| |
Quite addictive this RiscPC stuff. |
| |
This is a long thread. Click here to view the threaded list. |
| |
|
Bonez |
Message #120999, posted by Bonez at 00:07, 8/9/2012 |
Member
Posts: 50
|
I'm finding the more I read about RiscPCs and play with the damn things the more I like the overall concept. I've just aquired a StrongArm machine to go with my PC600 and PC700 which the seller kindly upgraded the hdd to a 20gig unit. Also recieved a second slice and a few podules as well for $1 each.
Am I sick or is this quite normal? |
| |
 [ Log in to reply ] |
| |
|
Martin Bazley |
Message #121001, posted by swirlythingy at 00:24, 8/9/2012, in reply to message #120999 |
Posts: 460
|
I understand Acorn wanted to make a 'future-proof' computer when they designed the RiscPC, which could be expanded with any upgrade without ever having to buy a whole new machine. (Thus demonstrating their legendary business sense once again.) Naturally, in practice, it didn't quite work out like that - you can't make everything replaceable, and sure enough, there are bits of the RiscPC which are a serious drag in this day and age, notably the memory bus.
But the multiple-slices idea was (AFAIK) unique, and made the RiscPC a really memorable machine. If you're a fan, have you tried Make-A-RiscPC? |
| |
[ Log in to reply ] |
| |
|
Jason Togneri |
Message #121004, posted by filecore at 05:46, 8/9/2012, in reply to message #121001 |
Posts: 3868
|
I presume you've already seen this: http://www.john-ward.org.uk/personal/john/computers/html/rocket.html
Yeah, the slices was an awesome and unique idea. The closest thing I've seen in modern times was specialised single-purpose blades in a rackmount server, but it's not entirely the same thing. |
| |
[ Log in to reply ] |
| |
|
Steven Gregory |
Message #121010, posted by thecellartroll at 14:27, 8/9/2012, in reply to message #121004 |
Member
Posts: 135
|
For some reason I seem to be the only Risc PC user that finds the slice business to be a bloody awful design. Unique and interesting certainly, but practical and sensible it is not!
It would be OK, but whoever designed it forgot that it would have to have cables inside.  |
| |
[ Log in to reply ] |
| |
|
Bonez |
Message #121012, posted by Bonez at 19:36, 8/9/2012, in reply to message #121010 |
Member
Posts: 50
|
What other personal computer of that era can double as a coffee table with enough slices added to it?
That said taking the lid off a PC clone and adding extras bits n bobs really doesn't take that much effort. IRQ conflicts are a pain though. Finaly managed to get 486 mobo with a few PCI slots. |
| |
[ Log in to reply ] |
| |
|
Steven Gregory |
Message #121013, posted by thecellartroll at 19:53, 8/9/2012, in reply to message #121012 |
Member
Posts: 135
|
Hmm. Looking at all the nasty ring-stains on my various tower cases, I'd say a few manage that
No nasty stains on my RPC though. Actually to be fair they seem to have made them from a plastic that resists the horrid retro yellowing quite well |
| |
[ Log in to reply ] |
| |
|
Bonez |
Message #121014, posted by Bonez at 20:10, 8/9/2012, in reply to message #121013 |
Member
Posts: 50
|
I've noticed that too and certainly better than the stuff which Apples were made of which gets brittle with age.
Have you tried any of the "Retro Bright" processes at all? Some good results seem to be had if you are careful.
[Edited by Bonez at 21:14, 8/9/2012] |
| |
[ Log in to reply ] |
| |
|
Gunnlaugur Jonsson |
Message #121025, posted by Gulli at 11:36, 9/9/2012, in reply to message #121010 |
Member
Posts: 138
|
For some reason I seem to be the only Risc PC user that finds the slice business to be a bloody awful design. Unique and interesting certainly, but practical and sensible it is not!
It would be OK, but whoever designed it forgot that it would have to have cables inside. Nah, I loved too at first and then I wanted to upgrade and realised it was not a very practical design. |
| |
[ Log in to reply ] |
| |
|
Jason Togneri |
Message #121034, posted by filecore at 06:00, 10/9/2012, in reply to message #121025 |
Posts: 3868
|
For some reason I seem to be the only Risc PC user that finds the slice business to be a bloody awful design. Unique and interesting certainly, but practical and sensible it is not!
It would be OK, but whoever designed it forgot that it would have to have cables inside. Nah, I loved too at first and then I wanted to upgrade and realised it was not a very practical design.
If they had some sort of cable routing (and standardised cabling) down the side of the expansion backplane area, that would have solved a lot of it. I'll admit that the cabling inside the RPC was far from ideal. It's more the concept of infinite slice-based expandability that was cool, despite there being a practical maximum of four slots on the backplane; that said, there was even once an experimental (and working, as far as I recall) 8-slot backplane that would have allowed for a fully-functional 8-slot machine.
But yeah, my Kinetic + 4.03 was probably the best upgrade I could ever have possibly made to the RPC. I just keep wishing that the second CPU slot could have been for a genuine dual processor setup rather than just a slot for an alternative processor. |
| |
[ Log in to reply ] |
| |
|
Steven Gregory |
Message #121043, posted by thecellartroll at 09:23, 11/9/2012, in reply to message #121034 |
Member
Posts: 135
|
If they had some sort of cable routing (and standardised cabling) down the side of the expansion backplane area, that would have solved a lot of it. I'll admit that the cabling inside the RPC was far from ideal. It's more the concept of infinite slice-based expandability that was cool, despite there being a practical maximum of four slots on the backplane; that said, there was even once an experimental (and working, as far as I recall) 8-slot backplane that would have allowed for a fully-functional 8-slot machine.
Ah yes, some kind of integrated cabling solution could have made a lot of difference. Placing the processor slots a couple of cm further back in the machine would also have improved things.
I'm also pretty dubious as to whether that PSU could handle all 4 podule slots being populated along with the two processors, let alone 8. I don't know if they ever made an uprated PSU, but they certainly didn't think to make a solution for adding another PSU into the super-stacked RPCs we all dreamed of when they were first released.
What would have been nice would have been for the two processors sockets to have both been used for ARM processors, which would mean that RISC OS would have stood some chance of having multi-core support, along with a way of fitting a full x86 mainboard inside the case, with in-built monitor switching and drive sharing. |
| |
[ Log in to reply ] |
| |
|
keith dunlop |
Message #121052, posted by epistaxsis at 15:27, 11/9/2012, in reply to message #121043 |
Posts: 159
|
Actually you could add extra power supplies and the second processor slot could, in theory, have an ARM card plugged in.
My Risc PC ended up being a four slicer with two power supplies and loads of drives.
The 8 slot backplane never worked though
Also there was a "certain technique" dealing with the drive cables even on a single slicer - I learnt it when I used to make Risc PCs for a living... |
| |
[ Log in to reply ] |
| |
|
Charles Justin Ferguson |
Message #121069, posted by gerph at 18:44, 13/9/2012, in reply to message #121001 |
Member
Posts: 48
|
I understand Acorn wanted to make a 'future-proof' computer when they designed the RiscPC, which could be expanded with any upgrade without ever having to buy a whole new machine. Most contemporary hardware was equally expandable without needing to buy a new machine. The most significant gain it had over other systems was that the system was easy to take part due to the plastic locking pins, and the fact that the case was plastic meant that it was significantly lighter than the A5000, or the contemporary PCs. Those are about the only redeeming feature I can think of about the design.
Remember that the Amiga had a plastic case, and was a bit more compact than the RiscPCs which it predated. I think Acorn were trying to retrieve the 'easy to expand and control things' image that the BBC had.
Otherwise, the design was pretty limited. Although you could take the machine apart easily when it was a single slice, a two or three slice system became unwieldy very quickly. The two most common upgrades - memory and hard discs were the most difficult to get to. The primary drive was at the bottom, and you have to remove all the slices (and with them any podules that were screwed to them - because you had to screw them in, otherwise a knock could easily loosen their connections), just to get to it. Memory was exactly the same, hidden under the floppy drive, if I remember rightly.
Extra harddiscs sat in the slices and (if my sketchy memory is anything to go by) were not easy to fix in. Together with the cables that came off the podules to make the extra drives work, it became something of a challenge to get the cable to the right place. Being IDE, if you were adding a new drive you had to remember that the drive that was on the cable initially needed to be master, and not to use cable select (or both use cable select), which invariably meant having to dig down to the second drive in order to check it. And you *had* to remove the drives because those jumpers are so fiddly you cannot remove them in situ - too many times of nearly managing to do so, but then dropping the jumper down under everything (and requiring the entire machine to be taken apart anyhow).
Fitting a new processor was relatively easy, so long as you didn't have any podules in the way, but you did this only a couple of times in the lifetime of the machine (well, normal people - as a developer it was common to switch out the processors, which was nice, but ... not a great selling point).
Fitting new podules wasn't too bad, so long as you had a backplane large enough for the podules themselves - if you needed to fit a new backplane then you had to again remove all the podules. Of course this could take some juggling, because only the lower 2 ports (again, if I'm remembering rightly) support DMA.
With the extra cables flying around from additional drives this introduces more electrical noise into the system, which - if you've got a StrongARM - you were careful of. The larger podule backplane caused problems with some StrongARMs anyhow.
In regard to the StrongARM upgrade, it was a great upgrade - and one that was really needed - but it wasn't really one that the RiscPC was designed for. The main system isn't designed for that kind of speed; it was designed with an expectation of faster processors, but they hadn't expected such a jump and so it's pretty impressive that it works.
Of course, fitting the ROMs for the SA (or 4.x) upgrades was a matter of digging through the machine to get to the bottom. Being regular ROMs, you needed to get some leverage under them, either through a tool, or a screwdriver under them. The extraction tools made it easier, but they still needed care if you were trying to remove chips whilst the podule backplane or processor was present.
On the external design, the slice's sliding front panel was nice, but like the doors on the fronts of PCs, they were often just left open, as you needed to get to the CD or floppy disc drive. The springs were tough, but the plastic that they fixed to were easy to break with regular use, resulting in the panel staying open all the time.
Additionally, ejecting a CD from software would just push against a closed panel, and would require you to open it. Depending on the CD drive fitted, the disc may be retracted, or the tray might just stay in its half-out position.
The power button on the front was obvious and barely worthy of note, but the fact that the reset button was on the back of the machine meant that you rarely used it. Especially when multiple slices were present it became more difficult to reach around to the button - and the most accessible side for the button had the power cable in the way, making it harder to reach.
If you placed the monitor on top of the machine, you found that reaching over the top to the reset button was hard - TFTs spoil us these days, as most of the top of the machine would be taken with the bulk of the CRTs body.
Once you got multiple slices you generally couldn't place any monitor on top of the case, as it ended up too high anyhow. The case could take it but neck strain said 'no'.
Using the machine on its side (without the monitor on top) was reasonable with a single slice, but with multiple slices it tended to not distribute the weight very well.
With the large IDE ribbon cables that had to trail across the machine from either the base or the podules to their drives, the airflow tended to be obstructed, which led to overheating in some cases. In general I don't think that airflow was as much of a design consideration as it is with modern systems. It could get quite warm in a populated system, and there's no case fan - or even a place for fitting one.
As a design, it might be memorable, but it had many detractors, both practical and technical. These days, the 'ease of upgrade' is a completely different league on modern hardware - drives are easily hot-swappable on server and some consumer kit. The systems I deal with these days show you how easy it is to replace parts - flick out the handle, pull machine out of chassis and place one bench, press clip to slide off top, and then you can just push your new expansion card in, or add more memory right in front of you. Your biggest problem when replacing memory is remembering which sticks you took out compared to those that you shove in. You don't tend to replace processors these days, so that part isn't even an issue.
Even desktops have sliding side panels to open the case, and clip in bays for new drives when they're not hot-swapable. The RiscPC's design shows its age significantly in the face of these, and experience has shown that the expandability it offered was really a bit fake.
________
--
Gerph |
| |
[ Log in to reply ] |
| |
|
Bonez |
Message #121081, posted by Bonez at 20:14, 14/9/2012, in reply to message #121069 |
Member
Posts: 50
|
Ok it's all academic know but one thing I find strange is why was the ps/2 mice was not introduced with them. Seemed logical as they had ps/2 keyboards. |
| |
[ Log in to reply ] |
| |
|
Jon Abbott |
Message #121083, posted by sirbod at 20:56, 14/9/2012, in reply to message #121081 |
Member
Posts: 563
|
Ok it's all academic know but one thing I find strange is why was the ps/2 mice was not introduced with them. Seemed logical as they had ps/2 keyboards. Depending on the impementation, a PS/2 mouse could go through the keyboard PS/2 socket with a driver. As a PS/2 device has to request access to the bus, you could also share the socket via a splitter. Some modern laptops use this technique.
One of those things that's probably never been tried, although I suspect the chip that handles it, is specific for keyboards.
Digressing slightly, the A5000 has PCB allocated for PS/2 which isn't implemented - why? |
| |
[ Log in to reply ] |
| |
|
David Holden |
Message #121088, posted by apdl at 09:04, 15/9/2012, in reply to message #121081 |
Member
Posts: 138
|
Ok it's all academic know but one thing I find strange is why was the ps/2 mice was not introduced with them. Seemed logical as they had ps/2 keyboards. Although the PS2 mouse arrived in 1987 it was only 2 button. When the RPC appeard in 1994 most PCs were still using a serial mouse and mass produced generic 3 button PS2 mice didn't really exist. Also Acorn had a shedload of quadrature mice they'd bought for the A3010, which hadn't sold too well. So there *might* have been some logic at the RPC launch. However a year or so later when the A7000 was launched they'd seen sense, although 3 button PS2 mice were still rare.
The really stupid thing was that at the same time as the A7000 they re-designed the RiscPC motherboard for the 700 and it would have been pretty simple to change to a PS2 mouse at that time. After all, they'd already had a load of 3 button PS2 mice manufactured for the A7000.
But then with logic there's Sensible, Acceptable, Dumb and Acorn. |
| |
[ Log in to reply ] |
| |
|
Bonez |
Message #121094, posted by Bonez at 20:59, 15/9/2012, in reply to message #121034 |
Member
Posts: 50
|
If they had some sort of cable routing (and standardised cabling) down the side of the expansion backplane area, that would have solved a lot of it. I'll admit that the cabling inside the RPC was far from ideal. It's more the concept of infinite slice-based expandability that was cool, despite there being a practical maximum of four slots on the backplane; that said, there was even once an experimental (and working, as far as I recall) 8-slot backplane that would have allowed for a fully-functional 8-slot machine.
But yeah, my Kinetic + 4.03 was probably the best upgrade I could ever have possibly made to the RPC. I just keep wishing that the second CPU slot could have been for a genuine dual processor setup rather than just a slot for an alternative processor. Maybe someone can help this dude out- http://www.old-computers.com/forum/forum_posts.asp?TID=6806
[Edited by Bonez at 22:24, 15/9/2012] |
| |
[ Log in to reply ] |
| |
|
Philip Webster |
Message #121099, posted by pwx at 00:36, 17/9/2012, in reply to message #121034 |
Member
Posts: 227
|
The 8 slot backplane was an actual product from Atomwide. I own one. 8 podule slots fill 4 slices - 2 podules per slice. The PSU can't handle it! |
| |
[ Log in to reply ] |
| |
|
Bonez |
Message #121101, posted by Bonez at 00:48, 17/9/2012, in reply to message #121099 |
Member
Posts: 50
|
The 8 slot backplane was an actual product from Atomwide. I own one. 8 podule slots fill 4 slices - 2 podules per slice. The PSU can't handle it! These days you can get uprated 200w PSUs how would they cope? |
| |
[ Log in to reply ] |
| |
|
David Holden |
Message #121103, posted by apdl at 06:22, 17/9/2012, in reply to message #121101 |
Member
Posts: 138
|
The 8 slot backplane was an actual product from Atomwide. I own one. 8 podule slots fill 4 slices - 2 podules per slice. The PSU can't handle it! These days you can get uprated 200w PSUs how would they cope?
The problem isn't primarily the PSU, it's the backplane itself. The data lines were just too long and put too great a loading on the bus. This is compounded by the fact that many podules date back to pre RiscPC days and were 'old technology' requiring higher drive currents. If you've got more than 4 podules then, inevitably, some of them will be in this category. Sometimes even with a 4 slot backplane the combination of 'old' and 'later high speed' podules causes problems with some machines. |
| |
[ Log in to reply ] |
| |
|
Bonez |
Message #121121, posted by Bonez at 06:31, 22/9/2012, in reply to message #121103 |
Member
Posts: 50
|
I'm curious to see how things work out. Got another StrongArm machine that's a bit flakey. I suspect battery leakage and shonky battery replacement, as it's not holding charge keeping time when switched off(charging circuit is ok though) and internal speaker is crackly continuously when switched on.
I'll use this as an experimental beasty. I've already transplanted the 586 card from it to another machine. For some reason the PC emulation software just hiccuped, but works fine on the other machine. Installed PC Dos 6.3 just fine and have tested some dos software.
If things get too bad with "el skonky one" I'll just part out the vram, ram, network module, roms, psu and hdd etc.
It came with a horrible looking Logitech mouse.
I've just been reading some of your old Riscworld articals Dave. I have a few socket 7 cases with AT psus laying about and may even toss the system in one of those as a rainy day project.
Also have a local contact(New Zealand)that knows RiscPCs inside and out which I can glean advice from. Also bookmarked plenty of useful RicsPC related web pages for future reference as well.
Note- Just offered another complete RiscPC600 so will put the good bits in that. Found this interesting artical mentioning the video wasn't that great on later boards compared to the earlier revision. Though fixed straight foward enough if you have the skills http://www.jrmiller.demon.co.uk/rpc/mods.html and a simple fix for the broken spring tab.
[Edited by Bonez at 22:08, 22/9/2012] |
| |
[ Log in to reply ] |
| |
|
Bonez |
Message #121131, posted by Bonez at 01:38, 29/9/2012, in reply to message #121121 |
Member
Posts: 50
|
Got the PC600. Came with a Cumana panasonic interface cdrom and pcb. Though the cdrom was connected corectly and driver correctly installed it'd just hang on startup with device 21 error.
Just booted holding the rh shift key and deleting the driver got it back booting from the hdd ok.
I suspected the later model IDE cdrom may not be picked up cause it's got ROS 3.6 I decided "what the hell" and had a shot anyway. The standard "CD is faulty". It noticed when a cd is not in the drive. Out of curiosity loaded the panasonic drivers. Now no hang up on start up but still no cd access though.
Swapped cdroms, cables, checked HDD was set to Master. No joy. I remembered I'd received a 20 gig hdd to replace the StrongArms drive which was ment to be iffy with RiscPCs. Booted it up, set the desktop up, turned it off, then hooked up the cd and restarted the machine-instant access to the cdroms.
Id still like to sort the issue out with the install on the Conner CFS420A drive though.
[Edited by Bonez at 03:51, 29/9/2012]
[Edited by Bonez at 06:13, 29/9/2012] |
| |
[ Log in to reply ] |
| |
|
David Holden |
Message #121132, posted by apdl at 07:02, 29/9/2012, in reply to message #121121 |
Member
Posts: 138
|
I'm curious to see how things work out. Got another StrongArm machine that's a bit flakey. I suspect battery leakage and shonky battery replacement, as it's not holding charge keeping time when switched off(charging circuit is ok though) and internal speaker is crackly continuously when switched on. Do you mean the battery isn't holding charge (in which case it's either a faulty battery or, sometimes, a leaky C70) or that the clock stops when the power is off? The latter is a common problem if the machine has been left for a couple of years without power and with a completely flat battery. Sometimes it can be coaxed back into life but often needs a replacement CMOS IC. |
| |
[ Log in to reply ] |
| |
|
David Holden |
Message #121133, posted by apdl at 07:12, 29/9/2012, in reply to message #121131 |
Member
Posts: 138
|
Swapped cdroms, cables, checked HDD was set to Master. No joy. I remembered I'd received a 20 gig hdd to replace the StrongArms drive which was ment to be iffy with RiscPCs. Booted it up, set the desktop up, turned it off, then hooked up the cd and restarted the machine-instant access to the cdroms.
Id still like to sort the issue out with the install on the Conner CFS420A drive though. Dump the 420A. It's the best part of 20 years old and although they were very robust drives they are slow, even by RiscPC standards. There were also problems with the 210A and 420A sharing a bus with many makes of CD ROM which I suspect is what your problem was.
The problem with these Conner drives was because many of them weren't standard drives, they were originally supplied to Olivetti who had the firmware changed for some reason and Conner denied all responsibility. Acorn was owned by Olivetti at this time and the drives were given to Acorn for use in RiscPCs. They were OK on their own but quite often 2 of the 201As wouldn't even work together as Master and Slave. |
| |
[ Log in to reply ] |
| |
|
Bonez |
Message #121134, posted by Bonez at 08:21, 29/9/2012, in reply to message #121133 |
Member
Posts: 50
|
Came to the same conclusion Dave after a few tries at sorting it. Funny you mentioned Olivetti, I have a PCS 40 with a 486SX33 in it I was given. I'll slap the Conner drive in. I does actually run quite well. It's got a Sound Blaster card with cdrom attached and should work out nicely.
As for the SA mobo I'm not going to bother repairing it. Not really worth the effort. I'll just use it as a hack/spares machine.
[Edited by Bonez at 09:27, 29/9/2012] |
| |
[ Log in to reply ] |
| |
|
Bonez |
Message #121142, posted by Bonez at 04:33, 1/10/2012, in reply to message #121134 |
Member
Posts: 50
|
Found an 8 gig Quantum Fireball dive which appears to be liked. No joy with a few other makes I have laying around. Once inilised and recognised on the desktop I did the Configure ADFSBuffers 0 thing and installed UniBoot using ZipEE/ff8. So it's bootable now which is cool.
Still moans about the verious cddrives though with, "cdrom faulty", "no cd drive present" and few other errors depending on the cd drive I try out. No biggy though.
[Edited by Bonez at 03:16, 2/10/2012] |
| |
[ Log in to reply ] |
| |
|
Bonez |
Message #121696, posted by Bonez at 04:26, 23/12/2012, in reply to message #121142 |
Member
Posts: 50
|
Tracked down a 16 sound module for the PC600 mobo, also another suitable bus mouse so no need for a ps/2 adapter now. |
| |
[ Log in to reply ] |
| |
|
Bonez |
Message #121862, posted by Bonez at 07:36, 3/2/2013, in reply to message #121696 |
Member
Posts: 50
|
Another addition to the family will a 410/1 with Risc Developements IDE podule and ARM3 update. Not too sure on how much ram it's got though. It'll be interesting to get this up and running. Need to get an RGB to VGA adapter though, as I don't have an RGB monitor. I know about the green on sync issue but not too worried about it at this point.
Unsure of what ROMS it has but I'm suspecting at least 3.1. I'd like to get an ethernet podule for it at some point.
Where the podules in these compatible with RiscPCs?
Also eying up an A4000 and a few other goodies. |
| |
[ Log in to reply ] |
| |
|
Bonez |
Message #121927, posted by Bonez at 21:01, 16/2/2013, in reply to message #121862 |
Member
Posts: 50
|
Picked up a network podule for the 410/1 and the PC600. Also gotten hold of an unused Risc OS 3.7 updrade kit designed for upgrading to a RiscPC to SA. It'll sort the CDRom issue. This also gives the option of soft loading Risc OS 4.02.
I may have an A4000 next week. I was also lucky enough to pick up an A3 touch pad keyboard with a couple of templates for use on user ports.
Time to flog of some x86 stuff................. |
| |
[ Log in to reply ] |
| |
|
Bonez |
Message #122009, posted by Bonez at 00:36, 9/3/2013, in reply to message #121927 |
Member
Posts: 50
|
Got a A4000 c/w nic. The hdd seems flacky but no big deal as my intention is to stick a IDE-CF convertion this one anyway. I'm picking up a manual VGA switcher to share the Trinitron monitor with my DECpc. That should be here by the end of the week. |
| |
[ Log in to reply ] |
| |
|
Bonez |
Message #122064, posted by Bonez at 19:12, 16/3/2013, in reply to message #122009 |
Member
Posts: 50
|
Hmmm won auction for another A4000, so now have a good supply of spares. The VGA switcher arrived and does the trick. |
| |
[ Log in to reply ] |
| |
| Pages (2): 1
> >|
|
