruction. I guess if you wanted to leave the machine sort of clean this would get you partway there. It is distinctly tricky to wipe yourself out leaving nothing but zeros. Think about how to do this.

Now that I think about it more I am almost certain that BOOT was found on DSKMON on 4k machines. On a 4k machine it would leave the whole core cleared. This feature was less useful on machines with more than 4k.
 
gnupublic said:
I tryed to install the serial disk handlers in the images, but I noticed some problems. They do not install into OS/8 V40 or OS78.
At least the combination of RL drivers from OS/78v4 into OS/8v3Q also accepted the handlers from os8diskserver.

I would like to know the complete version tree of OS/8. In the headers of some source files of OS78 there are some hints for version numbering, but this does not seem to be the complete story.
Click to expand...
What does it do when you try to install on V40. Where does V40 come from? I was under the impression that V3D was the last DEC release. BUILD has poor error handling. If you have too many devices loaded it doesn't tell you. It also will let you mark more than one SYS device as active. I don't know what happens in that case. Remove everything you aren't using.

I would really like to get ahold of an untainted release set. I have a directory tree of V3D from Charles Lasner's ibiblio account. But this is not trustworthy because it is not a DECtape image. The files have been converted from DEC ASCII (8th bit high) to the ASCII used today. I want to look at the real thing. Not something that has been massaged to remove nulls and form feeds. If you have to ignore white space and case on a diff then someone messed up the file.
 
The floppy that came with my 8/a shows:

.VER
OS/8 MONITOR V40 - CCL V4B5
.R PIP
*/V
OS/8 PIP V40B
*^C
.BOOT/VE
V5A
.RUN SYS BUILD
$VER
BUILD V40A

The moment I load the os8diskserver handlers, the system crahes. No RUN light any more.
This seems not to be a problem of any space/device or handler limit. I unloaded enough.

Yet I do not know much about the differences between OS/8V3 D/ F/ Q/ T/ 40 .... OS/78V1/2/3/4
 
Last edited:
gnupublic said:
The floppy that came with my 8/a shows:

.VER
OS/8 MONITOR V40 - CCL V4B5
.R PIP
*/V
OS/8 PIP V40B
*^C
.BOOT/VE
V5A
.RUN SYS BUILD
$VER
BUILD V40A

The moment I load the os8diskserver handlers, the system crahes. No RUN light any more.
This seems not to be a problem of any space/device or handler limit. I unloaded enough.

Yet I do not know much about the differences between OS/8V3 D/ F/ Q/ T/ 40 .... OS/78V1/2/3/4
Click to expand...
That is interesting. Must have been a release specifically to support the RL drives on the 8/a. The controller is a hex board so won't fit in a normal 8/e setup. I have an 8/a with RL01 drives but it only had the RL01 drives and I have never booted it from the RL01 drives. I have not cleaned the drives and packs up to the point where I would be comfortable loading the heads.

What exactly do you mean when you say "The moment I load the os8diskserver handlers"??

Is this in BUILD when you do the LOAD command to read the binary? That would imply a header issue with the binary.

Do both the SYS and non sys handlers cause the issue?

I am confused!
 
DougIngraham said:
Is this in BUILD when you do the LOAD command to read the binary? That would imply a header issue with the binary.

Do both the SYS and non sys handlers cause the issue?
Click to expand...
Yes and Yes!
This happens when LAOD SYS:SDSKSY.BN in BUILD. I tryed with DSKSYS.BN, the first version also.

You find more about OS/8 V40 here: http://www.vandermark.ch/pdp8/index.php?n=OS8.OS8-V40

The RL01 handler came first with OS8/V3D device extension. (OS/8 Device Extensions User's Guide AA-D3319-TA) (bitsavers/DEC/pdp8/OS8/). There are different handlers for RL02, some only reach 3 Partitions. I only found the handlers from OS/78 to access 5 partitions of RL02.
 
Volker,

Kyle used the RK05 sys handler as a template when he made the serial disk handler. Unless DEC made some big changes to OS8 I don't see any reason why BUILD should crash unless it is running out of space for its internal tables. If you could do a screen capture of the following commands I would be interested in the output.

.RES SYS:/E

.RUN SYS BUILD
$VER

$PRINT

I will look at this and see if I can figure something out.

Doug
 
sim> b rx0

.VER
OS/8 MONITOR V40 - CCL V4B5

.RUN SYS BUILD
$VER
BUILD V40A
$PR

RLSY: SYS RL0A
RLC : *RL0C RL1C *RL2C RL3C
KL8E: *TTY
RX11: RXA0 *RXA1
RXX1: *SYS
LAXX: *LPT
RL0 : *RL0A *RL0B
RL2 : *RL2A *RL2B

DSK=RXX1:SYS
CORE= 7
$LO SYS:SDSKSY.BN

HALT instruction, PC: 27603 (AND 12)
sim>

Doug,
this is now done with simh, but with the 8/a there's no difference. It's the same with DSKNSY, DSKSYS or SDSKNS.BN.
And sorry, there is no RESORC on the disk. I have no other V40 disk.
 
gnupublic said:
Doug,
this is now done with simh, but with the 8/a there's no difference. It's the same with DSKNSY, DSKSYS or SDSKNS.BN.
And sorry, there is no RESORC on the disk. I have no other V40 disk.
Click to expand...
That is interesting. It halted at what would be in the second page of a 2 page system handler. Presumably the RXX1 device is a 2 page handler. RESORC would have told us this. Do you have sources for the RXX1 handler? All the RL drivers I have are built as RX8E not RXX1 or RX11.

How about an experiment. I don't expect any difference but it doesn't hurt to try.

.RUN SYS BUILD
$PRINT
$UNLOAD RLSY
$PRINT
$LOAD SDSKSY.BN
$PRINT

I expect it to still crash on the LOAD so you wont get to the last PRINT. Any chance those binaries got corrupted somehow?
 
Even when I unload all handlers it crashes. I transferred the PAL code with K12MIT to the floppy and compiled it with PAL8 that was on the RL02 device in the 8/a.
Later I took an image from the floppy with PUTR.EXE on my DOS PC. I added PAL8 and tested compiling again. No changes.
 
I found an RXX1SY.PA on the Vandermark site mentioned earlier. It is not a 2 page system handler. It is almost identical to the DEC RX01 handler except it has a DATAPLAN GMBH copyright. It shows as a version M. I wish you had RESORC on your floppy so we could see the version number.

Since this one is not a 2 page handler it probably isn't the one on your system or it is a coincidence that the halt occured on the last page of Field 2.

Do you have more than 8k on your 8/a?

I am going to look a little more but I have spent all my free time today on this.
 
There is 32K in the 8/a. I put a RESORC.SV on the floppy image:

*/E

29 FILES IN 379 BLOCKS
59 FREE BLOCKS (3 EMPTIES)

# NAME TYPE MODE SIZ BLK KIND U V ENT USER
01 SYS RX8E RWF 494 SYS 0 M 07
02 DSK RX8E RWF 494 SYS 0 M 07
03 BO 31 RWF 16+ M 50
04 DO 31 RWF 16+ M 60
05 TTY TTY RW 17+ KL8E M 176
06 RXA1 RX8E RWF 494 20 M 34
07 LPT LPTR W 21+ LQP C 176
10 BM 26 RWF 22+ A 44
11 BN 26 RWF 22+ A 40
12 DM 26 RWF 23+ A 46
13 DN 26 RWF 23+ A 42

FREE DEVICE SLOTS: 04, FREE BLOCK SLOTS: 02
OS/8 V40
 
You must log in or register to reply here.
Back
Top