Important Note
All modifications are COMPLETELY at your risk. I have not the slightest idea if any of them work. If you decide do to any of this stuff, and it damages your hardware, that's on you. Proceed at your own risk.
All modifications are COMPLETELY at your risk. I have not the slightest idea if any of them work. If you decide do to any of this stuff, and it damages your hardware, that's on you. Proceed at your own risk.
Memory Related Modifications:
Storage Related Modifications:
Video Related Modifications:
Other Modifications:
Other:
This details the conversion, and installation, of a Model III Hires board, Cat # 26-1125, to a Model 4 computer.
Ted Haigh - Radio Shack technical bulletin - April 25, 1984
First, Modify the Model 4
Next, Modify the Graphics Board
Next, Install the Graphics Board
Next, Test The Board
The Problem
Fairly often I have been asked, "Gosh, why won't my hard disk and hardware clock work together? The hardware vendors say they should 'cause they are addressed differently, but they don't."
The Cause
On the Model 3 and 4, there is a signal line that is used to reverse the data bus on the external I/O bus so that data can be input to the computer. Normally, the data bus only "points out". When an external device "realizes" that it has been addressed, it can pull this line low to send data to the computer.
There are two basic methods by which the external device can control the lead. First, a "normal" TTL or LS-TTL output. This is the one that causes the problem. Second, an "Open Collector" output, which is the correct way.
Open collector outputs know how to share. Any device can pull the lead low without conflicting with any other device on the bus. Each may pull the line low, or do *nothing* to the lead.
"Normal" TTL signals, on the other hand, may pull the lead low, as above. The problem is that when a "normal" output is not pulling the lead low, it is pushing the lead high, as opposed to just leaving it alone. Now, when there is more than one device in the system, and one of the devices becomes active, it tries to pull the lead low while the rest try to push it high. The result is "contention" and somebody wins. Who wins is up for grabs. If the right party wins, the whole system may seem to work... for now. As the temperature of the components change, the system ages, etc., this can change.
If one or more devices in the system are open collector, and some are normal TTL, then some combinations of devices may work and others won't. If only one is TTL and the rest are open collector, the system may seem to work but that TTL device is eventually going to burn out.
A technically competent person armed with the schematic for the device in question should understand the problem if given this whole description. They should then be able to look on the schematic for the device and tell whether or not an open collector output was used. If all devices pass this test, then they may all be used together without conflict as long as they are addressed differently.
The Solution
A 7405 open collector hex inverter can be used to buffer or replace the output device used. Note that the 7405 will invert the logic of the signal, so you must either pick off the signal at a point that it is inverted to begin with, or run the signal through two sections of the 7405 with a 2.2K pull-up resistor on the output of the first section. The pull-up for the final output is already in the Model 3 or 4.
The 7405 is a fourteen pin chip, and may be piggy-backed to any existing fourteen pin chip in the device that uses pins 7 and 14 for ground and power. All leads except pins 7 and 14 are bent ninety degrees so that they point straight out. It is then placed over the "host" chip and pins 7 and 14 are soldered to the corresponding pins on the host chip. Wire wrap wire may be used to connect the 7405 inputs and outputs as necessary. If two sections of the 7405 are to be used (if the inverted sense of the output signal is not available), the pull-up resistor may be run from the output of the first section to pin 14 on the chip.
One trace cut should be necessary, to separate the output lead from its existing driver chip.
Disclaimer
As they say on TV, "This kind of stuff is best attempted by experienced people taking all the necessary safety precautions." If the above explanation and procedure did not make perfect sense to you, I recommend that you do *not* try this yourself. Find a local hardware hacker to help you, or take this explanation and the matter up with your various hardware vendors.
Background
Here's a simple modification to speed-up some early Model 4s. This does not work on all Model 4's, but it may work on yours. There were a few variations of the Model 4 circuit board, but I've heard that this works on MOST of the early boards. I haven't tried it, so do it at your own risk! If you're successful, you'll eliminate the wait states and your Model 4 will run faster. This speed difference is very noticeable with Orchestra-90.
How To Do It
Conclusion
That's all there is to it! This supposedly eliminates the wait-states and gives you a "true" 4 Mhz. Perhaps it will work on your board. Larry Payne did this to his system and said that it worked perfectly, and I guess you could call him somewhat of an expert on Model 4's!
That's all I know about this little trick. Perhaps some others can tell you more about it.
Additional Comments from Ray Pelzer
On the later machines (like MINE), you can pull out the Z-80 itself, and bend up pin 27. That cuts it off right at the final in-point.
For the earlier 4P's with the white screen
For the Gate Array's with the green screen.
Don't rely on page 132 of the 4P Service Manual #26-1080 for the correct IC labels because there are a few typo errors. Change these labels on page 132 before doing the work:
The following pin configuration will effectively translate the 34 pin cable required by the 5" disk drive to the 50 pin cable required by an 8 inch disk drive. In order to run eight inch drives, however, you will need to install a disk controller capable of eight inch operation. Homes Engineering and Micro Mainframe sell such disk controllers.
34 pin connection 50 pin connection pin 1...............pin 17 pin 2...............pin 22 pin 3...............pin 19 pin 4...............Cut pin 5...............pin 21 pin 6...............pin 32 pin 7...............pin 23 pin 8...............pin 20 pin 9...............pin 25 pin 10..............pin 26 pin 11..............pin 27 pin 12..............pin 28 pin 13..............pin 29 pin 14..............pin 30 pin 15..............pin 31 pin 16..............pin 18 pin 17..............pin 33 pin 18..............pin 34 pin 19..............pin 35 pin 20..............pin 36 pin 21..............pin 37 pin 22..............pin 38 pin 23..............pin 39 pin 24..............pin 40 pin 25..............pin 41 pin 26..............pin 42 pin 27..............pin 43 pin 28..............pin 44 pin 29..............pin 45 pin 30..............pin 46 pin 31..............pin 47 pin 32..............pin 14 pin 33..............pin 49 pin 34..............Cut
Much ado has been made of problems with the solder joints on the power supply boards where the output pins are located. But it's very easy to overlook the harness pin connectors!
In the 4P, the p/s connector is well out of the main flow of ventilation. Of course, it shouldn't heat up at all, but a bad connection produces resistance and will cause heat to be generated. This tends to weaken the spring tension of the female connector, the weakened tension causes an even poorer connection, resistance increases, and the problem worsens. You go in, look for the problem, wiggle the wires and the problem disappears. And because the 12v video connector pins are different than the others, they are especially susceptible. You can't determine the problem without actually removing the pins from the plastic connector housing.
After this problem became a nuisance on BOTH my 4Ps about a month ago, I really HAD to do something! Eventually I ended up removing the pins, cleaning them with pink pearl and alcohol, then bending them slightly with a small needlenose for more tension. Problems have completely disappeared in both machines.
If you happen to have the non-gate array mother board in your 4/4P and the video screen is "shakey" in the model 4 mode or speed and is NOT while in the model 3 mode, the video VCO (voltage controlled oscillator) which generates the 80 character by 26 line signal is probably on the edge of stability. This sometimes shows up more as the 'puter warms up.
On the 4P: in the front of the board, between U148 and U149 there is a variable capacitor (usually orange in color) designated C231, but you probably can't read the designation. It is about 3/8 inch in diameter, and has a whole in the top with what looks like a screw slot in the hole, you are looking at what is probably the culprit!! Using an insulated shaft TUNING tool that will fit into the hole, with the 'puter in the model 4 SYSTEM (FAST) and the 'puter warmed up so the symptoms appear, adjust (just a WEE TAD) so the video is again stable. Do this gently or you'll hate me!
On the model 4 the technique is the same, but the location is different. The capacitor is designated as C210 and located in the upper left quadrant of the board near the FDC conector, very close to U5.
There is the possibility that something else may cause the problem, but this usually accomplishes the desired result: a good working computer.
The USRobotics Courier 2400 Modem uses pin 12 of the RS-232C to indicate 'High Speed' to the computer. When it is set at the 2400 baud rate, (default), the HS LED in the front goes on and pin 12 of the RS-232C interface goes high.
The RS-232C interface of the TRS-80 Model 4 does not recognize pin 12 at all: it is connected to ground.
When the Courier sends the HS signal, the TRS-80 Model 4 sends it to ground, pin 12 goes low, and the HS LED does not go on. However, transmission is not affected.
An easy cure to have the HS LED function properly is to remove the wire that connects pin 12 at either end of the cable.
I bought an ATARI 400 and quickly got sick of that damn monopanel keyboard. Even with communications gear, it was a hit or miss proposition getting copies of programs off CompuServe, because with the small memory I must use unsophisticated terminal software. Little errors in file transfer were causing big problems.
To overcome that I tried to find a way to use my TRS-80 Model III for the download portion, because it supports the CompuServe A protocol, has gobs of memory, and runs circles around the 400 except in the areas of sound and graphics, where the Atari makes the TRS look mighty poor. I assume similar transfer between an ATARI 800 and Model III is possible too, but I doubt there is any compelling need.
So, I made up a cable to go from port #1 on the 850 interface to a black box, which looks fancy, but merely provides a place to hang two DB 25 connectors and some wire.
Here is how to wire two female DB 25's together, and how to make the cable from the 850 to the black box.
The process itself was not difficult, but I did encounter a couple of 'pitfalls' that resulted in not getting a working computer back until Tuesday evening (yes, I was getting worried!).
To start at the beginning, when I visited the tent sale I noticed a large display of cassette based Model 4's on sale for $50.00. After picking out one that seemed to have all of its parts intact and lugging it into the store and plugging it in to see that it brought up ROM BASIC ok as evidenced by the familiar 'Cassette?', 'Memory?' and 'Ready' prompts, I paid my money and took my prize home.
My first thought was to simply install the Model 4 motherboard and keyboard in my Model III. After some consultation and reflection, however, I decided it would make more sense to move the drives, drive controller, and RS232 board from the 3 to the 4. This offered the benefit of taking advantage of the newer electronics in the 4 and it would LOOK like a 4. With this plan in mind, I disassembled both machines and set about the task of removing and installing parts. The first difficulty I encountered was the fact that the ribbon cables that connect the floppy drive controller board and the RS232 board to the motherboard were different for the two machines. On the III, the connectors at each end of the cable are exposed on the same side, while the 4 required that the connectors be exposed on opposite sides. This is a function of a difference in the plugs on the motherboard.
The plugs on the III have teeth on both sides of the plug, while the 4 plugs had teeth on only one side. The original cables are also a bit flimsy and don't lend themselves to successful modification. This meant putting everything away for the night with plans to call National Parts the next morning. The young lady I spoke with at National Parts was quite helpful and looked up the part number for me. She explained that they could ship the new cables that day and I would have them in two or three days. I explained to her that she was speaking with a desperate man and asked her to try to visualize my den floor with parts from two computers scattered about. She very kindly arranged for me to pick up the cables that afternoon at the Northside Drive outlet store (next door to National Parts). After picking up the cables (about $7.00), I headed home confident that my 'new' Model 4 would be soon be up and running!
Here's where the REAL problem began! After reassembling everything and double checking all the connections, I plugged in the 4 and hit the power switch. No smoke, but no drives either! The Model 4 simply refused to recognize that the drives were there! I disassembled EVERYTHING and put it all back together slowly, rechecking every connection as I went - with the same result! The machine would come up in cassette basic just fine, but the drives wouldn't spin. More phone calls and consultations. Finally, it dawned on me that I had forgotten one of the basic rule of electronics (though I had been warned): Make sure nothing is grounded where it shouldn't be!
The culprit in this case was the floppy drive controller board! It didn't have the insulators on the back. The mounts on the Model 4 chassis are slightly different from their counterparts on the Model III chassis - enough different to cause a ground (short). I fabricated suitable insulators, reinstalled everything and the 4 booted just fine with drive 0 spinning and the 'Diskette' message appearing on the screen.
I did have to make one more trip, this time to the auto parts store to get a supply of wire ties before putting everything back together (those wire bundles looked like they could be a real problem if not securely fastened).
Would I do it again? You bet, but I'd do it a little differently of course as a result of what I learned in the process. The end result is very much worth the investment of $57.00 and the time required to swap parts. I've only been using the 4 for a few days now, but I LIKE it! From a hardware standpoint, the Model 4 keyboard is greatly improved over the 3 and of course the 80 by 24 display is much more convenient. Model 4 software (public domain) seems to be much more plentiful as well and you can still run Model III software in the Model 4's 'Model III' mode.
If you decide to tackle such a project, I've included some tips below that you might find helpful.
BUT ....
According to page 5 in the hardware section of the manual, a motherboard configured for 16K operation (jumper positions and capacitors) uses three voltages (+12, +5, -5) for 16K RAM Chips while 64K rams require only +5. When power is applied to the replacement motherboard with the jumpers and capacitors positioned for 16K operation, the 64K rams will get 'zapped' when hit with +12 volts. Modify the board to 64K operation before using non-16K chips or you WILL lose all your RAM chips!
Model III - PDP11 Hacker
If you are stuck for a keyswitch, move the bad ones onto the numeric pad, and use the ones from there (which tend to have little wear) onto the main part of the keyboard. That gets you 12 good switches.
Do NOT use a spray cleaner, or attempt to pop off the keys like on a Model 1. These solutions do not work for a Model 3.
This is a text file on how to put the SmartWatch Clock board in the Model 4 with a Hi-Res graphics board. First you have to purchase the SmartWatch from your local Radio Shack store. Stock number 25-1033. Also required is the program "SMTWATCH/ARC" found on this board. Unarc SMTWATCH/ARC and follow the instructions given therein. If you have a Hi-Res board for your Model 4, you must follow these added instructions as well.
IF YOU ARE A NOVICE, PLEASE GET ASSISTANCE FOR THIS!!
First, Remove the screws on the bottom of the Model 4 and remove the top, unplug the large cable and the ground cable from the monitor and set the monitor (top part of computer) aside out of the way. Next, remove the shield from the back of the computer and set it aside, Next remove the Hi-Res board to get to the "C" ROM chip. It should be U70 on the motherBoard, now CAREFULLY remove the mother (large) board that the Hi-Res board plugged into. You have to remove the cables on the sides of the board plus the ones along the top. To remove the cables on top, grab hold of the plastic cable in the center and pull gently toward yourself. They should come out with little effort. Now remove the cables on the top right of the main board in the same manner, except you grab them on the sides of the connectors (never pull on the wires!). You also have to remove the cable that goes to the sound board, it will come off like the ones on the right. The last cable you have to remove is the keyboard cable, remove the 2 screws that hold the shield around the connector and pull it off.
Now, with all connecting cables removed, remove the Phillips-Head screws that hold the board in place. Once this isdone, Lift the board out VERY CAREFULLY! Set the board on a not static surface and remove ROM C from the board by prying on BOTH ends carefully (one side at a time or both sides at the same time). Put the ROM C chip in the black foam-rubber that the Smart Watch is packaged in (it is a special anti-static foam rubber). Now get out your soldering gun (small) and solder-sucking tool. Carefully remove the empty I.C. socket that the ROM C chip WAS in. When you have it out and all the holes cleared out of excess solder, CAREFULLY solder pin 26 to pin 28 on the SmartWatch unit. Now clip off pins 1-2 and bend pin 28 out to the side. Clip off pin 27 and insert the now modified SmartWatch into the socket where the ROM C socket WAS at (make sure to get pin # 3 on the SmartWatch into Pin #1's place on the board ).
Now solder the SmartWatch in place (very carefully, that is!!). Double check all connections as some are also soldered in on top of the board (it is a DOUBLED SIDE board). If you do a good job of flowing the solder around the connections, it will solder both sides at one time. Now check your work for solder bridges VERY CAREFULLY, if no solder bridges can be found, Plug in the ROM C I.C. into the top of the SmartWatch. Now you have to make a slight modification to the case back (monitor housing). The screw plate in the center of the case MUST be removed (has a black metal piece with a hole in the center) Make sure you locate the right bracket first, it is the one ALMOST in the exact center of the back of the computer, behind the CRT. It is below the center screw close to the top of the back panel. The center bottom screw will fit into it.
Take a pair of side-cut pliers and cut the metal part out. It is best to trim the plastic away and remove the metal piece. After the metal piece is removed, cut about 1 inch of the plastic away to make it about 1 inch SHORTER than it was. When you think you have enough removed from the bracket, put the Mother board back in and insert all the screws back into there proper place. Attach all cables, insert Hi-Res board and see if the monitor will fit back onto the machine. If it does fit right, be sure to attach the monitor and ground cables back where they go. If it does not, trim some more off the center support bracket till it does.
All the above in 1 paragraph or less......... Open computer, remove CRT cable, remove shielding and Main circuit board,remove Hi-Res board, remove ROM C's I.C socket, Solder SmartWatch in place of socket (after removing pins 1-2 and 27, attaching pins 26-28, bending pin 28 to side.), Soldering Unit in place, modifing case, putting it all back together again!
SmartWatch Pinout follows.......
1-. * .-28 2-. front .-27 3-. of .-26 4-. I.C. .-25 5-. .-24 6-. .-23 7-. .-22 8-. .-21 9-. .-20 10-. .-19 11-. .-18 12-. .-17 13-. .-16 14-. .-15
The front of the I.C. is a notch or other type of marking on one end.
If in this document I have made a omission, I am sorry, please read through once or twice to get a good understanding of what you must do before opening the computer. This was done on the OLD model 4 (white screen type) and may or may NOT apply to all Model 4's. Take your time and do it right the first time and you will NOT have to set the time or date again on the Model 4!!!
You MUST get SMTWATCH/ARC as it has the driver programs and primary instruction docs for the installation.
On the front of the stock Radio Shack Hard Drives rated at 5 or 15 Meg there are two lamps. The red lens monitors the status of the hard drive write-protect, and the green monitors I/O activity. Most of us will never have any serious problem with the red one, since it is rare that we would use the write protect feature. Not so with the I/O lamp. Mine was going on and off so often, it finally burned out. The replacement can be expensive and long, by taking it to the repair center, or it can be done cheaply and easily in just a few minutes at home.
The lens snaps on or off with a simple flat screwdriver placed such that it levers against the top of the space just above or below the green lens, prying carefully. Small slots in the top and bottom of the green lens engage nibs in the lamp assembly. It's easier to see than to describe. With needle nosed pliers or tweezers pull the lamp straight out; notice it is a small bulb with two wire leads, much like an LED. Don't try to replace it with an LED, the voltage present is nearly 5VDC.
I don't know the part number of this lamp, nor the exact price through National Parts. But I've never seen another bulb like it anywhere. I did hear a rumor that it might cost about $9, and probably take a few weeks to get from Parts. There IS a cheaper way to do it, and will take just a few minutes of your time.
From a Tandy Radio Shack store in your neighborhood, while they still have them, get a miniature lamp with wire leads rated at 6VDC, 25 ma. (272-1140). There's about 6" excess leads on this little critter; cut off all but about 3/16" past the shrink tubing that secures the leads to the bulb. Strip and tin each of the leads with solder, straighten out the leads, insert in the socket, replace the cover, and call it a job well done. For 99 cents!
It took me almost as much time to do it as it did to tell you about it. Polarity means nothing to a bulb, so don't worry about that aspect. And if the lamp is too bright, or you want a really long-lived one, use the 12 volt version; it might not burn bright enough to shine through the white mask. Throw it away if you'd like.
Maurits van Wijland has documented the steps needed to upgrade a 16K Model 4 to 64K, complete with pictures. You can download a PDF here.
Both the Model III and Model 4 will give garbage screens if there is a dead memory chip or processor. There are many many reasons why they might not be running that various parts and tests would need to be done to track down the culprit, but memory chips and processor are a good place to start.
Model 4 memory is a 4164 and may be available on the web without resorting to eBay.
I had a terrible visit from Mr. Murphy -- the Murphy of Murphy's Law fame. Murphy left after making sure that I'd torn into the 4P that was thought to be working at 6.177 MHZ. Lo and behold, I'd lied to the only person one should NEVER lie to: myself! The poor old hard working 4P had been operating at a little over 6.5 MHZ for over a year!
Well, when changed to 6.177 MHZ the unit would work fine, but when taken a wee hair above 6.6 MHZ the unit began to fail realiably. The keyboard would apparently have a case of "key-bounce" putting some garbage on the screen whenever some keys were struck. I finally traced it to U86 (my 4P is the gate-array), which is a 74LS245 - a bi-directional transceiver chip which handles data (lines) between the Z80 to keyboard, memory and the I/O (50 pin edge connector).
Now for the new probem. No 74LS245's; no 74LS645's either, they would have worked also, but I'd let my stock deplete and had not re-filled!
At this point my sweetheart, Miss Serendipity rang my gong. I had left over from an earlier project one 74HCT245 chip. Looking again into the spec manuals to check it out, I found that the HCT version was faster than either the 'LS245 or '645. Let me do a little explaining: HC stands for Hi-speed CMOS, and the T means that it is compatable with the TTL type of chip. The very type used in our computers. It worked. Oh JOY,JOY! No more keyboard "garbage". In the gate array 4P this is a plug-in conversion as it is on the non-gate array 4P: U71. BUT, since I had ONLY ONE, I had to shift it to the non-gate array 4P -- it worked! This is U71, also a socketed (plug-in) chip.
The conversion also worked on the gate array 4 and the 4 non-gate array that I have. On the non-gate array 4 board the chip number is also U71 -- also a plug-in -- at least on the board that I have. I don't know about the earlier revisions. Sorry. It is possible that they also may be of the "socketed" (plug-in) variety.
The top speed that I tried on the non-gate array was 5.xxx MHZ. Didn't have the time or the parts (remember: MY 4P WAS DOWN!) to do any more.
The chip number on the gate array version of the model 4 was U28. That now brings us the pain in the Zorch! The very thing of which so many have a fear: SOLDERING! I'll provide you with the information, but the rest is up to you!!
For this type of work one needs to have the CORRECT tools. While I'm not a complete convert to Radio Shack, there are SOME tools they have which are adequate and good enough to do the job. The list follows:
Soldering iron (cool handle) .............. cat# 64-2080 $6.49 23 or 27-watt heating element ............. cat# 64-2081 $5.99 2 Long Life Iron-clad tips Package ........ cat# 64-2089 $4.99 "One-hand" Desoldering Iron ............... cat# 64-2060 $7.95 Desoldering Braid ......................... cat# 64-2090 $1.69 Soldering iron holder & cleaner (option)... cat# 64-2078 $6.29 Recommended solder: 63/37 .050", 1.5 oz.... cat# 64-015 $2.49
A word of explanation for those who may not know: the 63/37 figures for the solder refer the alloy 63% tin and 37% lead which provides one of the lowest melting temperature solders for electronic work. There are others, but these are exotic alloys and expensive as all get out! The word that is applied here is EUTECTIC, which simply means that it melts at a lower temperature than any of its parent metals. The .050" refers to the outside diameter of the solder. This smaller size permits you to get-on and get-off more quickly than otherwise possible with larger diameter solders.
The "one-hand" desoldering tool is very good for removing (literally, sucking out) solder from the holes. However, a word of caution: if the "hole" is not completely filled with solder to the point of being "over the BOARD SURFACE", a good idea is to ADD a little solder in order to have some solder available to conduct the heat THROUGH the solder in the hole before applying the "SUCK(TION)" to the now melted solder.
The De-soldering Braid you will find useful to clean-up some "stubborn" holes, as one cannot always get all of the solder out. Again, adding a little solder can be helpful at times to completing the cleaning job.
Now to the "surgery". Since you may not have access to a $500 to $600 de-soldering tool that I have in my shop (bedroom #2), I would suggest that you use a technique that I OFTEN resort to: Take a pair of NARROW-NOSED wire cutters and clip the IC leads as close to the sides of the chip as possible --- ALL of them!! Make sure you note the orientation of pin ONE :a "dimple", notch, or a similar marking that would indicate the end that pin one is on. Note that the board should indicate the "notch" also. Take an ordinary pair of TWEEZERS (don't let the wife catch you!), clamp onto the cut pins one at a time and from the reverse side of the board apply the soldering iron. When the solder melts, the tweezers will easily extract the cut pin. Do this to all of them. Upon the removal of all of the pins, you will need to remove the solder that is still in the holes. One can apply the de-soldering braid and use up a LOT of it, or do it correctly as described above. Using the de-soldering braid alone includes the risk of damaging the board. But after all, IT IS YOUR BOARD!!!
Now to something I should have mentioned above among things needed: Sockets! I would think that one would look ahead and get sockets. But my point is: I DO NOT recommend the low profile sockets that Radio Shack might have in stock (part number below), UNLESS they LOOK EXACTLY like the sockets that are used for the Z80 and memory sockets. Though Tandy tries to do things as inexpensively as possible, Tandy did not use poor sockets on their later (4/4D/4P) boards. The 20 PIN sockets are made by the AMP Corporation, and though they may be tin-plated, these are fine sockets. The usual low-profile sockets Radio Shack had to sell (if not as described above) were made by TI and others. Check cat# 276-1991 (20 pin socket) and be sure that they are as described above! The price on these sockets is approximately $0.59 each at Radio Shack.
If they do not have the correct sockets, I'd recommend soldering in the 74HCT245 chip; it can be obtained from JAMECO for about $0.69 or JDR MICRODEVICES for $0.99. However, JAMECO has a minimum order of $20 and JDR MICRODEVICES has a minimum order of $10. JDR MICRODEVICES has a 800-538-5000 number, free in the USA and Canada, and accepts credit cards. JAMECO also accepts credit cards but does not have a toll free number. If on the other hand you wish to write either of them: JDR MICRODEVICES, 110 Knowles Drive, Los Gatos, CA 95030. JDR has a BBS where you can place your order; the number is 408-374-2171. The address of JAMECO ELECTRONICS, 1355 Shoreway Road, Belmont, CA 94002. Their hotline (24 hours) number is (415)-592-8097.
If any you who may be hardware oriented and would like to experiment further, feel free to do so. However, I would suggest a bit of caution: There is actually a great deal more to speeding-up a computer than merely changing the microprocessor, the speed of the microprocessor, and memory chips. Remember, in TRS-Link #12 we exchanged the Z80A for a Z80B (6 MHZ), and Tandy's 200 nano-second memory chips with 150 nano-second memory chips.
One can run into ALL sorts of timing problems. One could possibly run a whole gamut of experiments without having any problems, but on the other hand... I will continue to work on the speeding up of our computers, and I will keep you posted on the speed-up mods I'm doing in TRSLINK; however, for ease of installation, I still recommend the speed-up plug-in board from Anitek for only $35.00 USA. One can easily continue from there!
This article was written to make the SMUG membership and clubs we exchange newsletters with, aware of a problem that appears to be becoming commonplace with the Model 4 computer.
Phil Brown mentioned to me some time ago that he was having problems with his Model 4 in respect that the screen would flicker occasionally. We came to the conclusion that there was probably a poor connection somewhere between the power supply or on the video circuit board itself. As this fault appeared very randomly, we thought we would leave it until it got worse. (Hopefully making it easier to find.)
At about the same time, Bill Vanderstelt started having trouble with his Model 4. Although the problem appeared the same with the screen flickering, it acted very much different than Phil's computer did. It seemed to Bill that the problem occurred around the same time of the day, about 4PM, just when people would start preparing the evening meal. With the heavy load on the power lines, the line voltage would start dropping and, as we all know, if the voltage drops low enough, it causes computer screens (T.V. screens) to shrink. This is what appeared to be happening with Bill's computer. The screen would shrink briefly, then go back to normal size, as if a heavy current draw had occurred. Bill spoke to the power company, but I don't know if they ever came and checked the lines in Bill's area.
Phil's Model 4 got worse with time just as we predicted. We had a look at it a few days ago to see if we could find the fault, and came up with the answer in just a few minutes. The problem was in the power supply, where the connectors plug onto the power supply pins. We found that the solder that held the 12 volt video supply pins to the foil on the back of the power supply had developed a hairline crack around several of the pins. The solder would make and break its connection at random times and cause the screen to go wild. We resoldered all the connection pins to ensure good solder joints. To date, Phil reports that he has had no more problems with screen jitters!
Bill brought his Model 4 over the next day to see if we could find any fault with his machine. It turned out that the problem was in the identical spot as Phil's computer. We resoldered all the pins on the power supply in Bill's computer and it's been working great ever since.
I did my own computer several months back, with the problem being the same and in the identical place. That makes three machines that had a similar problem in a very short time. If any of our readers are experiencing problems with the screen flickering, I would suggest checking the solder at these points. It's very easy to check and repair this problem. With the computer top off, unplug the connectors to the power supply, remove the 4 screws that hold the board in place. Then check the solder at these points. Use a magnifying glass if needed and be in good light. If you see cracks in the solder around these pins, simply resolder them and your problems should be cured.
Well I guess I had better deliver now what was promised: How to prevent you 4/4D/4P from sounding like a mosquito passing gas! This idea came about when I was trying out "13 ghosts". The music sounded LOUSY and it sure was improved with the following modification.
The 4P is quite easy and simple because the "cassette out" signal (yes, I know that the 4P does not have a cassette port) goes through the "sound port" also. Because of this all you have to do is but two items: Change one resistor, and change the "speaker" along with some additional wire.
The 4 non gate array: the one with the RS232 connector out the bottom is just this side of exasperating. That is because you have a separate sound board which announces errors. It is located just behind the floppy drive support "towers" and must be removed. Some were held in place with double-sided foam sticky tape, some used a "Ty-rap".
Hams (Amateur Radio Operators) can appreciate not having to hook up a amplifier-speaker combination to their 4. I use it to practice Morse code.
The Gate array 4 (RS 232 connector out the back) is a lot easier.
Now play the games and enjoy! Remember, in all cases your errors are announced more loudly as well.
A friend did the mod on his wife's non gate array 4. What he didn't tell me was that he had lost the speaker that I'd given him and had installed a 6x9" speaker with a heavy magnet on it into the machine. He was busy when I visited, so he asked me to "check-out" the mod that he'd done. The sound was so loud that it nearly blew me out of the chair!
In addition, one could add a internal/external speaker jack to the computer. I assure you, you'd be amazed at how loud 25 to 50 milliwatts of audio is when put it through something like Mach One or Mach Two speaker!!
Well for the 4 we will Need 18 TTL Chips, a roll of Wire Wrap wire and tool. And 16 to 32 120ns 256K rams. (Yea I know 150ns will work Just a well but... If you want to speed up the machine faster that 4 Mhz then get the 120ns) I use NEC and have had good luck. Now I said 16 to 32 or 512K to 1 Meg. the addressing is good to 2 meg. I have had up to 1.5 meg in my Model 4.
We will need to add a New port and it was nice of Tandy to leave two decoded address on U50. I used U50-10 (94H) as the Ram control port.
To be fully compatable with all model 4 software port 94H Must have at least 1 of the lower 5 Bits set. For example if you set 94H to a 1 then this will enable the second group of 64K addressing in the first 256K ram to be used for the Alt 64K ram bank (32K high & 32K low).
The way this works is simple. In the model 4/4P the Alt 64K ram is accessed using port 84H, this will not change after the mod. Port 94H will select which 64K group within the 2048K Max Ram will be used as the alt 64K ram bank that is accessed by port 84H. So we have 5 bits in port 94H D0-D4. This will give us 1 to 31 pages of 64k Ram that we can select for port 84H to access.
To get more than 512K ram we will need to stack ram chips and bend up pin 15 of the stacked chips. This is the CAS* control line and will be connected to the 74S08 chip in the drawings. You can get 1 meg in a model 4 without cutting any metal.
The addition of the HD64180 will require the following
What did he Say Hmm.. Well the best way to Have your cake and eat it too, is to make two options avalible for address decoding. We can use the "Z80 Bank Method" in which you use the port 148 to select the starting address for the 64K alt memory.
Port 148 - D0 D1 D2 D3 D4 D5 D6 D7
A16 A17 A18 A19 A20 WP Z80/HD64180
Mod 4
When D7 is Set HIGH then the HD64180 MMU will control the first 512K of addressing and we can select the starting address for the Upper 256K Ram address area of the MMU.
Port 148 - D0 D1 D2 D3 D4 D5 D6 D7
Not Not Not A19 A20 WP Z80/HD64180
Used Used Used Mod4
Now D6 is Reserved for the model 4 Z80/HD64180 mod because the model 4 doesn't have a Write Protect Bit for the Lower Ram if Used for Rom Image.
Ok A16-A20 is as follows.
A16 65536-131070 A17 131071-262140 A18 262143-524287 A19 524288-1023999 A20 1024000-2047999
The Parallel Connection List on Page 2 is a wiring diagram for the daughter borad. It works like this. Get an adapter socket like the SAMTEC Model APA-640-G-D for the Z80 then get a Robinson Nugent ICS-6475-2-TG30 for the HD64B180R0P. The Wire Wrap Adapter plugs into the Z80 socket and will fit through a perf board. Then wire across Per the Parallel connection list.The DUP1 is the HD64B180R0P.
The Last thing to do is add the Crystal freq. We used DCLK on U3-16 and connected it to pin 3 of the HD64180. This has a freq. of 10.1376 Mhz used in 64x16 Mode and 12.672 Mhz for 80x24 Mode is found at U17-12. The Model 4P doesnt provide the seperate clocks before being switched like the model 4.
The reason we use the 10 Meg clock is that the Model 4 has some slow PAL chips and the placement of the MREQ in relationship to T1 is different for the HD64180 than the Z80. I believe that if the PALs were replaced with some fast 74F chips then we could run at 6.144.
See Readme/txt for Limitations..
I am sorry but I have not put together a step by step procedure for modifying the model 4 so if you can't figure out what must be done then PLEASE don't do it.
These mods were done because the people who manufacture add on boards would not provide Schematics. I was not going to BUY a $200 or $300 board and not get schematics.
I have tried several times to contact a Certian Company for some technical info but was always lost in the answering service. I would much rather plug in a PC board than wire wrap this mod but if it dies and you don't have schematics the so do you.
Parts List:
Chip # Piggy back # Type 1 U41P 74LS244 2 U37P 74LS273 3 U35P 74LS157 4 U63P 74LS157 5 U25P 74LS32 6 U53P 74S74 7 U39P 74LS32 8 U12P 74S86 9 U27P 74S08 10 U14P 74S08 11 U75P 74LS157 12 U76P 74S42 13 U63P 74LS157 14 U64P 74S08 15 U29P 74S260 16 U31P 74S11 17 U61P 74S04 18 U20P 74S32 DUP1 On Daughter Bd. HD64B180R0P
The 74LS Chips could be 74S Chips. I would advise spending the money for them however I have used LS where shown and they work. However some are close.
Now I am sure that when you buy "Quality" chips from the discount hous that they are "PRIME". I have found that there is a large difference between the Tol. of the S and LS series. So if you get a "SLOW" LS chip And donot have a good SCOPE (At least 40MHZ) you will go nuts.
So do the mod in pieces.
Make the daughter board large enough for the Z80 adapter,HD64180 socket, and the nessary 1488,1489 level converters for the 2 serial ports. ( You can design this as I am tired of typing )
TRS-80 Model 4 HD64180 and Memory Modification MOD4FIG1/TXT - Updated 01/21/87 Fig 1 All chips are piggybacked and use Vcc & Vss on the host chip. |--------------| +5 >-----| | 74LS244 | |----|----| See Note 3 D0 >-------|12 Addr 16 8|---------------------|2a0 15 | Pin 1-A8 D1 >-------|16 Addr 17 4|---------------------|5b0 d 9|---> 256K Rams D2 >-------|3 Addr 18 17|---> HDA16 >--|3a1 c12|---> A18A D3 >-------|18 Addr 19 2|---> HDA17 >--|6b1 b 4|----------|A17 D4 >-------|5 Addr 20 15|---> HDA18 >--|13c1 a 7|-------| |A16 D5 >-------|14 Unused=0 6|--> BK 64 U14P-8 >-|14c0 | | | D6 >-------|7 W Protect13|--> | 74LS157| | | D7 >-------|9 Z80/MMU 11|---------------------|1M 3-U35P| | | | 1-U41P 1&19|--| |----|10d1 Mux | | | |--------------| | | |--|11d0 | | | ________ 5-U25P | | | |---------| | | U50-10 Port94 >10|----\ | | |----------------| | | __ | LS32|8-----| *---1|----\ | | | U48-4 IN >-----9|----/ OR | | S08 |3------| | | |--------------| U14P-11>-|---2|----/ AND | | | 74LS273 | | 9-U27P | | D0 >-------|13 Addr 16 12|-< U41P-8 | |---------| | | D1 >-------|14 Addr 17 15|-< U41P-4 |-----|4a 3a1|------| | D2 >-------|17 Addr 18 16|-< U41P-17 | 2a0|---------| D3 >-------|18 Addr 19 19|-< U41P-2 | 4-U63P | ____ D4 >-------|3 Addr 20 2|-< U41P-15 +5 >--|5 LS157 7|--> ICAS D5 >-------|4 Unused=0 5|-< U41P-6 Gnd >-|6,15 1 | See Note 2 D6 >-------|7 W Protect 6|-< U41P-13 |-------|-| D7 >-------|8 Z80/MMU 9|-< U41P-11 |->Mux See Note 1 U17-8 Rese>-|1 11 | ____ |------------| |----------|---| MREQ |--------| | 2-U37P | On all 157's U38-12 >----|2 D Q |---| _______ | pin 15 is to Pclk |6-U53P 4|--< +5 Port 94 5-U25P | be grounded HD64180-64>----|2 Ck 1|--| U50-10 >--1|----\ | Use Z80-6 if |--------| | | LS32|------| the 64180 is___ S74 | U49-4 >---2|----/ not used. REF U60-4 >----| ___ OR OUT |----> AD18 U37P-16 |-5|-----\ AD19 U37P-19 >-|-----\ | LS32 |6----9|----\ | LS32 |3-----4|-----/ | S08 |8--> U35P-14 AD20 U37P-2 >-|-----/ 7-U39P |10|----/ Bank 64K 7-U39P OR | 10-U14P OR | AND U55-18 >----1|----\ | A15 | S86 |3------12|----\ | |--2|----/ | S08 |11:-*--> U27P-2 U37-6 >--| 8-U12P |---13|----/ Bank FXUPMEM XOR | 10-U14P U37-2 MB0 >-4|----\ | AND The Only Cutting Needed: | LS32|6--| U37-5 MB1 >-5|----/ | Note 1. MUX - Cut trace from U18-7&13 to 5-U25P | feed through just below U18-8 OR | Connect U63P-1 to Feed through Note 3. A8 - Remove R44 & the | Note 2. ICAS - Cut Trace from U18-12 to Jumper between E11 & E12. Connect a wire | feed through under R53 from U35P-9 to E12 & also to the RAM | Connect U63P-7 to Feed through side of R44 pad. | under R53 =============================================================================== TRS-80 Model 4 HD64180 and Memory Modifications MOD4FIG2/TXT - Updated 01/21/87 Fig 2 U58-19 Cas >-|12|----\ See Note 4. | | S08 |11--*---------10|----\ *13|----/ OR | |LS32 |8--> CAS1x 0-256K U14-1 9-U27P | A18A>--*-9|----/ Note 4. Remove U58 & bend up | | 5-U25P pin 19 then replace it. |------------------------| | OR Connect U58-19 to U27P-12&13. | |-----------------------| Then lift U72 bend up pins | |----|----| 12&19 then replace it. They |----|3a1 1 | |----------| connect to nothing. | 13-U75P | |----------|12 1|--> CAS2x 256K-512K +5 >---|2a0 4a|------| | 2|--> CAS3x 512K-768K | 74LS157 | A20 U37P-2 >---|13 3|--> CAS4x 768K-1024K Gnd>-15|---------| A19 U37P-19 >---|14 4|--> CAS5x 1024K-1280K | 74LS42 5|--> CAS6x 1280K-1536K |-------| |12-U76P 6|--> CAS7x 1536K-1792K AD18 U37P-16 >-|2a0 a4|----------|15 7|--> CAS8x 1792K-2048K A18A U35P-12 >-|3a1 | | 9| Unused |13-U63P| | 10| Unused |74LS157| | 11| Unused Z80/MMU U41P-11>|1 | |----------| ___ |-------| ___ REF >---------2|----\ ____ REF >----9|----\ _____ | S08 |3---> Ram Bank 1 CAS1 | S08 |8---> Ram CAS1x >--------1|----/ See Note 4. CAS2x >--10|----/ Bank 2 10-U14P Cas1&2 connect to the feed 9-U27P ____ ___ AND holes just below the ram banks AND CAS2 REF >---------4|----\ that went to U72 12&19 ____ _____ | S08 |6---------------------------> Ram Bank 3 CAS3 CAS3x >--------5|----/ 14-U46P Note: Cut Traces for Cas lines at the feed hole ___ AND just below the ram banks going to U72 if the REF >---------1|----\ ram appear to be unstable____ _____ | S08 |3-------------> Ram Bank 3 CAS4 CAS4x >--------2|----/ ----------------------------------------------------------------------- Write Protect for Ram Based Operation - 0 to 14K Tandy Mod-This Res. is not on U55-9 A14 >------1| __ the Sch. 56 Ohm U55-18 A15 >-----2|----\ WR All 256K Rams <-----/\/\/\---| U35P-4 A16 >-----3| S260|5---------------1|----\ R63 | U35P-7 A17 >----12|----/ |---------2| S11 |12----13|----\ | U14P-9 A18-A20 >13| 15-U29P | |---13|----/ |LS32 |11--| NOR | | 16-U31P |-12|----/ R63 | | AND | 7-U39P Memory U72-6 Fxupmem >--4| | | | OR Write U25P-6 MB1&0 >-- 8|----\ | | | For Ram *-------9| S260|6-------| |------| |---------| *------10|----/ | ___ | *------11| 15-U29P U59-2 SEL0 >---| U38-11 MWR >--| | NOR Note 5. |------------------------------------| Find Resistor on back of | main board between U38-11 | \ | and R63. Remove and mount | \ | on the front side with U37P-6 Write >------5| S04 O-----------------| above connections. Protect Bit | / 6 the 56 Ohm D6 | / 17-U61P resistor ==================================================================== TRS-80 Model 4 HD64180 and Memory Modification MOD4FIG3/TXT - Updated 01/21/87 Fig 3 Address Mapping Defeat A16 U35P-3 >---1|----\ | S32 |3--------4|----\ A17 U35P-6 >---2|----/ | S32 |6-------1|----\ 18-U20P |---5|----/ | S08 |3-----| HDA18 U35P-13 >---------------| 18-U20P |----2|-----/ | | 9-U27P | Z80/HD U37P-9 >-------------------------------| | | |--------------------------------------| | If we could replace U37&59 with | a faster PAL or 74F logic | then we could run at 6.0Mhz | or faster U37-12 >-------------------|-----9|----\ | | S32 |8------------> U31P-13 Lift U37 bend up Pins |----10|----/ Sel 0 12&15 then replace it. | 18-U20P Connect to U20P |----12|----\ | S32 |11-----------> U59-1 U37-15 >------------------------13|----/ Sel 1 Note : Remove U59 then This circuit will disable address bend up Pins 1&2 decode when the MMU is accessing memory on U59 and Replace it. beyond the normal 64K map. ===================================================================== Z80H - Adapter - HD64180 Fig. 4 Parallel Connection List MOD4FIG4/TXT This is a Parallel listing of the common connections between the Z80 Socket/Z80H cpu/HD64180 Cpu. -------------| |---------------| |-------------| Z80 Description | Z80Soc | | DPU1 | | Z80 Socket | | HD64180 Cpu| Pin/Desc | | Pin/Desc | | Pin/Desc |Comments 1 A11 |----| 1 A11 |---| 24 A11 |Address Line 2 A12 |----| 2 A12 |---| 25 A12 | " 3 A13 |----| 3 A13 |---| 26 A13 | " 4 A14 |----| 4 A14 |---| 27 A14 | " 5 A15 |----| 5 A15 |---| 28 A15 | " 6 U3-16 10 Mhz Freq. 3 CLK Clock HD64180 7 D4 |----| 7 D4 |---| 38 D4 Data Line 8 D3 |----| 8 D3 |---| 37 D3 " 9 D5 |----| 9 D5 |---| 39 D5 " 10 D6 |----| 10 D6 |---| 40 D6 " 11 +5 Volts |----| 11 +5 Volts |---| 32 Vcc 5 Volt Power 12 D2 |----| 12 D2 |---| 36 D2 Data Line 13 D7 |----| 13 D7 |---| 41 D7 " 14 D0 |----| 14 D0 |---| 34 D0 " 15 D1 |----| 15 D1 |---| 35 D1 " 16 INT |----| 16 INT |---| 9 INT0 Maskable Int 17 NMI |----| 17 NMI |---| 8 NMI Non Maskable Int 19 MREQ |----| 19 MREQ |---| 59 ME Memory Enable 20 IORQ |----| 20 IORQ |---| 58 IOE I/O Enable 21 RD |----| 21 RD |---| 63 RD Read Enable 22 WR |----| 22 WR |---| 62 WR Write Enable 24 WAIT |----| 24 WAIT |---| 4 WAIT Wait Input 26 RESET |----| 26 RESET |---| 7 RESET CPU Reset 27 M1 |----| 27 M1 |---| 61 LIR Load Inst. Reg 28 REFRESH |----| 28 REF | | 57 RFSH Refresh 29 GND |----| 29 GND |---| 33 Vss Ground 30 A0 |----| 30 A0 |---| 13 A0 Address Line 31 A1 |----| 31 A1 |---| 14 A1 " 32 A2 |----| 32 A2 |---| 15 A2 " 33 A3 |----| 33 A3 |---| 16 A3 " 34 A4 |----| 34 A4 |---| 17 A4 " 35 A5 |----| 35 A5 |---| 18 A5 " 36 A6 |----| 36 A6 |---| 19 A6 " 37 A7 |----| 37 A7 |---| 20 A7 " 38 A8 |----| 38 A8 |---| 21 A8 " 39 A9 |----| 39 A9 |---| 22 A9 " 40 A10 |----| 40 A10 |---| 23 A10 " ================================================================ Model 4 Memory Modification for Non Green Screen types Parts List MOD4FIG5/TXT Chip # Piggy back # Type 1 U41P 74LS244 2 U37P 74LS273 3 U35P 74LS157 4 U63P 74LS157 5 U25P 74LS32 6 U53P 74S74 7 U39P 74LS32 8 U12P 74S86 9 U27P 74S08 10 U14P 74S08 11 U75P 74LS157 12 U76P 74S42 13 U63P 74LS157 14 U64P 74S08 15 U29P 74S260 16 U31P 74S11 17 U61P 74S04 18 U20P 74S32 DUP1 On Daughter Bd. HD64B180R0P
Think of a "null modem" as a way to fool two computers into "thinking" they're connected to modems, when in fact they are connected to each other through their serial ports.
The RS-232 is one standard used for serial communications. The most common connector used for this standard is the DB-25 (male and female) (but ATs and others use a 9-pin connector; if this is your case, see the note at the end).
Two types of equipment are connected:
DCE normally uses female connectors, but for DTE it varies. In most cases, the computer is DTE and the modem is DCE.
The following is a list of connector pins used in the project and their use. Normal modem communications use also pins 12 and 22, Speed Indicator (SI) and Ring Indicator (RI).
A "null modem" makes each computer (DTE) believe that the other is DCE, by pairing the lines and crossing the signals used.
For the construction of the "null modem" let's assume that each computer has a cable connected to its RS-232 serial port, and that there is a male connector at the free end of each cable. Since you would normally connect that end to a modem, and since most modems (DCE) have female connectors, that should be the case in most instances, but do not underestimate manufacturers.
Connect pin 1 of connector A to pin 1 of connector B " 2 " A " 3 " B " 3 " A " 2 " B " 4 " A " 5 " B " 5 " A " 4 " B " 6 " A " 8 " B " 7 " A " 7 " B " 8 " A " 6 " B " 6 " A " 20 " A " 6 " B " 20 " B(Note that in the last two steps, pin 6 is connected to pin 20 of the SAME connector)
You have built a "full" null modem. There are other types, such as the "standard" null modem. To build the "standard" type:
Connect pin 1 of connector A to pin 1 of connector B " 2 " A " 3 " B " 3 " A " 2 " B " 6 " A " 20 " B " 20 B " 6 " A " 7 " A " 7 " B " 8 " A " 4/5 " B " 4/5 " A " 8 " B " 4 " A " 5 " A " 4 " B " 5 " B
(Note that in the last two steps, you are connecting pins 4 and 5 of the same connector)
Depending on the two computers being connected, one of the two types indicated could be more suitable than the other. Try both before giving up!
Uses
A common use for "null modems" is the transfer of files from one machine to another. I will cover it briefly.
Let us assume you have a TRS-80 Model 4 and a Tandy 2000, and that you have terminal programs for both. (If we had double sided 80 track drives in your Model 4, we could do it with SX4, but let us assume we don't).
Note on DB-9 (nine-pin) Connectors
The AT and similar computers (and the Model 1400-LT!) don't use the DB-25 connector: they use a smaller, 9-pin version. Since in most microcomputers only 9 of 25 pins in the DB-25 connector are used for the RS-232, it makes a lot of (painful!) sense.
Here are the equivalent signals for the 9-pin connector:
Pin 1 is CD, equivalent to pin 8 " 2 " RD " " " 3 " 3 " TD " " " 2 " 4 " DTR " " " 20 " 5 " GROUND " " " 7 (and maybe 1) " 6 " DSR " " " 6 " 7 " RTS " " " 4 " 8 " CTS " " 5 " 9 " RI " " " 22
How about expanding your model 4P to 512K ram or even to 2 Meg. Sound expensive? Well for the 4P we will Need 10 Chips, a roll of Wire Wrap wire and tool. And 120ns 256K rams. (Yea I know 150ns will work Just a well but... If you want to speed up the machine faster that 4 Mhz then get the 120ns) I use NEC and have had good luck. The last set I bought were $3.54 ea. So a 512K system should cost you $75.
This Mod is for the GATE ARRAY board ONLY. If you donot have a gate array board and have schematics that logic used here will also apply. The chip numbers will change. Actually the original memory mod was first done on my old Model 4 which has 1 Meg installed
We will need to add a New port and it was nice of Tandy to leave two decoded address on U28. I used U28-10 (94H) as the Ram control port.
To be fully compatable with all model 4 software port 94H Must have at least 1 of the lower 5 Bits set. For example if you set 94H to a 1 then this will enable the second group of 64K addressing in the first 256K ram to be used for the Alt 64K ram bank (32K high & 32K low).
The way this works is simple. In the model 4/4P the Alt 64K ram is accessed using port 84H, this will not change after the mod. Port 94H will select which 64K group within the 2048K Max Ram will be used as the alt 64K ram bank that is accessed by port 84H. So we have 5 bits in port 94H D0-D4. This will give us 1 to 31 pages of 64k Ram that we can select for port 84H to access.
To get more than 512K ram we will need to stack ram chips and bend up pin 15 of the stacked chips. This is the CAS* control line and will be connected to the 74S08 chip in the drawings. In the 4P however you will need to cut away some of the metal sheld so the rams will clear. You can get 1 meg in a model 4 without cutting any metal.
I don't own a Model 4P (Don't get Worried) so Jerry Johnson's 4P was the first ginnypig. Jerry says it nice to have 352 Grans (512k) for a system disk.
The instillation for the 4P is as follows:
#1 74LS32 - U105P - Bend all pins Except 7&14 #2 7486 - U24P - " " " 7&14 #3 74S08 - U129P - " " " 7&14 #4 74LS32 - U87P - " " " 7&14 #5 74S08 - U114P - " " " 7&14 #6 74157 - U111P - " " " 8,15&16 #7 74157 - U110P - " " " 1,8,15&16 #8 74LS244- U125P - " " " 3,5,7,9,10,12,14,16,18,20 #10 7442 - U116P - " " " 8,16
U85-16 to U24P-2
U85-12 to U105P-9
U85-15 to U105P-10
If you Get Trash on the Screen Check the Wiring You GOOFED.
If it runs but is unreliable then you will need a good scope and look at the RAS line. Check for Ringing. Add resistance to the series resistor if you have ringing. If the transitions are a good 4.75V then reduce the resistance.
Check the Cas line also. 47 Ohm maybe to much.
On My model 4 I have No resistors in the Ras or Cas.
If you Use NewDos/80 then I have a RamDisk program that will add a 5th drive and use the memory we just added.
The RamDisk uses the fourth Pdrive, so if you have 512K ram then set Pdrive #4 as follows:
Then type:
Ramdisk 4 ' I renamed it to RM so it would be " RM 4 "
If the NewDos/80 4P system has two Directorys then we need two steps (I calculated the Directory starting and ending points so I can return Directory Protected Status)
Copy 0 4,,fmt cbf ndmw /sys
Copy 0 4,,nfmt cbf usr ndmw
If you Made the ModelA/III a SYS file then a full disk copy will work
Copy 0 4,,fmt cbf ndmw
then to mount as drive 0:
Ramdisk 0
thats it. ( Yes Jack we "Can" CHAIN through this Version )
The Ramdisk program will support a standard Model 4 or 4P with Just 128K if You set the Switches accordingly
Model 4P Hd64180 Mod.
I have Mods for the Model 4 and Have 1 Meg in mine But the design incorporates u sing a HD64180. This Makes the Mod a little Complicated. The Model 4 Version board that I have is one of the older Non Green screen type and I had to Redo the the RAS/MUX/CAS circuits in order to Run the 1 Meg rams . However if there is some interest for the model 4's I will send in the Mods I have done. interest for the model 4's I will send in the Mods I have done.
This is an additional mod to add the HD64180. If you are doing this all at once then you must use the MOD4PHDx/TXT files for reference. The difference between the two mods is really the need to:
What did he Say Hmm.. Well the best way to Have your cake and eat it to is to make two options avalible for address decoding. We can use the "Z80 Bank Method" in which you use the port 148 to select the starting address for the 64K alt memory.
Port 148 - D0 D1 D2 D3 D4 D5 D6 D7
A16 A17 A18 A19 A20 WP Z80/HD64180
Mod 4
When D7 is Set HIGH then the HD64180 MMU will control the first 512K of addressing and we can select the starting address for the Upper 256K Ram address area of the MMU.
Port 148 - D0 D1 D2 D3 D4 D5 D6 D7
Not Not Not A19 A20 WP Z80/HD64180
Used Used Used Mod4
Now D6 is Reserved for the model 4 Z80/HD64180 mod because the model 4 doesn't have a Write Protect Bit for the Lower Ram if Used for Rom Image.
Ok A16-A20 is as follows.
A16 65536-131070 A17 131071-262140 A18 262143-524287 A19 524288-1023999 A20 1024000-2047999
The Parallel Connection List on Page 2 is a wiring diagram for the daughter borad. It works like this. Get an adapter socket like the SAMTEC Model APA-640-G-D.
This is a Wire Wrap Adapter that plugs into the Z80 socket and will fit through a perf board. Then wire across Per the Parallel connection list. The Last thing to do is add the Crystal freq. We used DCLK on J7-10 and connected it to pin 3 of the HD64180. This has a freq. of 10.1376 Mhz in 64x16 Mode and 12.672 Mhz in 80x24 Mode. Now if you use NewDos/80 then the 64x16 mode is the most used. But if you want 80x24 then a different clock must be obtained. All you need is a crystal for 10 Mhz and place it across pins 2&3 of the HD64180.
Model 4P Memory Modification for Gate Array Boards MOD4PFG1/TXT |---------------------------------| Fig 1 U28-10 >---|--4|----\ |----- 9|----\ Port 94H | #4 6---| | #4 8-----| U30-5 >------5|----/ | U29-5 >------10|----/ | OUT* U87P OR | IN* U87P OR | 74LS32 |----| 74LS32 |---| |-----|-----| |--------|--| D0 >-------| 11 2|-------- Addr 16 --------|2 1&19 |64-128K D1 >-------| 5|-------- Addr 17 --------|17 |128-256K D2 >-------| 74LS273 6|-------- Addr 18 --------|4 74LS244 |256-512K D3 >-------| U85P 9|-------- Addr 19 --------|15 U125P |512-1024K D4 >-------| 12|-------- Addr 20 --------|6 |1024-2048K D5 >-------| 15|-------- Unused --------|13 | D6 >-------| #9 16|-------- W Protect ------|8 #8 | D7 >-------| 1 19|-------- Z80/MMU --------|11 | |----|------| |------| |-----------| | Mux >---1| 74157| Note Pin 1 #7 is Bent Down Reset* >-------| Addr16>-2| U110P| Addr17>-3| #7 |4----------5|----\ U27-14 >----1|----\ |------| | #3 6----> A15 | #2 3---------1|----\ |----4|----/ Pin 1 U85-16 >----2|----/ | #3 3----|-Bank---| U129P All 256K Fxupmem U24P |----- 2|----/ | 74S08 Rams 7486 | U129P | AND XOR | 74S08 |----9|----\ U85-12 >----9|----\ | AND | #3 8---->>MUX 256 MB0 | #1 8--| |---10|----/ U85-15 >---10|----/ | MB1 #1= U105P U85P-12 >---5|----\ | 74LS32 A20 | #1 6--| OR |-------4|----/ U85P-6 >----1|----\ | A18 | #1 3----| U85P-9 >----2|----/ A19 U115-6 >----2|----\ Cas1A Cas* | #4 3-----------------------------1|----\ 47 Ohm MUX256>>-|--1|----/ | #5 3---/\/\/--> | U87P U87-5 >-----2|----/ Cas1* | 74LS32 Ref* U114P 0-256K | OR 74S08 AND |-------| |-------|9----->1792-2048K Cas8A |---1---| | |7----->1536-1792K Cas7A U115-6 >----3| U111P | | U116P |6----->1280-1536K Cas6A Cas* | 74157 |4-----------12| 7442 |5----->1024-1280K Cas5A +5V >---2| #6 | | |4----->768-1024K Cas4A |-------| Addr20 >-13| #10 |3----->512-768K Cas3A Addr19 >-14| |2----------5|----\ 47 Ohm Additional Memory Addr18 >-15| | Cas2A | #5 6--/\/\/--> CasXA >---|----\ |-------| |--4|----/ Cas2* | #X --->Pin 15 256K Rams | U114P 256-512K U87-5 >---|----/ Each New bank U87-5>-----| 74S08 Ref* CasX* Ref* AND ========================================================================= Model 4P Memory Modification for Gate Array Boards MOD4PFG2/TXT Fig 2 Lift R23 Closest to U115-11 U115-11 >--*-------------------X |-/\/\/---*--->RAS0 | | 27 Ohm | |---12|----\ | | | #5 11---------| | |---13|----/ | | U114P | | 74S08 | | AND | U115-3 >---*-----X -/\/\/- X----------------|--->RAS1 Remove R21 1 U105P - 74LS32 2 U24P - 7486 3 U129P - 74S08 4 U87P - 74LS32 5 U114P - 74S08 6 U111P - 74S157 or 74157 7 U110P - 74S157 or 74157 8 U125P - 74LS244 9 U85P - 74LS273 10 U116P - 7442 11 U117P - 74S08 For additional memory 12 U96P - 74S08 For additional memory