Re: George!
It looks like you've refactored/consolidated a bit - or were there components you hadn't installed at VCF-E? We installed the patch panels. No other change since VCF East.
Re: Megaprocessor - built from individual transistors
On June 24, 2015 1:27:56 AM PDT, Pontus Pihlgren wrote: > On Wed, Jun 24, 2015 at 06:10:38AM +, d...@661.org wrote: > > On Tue, 23 Jun 2015, Phil Budne wrote: > > > > >If I were going thru the trouble, I'd want build a TX-0 clone! > > > > I think it would me more interesting to build a replica of a pdp-8 > > straight-eight using significantly-reduced flip chips with > > surface-mount parts. > > > > This exact thing has been on my TODO-list for a while. It will > probably > remain on the TODO-list for some time, but I really want to do it! > > I like to think that you could shrink the computer by, at least, a > factor of four. Probably smaller. > > (make the flip chips double sided and double the function of each flip > > chip). > > /P I don't think thatt making the flip-chips double-sided will be desirable. I want to shrink the machine and minimize any need to reengineer the backplane. By the way, can the backplane of a straight-eight be realized as a PCB? Replicating that component without that will suck. -- David Griffith d...@661.org
Re: organizing a trip to Cuba
On 24 June 2015 at 14:19, Johnny Billquist wrote: > Oh, I know. I'm from Sweden. We had a very big scandal where 5 containers > with a VAX-11/782 and peripherials or something like that was found under > strange circumstances. When the whole thing started to be investigated > suddenly no one seemed to know or own those containers. The system was > unclaimed for years, and it became a question of what to do with it, since > no one seemed to claim it. I think it was eventually decided that since DEC > made it, it was returned to them. The original shipping destination was of > course somewhere in Soviet Union. This was in the early 80s... I'm sure > someone can find the full story online somewhere. It's mentioned in the Datasaab article on Wikipedia: https://en.wikipedia.org/wiki/Datasaab -- Liam Proven • Profile: http://lproven.livejournal.com/profile Email: lpro...@cix.co.uk • GMail/G+/Twitter/Flickr/Facebook: lproven MSN: lpro...@hotmail.com • Skype/AIM/Yahoo/LinkedIn: liamproven Cell/Mobiles: +44 7939-087884 (UK) • +420 702 829 053 (ČR)
Re: cctalk Digest, Vol 12, Issue 5
Good info David. Thanks once more. - "J. David Bryan" said: Subject: Re: HP 2113e Battery resistor >I might put NiMH batteries instead That may not be advisable, given the continuous constant-current trickle charger in the CPU power supply. The Panasonic "Nickel Metal Hydride Technical Handbook" recommends charging for no more than 10-20 hours, saying: "The overcharging of nickel-metal hydride batteries, even by trickle charging, causes a deterioration in the characteristics of the batteries. To prevent overcharging by trickle charging or any other charging method, the provision of a timer to regulate the total charging time is recommended." Panasonic's "Nickel Cadmium Batteries Technical Handbook," on the other hand, says explicitly that continuous trickle charging for Ni-Cds is a recommended charging method. -- Dave
Re: Megaprocessor - built from individual transistors
> On Jun 25, 2015, at 7:16 AM, David Griffith wrote: > > ... > I don't think thatt making the flip-chips double-sided will be desirable. I > want to shrink the machine and minimize any need to reengineer the backplane. > By the way, can the backplane of a straight-eight be realized as a PCB? Of course. The only reason a PCB backplane would be an issue is if there are signal integrity concerns, but at DEC speeds that’s not going to be a problem. Especially if what you’re aiming to replace is wire wrap which is in general worse at signal integrity than PCB. Now if you were looking to replace the chassis wiring of a 6600 with a PCB, that would be a different matter, then you’d want to be a bit more cautious. Even then it’s clearly doable, if you use transmission line design on the PCB signals. paul
Re: Megaprocessor - built from individual transistors
On Thu, Jun 25, 2015 at 04:16:03AM -0700, David Griffith wrote: > > I don't think thatt making the flip-chips double-sided will be > desirable. I want to shrink the machine and minimize any need > to reengineer the backplane. By the way, can the backplane of > a straight-eight be realized as a PCB? Replicating that > component without that will suck. I've not chosen that restriction and would not mind rerouting the backplane if it means 1. that I can shrink the design and 2. if it means I can fit it on a multilayer PCB of suitable size. Now, I've looked at the backplane of a PDP-12 and it looks to be 4 or 5 levels deep at the most and fairly "roomy". So, with my limited experience it doesn't seem impossible to recreate in a 4 layer PCB. I'm not sure what would be a suitable edge connector though, that would probably dictate the size of the flip chips more than anything. Please correct me if I'm talking complete bulls**t. This is all new to me. /P
Re: UNIBUS extension card/cable sets
> From: Alan Perry > FYI, in my 750, the UNIBUS expansion has a L0010 in the main cabinet > and a M9014 in the expansion cabinet. > From: tony duell > If it's any help I pulled an M9014 out of the Unibus Out slot of my > 11/730. There's currently an M9302 in there. Very interesting! Thanks for the data. This seems to indicate that the M9014 could function as either end of the cable. I was wondering if maybe the M9015 was an M9014 with termination resistors, or something (the way the QBUS versions come with and with termination), so I tried to find a picture of one, but... I can't even find a picture! Has anyone even seen an M9015? >> I see 'three' different kinds of 'UNIBUS to cables' cards listed: >> M9014 UNIBUS to 3 H854s >> M9015 3 H854s to UNIBUS >> M9031 UNIBUS to 3 3M cables for 11/74 >> M9042 UNIBUS to 3 H854, Dual So I have compared an M9014 and an M9042; the former is a 'normal' height dual module, the latter is a 'short' dual module. I suspect that they have the same pinout on the Berg headers; I tried a couple of UNIBUS signals, and they led to the same pin on the Bergs on the two different units. If and when I get energetic I will make a complete pinout list for the two units (I haven't been able to find any documentation on any of them online). Some day I'll even try joining a BA11 to an -11 with a pair of M9014's, and see if it does indeed work. Also, does anyone out there with an 11/70 know what the M9031 is used for? (I.e. does your system have one/more, and if so, where are they?) Noel
RE: UNIBUS extension card/cable sets
>> If it's any help I pulled an M9014 out of the Unibus Out slot of my > > 11/730. There's currently an M9302 in there. > > Very interesting! Thanks for the data. This seems to indicate that the M9014 > could function as either end of the cable. Somewhere I have the expansion box that went with this VAX. The thing is that I moved some items into a storage unit for the house move and haven't moved them back again. I suspect the expansion box was amongst them. When I get a circular tuit of a suitable type I will take a look at what is in there. The M9014 is just connectors. There are 3 40 pin Berg headers, one row of each appears to be ground (or at least they're all joined together and if they go somewhere other than ground I will be very surprised) That leaves 60 pins for the unibus signals. Of course power will not be carried. > I was wondering if maybe the M9015 was an M9014 with termination resistors, > or something (the way the QBUS versions come with and with termination), so I would be very surprised. Unibus is normally terminated at the ends and not in the middle. I guess the 11/730 terminates the 'start', you put an M9302 in the 'end' I think I did read somewhere you could use a BC11A cable with the 11/730 (this does not surprise me), so that would not have extra termination. -tony
Re: Transmission lines Was: UNIBUS extension card/cable sets
> From: John Wilson > I chose the digital version of EE as my major precisely because I knew > I'd flunk Fields and Waves. Transmission lines are black magic as far > as I'm concerned! I too have a hard time with analog in general, but transmission lines I seem to be OK with. The way I think about them is to model them as pipes, and the signal as a sound (single pulse) sent down the pipe. Proper termination is like a piece of cotton at the end of the pipe, it sucks up the sound and you don't get a reflection. If you just cap off the end of the pipe (i.e. no termination), the sound bounces, and you get an echo. So if you have a small un-terminated branch, part of the pulse bounces off the end, and comes back out, and then propogates both ways, so the original pulse gets a messy trailer tacked on the back of it. Etc, etc. I dunno how accurate this model of mine is, but it seems to work OK! :-) Noel
Re: Transmission lines Was: UNIBUS extension card/cable sets
On 2015-06-25 12:02 PM, Noel Chiappa wrote: > From: John Wilson > I chose the digital version of EE as my major precisely because I knew > I'd flunk Fields and Waves. Transmission lines are black magic as far > as I'm concerned! I too have a hard time with analog in general, but transmission lines I seem to be OK with. The way I think about them is to model them as pipes, and the signal as a sound (single pulse) sent down the pipe. Proper termination is like a piece of cotton at the end of the pipe, it sucks up the sound and you don't get a reflection. If you just cap off the end of the pipe (i.e. no termination), the sound bounces, and you get an echo. So if you have a small un-terminated branch, part of the pulse bounces off the end, and comes back out, and then propogates both ways, so the original pulse gets a messy trailer tacked on the back of it. Etc, etc. I dunno how accurate this model of mine is, but it seems to work OK! :-) For the YouTube generation - there's Dr John Shive's 1959 video: https://www.youtube.com/watch?v=DovunOxlY1k Note, there's an intro of 1 min 20 seconds. --Toby Noel
Re: Megaprocessor - built from individual transistors
From: David Griffith: Thursday, June 25, 2015 4:16 AM On June 24, 2015 1:27:56 AM PDT, Pontus Pihlgren wrote: I like to think that you could shrink the computer by, at least, a factor of four. Probably smaller. I don't think thatt making the flip-chips double-sided will be desirable. I want to shrink the machine and minimize any need to reengineer the backplane. By the way, can the backplane of a straight-eight be realized as a PCB? Replicating that component without that will suck. I've laid out a few of DEC's backplanes as PCBs before, mostly for the later designs that use both sides of the edge connectors. They tend to require about 4 signal layers. It can be helpful to bus power separately (using wire much as was originally done), and to route on the diagonal, as the funky sockets have larger chaseways in the diagonal directions. Also, the device is relatively high current. Some of my earlier attempts used trace widths that, in retrospect, weren't wide enough (especially power traces), trying to keep the number of layers down. I can't particularly recommend the original backplane form factor. They are all based on a 1/8 inch grid instead of the industry standard 0.1 inch, and not at all easy to find or inexpensive these days. From: Pontus Pihlgren: Thursday, June 25, 2015 6:39 AM Now, I've looked at the backplane of a PDP-12 and it looks to be 4 or 5 levels deep at the most and fairly "roomy". So, with my limited experience it doesn't seem impossible to recreate in a 4 layer PCB. Agreed. 6 with power and ground planes might work well, too. I'm not sure what would be a suitable edge connector though, that would probably dictate the size of the flip chips more than anything. To my mind, the obvious choice for a new edge connector would be 0.1" headers and sockets. They are easy to find and produced by the millions, which means they are also relatively affordable. Please correct me if I'm talking complete bulls**t. This is all new to me. The only other concern I've had during my thought experiments along this line related again to the current/power involved. The device is likely to become difficult to cool if you achieve a 4X volume reduction. Oh, and I guess the nightmare of soldering many thousands of surface mount discretes and transistors to manufacture such a thing. Vince
Re: Transmission lines Was: UNIBUS extension card/cable sets
> On Jun 25, 2015, at 12:02 PM, Noel Chiappa wrote: > >> From: John Wilson > >> I chose the digital version of EE as my major precisely because I knew >> I'd flunk Fields and Waves. Transmission lines are black magic as far >> as I'm concerned! > > I too have a hard time with analog in general, but transmission lines I seem > to be OK with. > > The way I think about them is to model them as pipes, and the signal as a > sound (single pulse) sent down the pipe. Proper termination is like a piece > of cotton at the end of the pipe, it sucks up the sound and you don't get a > reflection. If you just cap off the end of the pipe (i.e. no termination), > the sound bounces, and you get an echo. > > So if you have a small un-terminated branch, part of the pulse bounces off > the end, and comes back out, and then propogates both ways, so the original > pulse gets a messy trailer tacked on the back of it. Etc, etc. > > I dunno how accurate this model of mine is, but it seems to work OK! :-) Nice analogy. Those “bounces” occur at any irregularity, and in fact you can locate impedance fluctuations by the echoes they produce; this is called TDR (time domain reflectometry). I still have a photo copied out of the 1980s magazine RSTS Professional, which claimed to show how to convert thick to thin Ethernet. The simple answer is “with a coax connector adapter” since both are 50 ohm coax. The article instead used a thinwire T connector, with the terminator still on it. As Tony points out, terminators go at the end, not the middle (that’s what the word means). So that configuration would be somewhere between marginal and broken. paul
Re: UNIBUS extension card/cable sets
> From: tony duell > There are 3 40 pin Berg headers, one row of each appears to be ground Ah, hadn't noticed that! But then again, I hadn't looked at them closely yet! :-) Yes, they do connect to ground - all the UNIBUS ground pins are ganged together, and connected to the A-row Berg pins on all 3 connectors. So every other wire on the 40-conductor flat cables should be ground - that's even better than the classic BC11A, where almost every other line is, from what I can see, simply left floating (which is better than nothing, but not as good as grounding them, is my understanding). >> I was wondering if maybe the M9015 was an M9014 with termination >> resistors, or something (the way the QBUS versions come with and with >> termination) > I would be very surprised. Unibus is normally terminated at the ends > and not in the middle. Right, but the very similar QBUS does have terminations (of a sort - the rules for when you need terminations on QBUS extensions are so complex that I don't really grok them yet) 'in the middle', so... I just couldn't find out _anything_ about M9015's, so I was just guessing in the dark. Real data gratefully received. > From: John Wilson > I was kind of assuming that there's some impedance-matching (etc.) > problem with using ribbon cables for more than one hop. Hmm. Well, I dunno; that may be beyond my (minimal :-) level of analog expertise. I would have assumed that it's the _change_ from one impedance level to another that's the issue (you can get a reflection off the junction), so whether one's using long or short cables between a pair of M9014's, it shouldn't be _that_ big a deal (modulo propagation delays, which _are_ an issue with length). Perhaps someone else can opine? But I hope we can do fairly long runs with the 40-conductor (aka BC05L-xx), that could save us when we run out of BC11A's, if that strange Flexprint flat white cable the BC11A uses is no longer available. Noel PS: From what I can see so far (done half the pins), the M9014 and M9042 do have an identical pinout on the Berg connectors.
Re: UNIBUS extension card/cable sets
> On Jun 25, 2015, at 1:02 PM, Noel Chiappa wrote: > >> From: tony duell > > ... >>> I was wondering if maybe the M9015 was an M9014 with termination >>> resistors, or something (the way the QBUS versions come with and with >>> termination) > >> I would be very surprised. Unibus is normally terminated at the ends >> and not in the middle. > > Right, but the very similar QBUS does have terminations (of a sort - the > rules for when you need terminations on QBUS extensions are so complex that I > don't really grok them yet) 'in the middle', so… That doesn’t seem likely. While DEC engineers on average didn’t understand transmission lines as well as, say, Cray engineers, the definition of “termination” is something that you learn in EE 101 and are unlikely to forget. paul
RE: UNIBUS extension card/cable sets
[M9014] [One row on each Berg is ground] > Ah, hadn't noticed that! But then again, I hadn't looked at them closely > yet! :-) Yes, they do connect to ground - all the UNIBUS ground pins are > ganged together, and connected to the A-row Berg pins on all 3 connectors. I am not at all surprised they are grounded, it's the only logical place for them all to end up. > So every other wire on the 40-conductor flat cables should be ground - that's > even better than the classic BC11A, where almost every other line is, from > what I can see, simply left floating (which is better than nothing, but not > as good as grounding them, is my understanding). I am surprised. DEC didn't waste copper like that. It's been a long time since I worked on a BC11A, but I thought alternate wires were grounded. Maybe a track right along the edge of the PCB where the cable comes off (so you can't see it). I will check. > Right, but the very similar QBUS does have terminations (of a sort - the > rules for when you need terminations on QBUS extensions are so complex that I > don't really grok them yet) 'in the middle', so... I never really understood the Qbus rules either. The ones I saw make very little electrical sense (unlike Unibus, which is designed as a transmission line). I don't think I've ever taken Qbus outside one backplane, as a result just about anything will work. On Unibus too, you can generally get away with no terminator at the far end if you are in one backplane, or even one mouting box. Can be useful for testing... > Hmm. Well, I dunno; that may be beyond my (minimal :-) level of analog > expertise. I would have assumed that it's the _change_ from one impedance > level to another that's the issue (you can get a reflection off the > junction), so whether one's using long or short cables between a pair of > M9014's, it shouldn't be _that_ big a deal (modulo propagation delays, which > _are_ an issue with length). Perhaps someone else can opine? Well, Unibus is terminated into 180 Ohms and 390 Ohms, isn't it? The thevenin equivalent is thus around 123 Ohms. Most ribbon cables have a characteristic impedance when used with alternate wires grounded of around 100 Ohms (I seem to remember that is certainly right for the twist-n-flat ones). That's a small mismatch, but I don't think it is going to cause big problems. -tony
Re: organizing a trip to Cuba
On 2015-06-25 13:31, Liam Proven wrote: On 24 June 2015 at 14:19, Johnny Billquist wrote: Oh, I know. I'm from Sweden. We had a very big scandal where 5 containers with a VAX-11/782 and peripherials or something like that was found under strange circumstances. When the whole thing started to be investigated suddenly no one seemed to know or own those containers. The system was unclaimed for years, and it became a question of what to do with it, since no one seemed to claim it. I think it was eventually decided that since DEC made it, it was returned to them. The original shipping destination was of course somewhere in Soviet Union. This was in the early 80s... I'm sure someone can find the full story online somewhere. It's mentioned in the Datasaab article on Wikipedia: https://en.wikipedia.org/wiki/Datasaab Hmm. Some interesting information on that page. But anyway, no, the smuggling story in that article is not the one I was referring to. That story is about some Swedish computers for ATC (also usable for military purposes obviously) which contained some American components, for which Sweden did not have an export license, so they somehow got to the SU without permission, which was later revealed, and was not a nice story, since the Swedish government was involved. Search for "containeraffären" in Google (unfortunately I only manage to find Swedish texts about it, but Google translate is your friend). In short, a private businessman in Sweden was involved a the shipping of a VAX-11/782 and peripherials via South Africa and Sweden (maybe Switzerland was also involved). It was caught by the Swedish customs, and the stuff never reached the SU. I guess that sortof repaired the damage of the previous "Saab-affären" story in the eyes of the USA... Johnny
RE: Transmission lines Was: UNIBUS extension card/cable sets
> >> From: John Wilson > > > >> I chose the digital version of EE as my major precisely because I knew > >> I'd flunk Fields and Waves. Transmission lines are black magic as far > >> as I'm concerned! I really do fail to see how you can possibly understand or design digital systems without understanding analogue electronics, in particular transmission lines. FWIW the Motorola ECL databooks were pretty good at giving an introduction to this, simpy because with ECL you have to design just about every interconnection as a transmission line. Do it right and the design works, and unlike some other logic families where you hope for the best with the interconnections, ECL stays working. > I still have a photo copied out of the 1980s magazine RSTS Professional, > which claimed to show how to convert > thick to thin Ethernet. The simple answer is “with a coax connector adapter” > since both are 50 ohm coax. The Err, yes :-). The BNC-N adapter is very useful :-). More seriously, I've seen thickwire transceivers that had a pair of N connectors (not a beesting tap) fitted with BNC-N adapters and ues on thinwire. Technically that is wrong, there is a minor difference in the transceiver spec (I forget what, but the data sheet for at least one of the transceiver ICs pointed it out), but in will work. > article instead used a thinwire T connector, with the terminator still on it. > As Tony points out, terminators go at ARGH! Ethernet is more touchhy than most as IIRC the transmitter is a current source, the receiver effectively senses the voltage across the terminator. A collision is too high a voltage. So ethernet can't work with incorrect termination. That's why DEC had you put 2 terminators on a T piece on the ethernet BNC connector of a VAXstation (or whatever) to get it to pass diagnostics. WIth a network that short you are not going to get detectable reflection problems, but if you only had one terminator, every transmission would be a collision. -tony
Re: UNIBUS extension card/cable sets
> From: Paul Koning >> Right, but the very similar QBUS does have terminations (of a sort - >> the rules for when you need terminations on QBUS extensions are so >> complex that I don't really grok them yet) 'in the middle' > That doesn't seem likely. ... the definition of 'termination' is > something that you learn in EE 101 and are unlikely to forget. I agree with your sentiments, _but_ all I know is that if you look at, e.g. the 1982 'microcomputers and memories' handbook, pg. 251, you'l see a three-backplane system, with terminations at the start _and_ end of the first backplane (and also at the end of the third backplane). I've seen other similar diagrams elsewhere, that's merely the first one I ran across in a quick search for this post. Like I said, I cheerfully admit that I don't really fully understand what's going on there in the analog domain, but I've seen this in more than one place in the DEC QBUS multi-backplane configuration instructions. Noel
Re: Transmission lines Was: UNIBUS extension card/cable sets
> On Jun 25, 2015, at 1:49 PM, tony duell wrote: > ... > >> I still have a photo copied out of the 1980s magazine RSTS Professional, >> which claimed to show how to convert >> thick to thin Ethernet. The simple answer is “with a coax connector >> adapter” since both are 50 ohm coax. The > > Err, yes :-). The BNC-N adapter is very useful :-). More seriously, I've seen > thickwire transceivers that > had a pair of N connectors (not a beesting tap) fitted with BNC-N adapters > and ues on thinwire. Technically > that is wrong, there is a minor difference in the transceiver spec (I forget > what, but the data sheet for at > least one of the transceiver ICs pointed it out), but in will work. I see a slight difference in the input current spec of the receiver part of the transceiver. 250 uA instead of 25 uA. That makes some small amount of sense given the smaller station count. > >> article instead used a thinwire T connector, with the terminator still on >> it. As Tony points out, terminators go at > > ARGH! Ethernet is more touchhy than most as IIRC the transmitter is a current > source, the receiver effectively > senses the voltage across the terminator. A collision is too high a voltage. > So ethernet can't work with > incorrect termination. > > That's why DEC had you put 2 terminators on a T piece on the ethernet BNC > connector of a VAXstation (or > whatever) to get it to pass diagnostics. WIth a network that short you are > not going to get detectable > reflection problems, but if you only had one terminator, every transmission > would be a collision. Good point, collisions are not reflections, though reflections will be interpreted as collisions. Your comment about measuring voltages reminds me of a network monitoring device DEC built, and almost turned into a product. This was mainly for thickwire, though it would work on thinwire too. The intent was to map stations by their physical position on the cable. The approach was to put voltage measuring devices at both terminators, and record the observed voltage levels for packets from a given MAC address. The voltage ratios at the two ends would tell you where the station is, provided the coax is reasonably uniform. And the ranking of those ratios would pretty accurately show you the station order even if the cable isn’t all that uniform. I’m not sure why this wasn’t shipped. Perhaps it was around the time that structured wiring and star-based wiring with lots of repeaters started to come out, displacing the long bus topologies of the original Ethernet. paul
Re: UNIBUS extension card/cable sets
> On Jun 25, 2015, at 2:01 PM, Noel Chiappa wrote: > >> From: Paul Koning > >>> Right, but the very similar QBUS does have terminations (of a sort - >>> the rules for when you need terminations on QBUS extensions are so >>> complex that I don't really grok them yet) 'in the middle' > >> That doesn't seem likely. ... the definition of 'termination' is >> something that you learn in EE 101 and are unlikely to forget. > > I agree with your sentiments, _but_ all I know is that if you look at, > e.g. the 1982 'microcomputers and memories' handbook, pg. 251, you'l see a > three-backplane system, with terminations at the start _and_ end of the first > backplane (and also at the end of the third backplane). What’s between the backplanes? If it’s a repeater, that would be correct (because you have two separate transmission lines). If it’s a cable, then this is wrong. But as Tony said, it might be that it’s not so wrong that things break, thanks to the small scale and slow speed of Qbus. paul
Imaging TRS-80/III single-sided disks
Dear sirs, Imagedisk is my savior, and I image all kind of disks I know with it :) But now I got a pair of TRS-80 model III single-sided disks. How do I image it using imagedisk? Can I use a double-sided floppy drive to image it? Or do I need to put the single-sided drive on my PC? Please, help! :) Thanks Alexandre --- Enviado do meu Apple IIGS (pq eu sou chique) Meu site: http://www.tabalabs.com.br Meu blog: http://tabajara-labs.blogspot.com
Re: Imaging TRS-80/III single-sided disks
On Thu, 25 Jun 2015, Alexandre Souza wrote: Dear sirs, Imagedisk is my savior, and I image all kind of disks I know with it :) But now I got a pair of TRS-80 model III single-sided disks. How do I image it using imagedisk? Can I use a double-sided floppy drive to image it? Or do I need to put the single-sided drive on my PC? Please, help! :) IMD will image it just fine on a double-sided drive. To be sure, set IMD to single-sided before you start. Mike Loewen mloe...@cpumagic.scol.pa.us Old Technology http://q7.neurotica.com/Oldtech/
Re: Megaprocessor - built from individual transistors
On Thu, Jun 25, 2015 at 10:22 AM, Vincent Slyngstad < v.slyngs...@frontier.com> wrote: > Please correct me if I'm talking complete bulls**t. This is all >> new to me. >> > > The only other concern I've had during my thought experiments > along this line related again to the current/power involved. The > device is likely to become difficult to cool if you achieve a 4X > volume reduction. > > Oh, and I guess the nightmare of soldering many thousands of > surface mount discretes and transistors to manufacture such a > thing. > The easy way to do this is to make one board with a small CPLD connected to all the pins and program it to match the logic of the card you are emulating. That way you can have a couple of hundred boards made up for you at a reasonable cost that are all the same but ends up with different functionality. Just stick a label on the handle to tell what the card was programmed as. I don't think you would have heat issues unless you actually tried to use the original circuits. The Straight 8 logic uses pulses for some things instead of levels making it not really compatible with more modern logic. At least not a direct replacement. The later models would be better targets. An I for example might be the best choice. You can't shrink the size of the Front Panel because the switch spacing is already just a little narrow for people with large fingers. It would be better to make it a little oversize in fact. I don't have large hands and it would still be better for me if the spacing was slightly wider. -- Doug Ingraham PDP-8 SN 1175
Re: Imaging TRS-80/III single-sided disks
On Thu, 25 Jun 2015, Alexandre Souza wrote:
Dear sirs,
you'll have to settle for us.
Imagedisk is my savior, and I image all kind of disks I know with it :)
But now I got a pair of TRS-80 model III single-sided disks. How do I
image it using imagedisk?
Same as every other disk that you have imaged. However tell Imagedisk
that you only want side A. That shouldn't be a big deal, but many people
with single sided drives reFORMATed, without erasing what was on the
second side. Therefore, it is possible, if they were using used disks,
that side B might still contain side B from some previous use, such as PC
or Apple. And, some people using single sided drives punched extra holes
in the jacket to be able to flip the disk over ("FLIPPY disks") to use the
other side as if it were another disk. so, if you see an extra write
enable notch on the wrong edge of the disk, and/or amateur holes made to
make index hole access symmetrical, then the disk may have been flipped
over and written with a single sided drive. "Berkely MicroComputer
Flip-Jig" was the best jig ever made for marking diskettes for making
flippies.
The model 3 TRSDOS was 256 bytes per sector, with 18 sectors per track.
HOWEVER, besides TRSDOS, there were other operating systems that could be
run on it, and some of the CP/M conversions used other sector sizes.
Can I use a double-sided floppy drive to image it?
YES.
Although there have existed some drives where SS and DS were actually
incompatible, such as the difference in index hole placements on 8", Side
A of a Single Sided 5.25" disk is the same as Side A of a DS disk.
360K drives will work just fine.
NOTE: If you use a 1.2M drive, then you will encounter exactly the same
issues as when you use the 1.2M on a 360K PC disk.
MORE IMPORTANT NOTE: TRS-80 used a Western Digital FDC; IBM uses NEC.
(TYPE of FDC, disunirregardless of BRAND used)
TRS-80 starts the first sectors of the tracks earlier on the track than
IBM. You will sometimes encounter TRS-80 disks on which the first sector
of each track can't be found/read by the PC! That can actually be worked
around fairly easily. If you interrupt the index signal, then a disk can
still be read or written, but can not be formatted (and any errors will no
longer be properly understood by the BIOS).
You can interrupt the signal by making a special floppy cable without that
wire, or with a toggle switch on that wire, or maybe putting a tiny piece
of electrical tape on the card edge, so that that pin of the cable doesn't
connect. Will you remember to set things right afterwards?
Or, you could clip the wire of the index sensor (in a reversible way!!!).
OR, you can try just putting a small piece of opague tape (such as a write
protect tab!) over the index hole access hole of the disk jacket. On some
drives that are too smart for their own good, such as the otherwise nice
Teac 55, that will be misinterpreted as drive not ready! But it works
great on the original full height Tandons used by TRS-80 and by PC/5150.
Do not let your tape fall off into the drive.
Sometimes it may work to slow down the motor of the drive, and run it at
too slow a speed.
Or do I need to put the single-sided drive on my PC? Please, help! :)
NO.
In fact 160K/180K PC drives (the SS version used before DOs 2.0 provided
double sided capability) were often the SAME BRAND, SAME MODEL drives as
TRS-80 model 3.
The 360K DS drives (since DOS 2.0) were still the same, with the addition
of a second side.
Later, IBM switched to other brands and models of drives, such as "half
height", but they remained completely compatible.
WARNING: If you DO put the TRS-80 drive into a PC, you will need to change
some jumpers! The original normal/standard configuration for such drives
("SA400 interface") uses pin 10 for drive select of drive A, pin 12 for
drive select of drive B, pin 14 for drive select of drive C, and did not
provide for a fourth drive.
Radio Shack did not consider their customers, NOR THEIR STORE PERSONNEL,
to be competent to change jumpers of the drives for TRS-80 model 1. So,
they put in ALL of those jumpers, including using another "unused" pin for
drive D (which later drives used for SIDE SELECT!, and then they removed
the pins in the connector cable for each position so that each drive
connector of the cable only had one drive select pin connected. Then, the
drive would automagically be whichever drive that position of the cable
was for. Therefore, connecting a TRS-80 drive to a PC requires
changing the jumpers.
IBM did not consider their customers, NOR THEIR STORE PERSONNEL, to be
competent to change jumpers of the drives for PC/5150. So, they jumpered
every drive as if it were drive B (pin 12), and then they put a twist in
the cable so that pin 10 of the FDC board connected to pin 12 of drive A,
and pin 12 of the FDC board connected to pin 12 of drive B. That twisted
cable also made chang
Re: Megaprocessor - built from individual transistors
From: Doug Ingraham: Thursday, June 25, 2015 4:10 PM On Thu, Jun 25, 2015 at 10:22 AM, Vincent Slyngstad wrote: The only other concern I've had during my thought experiments along this line related again to the current/power involved. The device is likely to become difficult to cool if you achieve a 4X volume reduction. The easy way to do this is to make one board with a small CPLD connected to all the pins and program it to match the logic of the card you are emulating. That way you can have a couple of hundred boards made up for you at a reasonable cost that are all the same but ends up with different functionality. Just stick a label on the handle to tell what the card was programmed as. I've been thinking lately that whole Omnibus cards could be generalized that way. I don't think you would have heat issues unless you actually tried to use the original circuits. I had assumed, since the subject contains "built from individual transistors", that something akin to the original straight-8 circuits was what was being discussed. Straight 8 signals typically run 8 mA when low and 10 mA when high, and that's from the -15V supply, per signal. I estimate that's a few hundred watts for the whole thing. Reducing the supply voltage to -5V would presumably cut the power (and hence the heat) to a third, by eliminating most of the waste heat from all the pull-downs. The Straight 8 logic uses pulses for some things instead of levels making it not really compatible with more modern logic. At least not a direct replacement. The later models would be better targets. An I for example might be the best choice. I like the 8/I too, though it's not "built from individual transistors". You can't shrink the size of the Front Panel because the switch spacing is already just a little narrow for people with large fingers. It would be better to make it a little oversize in fact. I don't have large hands and it would still be better for me if the spacing was slightly wider. I completely agree, at least if you're actually going to use the darn thing a few hours at a time. Some models of the PDP-11 did use much smaller (still available) switches, but fitted them with honking paddles to make them more comfortable to use. Vince
Re: Imaging TRS-80/III single-sided disks
A few more gotchas to be aware of:
Some of the operating systems on TRS-80 started the sector numbering at 0.
Therefore, depending on which operating systems were used, you could have
sectors numbered 0 - 17 , or 1 - 18. That can be confusing if your PC
can't read the first sector, but tells you that it successfully read
sector #1.
Model 1 TRS-DOS is SINGLE DENSITY. IMD can handle that IFF the FDC in
your PC is one of the ones that supports FM/SD. NOTE: some/many? TRS-80
users installed additional hardware for double density, but the OS might
still require FM/SD for track 0. Disks written on a stock model 3 for
use in a model 1 are not quite the same format as model 1 disks.
Model 1 TRS-DOS uses some "NON-STANDARD" DAMs (Data Address Marks). I
don't know how to read those with PC FDC. But, they can be read with
Catweasel, OptionBoard, or other "flux transition" boards.
Or design and build a PC FDC board with a 1771 WD FDC and a WD 179x FDC
just to be able to handle disks with NEC/WD incompatabilities - NOTE: that
disk controller should ALSO have an NEC FDC, or you won't be able to have
full PC compatability.
Model 1 TRS-80 has 35 tracks. They started off using the Shugart SA-400
floppy drive, which is 35 track. I think that most/all? of their later
drives could do a full 40 tracks, BUT don't count on the software
supporting that. Expect 35 tracks for TRS-DOS, although most/all? of the
after-market operating systems supported 40 track, and some supported
double sided drives ("Gee, didn't expect THAT!"), and some users even used
80 cylinder drives, for 720K capacity. Yes, there were also hardware mods
for TRS-80 to use 8" disks. Yes, you can put 3", 3.25", or 3.5" drives on
a TRS-80.
Not all TRS-80 diskettes were of the highest quality. Be prepared to
clean the heads of your drive after you finish, and maybe periodically
part way through. Some TRS-80 users considered Wabash to be "premium"
diskettes.
--
Grumpy Ol' Fred ci...@xenosoft.com
Re: Northtop Flying Wing, inflight computing / was Re: XH558 - was Re: using new technology etc
>From memory, so please forgive a mistake or two: The TB-49 Wings would Yaw (side to side motion) while in flight, sometimes just enough to make the crew seasick, sometimes enough to be dangerous when in formation with other aircraft and always unable to stay on track to be a useful bomber. I recall someone saying the yaw was several wing spans in length in each direction. The autopilots of the time couldn't dampen it fast enough let alone keep it under control. Jack Northrup and his team knew they would have to wait for something both programmable, more data inputs and faster. there were about 4 or 6 piston engined, and 4 or 6 jet engine versions. Stored on the ramp at Ontario California airport for many years and then sold for salvage, I think in the late 60s or mid 70s. Jack Northrup continued to be enthusiastic about the tail-less design even in retirement. Much later, Jack was given a vip tour of the secret B2 factory and presented with a model of the design in Lancaster CA before his death.Im pretty sure there is a book on this, perhaps from the Smithsonian Air and Space Museum. On Fri, Jun 19, 2015 at 2:40 PM, Brent Hilpert wrote: > On 2015-Jun-19, at 9:07 AM, Christian Gauger-Cosgrove wrote: > > > > Bringing this topic full circle, does anyone know if any minicomputers > > (DEC PDP-8s or 11s, DG Novæ, HP 21XXs, et cetera) were ever used on > > aircraft? Not transported by one, but I mean setup and used on one. > > > On 2015-Jun-19, at 12:09 PM, Toby Thain wrote: > > On 2015-06-19 3:05 PM, geneb wrote: > >> On Fri, 19 Jun 2015, Toby Thain wrote: > >>> > >>> "in 1949 the Air Force ordered all the flying wings destroyed, all > >>> the jigs and tools destroyed, every trace of the flying wing > >>> eradicated. A few years later even the engineering drawings were all > >>> destroyed by new Northrop management." > >>> > >> I don't know why they went to those lengths, but it's my understanding > >> that the program was cancelled because at the time, the USAAF (USAF?) > >> mandated stall testing as part of their development programs. Without > >> serious flight control computers, stalling a flying wing just ends up in > >> a freshly planted aluminum tree. Even WITH good computers, stalling a > >> flying wing is a Bad Idea(tm). AFAIK, the B-2 has never been stalled > >> (on purpose), even during development. > > > > Thanks. I knew there must be more to it... I wonder if the cited book > covers this angle. > > > > To tie these two lines of question together (and bring it back very much > on-topic), the BINAC (amongst the first stored-program computers, 1949) > was supplied to Northrop for research into airborne flight control (quick > web search says part of the Snark missile project), > > I'm not suggesting the BINAC and YB-49 (the flying wing) were connected, > but it's interesting they were contemporary projects both at Northrop, and > computer control was just what the flying wing needed. > >
