brooks@maddog.llnl.gov (03/16/89)
Before I get a hundred replies for inaccurate memory of vacuum tube computers, it appears that my memory of what media was used for delay line purposes was not correct. These lines were also called "mercury lines" because mercury was used to create a suitably slow propagation speed for those suitably slow computers of the day. Just to remind us of where we have been, does anyone who has access to a suitable article want to post a one page synopsis of just how these delay lines worked? The past is often useful in forming the future as new technology spawns. Is the news software incompatible with your mailer too? brooks@maddog.llnl.gov, brooks@maddog.uucp, uunet!maddog.llnl.gov!brooks
haynes@ucbarpa.Berkeley.EDU (Jim Haynes) (03/16/89)
Mercury delay lines, as used in EDSAC and UNIVAC I, I believe used quartz transducers and pulses of RF propagating in the mercury. Hence the binary data modulates (on/off) a radio-frequency signal. It's detected at the receiving end and reconstituted into data, and also re-introduced at the sending end so that the stream of bits in the mercury recirculates forever. Mercury delay lines were also used in radar and in distance measuring equipment for air navigation. Later there was a technology using solid wire as the delay medium, with the data being represented as a torsional pulse in the wire. This didn't use the RF carrier; the data bits went directly into the wire. I've seen this technology used in transistor computers: Packard-Bell 250 and Royal Precision RPC something or other, I forget the model number. The transducers were magnetostrictive. There were also some very high speed delay lines that used something like a glass prism as the delay medium. Signal was introduced at one face, traveled through the material to another face, where it was reflected at an angle to another face, and so on through several more reflections until it emerged from the output face.
rfm@sun.com (Rich McAllister) (03/16/89)
In article <21976@lll-winken.LLNL.GOV>, brooks@maddog writes: >Just to remind us of where we have been, does anyone who has access to a >suitable article want to post a one page synopsis of just how these >[mercury] delay lines worked? Here's a shortend quote from _IBM's_Early_Computers_, by Bashe, Johnson, Palmer, and Pugh, MIT Press, 1986, p110. Highly recommended, especially for people who think IBM never invents anything. The idea of using a mercury tank acoustic delay line for storage in computers had originated about [1944] at the University of Pennsylvania, where [Eckert and Mauchly were working on the EDVAC.] Eckert had invented the mercury delay line [for radar applications.] He recognized in the delay line a device capable of storing a large number of binary digits while requiring relatively few vacuum tubes. His idea was to introduce a train of electrical pulses and gaps representing 1s and 0s, respectively, to a quartz-crystal piezoelectric transducer at the front of a mercury tube or tank and allow the acoustical waves generated by the transducer to propagate through the mercury. At the far end the train of electrical signals was reconstructed by another piezoelectric transducer .... Each delay line could store hundreds of bits [!] at the cost of a few vacuum tube circuits and a temperature-controlled tank. The temperature control is necessary to regulate the speed of sound in the mercury, so the same number of bits always have the same delay. Rich McAllister (rfm@sun.com)
lfoard@wpi.wpi.edu (Lawrence C Foard) (03/17/89)
Is data transmission with ultra sound practical? You can get 10Mhz ultra sound in water, and it travels very fast. I wonder if connecting PC's with garden hoses would be cheaper than ethernet cable, no electrical noise but stepping on the hose could cause problems ;) Does any one have a list of prices and speeds of available risc chips? -- Disclaimer: My school does not share my views about FORTRAN. FORTRAN does not share my views about my school.
beyer@houxs.ATT.COM (J.BEYER) (03/18/89)
In addition to delay line memory, some machines used magnetic drum memories (e.g., IBM 650). The trouble with delay line memory and most drum memory is that you have to wait until the address you want comes along out of the memory. By Murphy's law, the one you want just went by, so you have to wait until it comes around again on the guitar. The larger the memory, the longer it is and the longer you have to wait. There are ways around this, sort-of. What was done to the 650 was that each instruction had 2 addresses, one where the data came from or went to, and one, the address where the next instruction was to come from. A clever assembler was used that knew the execution times of the instructions, and arranged to place the next instruction at a location that would just be coming around as the current one finished. Fun, huh! -- Jean-David Beyer A.T.&T., Holmdel, New Jersey, 07733 houxs!beyer
dave@stcns3.stc.oz (Dave Horsfall) (03/22/89)
In article <1365@wpi.wpi.edu> lfoard@wpi.wpi.edu (Lawrence C Foard) writes: | Is data transmission with ultra sound practical? You can get 10Mhz ultra sound | in water, and it travels very fast. I wonder if connecting PC's with garden | hoses would be cheaper than ethernet cable, no electrical noise but stepping | on the hose could cause problems ;) I can see it now... ``The Nylex Corporation, a manufacturer of garden hoses and sprinkler accessories, today announced a product aimed specifically at the fast-growing PC LAN market. Called HoseNet, it is based upon a cheap flexible medium utilising the latest techniques in transmission technology...'' Sorry, but I thought a little levity would be appropriate at this point. -- Dave Horsfall (VK2KFU), Alcatel STC Australia, dave@stcns3.stc.oz dave%stcns3.stc.oz.AU@uunet.UU.NET, ...munnari!stcns3.stc.oz.AU!dave Self-regulation is no regulation
maguire@cs.columbia.edu (Gerald Q. Maguire) (03/25/89)
I first gave a talk which described using optical delay lines for data storage in spring 1986 in Belgium and in The Netherlands. Following a later talk which included this idea - I was ask by Dave Sincoskie of BELLCORE if I had compared it to a disk, I had not at the time but since have. This comparison and the limits to this technology are described in a paper: @Inproceedings[Maguire89c, Key=<Maguire>, Author=<G. Q. Maguire Jr. and P. R. Prucnal>, Title=<High-density optical storage using optical delay lines: Use of optical delay lines as a disk>, Organization=<SPIE - The International Society for Optical Engineering>\, Booktitle=<Medical Imaging III>, Year=<1989>, Month=<January>, Volume=<1093>, Pages=<6 pages> ] For those who are interested and can't wait for the proceedings, send me e-mail and I'll mail you the PostScript version of this file (which is 42173 bytes long). This file includes the diagrams. I hope to find a student interested in building a prototype which will act much like a SCSI disk (in having logical sectors), but with considerably faster access time. I anticipate a major application of this technology to be in local swap space (as there are no moving parts and 500MB will fit within a 9 liter volume using a single frequency and much less volume if you freq. multiplex - but you then increase the size and complexity of the optics) and for image processing it makes an image from among 100T bytes of image data available within 1 sec. Chip
david@linc.cis.upenn.edu (David Feldman) (03/27/89)
> From: rfm@sun.com (Rich McAllister) > Here's a shortend quote from _IBM's_Early_Computers_, by Bashe, Johnson, > Palmer, and Pugh, MIT Press, 1986, p110. Highly recommended, especially for > people who think IBM never invents anything. > > The idea of using a mercury tank acoustic delay line for > storage in computers had originated about [1944] at the > University of Pennsylvania, where [Eckert and Mauchly > were working on the EDVAC.] Eckert had invented the We had a mercury delay system laying around the Digital Systems Lab here at the Moore School at Penn. According to Eckert, who came to Penn for publicity, there were piezo quartz elements at both ends of the delay lines. This unit had five or six tubes (now emptied of mercury) arranged cylindrically. It also had some sort of interfaces in the middle of the delay tubes with vacuum tube amplifiers attached. Eckert told me that (if his memory was correct, he said) this unit was designed to test variable delay delay lines. We had some more old stuff laying around too. There was a drum memory with 256 read heads. Pieces of the ENIAC that weigh nearly 1000 pounds each. Bits of the old differential analyzer. An original PDP 8 made of discrete transisters and diodes, with a polar vector scope used (among other things) to play the original "Space Wars". All of this stuff has been thrown out now. I consider myself lucky to have seen it. _ /| Dave Feldman \'o.O' david@dsl.cis.upenn.edu =(___)= Ok, cough! U DSL - land of wonder and enchantment ACK! PHHT!