2DRIVE.THD --- Copyright 1988 by Phil Wheeler An original compilation of Compuserve Model 100 Forum messages for use by Forum members only. The file 2DRIVE.TDD in Lib 9 discusses developing a dual-TDD system for the Tandy computers. This set of messages discusses the implementation of such a system, and results from the first users. Additional THD files on this topic are expected! Message range: 175332 to 175547 Dates: 10/8/88 to 10/14/88 Sb: #Two Drives Online! Fm: bob scott 73125,1437 To: SYSOP Tony 76703,4062 Built the two drive interface. Chip version was too complex and flaky. Switched to micro relays. Device is the size of a cigarette pack and runs off a plug in xformer. Total cost was about $25 at RS. Since TS-DOS/Random won't allow raw sector access to non RANDOM disks, I've DL'ed POWRDOS to generate a disk to disk copy utility that can occur untouched by human hands....(faint cheer). Will keep you posted if you're interested. Bob P.S. only need to switch six lines! Fm: Tony Anderson 76703,4062 To: bob scott 73125,1437 Congratulations on getting it to work before I did. I'm still waiting for my second drive to arrive... it's in the UPS pipeline. (grin) Since you mention POWR-DOS, it appears you're using a 100/102... you're lucky in that respect; most of the program you need for disk-to-disk copy is already available; just some modification to be made. I'm doing it on a 200, using POWR-DISK (200), the predecessor to POWR-DOS, and less capable. It can still be done, and I'm only waiting for the second disk drive to arrive to test it out. I built mine in a 7x11 metal chassis, so the two drives sit on top of it, side by side, making a neat dual-drive setup. Also used small relays and external power pack. I had many of the parts on hand. You will also find that with a little judicious work on your part, you can use the relay interface with an external modem and disk drive to simulate download to disk capability. Connect modem to port #1, and drive to #2. (#1 is relays off, #2 is relays on) Download a file to RAM, jump into BASIC, witch the relays, use a BASIC program to copy the file to disk and kill it from RAM, switch the relays back, then jump back into TELCOM for more. You can even download large files (larger than RAM) by using ^S to stop somewhere in the middle, save the part you have to disk, get the second part (^Q), and APPEND it to the first part already on disk. External modem needs to hold the connection while you are off in BASIC doing all the disk activity. Why don't you outline your construction and operating technique, and give us a file on the subject. 2DRIVE.001 would be a good filename. Fm: bob scott 73125,1437 To: Tony Anderson 76703,4062 Righto! I used mine for something similar: switching between a serial printer and a PDD. Looks like this is going to be a fun toy. Just got POWR-DOS up and running (quite a hhassle makingit get along with MENU and RANDOM in ROM). Any it just finished it's first disk-to-disk backup and it worked! Bad part is that it took about 25 minutes. random sector writes were SLOOWWWW. I'm going to DL the POWR-copier and see if I can modify that and speed the process up a mite. The everything in one box setup is a good idea. Raises the drives up. I've got mine stacked right now. The version I've got is hreally more an outgrowth of the breadboard test version (didn't really expect it to work!) and I'd do a finished version in a much neater fashion. I think I'd try to get some of the DIP header's that RS uses for the drive end of the cable to make up special drive only cables, and have parallel DB-25's for modem or printter/disk use (too much wire on the desk now). Also think I'd find some DPDT MAG-Latch relays to allow powering from the M100 (only need juice to switch states). But this doesn't look too bad,and it works, so you can guess how far down the road MOD 0.1B is! Stay tuned. Bob. Fm: Tony Anderson 76703,4062 To: bob scott 73125,1437 I got a couple of latching mini-relays from Peter St. Clair; one of our early, and long lamented hardware hacking geniuses here. They're like 8-pole, double throw, latching. He had a circuit made up with several of the relays and an "addressable" line monitor. You could send a command down the line, and the monitor would select which relay you wanted; then you issue a close or open command, and it would act only on the selected relay. Worked real well. Unfortunately, Peter was not real good on writing down his developments, and when he moved on to a more sophisticated computer, most of his know-how moved on, with him. Problem with using latching relays, is you need a flip/flop toggle of some sort to select the "other" coil, when the relay is in (whichever) normal state. Having been unable to find or develop such a circuit over the years is what's prevented me from documenting the hardware approach to multiple RS-232 ports. The "drive-only" cables is a good idea, if you can figure the cable out. The connector schematic is in the manual (for the TDD-1), but I had difficulty testing a complete circuit between both ends of the standard cable when I tried to figure that out myself. It would make a "neater" semi-permanent setup. My original idea was "stacking" the controller and two drives. But I opted for the horizontal approach for a couple of reasons; 1) I could front-mount a couple of LED's, one under each drive, which would indicate which drive was "selected", and 2) I couldn't find a box that was the right size to use in a vertically stacked configuration. While I have 8 R-S stores up here, only one non-RS store, and it's not well stocked. I'd have had to mail-order something. I'd like to have used a plastic box, about the same size as the TDD, only about half as high. That would have made a neat package. It also occurred to me to use a 17 inch chassis, and include a cassette machine in the package. - Haven't done that yet, but it's a possibility. Fm: bob scott 73125,1437 To: Tony Anderson 76703,4062 Tony: I think if this thing starts looking like a permanent installation I'll redo it along the lines you've outlined. For permanent drive cables, I was thinking of just ordering two TDD cables from Natl Parts and hacking them up. Also solves the controller to drive cable problem (twice!). Anyway, the latest flash is that POWR-COPY was pretty easy to modify and works MUCH quicker than the sector approach. The latter may be the hot ticket for copying odd disks (like SARDINE dict.) though. Full disk to disk copy took a bit over 9 minutes with an essentially empty RAM a bank (0MENU, POWR-DOS and theCopy pgm). Needless to say in those nine minutes I was free to fly to Europe if I wished. Nothing like the happy clicking of a drive controller. I'm going to try a mass backup of my disks this weekend using this approach and if it works out, I'll upload a software/hardware description to DL9. Good point on the latching relays. I've never used one and sort of thought it was like a D flip flop. Seperatecoils would be a real bother. I initially tried using tri-state buffers (since the TDD uses TTL levels out and relies on the M100 to pull the 0 volts to -5) but never got solid results. Might work if I had added RS-232 line drivers to the circuit but then you've got 4 more chips, 3 power supplies...Yuk! 'Lot to be said for relays.... Bob Fm: Tony Anderson 76703,4062 To: bob scott 73125,1437 That's a good idea... getting two replacement cables from Nat'l Parts and hacking them to fit. That also saves the original cables for portable use when you want them. And speaking of permanent installation; what do you think about adding a larger power supply, maybe 12 volts at 1.5 amps (readily available), and stepping it down, via 7806 regulators, to power the two TDD's when used in this mode? One could save a lot of plug space that way. I'll bet the clicking relays, and whirring disks are a sound to be heard and appreciated! I'm envious that you got there first. Indeed, modification of POWR-DOS's COPY.PWR program was the logical way to do it for 100/102 users. A bit more complex for the 200 users since COPY isn't available in a 200 version, and Power-DISK doesn't support sector access. (Documented support anyway) While 9 minutes is SUPER, even 25 isn't bad for a completely automated disk backup. I'd be happy with that. NO DISK SWAPS!!!! I wish latching relays were that simple; but most of them have a "latch" and an "unlatch" coil. Meaning a single pulse line can't switch it to the alternate state. That would be ideal. I've never been able to find that, though. The nearest I've seen is a pulsed, stable, flip-flop, with the relay tied to one of the output lines. I haven't found a reliable circuit, though. In theory it's possible, but I just haven't found one. In a "chip" approach, I was thinking of 4606 bilateral switches. Each chip has four, so you'd need four chips to handle on/off - off/on for 8 lines to 2 devices. Problem is, you'd still need a relay to switch them to on/off states. Another problem, with only +/- 5 volts in the line, internal resistance of a couple hundred ohms may present a significant voltage drop. -- May not work. Fm: bob scott 73125,1437 To: Tony Anderson 76703,4062 I like your idea on the "mothership" cabinet. My thought would be to put in a small 12 xformer and diode bridge to run the relays and a slightly heavier 6v unit with filtering and voltage regulation to run the drives and the M100. A power switch on the front would kill the whole works. I'd like to keep the relay power separate just to avoid a problem with back EMF from the relay coils frying something in the drives or M100, queching diodes aside. This would make a nice desktop unit and would need (faint but sincere cheer) only ONE outlet. Then I could leave my AC adapter(s) in my travel bag. Normal ops would likely be the TDD in bay one, and a 2400 bps modem in bay two. Next step is to take a varient of your telcom idea nad modify XMDPWx to optionally DL to a buffer as big as free memory for both text and protocol xfers, then ^S the host and APPEND the file to disk. Downloading a 100K file would be a snap. Incidentally, I'm only switching DB-25 pins 2,3,4,5,6 and 20. Pin 7 is tied common and all others are left hanging. I looked at CMOS switches, only problem I see is that you'd need a bipolar 6v supply to run them (input voltage to the switches must stay between Vss and Vdd) but that's not impossible. My attempt using bus transceivers was pretty much a disaster so I'd be interested at hearing what you find out about CMOS if you try it. The ON resistance shouldn't be a big deal, with 12v between Vss and Vdd it will only be about 100 ohms (for the 4066) and the current flow is pretty low. Of course if you're modem or printer puts out a full +/- 12 volts..yikes! Were I to fool with chips again, I think I'd just bite the bullet and put in RS-232 line drivers and tristate bus tranceivers. The relay approach seems to work, however. Only problem is the settle time but a FOR J=1to10: NEXT: or a BEEP seems to be enough time. Since part of the speed problem is the cassette relay, I don't know that chips would be all that much faster anyhow. So, what do you think of patching SARDINE to switch MOTOR ON when executed and motor off when done? Full time online spell checker for TXT? Bob Fm: Tony Anderson 76703,4062 To: bob scott 73125,1437 You raise many interesting ideas for this relay controller. I've been talking about this project for some time; since folks found out they couldn't download directly to the TDD disk. But it looked like the stumbling block was that many folks did not want to upgrade to an external modem. It's just recently that I started thinking of it in terms of a 2-drive system. (I HATE disk swapping!) Well, I'd accept two power supplies, built-in; one for relay control, one for the drives. And the idea of a modified Xmodem/Terminal program to handle a buffered download isn't too hard to handle. (Boy, wouldn't that be something!) However, don't forget the 64K file size limit. You can't download 100K into a single file. (grin) I believe the 4066 will work from a single-ended supply... I'd have to go back and check some manuals. Wasn't aware that they would only pass voltages lower than Vdd. Maybe that's another situation where relays are still superior. -- I knew we would need a delay for settling time, and assumed it would be about 1/2 second, which is redily available as a CALL in the TELCOM dialing routines. I suppose SARDINE could be patched, but not having ever received my promised copy, I wouldn't be able to explore the concept myself. Technically, any program could be patched to switch the relays; either by BASIC statement or ROM CALL. And of course, most folks who need spelling checkers have moved on to the multi-ROM version; which seems to be superior. But _any_ program can be set up to switch the relays, where it would be useful. Fm: Tony Anderson 76703,4062 To: bob scott 73125,1437 One little thing I noted in your message when reading it offline... a potential problem if you plan on running both the drives and the M100 from a common power supply with an on/off switch. ... If the AC supply plug is installed in the external supply jack on the computer, and you switch off the external power, it's like you've gone into a power failure condition, and the CMOS memory will start to draw power from the internal nicad. This may not be a problem, unless you have an older M100, where the nicad stack is no longer able to hold a charge, or where it's not functional at all. A condition that might well be masked if you regularly use AC and fresh batteries. So... in this case, I'd suggest your power switch only interrupt supply voltage to the relays and drives, but leave power being supplied to the computer. ... Or, install the "Uninterruptable Power Supply" modification in the file UPS100.200 in Lib. 8. Fm: bob scott 73125,1437 To: Tony Anderson 76703,4062 Good point on the 64k. Let's say "100k down in an unspecified number of files" . I'm ready to rip into Mod 0.2, but the thought of slicing my hands to ribbons trying to chain drill or nibble a db-25 hole (or two) in a cabinet isn't appealing. Gotta check around for a prepunched box, or maybe the mat shop a work has a punch (fat chance as we're just getting rid of paper tape in the Navy!). As you say, thhe external modem business is looking up though. One thing I was sorta surprised at is that you don 't need a bunch of p\ull up/down resistors on the drives to keep them from going screwy when you switch. Seems to work fine the way it is. Was thinking that a DTR modem disconnect could be preserved with a high value (100k?) pull resistor which held the carrier when the relay switched to the drive. The value would have to be hig enough that the M100 could over power it to bring down though. (Just a thought as I'd likely strap the modem too ignore the DTR and use S/W disconnect.) Bob Fm: bob scott 73125,1437 To: Tony Anderson 76703,4062 Yeah I was thinking about that. On the one hand, I suppose it's good to execcise the little critter (it's supposed to hold for a couple of weeks or something with 32k, right?) On the other hand....Might be better to switch just the relay and drive power supplies, eh? Bob Fm: Tony Anderson 76703,4062 To: bob scott 73125,1437 Well, the time has been quoted variously between 8 days and a month. I'd suspect that since the nicad is constantly float-charged, it has very little memory capacity, and probably a matter of hours is more like it, in practical terms. I think I'd opt for switching the relay and drive supplies only. One neat thing there is that powering down, will always power back up in the default condition selecting drive #1. (or whatever is connected as device #1 - relays off.) Fm: Tony Anderson 76703,4062 To: bob scott 73125,1437 Since a DB-25 chassis punch sells for about $125, I've found a Dremel tool to be the easiest way to cut those little slots for such plugs/sockets. Another thing I thought of today, is that if you include both the DB-25 sockets and the dedicated drive cable, as we discussed, if you have two computers, like I do, you can connect to the second drive directly through the DB-25 socket from the second computer. i.e. one drive per computer, with the additional advantage of the disk-disk copy capability. I'd have to look at the pull up/down resistors to see what's happening there, but don't look a gift horse in the mouth... if it works, don't fix it. (grin) I have no difficulty with my external modem, it holds the line even when physically disconnecting it from the port. Some will, some won't. But there is probably a way to make all of them owrk that way. I don't know if the 100K pull resistor would work... that seems sort of high for a pull. I'm used to seeing 1K or 1.5K. A DTR modification would be the most reliable. Fm: bob scott 73125,1437 To: Tony Anderson 76703,4062 Yeah, I was just sort of whistling in thedark, there. Actually most modems have a switch that will cause them to ignore DTR so it's really no big deal. Hmm....$62.50 per hole. Might take a while to amortize the cost of a DB-25 punch unless I start banging these things out for profit! Still haven't turned one up, and the thought of showering my self with steel filings is still not too attractive. I can usually talk myself into one such undertaking, but it's the second go round that I hate; panel opening one is _always_ better looking! Bob Fm: Tony Anderson 76703,4062 To: bob scott 73125,1437 If you really want one of those DB-25 chassis punches, I can look up the address of the suppler and let you know. Maybe you could "rent it out". (grin) Fm: bob scott 73125,1437 To: Tony Anderson 76703,4062 Uh, yeah. "Well I tried this damn thing out, and it just don't work!" There's a big installation goat-rope going onat my work site, so I'm hoping that one of the contractors has one languishing at the bottom of a bag. Onegood thing about procrastination is I had yet nother idea. Power outages seem to be pretty popular here in PR. I'm thinking of making a mini-ups using 5C or D cells and a diode and building that into the box as well. I started to get into ni-cads or lead-acid and recharger circuits but then I realized that the 4 alkaline D's I bought 16 months ago to run my M100 when it was both away from AAC and near the travel bag haveyet to expire. so every couple of years I'll go wild and buy four new batteries for the thing. Ought to keep the box from sliding about on the desk, as well. Maybe even give me an excuse to boost the LED/$ ratio of this thing (a precious figure of merit for me). Bob Fm: Tony Anderson 76703,4062 To: bob scott 73125,1437 Well, that's a good idea. The UPS part. (grin) One of the things we often overlook, but which has been proven a real winner over the years, is a common 6-volt lantern battery. They're documented to run over 300 hours, full-tilt; and might be ideal as a UPS supply, since they have a long shelf life. What changes did you make in the Power-DOS COPY program to make it handle the relay switching? Where at? What line numbers?