test-drb July 2014

test-drb@ccmp.vtda.org

181 participants
190 discussions

by marco＠familie-rauhut.eu

Hi i have here a VAXstation 4000/60 with a VT100 Terminal attatched the serial console. In >>> console mode are some commands not availible and i can not figure out why. I want top set kbd to 1 . >>>set kbd answer is ?23 ILL CMD. Also by typing set kbd 1. The test / util commands do not work also... I tried to set a password. Type set pse 1. Login. But there is even the same isue. Any Ideas what is happen? Marco

10 years, 5 months

Counterfeit ICs, was Re: 6809 vs. 6809E

by mike＠fenz.net

In most cases, I find the remarked ICs are actually OK. They're obviously fakes still (commonly the origin printed on the top is different to the origin in the casting on the bottom), but functionally they're fine. This puzzled me initially, but I asked one of the sellers about this, and he said that it was done because the process of removing the ICs from the boards often marks or damages the printing. So the standard process involves remarking, straightening the legs to factory original angles, and re-tinning. It results in used ICs that look new. Sometimes they're sold as 'remanufactured' (which is sort of accurate), other times they're sold as new, which is obviously untrue. I'm not that bothered so long as they work, though I probably should complain about it. On 3/07/2014 9:03 a.m., Tony Duell wrote: > This is becoming real problem. The remarked devices, of course, have no relation to what they claim to be, they can't possibly work. Heck, I once enqired about a replacement for an HP custom IC and was asked 'what package, and how many pins'. Given this device had only ever existed in one pacakge, with one particualr pin count, it wa obvious I was goign to get a fake. I do wonder why they botyher. They can't make that much money selling to hobyists, surely no commercial company wants such fakes. > >> I spent a while looking >> over the datasheets of the 68B09 as compared to the 68B09E. It's obvious >> that there's some difference in the way the timing works, but I haven't >> been able to come to any firm conclusions what the exact difference is, >> and what purpose it serves. The best I can do is that it's related to >> synchronization with some external devices. > > he difference is where hte clock circuitry is. > > The 6809 (normal and E_) need a pair of clocks in hase quadrature. The E > clock (Enable) and the Q clock (Quadrature). FWIW, the 'clock rate' of > the 6809 is specified as the frequency of these clocks > > The 6809 (on E) has an internal clock generator. It can be driven by a > crystal of 4 tiems the CPU clock rate (soe a 68B09 can use an 8 MHz > crystal). The E and Q pins o nthe chip are outputs > > The 6809E (E=External) has no internal clock generator. The E and Q pins > are inputs, to be driven by an extgernal circuit that produces the > approrpatie signals in phase quadrature. > > IIRC, i f you use the 6883 SAM (DRAM controller) you prettty much hace to > use the 6809E. The SAM generates E and Q, and expsects to eb able to > synchroinse the CPU to them. > >> So - can anyone explain >> the difference, in terms that aren't too hard to understand? And - is >> there any hope of modifying a machine that would normally require a >> 68B09E to be able to use a 68B09? > > That depends on the circuitry around the CPU. However, if it sues the > 6809E CPU, it's a fair bet it is goign to do oomething odd with E or Q > (perhaps stretchign one of them when acessing slower memory or > something). In which case, tryign to use the plain 6809 is goignt to be a > lot of owrk. The 6809E was origianlyl chosen for a reason, after all (and > I doubt it was cost, the cost of the external clock curcuit would > outweight any difference in price of the 2 chips when new), so the reason > is more likey that said clock circuit does soemthing extra that can't > easily be done wit hthe plain 6809. > > -tony

10 years, 5 months

6809 vs. 6809E

by mike＠fenz.net

Yes, I'm certain it'd be possible. All the digital components are very standard, apart from the later WPC system, which includes an undocumented ASIC. I think producing a working CPU would be a good first step though. Then, if I'm keen, I could try integrating all the I/O etc and see what happens. On 2/07/2014 5:04 p.m., ben wrote: > On 7/1/2014 10:39 PM, Mike van Bokhoven wrote: > >> I don't think the requirement is there yet. 68B09Es are still >> available; they're getting a bit more expensive, but you can still pick >> them up for under $10 without too much trouble. They don't fail that >> often either, failures usually seem to be in the RAM, 6821s, sockets >> etc. When you say 'use a small FPGA development board', do you mean to >> emulate the entire controller board, or just the CPU? Just curious what >> you had in mind. > > I have seen the 6809 color computer emulated as well as the SWTP 6800 > on a FPGA. The only thing lacking is GOOD I/O for the devices as > most FPGA's have just a SD card interface for I/O. > Ben.

10 years, 5 months

HP3000 Available

by leec2124＠gmail.com

Hi all, FYI I am starting to trim down my collection and have listed my running HP3000 Micro3000XE system on EBay, auction number: 221481360061 Classic 3000 with lots of spares and software included. Email with questions or for a detailed spreadsheet of items. Thanks! -- Lee Courtney

10 years, 5 months

cleanout in Uruguay

by sales＠elecplus.com

. The man's name is Jonathan Gill, email <mailto:jmhgill at hotmail.com> jmhgill at hotmail.com. . I will be out of pocket for the next couple of weeks, please contact him directly if you are interested in any of his stuff before it becomes landfill. Cindy Croxton Electronics Plus 1613 Water Street Kerrville, TX 78028 830-792-3400 phone 830-792-3404 fax sales at elecplus.com AOL IM elcpls

10 years, 5 months

OLD computer stuff in Uruguay, he will ship to end users

by sales＠elecplus.com

Hi Cindy, I'm having a big clearout of my shed here. There is all kinds of stuff from cp/m, xt, 286, 386sx, 386, 486sx,486, 586amd, k62,P1, pII,P3,P4,Athalon y Sempron. There are mothers, micros, ram, fdd, hdd, cd power supplies, cabinets -you name it I probably have it. So if you have a wish list of any old stuff let me know before it becomes landfill ! Please email me if you want his email address, he will ship. Will sell tested and untested. Also has 360KB and 1.2MB floppies, not yet tested, lots of them. Cindy Croxton Electronics Plus 1613 Water Street Kerrville, TX 78028 830-792-3400 phone 830-792-3404 fax sales at elecplus.com AOL IM elcpls

10 years, 5 months

Multi-Precision Unsigned Integer Multiply and Diivide

by jhfinedp3k＠compsys.to

Just to make sure for anyone who is reading, this code will be for a PDP-11 which supports 16-Bit values. About 10 years ago, I produced some rather complicated subroutines for multi-precision Unsigned Integer Multiply and used inverse Multiply to effectively produce Divide. Since they were mostly for a minimum of 128-Bit and larger arithmetic, I felt that they were reasonable as a first effort. More recently, I find that I now need a small and reasonably efficient routine (both in code size and efficiency - assume the extra multi-integer instructions MUL and DIV are present) to Encode 32-Bit (perhaps up to 48-Bit, but certainly no larger than 64-Bit) Unsigned Integers - which requires effective division by TEN. I already have very simple alternatives which can perform Unsigned Division of 32-Bit Unsigned Integers (and by extension 48-Bit Unsigned Integers); they either require lengthy one bit at a time manipulation for all 32 bits or a table for subtracting all of the 9 powers of TEN vs all of the 32 Bit Unsigned Integer. Since I already MUST have a subroutine to Encode a 16-Bit Unsigned Integer, that is the acceptable starting point at the present time. The alternative I am using right now is to (R0,R1 holds the 32-Bit Value): Cmp #4999.,R0 ;5000.*65536.=327680000. Bhis 110$ ;Branch if DIV will not overflow 10$: ; Subtract 100,000,000 up to 42 times ;..... 110$: Div #10000.,R0 ;Get two values to Encode ;..... As long as the 32-Bit value is less than 327,680,000 (which means that R0 is less than 5000 - if you divide 327,680,000 by 65536, you get the high order portion of the 32-Bit value - otherwise the quotient overflows), the Div #10000.,R0 can be used. That results in three cases for any Unsigned 32-Bit Integer: (a) Less than 65,536 which is just the remainder, R1 to Encode (b) Less than 327,680,000 which uses the DIV to produce both the quotient, R0, and the remainder, R1 to Encode (c) Less than 4,294,967,296 (which is one more than the maximum for an Unsigned 32-Bit Integer) for which I subtract 100,000,000 up to 42 times (rare and quite acceptable), Encode that value, then finish by taking care of the remainder of up to 99,999,999 using (b) above. Note that the number of subtractions can be greatly reduced (to 12 times or less) if the first subtractions are for 1,000,000,000 Is there some (relatively simple and code size efficient) algorithm that can use the SIGNED PDP-11 DIV instruction for values that are equal or greater than 327,680,000 so that ONLY 16-Bit values need to be Encoded? OR, if the 32-Bit value is first shifted, is there some value that I can use to MUL or DIV with to obtain the first value under (c), i.e. the upper two digits for the HUNDREDS of MILLIONS and BILLIONS without having to subtract up to 42 times (or even 12 times)? For example, if just the upper portion of the 32-Bit value, in R0, is divided by 1525, the result will usually be the required upper two digits, but as each multiple of 100,000,000 is closely approached, the result will (rarely) be one too high. Dividing by 1526 produces a result which will (rarely) be one too low. Perhaps there is an algorithm that will always be correct and the multiple subtract is not needed? Maybe someone has found a correction factor which uses the remainder of the DIV by 1526 plus the lower portion of the original 32-Bit value, in R1, to correct the result when that result is one too low. Just thought I would ask to see if anyone else already found a much better algorithm. Jerome Fine

10 years, 5 months

DEC VTs , VRs, and LAs

by useddec＠gmail.com

My crew is supposed to start on my last storage unit, 25 ft, next week or so and be done before school starts. There are about 30 or more VTs, VRs, and LAxxs close to the door, I would really like to sell most of them. Most worked about a year ago, except for the VT340s and VT340+s. There are several almost complete LA210s, and should be complete LA12s, 50s, 75s, 100s, 120s, LS120s and maybe others. There are no VT05, 50, 52, 55, 61, 72,or 100s at this time. Everything is newer except a few WT and / or WD78. most can be shipped, but cost a bit, especially vrc21s. I have a lot of LA120 and other parts. I also have new tubes for the VT330s and 330+s , all three colors, which may fit in other terminals. Please contact me off list. Shipping from IL, 62853. Pickup is great. Thanks, Paul

10 years, 5 months

HP-1000 Minicomputer System (w/DEC RL02) Guidance

by christopher.parish＠parishcomputers.com

I wasn't expecting such an enthusiastic response, thanks everyone! Right now, I'm focused on getting the RL02s back in proper working order. Both drives have worn spindle ground brushes (the little graphite conical things under the spindle) and I'm looking around for replacements. Graphite is neat in that it's a conductive, self-lubricating material. There are a few places that will manufacture custom graphite parts that I'm looking into, but it will likely need to be a bulk order to make it economically effective (and it may still be obscene). An alternative material is brass. That part is affordable in small quantities ($30ish for 1). It's still significantly softer than steel, but friction and galling could be a problem. > You may want to consider using clang instead of gcc for your compiler. Thanks for the info. After looking at the software architectures more closely, I really do like how de-coupled and abstracted everything is in LLVM/Clang. When it gets to that point, that's what I'll use. > Have you seen Reinhards work? Not quite what you are doing but > maybe it can be of help. > http://www.pdp11gy.com/indexE.html#file:///E:/homepage/indexE.html That was one of the first things I saw in Google when I originally researched the disk interface. There's plenty of good information there about the packs and the drive communication. Between that and the bitsavers technical manual, I think I have everything I need. I foresee some irony involving an emulated RL02 connected to an emulated RL02 controller... > But to be honest, I don't see the logic in using an RL02 on an HP-1000 system in the first place. It all comes down to price and technical challenge. Original accessories are hard to find, and the ones I have seen are very expensive. By comparison RL02s are everywhere and packs cost almost 10 times less, but that's not the *real* reason. By day, I am a freshly minted computer scientist, but I find far more enjoyment developing system level stuff. I love vintage computing, and what better way to spend my off-hours than using super-modern technology to add new features to these landmarks of computing? In order to develop an interface between the HP and the RL02, I've had to learn all sorts of things about Verilog, timing closure, proper design constraints, etc. For me, it's the best of both worlds. I get to play with computers that are older than I am (don't judge me ;) ) and learn marketable skills that should help me down the road. Christopher

10 years, 5 months

6809 vs. 6809E

by mike＠fenz.net

A question for someone with specialist knowledge in 6809s. I work on pinball machines a lot, and the 68B09E is a commonly used CPU. Unfortunately, it's getting harder to find 68B09Es (though by no means impossible quite yet). It seems that most of them on eBay are counterfeits; the last lot I bought were all fake, marked Japan on the front, but China on the back. They looked very convincing, with the markings etched very nicely, rather than printed. Unfortunately, they didn't work. Overall, it looks like getting 68B09Es is becoming harder, and I'm wondering if there's some alternative. I spent a while looking over the datasheets of the 68B09 as compared to the 68B09E. It's obvious that there's some difference in the way the timing works, but I haven't been able to come to any firm conclusions what the exact difference is, and what purpose it serves. The best I can do is that it's related to synchronization with some external devices. So - can anyone explain the difference, in terms that aren't too hard to understand? And - is there any hope of modifying a machine that would normally require a 68B09E to be able to use a 68B09?

10 years, 5 months

Jump to page:

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

2003

2002

2001

2000

1999

1998

1997

test-drb July 2014