Translate this page

Controlling The Real World With Computers

Questions, Answers And I Dunno

Q: I..am having some difficulty with compiling the first program (experi1.c). I am using Visual C++ 6.0 on Windows XP and recieve the error "unresolved symbol _imp" during the linking phase of the build. I have also tried building it under DOS 6.22 using a free compiler named the Pacific C compiler and recieve the error "undefined symbol" with the cursor positioned at the line that contains "((inp(switch_port) >> (input +3))&1 ^1", indicating that this is the line that contains the error.

A: If you read the main page, you will find that it says, among other things that

"The tutorial examples will work on most operating systems that allow direct access to ports (more on ports later). They will not work with NT, 2000, XP or other operating systems that do not permit direct port access. DOS, Win 3.x, Win95, Win98 and WinMe will work, among others."

One of your problems lies in the fact that XP does not permit direct port access.

If you have yet to download Pacific's manual, you need to do so at http://www.htsoft.com/products/pacific.html

Port access is explained on page 72. It's a little confusing with all of the references to assembly code, but just as useful as inp() and outp(). Use the form

port unsigned char io_port @ 0xe0;

from their example. Your port name and location will, of course, be different:
port unsigned char switch_port @ 0x318;

This would replace the two lines

#define base 0x300 // change if a conflict is found
#define switch_port base + 0x18

and you should be able to remove the inp():
return (switch_port >> ((input + 3) & 1)) ^ 1;

In other words, any time you make reference to switch_port after making the declaration, you actually make reference to the port address and get data from it.

Q: I want to know how to write the test routines in c to apply inputs like voltage to a resistor and measure the current from it.

A: You might try the following:
Control And Embedded Systems
The resistor should be a value that is significantely lower in value than any you expect to see in the circuit you are testing. Also, the board and the circuit you are testing must have the same ground. Above all, it should never be used in an attempt to measure AC line current!! Use it for low DC currents only.

Get the inputs for AD1 and AD2 (any two of the 8 AD inputs can be used). Use C procedures similar to those you can find in Experiment 9 to get the values for each side. Now find the difference of the two values. This will be related to the voltage drop across the resistor. Now use Ohm's law to get the current:
I = V/R
Thus, if R = 1 ohm and the drop is .1 volt, the current is .1 amp. You will need to determine the numerical value your particular converter returns for a given voltage. Use that ratio as a conversion to determine voltage rather than the idealized values used in the text.

Q: In Experiment 1, I couldnt figure out how the switch port address was calculated as 0x318.

A: The base address determination is convered on the hardware page

The devices on the board are 8 bytes away from each other, starting with the Analog to Digital Converter (also noted on the hardware page). You might want to re-read the hardware page and the sections dealing with HEX notation as it relates to actual bits on the address and data busses.

Q: In Experiment 2, when you say "Connecting pin 9 to pin 1 or 25 of Header 3 should indicate that switch 4 has been turned on." How exactly are we to connect pin 9 to pin1? I mean is this done physically by means of a jumper etc.?
A: The best way to do this would be with a ribbon cable similar to the one described in Experiment 1 and included with Experimenter's Kit #1 found on the order page. The wires are simply touched together.

Q: I am doing a project in building device drivers for a ADC and DAC. We [are supposed to] read back the motor speed through the ADC. The problem is how to convert the mechanical energy into analog signal? We will use a Frequency to voltage converter.
A: I don't do homework, but I'll give you a hint. Connect a low-value resistor from one of the motor's terminals to ground. Use an oscilloscope to look at that terminal while you vary the voltage to the other terminal.

Q: Please give me a hint about how to read the motor speed back into the computer through ADC.
A: What did you see when you ran the test I suggested? If you don't have a scope, try connecting the terminal through a capacitor to an audio amplifier (this presumes you are using a good, isolated power supply for the test -- not AC line voltage!!). Now vary the voltage on the other terminal. Now, what does a FV converter do?

Q: Great tutorial!!
A: Thank you. (not a question, but I sure like stuff like that!)

Q: On Experiment2 I was wondering where the (active low) CS signal that goes to IC3 8255A pin 6 actually comes from. This incoming signal is labeled (active low) PPI SELECT. I'm trying to figure out all of the necessary signals that are needed to actually write to the control register. I understand that 0x23 in base + 0x23 will set pins 8 and 9 high (A1 and A0, respectively), and the write signal to pin 36 is automatically taken low by the PC when the outp_ function is invoked but I'm not certain what roll the (active low) PPI SELECT plays in the control register setup. Thanks!

A: PPI SELECT comes from the 74LS138 and was actually shown there as the PIO SELECT. PIO means "Peripheral I/O" and is often used interchangeably with PPI, although PIO is more often used with the Motorola peripheral interface chip. I have changed the reference in Figure 1 to make it clearer.

Register select is a combination of address, select and read/write. See the tables in the 8255 data sheet and Experiment 2 for more detail.

Q: Do you know if it is possible to draw a small circle (about the size of a penny) on the screen and link it to an input byte (for exmple: a byte from Experiment1) and change the color (depending on bit status of the desired byte) of this filled circle from red (for off) to green (for on) using C ? I have checked my book "A Book on C" and didn't find anything on the subject and I also checked a few C boards and still had no luck. Can drawing even be done using the C language? Thank you again and again your time is once again appreciated.

A: Go to my Super Start site then to the Free Downloads section. Download the full zip. The test program does something similar to what you are talking about. It first queries the board (after it detects its presence -- I will discuss auto-detect for the board when I get into the analog sections), then prints switch numbers on the screen, with green meaning off, and red meaning on. That section is not graphics though, but colored text.

You can also draw in C, but it depends on the compiler. MIX has one of the best selections I have seen, which is the reason I used it for Super Start -- it uses a lot of graphics.

Q:Does the program actually place the Hex value 0x318 on the Address lines of the ISA Bus in the fashion that I have stated?

A:Everything eventually must get down to high or low bit indications on the bus. The binary is a closer text representative of the pattern on the bus.

Q:How does the PC know that a valid port address has been detected and does the PC actually force the AEN signal to go low when this valid port address has been detected?

A:IN or OUT commands cause the processor to make AEN low and either IOR or IOW low.

Q:Does the program statement
return ((inp(switch_port) >> (input + >3)) & 1) ^ 1;
actually force the IOR (active low) signal to go low?

A:The inp() part does, along with causing AEN to go low and setting up the port address.

Q:Does IC7 74LS244 Pin 19 (EOC/SWITCH SELECT) have to be manually placed low by the user?

A:No. This is a select line from the '138. A switch would have to be taken low by the user (could be a switch or something that looks like one). EOC stands for End of Conversion. It connects to the Analog to Digital Converter and signals the end of an AD conversion.

Q:I'm not sure what purpose IC11 Pin 19 actually surves in this experiment?

A:This line enables IC12 so data can be placed on the data bus in the case of inp(), or placed on a device on the board in the case of outp(). It also activates IC10 which then further decodes the selected address and selects the particular device on the board.

Q:Most importantly I am trying to understand how the PC actually knows that the inputs from Experiment 1 are at port address 0x318. It appears that IC11 74LS688 only compares address lines A6-A9, this would only check to see if DIP SW1-SW3 have been set to 001 which represents Hex value 0x300. How does this relate to the Hex value 0x318 that is addressed in the program?

A:IC 10 does the rest of the decoding.

Q:When an ISA card is used and port addressing is assigned, is it a known that 64 bytes will be consumed by the card, it appears that in all of your examples the range is 64 bytes (200-23f, 300-33f)?

A:Yes, it is known by the user. The computer does not automatically know anything. Being one of the dumbest things on earth ("only knows how to count from 0 to 1" wouldn't look too good on the ol' resume), it knows only what it is told.

Q:The address used in Experiment 1 is port address 0x318, the end result is a possibility of 8 bits (only 3 of which are being used in Experiment 1) being placed directly on the ISA Data bus. The ranges (Experiment 1 uses 0x318-0x31f), when broken down, for all of the experiments each have a decimal value of 8 ( 0-7 ). Is the port address what actually holds the Data Bus values from the ISA card?

A:The data bus will hold values for a length of time specified in the microprocessor data sheet. It is the responsibility of the hardware designer to insure that devices will be able to grab the data in time. The same thing applies to placing data on the bus.

Q: I am working with experiment 4. I have replaced the individual diodes with an inverter which is used to control the Port A lines. I am using a 7414 instead of a 7405. Currently I have a matrix that has the possibility of 24 closures. I am using dip switches as inputs. My control lines are Port A Bits 0,1,2 (rows), and Port B as input (columns). So each Port A bit strobes a row of 8 switches.

When I used individual diodes, the circuit worked properly, meaning that any number of switches that were or were not activated were read correctly by the program.

My problem arose when I replaced the diodes with the inverter. The circuit and program work correctly when any switch or switches within a row are activated. If a switch from another row is activated, the program returns incorrect results. Using a DMM, I have found that during this time, there is a reading of about .245 volts on more than one output line of the 7414.

Do you think that I may have a loading problem, possibly a timing problem, or that a 7414 is just the wrong device for the application. I have considered installing 1) a delay in the code or 2) a driver IC for the outputs of the 7414.

A: The description of the 74LS05 begins: "The 74LS05 (see 74LS05.PDF) and similar devices do not produce a high output. They only pull things low, commonly called sinking, which is what they can do with the pullups on Port B." The fact that the 74LS05 does not produce a high output is its most important characteristic insofar as this application is concerned. A 7414 (see 74LS14.PDF), on the other hand, typically produces about 3.4V, or at least will try to. The low and high outputs can still fight each other if two or more switches are turned on. On the other hand, the 74LS05 can only provide what the pullup resistor provides because it has no circuitry to produce a high output on its own. It has what is called an open collector output. Study the matrix very carefully and imagine devices that can produce high as well as low outputs connected to it. Now imagine two of the switches being turned on with one high and the other low. I think you will see that two of the outputs could be connected to each other. The outputs of devices that can produce both active highs and lows should never be connected to each other.

Q:I was doing experiment 2 as described on your site, using the same code except with base address 0x200. It compiled and ran but it shows that the 8255 inputs are constantly 1. I don't have this problem with the 3 basic switch inputs. I checked the circuit board with an ohmmeter and I get 30 ohms between the I/O posts on header 3 and ground. Is this a short circuit? I didn't see any visual signs of a short circuit. Any suggestions?

A:I noticed that I forgot to properly encode the "<" and ">" signs to HTML. That made the source code appear incorrect. You might want to check to be sure your #includes are correct, along with other things that use less than and greater than.

You are probably getting 0 from Port B which makes the program think a line has been pulled low. The 30 ohms pretty well confirms that. You should measure between 1000 and 20000 ohms with the board out of the computer, nothing connected to Header 3 and the negative lead of your meter to ground.

I would also check for bridges between any pins of Header 3 and/or array 2. Also, make certain pin 1 of array 2 (marked with a line or dot) is in the hole with the square around it. You should see close to 5 volts between the Port B section of Header 3 and pin 1 of header 3 with the computer turned on. Port B is on the odd-numbered pins from 9 through 23. The measurement should be to the bare ends of a ribbon cable connected to a header on the board. Measure bare ends only to romove any possibility that other circuitry is affecting the measurement.

Make certain all ICs are in the correct places and correctly oriented -- all have their notches pointing the same way. I never try to remove and replace any incorrectly placed ICs. It is better to cut the leads off of them and throw them away than to take a chance on hurting the board by attempting to remove one. It's a lot easier just to pick up an 82C55 at Jameco or someplace than it is to try to remove the thing nicely.

Q:I bought a kit from you a while back (just got the remaining parts, thx), and I began working on it last night. I'm finding that some joints just absolutely refuse to flow. I don't know whether the base plate isn't getting hot enough, or what the deal is, but the solder just sticks to the exposed portion of the part to be soldered, and won't bond to the base. This only seems to happen occasionally ... out of about 35-40 joints last night I think I ran into 3 or 4.

A:You might turn your iron up a little if it's adjustable. Sometimes the surface gets a little oil on it. Sometimes rubbing the iron on it will help. You can also try cleaning the pad on the board with a pencil eraser, but not too much too hard to take the silver tin off.

The most important thing however, is to make sure the tip is coated with solder and shiny, not dull. Clean it frequently while hot on a wet sponge.

Place the iron on the pad with it against the wire and run solder into where the two meet. Pull the iron up the wire as soon as the solder begins to flow and wet the joint.

Hope this helps.

Q:Is the Microsoft VC++ 6.0 compiler adequate for the tutorials? I've been told that chips are compiler specific.
A:You should be able to make a console application, providing you can have direct port access. NT, 2000 and others do not permit it.

For VC++, See the definitions for _inp, _inpw and _inpd
and _outp, _outpw and _outpd
in VC++ help.

Q:Is there gainful employment in embedded programming for someone without a CS or EE degree?
A:From what I have seen, it's experience or degree, but you get a little edge with the paper. I have a degree in Sociology with a minor in Psychology. That and a buck will get you a cup of coffee unless they find out you are a sociologist -- then they charge 2 bucks. I went back and got a CS, going in the evenings after work. Took 6 years but it was worth it.

Q:HI Joe
Your tutorial is one of the best ones I have seen on the net. It presents a very good entry level introduction for beginners like me.

Unfortunately the net speed here in india is very low and it takes a lot of time to get the pages. Can you please Zip all the tutorial matter and put it as a downloadable. It will be of a great help.

A:A zip of the site is about 3.4 megs, which would take quite a while to download. I'm not as concerned about that however, as I am about copyright issues. US copyrights are ignored in certain other countries. I will have to check to make sure permitting a download doesn't give someone the right to claim all of my hard work as their own. If anyone knows of a place where I can download an agreement people can click on to say they have read before downloading, please Let me know. Are there others who would like to download the site?

Q:I have finished working with experiment 3 and I tried a little something that seems straight forward enough but yielded an incorrect result. The experiment itself went well, though.

After writing the control register to set up the 8255, I immediately tried to read the control word from 8255. I did not get the results expected. In brief, the control word written was 82 hex. When attempting to read the control register, I continue to read either EF hex or FF hex.

Do you have any clues as to why I cannot read the control register? The data sheet that I have says the control register can be read as well as written.

A:Looks like you might be basing your experiment on experi3a.c, which is just a test. It leaves out an important step that is included in experi3b.c. The test does not need to establish the port addresses by calling get_port() because it doesn't use them. Any actual program however, must call get_port(). Also, please note the notes for get_port(). You need to match the switch settings with what you assign to base or it won't work, at least until get_port() is turned into an auto-detect procedure.

Q:Hey Joe,
Had a question or two for you. I am looking into making a seismograph recorder using the Super Start board and I noticed on one of your programs you have an oscilloscope program and I am thinking about using the scope as the reader but I wanted to make some modifications to it (Scan rate ect.) and I was wondering where the source code is or if it is available? The other question is, do you have any plans to add information on how to bit bang data on the tutorials?
A:For those who don't know, the board was originally designed for The Super Start Project, which is where it gets its name. Go to the QuickStart section of The Super Start Project at http://www.superstart. org/quickstart.htm. You can download everything there in a zip file. One of the sources is an assembly file. That's where a lot of the recording code is. See the recplay.c file, about line 971, for the scope routine. Also for those who haven't encountered it, bit banging is used when you don't have a serial port available that's suitable for the type of communication you need to do. Under such circumstances, a device such as the PPI might be used to assemble a stream of bits in a manner similar to a standard serial interface. I might be able to get into some of that when I discuss timer interrupts.

Q:Thanks for the information about the 8255. Now i can finish my thesis. I am from Indonesia. Sorry my grammar.

My thesis is about the 8255 as a controlled ADC 8 bit for Digital Multimeter thru PC. I am confused about automatically changing range (voltage) on my voltmeter. How to separate hot line current (power) from my PC. With optocoupler or what?

A:Don't worry about the grammar. After all, you can write in my language. I couldn't even say hello in yours! While I don't do people's school work, I can offer a few directions you might take. Something as simple as a transformer, full-wave rectifier and multi-tap voltage divider along with comparaters for range-checking should get you started. There is such a thing as an analog opto if you need high frequency response.

Q:I used Delphi 5 to wrote a program to control a PPI (peripheral interface) card. I tried the same thing in Turbo Pascal & Power Basic and it worked properly. I tried to use Turbo Pascal for Windows & Delphi 5 but there is no statement to control the PPI like in TP & PB above. Please inform me how to control a PPI card from Turbo Pascal for windows and Delphi or what kind of statement I should use.
A:Maybe one of the newsgroups such as those listed at http://www.inprise.com/delphi/newsgroups will have the answer. You might give them this site as an example of what you are trying to do. Be sure you tell them the operating system you are running under. You can't, for example, run some of the examples in the tutorial under the NT operating system because it does not permit direct port access.

Q: I notice there is a clock running at 14.31818Mhz but I don't see where is generated from.
A: The clock is on B30. I left it out before. Now it's included. Thank you.

Q: "pulled up": What does this mean Electronically/Electrically?
A: Unlike outputs, inputs to some ICs float, which is to say you can't count on them to be high or low. It's very common to use resistors to pull them up to high logic, which is 5 volts in this system. It's just as possible to have a pull down resistor to ground. I'll get into that more in a latter experiment.

Q: What are buffers?
A: Buffers are used to increase the output drive capability of a line, or to lighten the load seen by outside circuits. This is sometimes called repowering. Some buffers have a third line entering the side of the triangle. When it is not activated, the output of the buffer is removed from the circuit to keep it from interfering with other devices.

Q: So..., when we say (in programming/SoftWare) that there is something in a Buffer...: what does this mean Electronically? (from your explanation above)
A: If I am understanding this correctly, I would use IC7, the 74LS244, as an example. It has 3 inputs called SWITCHIN1 2 and 3. It is possible to read this buffer programmatically, and it is common to say they are "in" the buffer. When one is turned on, its normally pulled-up input is pulled down to ground. That information is considered in the buffer and can be read out of it.

Q: Are the switches Physical or just an Abstract concept?
A: These are just inputs (see above). A better phrase might be "The switch inputs are designated SWITCHIN1, SWITCHIN2 and SWITCHIN3 on the schematic and can be accessed on pins 5, 4 and 3 respectively of HEADER 1." (as a matter of fact, it is now)

Q: Well, by saying "switch inputs" do you mean that
1: That there is an INPUT
2: This INPUT acts as a SWITCH
OR a place where I can (Electrically/Electronically) connect (i.e. wiring) a SWITCH-device...
or...?
A: All of the above, but the most accurate would be that it is a place where a closure device can be connected.

Q: I have a question. I am currently working on the 64 switch matrix program and circuit. I intend on using dip switches for the matrix. My question is should I be concerned about the ability of the 8255 to sink current when a large number of the switches will be closed simultaneously. Is there a possibility of stressing the device under the circumstances I have mentioned?
A: I am going to get into more matrix information in experiment 4. Please note what can happen if you close two switches at once:

             |-----------------------|
 Port A bit 0| 4| 5| 6| 7| 8| 9|10|11|
 Port A bit 1|12|13|14|15|16|17|18|19|
 Port A bit 2|20|21|22|23|24|25|26|27|
 Port A bit 3|28|29|30|31|32|33|34|35|
 Port A bit 4|36|37|38|39|40|41|42|43|
 Port A bit 5|44|45|46|47|48|49|50|51|
 Port A bit 6|52|53|54|55|56|57|58|59|
 Port A bit 7|60|61|62|63|64|65|66|67|
             |-----------------------|
  Port B bits  0  1  2  3  4  5  6  7

You are checking rows by taking bits from Port A low. It is possible to connect one line of A to another one of its other lines by turning on two switches at once. If you have taken one of them low, the other will be high, causing them to "fight" each other and possibly stress the IC beyond normal limits. As it stands, the circuit will permit only one switch at a time to be turned on.

Experiment 4 will address this problem by placing diodes at each cross-point.

Q: i just happened to surf on some reseach related to my project and i found that your page provided me with more than sufficient information.Its really awesome.By the way there is a topic about how to use digital inputs to control motors....i would like to ask you when will that information be released.By the way keep up the good work!
A:Soon and very soon (if you know the reference). I hope to get into outputs in experiment 4. Actually, the matrix in 3 uses A as an output and shows how to set up the 8255 for output. Motors will probably be in 5 or 6.

Q: I'm really impressed by your description of this I/O card. It has made me choose as a project for my summer study. Since I stay in India, it won't be possible to meet you personally but I hope to get a good support from you through emails. I would be grateful to you if you could clarify a few points. The circuit is a bit complex. I find it difficult to understand the different parts of the card. Can you provide a block diagram ? Also, since I live far away from US, it won't be feasible to buy the stuff from here. It would be very kind of you if you could provide me with the PCB Layout.
A: The block diagram is a good idea, and I'm working on it. In addition to copyright problems however, I don't have a layout that could be used to make a board. I have the original films of course, but obviously can't let them go!

Q: I wonder if you could send me an more elaborative text/article (with programms) about INTERRUPTS: i.e. how the whole mechanism works LOGICALLY... Thanks!
A: This subject has yet to be covered. I don't recall who sent this, but apparently the following questions derive from the URL that was also in the message: http://www.internals.com/articles_main.htm

It's all interesing stuff, but not yet pertaining to my site. When I do get into interrupts I'll cover some of the following. Meanwhile, whoever sent this might consider asking the following questions of the folks at the above URL since they wrote the article from which the questions derive:

To hook an interrupt, a program need only to replace a specific interrupt table entry with the address of its own interrupt handler. Whenever this interrupt occurs, the handler is automatically invoked by the processor. A good programming practice is to place a call to the previous handler inside the interrupt handling function. This ensures that all previously installed hooks will get a chance to handle the incoming interrupt.
hook an interrupt:
How does a code for HOOKING an interrrupt look like in Assembler and C languanges?

address of its own interrupt handler:
How do you know the address of the Application's interrupt handler?

A good programming practice is to place a call to the previous handler inside the interrupt handling function. This ensures that all previously installed hooks will get a chance to handle the incoming interrupt:
Don't you think that this will cause some conflict, I mean if the handlers do different things...

Note that when loading an application which uses software interrupts to call operating system services, there is no need to patch the application code with the actual memory addresses of the OS services. Therefore, the operating system can freely relocate its services in memory by simply updating the corresponding interrupt table entries with the new addresses.

no need to patch the application code with the actual memory addresses of the OS services:

What do you mean by PATCH?

operating system can freely relocate its services in memory:
Why will the OS need to relocate its services?