5 Transistor PIC Programmer *Schematic Added to Step 9!

56K4160

Intro: 5 Transistor PIC Programmer *Schematic Added to Step 9!

Make your own PIC programmer for your computer's parallel port. This is a variation of David Tait's classic design. It is very reliable and there is good programming software available for free. I like IC-Prog and PICpgm programmer. Best of all, it uses just two voltage regulator and 5 transistors!

*** I added a pic of the final result, and pics of my new mini-programmer with a clear top. Click the smaller images below!

** This is a new variation and it didn't work 100% correctly on the first attempt. I guess I got ahead of myself.. I have built several variations, and I thought I was on top of things. :) There are a couple of changes, but everything worked out in the end. I had to add an additional npn transistor and change a couple of resistor values. These changes are already reflected in this list, but are not updated in all the pics. See step 7 for pics of the software I use and how to set up the programmer.

You need:
A male DB25 socket
4x NPN transistors, such as the 2n3904
1x PNP transistor, such as the 2n3906
1x 7805 voltage regulator
1x LM317 voltage regulator (and appropriate resistors to make 12.5V)
1x 10k SIP resistor network
4x 10k resistors
1x 22k resistor* update for step 3
1x 5k resistor
1x 1k resistor* update for step 3
1x machined-pin chip socket
soldering iron, protoboard, wraping wire, wrapping tool, glue gun.

STEP 1: Index Card

If you have copper tape, lay a strip down as a ground plane. If not, put a row of staples into the paper along one edge and solder them together.

Then bend the legs of the SIP resistor network, and glue as shown.

STEP 2: ICSP Port

Make an ICSP port with part of a chip socket, like this. Carefully bend the pins at a right angle.

Now glue port down.

Now is also a good time to glue your transistors on. You can also solder the emitter of your npn transistors to the ground plane, now. I have labelled each transistors purpose here.

The three npn transistors will be wired as inverters. They will essentially "take power away" from their respective pullup resistor when a current is placed onto their base pin.

The PNP transistor (upside down) will control the programming voltage. It is also going to invert it's signal.

**EDIT: I just realized an omission in this design. There should be one additional npn transistor that is used to drive the PNP transistor. This will buffer your computer port from the voltages at the pnp's base. My bad. This will also uninvert the signal. See step 8.

STEP 3: Base Resistors

I used 10k base resistors. Solder where circled. I messed up the pnp transistor in this pic. Disregard the whited out area.

**EDIT: the base resistor for the "data in" tranny should be 22k. Also, data out tranny should not be pulled up with the 10k resistor network. Instead, pull it up with a 1k resistor. I just realized that these two resistors will form a voltage divider, and if each is 10k data high will be 2.5V... no good. (Alternatively, you could just leave things the way they are, but connect Data Out transistor's collector to all remaining 5 10k pullups. This makes the divider 2/10, which should still suffice. On my particular circuit, that's what I did, and it registers 4.24V as high, which should be enough.)

Picture 2:
The pnp transistor gets two base resistors wired as a divider. Solder the 10k resistor between emitter and base. Solder one end of your 5k (actually I used 3.3k cuz I had it lying around) to the base. You can connect collector to Vpp pin, now, since it is close.

Eventually, you will be connecting the emitter to 12.5V source. The 10k resistor keeps the base high - thus programming voltage off. When pin 5 of your parallel port goes low, it pulls the base low, via the 5k resistor.

The schemmy I used also showed a 10k resistor between collector and ground. I'm not sure what it's for. I think it is to ensure that the PIC's MCLR pin does not float. But that would be silly, since MCLR is usually going to be connected to an external pullup, anyway. In addition, MCLR pin is an active sink of a few microamps. It doesn't float. At any rate, I have recklessly omitted this resistor. Bonus points for anyone who can tell me why this is bad idea.

STEP 4: DB25 Port

DB25 is the designation of a parallel port. As far as I know, they are synonymous. You want the male part, as your comp has a female plug.

You may glue it on the edge of the card, for now.


No wait! You glued it too soon! First make pins 18-25 common, as they will be common ground pins. Oh.. it's ok, cuz the card can bend.

Actually, a better way to do this part is to bend each pin over onto it's neighbor, then solder them down. I am just trying to illustrate how the connections should go.

STEP 5: DB 25 Connections

Ok. Pin 2 of the DB25 port is the data out pin. Connect it to the "data out" base resistor. The final result: when this pin goes high, the pic's RB7/data pin will receive a low signal. (what's the point of inverting things? A side effect of inverting a signal is that you buffer it, as well. Buffering the signals here, using an external power source, is the whole point of the npn transistors.)

Pin 3 is the clock out pin. Connect it to the "clock out" base resistor.

Picture 2: pin 10 is the data IN pin. Connect this to the pullup resistor of the "data in" transistor, as seen in blue circles.

Pin 5 is the programming voltage pin, or Vpp pin. See step 8. You will need to add a fourth npn transistor, and connect this line to it's base resistor. The transistor's collector will connect to the 5k base resistor of the pnp transistor. The emitter will connect to ground plane.


STEP 6: ICSP Port Side

In my setup, I chose to make clock bottom, data top, and ground, Vdd, and Vpp inbetween. This is completely arbitrary.

The ICSP data pin will connect to BOTH the pullup resistor for the "data out" tranny AND to the base resistor of the "data in" tranny. BLUE circles **EDIT: pull up Data Out with either a 1k resistor, or with all 5 remaining 10k pullups on the resistor network. Using just one 10k resistor will cause data high signal to be divided down to 2.5V.. That won't register as high, as CMOS parts running at 5V need about 3.5V to register high.

The Vpp pin will connect to the PNP transistor's collector.

The Vdd pin will connect with your network resistor pin 1. ORANGE circles If you want an on/off switch on the programmer, insert it between these points.

The ground pin will connect somewhere on the ground strip.

The clock pin will connect with the pullup resistor of the "clock out" transistor. YELLOW circles

STEP 7: New Pictures... Finished and Tested.

Here's the finished programmer. You can't tell in the pic, but I cut a piece of clipboard to the right size and used Elmer's to glue the card to the board.

I pulled out my LCD for a quick test. It reads, writes, erases. What more can you ask? Check the pics for a screenshot of how to set up ICProg or PICPgm programming softwares. Also check step 8 for detail of a couple of corrective measures that are featured here.

I added two lm317's for 5V and programming voltage.

STEP 8: Correction!!!

Here's the correction. Oops... update. See next pic.

You should have another npn transistor to buffer the port from the potentially hazardous voltages at the pnp's base. This is depicted in the top left. Collector does not attach to a pullup resistor. The pnp base is already pulled up to Vpp. Emitter is grounded. The collector connects to the 5k base resistor of the pnp transistor. I also show the 10k pull down resistor that I omitted earlier. I still don't know what it's for, though. :)

Because you are buffering with the use of inverters, when you use a TAIT compatible programming softare, you will need to go into the programmer settings and invert the clock, data out, and data in. Because you double invert the Vpp line, you will leave it alone.

FYI, the original TAIT uses DB25 pin 4 to control Vdd. I don't like this, because then you can't run your pic from the programmer's power source. I have added a manual switch in some of my other progammers, but it never gets used. Why would you go behind your computer to turn your circuit on/off? I just add a switch to my breadboard/circuit to control Vdd. You have to disconnect power or the icsp cable when not in use, though, so as to avoid shorting power and ground.

STEP 9: Schemmy, Using a 9V Battery! and a Gratuitous Kitty Photo :)

Pic 1: Just add an on/off switch to the battery, and this programmer is good to go. If your circuit draws more power than the wimpy battery can handle, add a different power supply between 9 and 12.5V (check if with a multimeter! 12V unregulated usually means 18-20V under low draw - and will kill yor pic). If your nearest wall wart gives more than 12.5V, then you will have to add another voltage regulator.

OR you could leave the 9V battery connected to the pnp transistor, but disconnect it from the 7805. Then insert your external power source, less than 35V, to the 7805. Well, now that you understand how the programmer works (ya do, right?), you can modify it any way you like from here. Adding some indicator LEDs might be nice?

Pic 2: Smurfy. Shhhh, she's sleeping.

60 Comments

i have made the hardware as shown in the schematic in last step but i am unable to get the desired results ..
is the schematic in the last step is fine??
Yeah, the schematic is verified by multiple people.

One thing about the selected resistor values is they're designed for low current draw. But this isn't ideal for ICSP with loaded lines. If you are trying to program a PIC out of circuit, recheck your wiring. If you're doing ICSP, you might want to decrease all the resistor values proportionally. Say 1/10 the values shown, just make sure your transistors can handle the current.
is the schematic updated so it works?
The Project is great but the cat is the best!!
... Hehe, yeah. This was an awesome picture. She fell asleep in my hand and I carried her around until she woke up. :) She had a litter and gained a few bazillion pounds since then, though. :)
Well, adding a socket it ok for 1 type of chip. But this can program a lot of different chips. Many have different pinouts. So I prefer to just leave a port where you can plug in a 5 wire cable. Anyway, AFAIC the best way to program a DIP chip isn't with a socket. I've tried that. During debugging, you'll end up programming a chip dozens of times. Regular sockets are a PITA to put the chip in and out. ZIF sockets are slightly less of a pain, but they are sorta fragile; they don't last long in my hands, anyway. I prefer to wire up the programming connections to a breadboard, permanently. It's easier to use than a regular socket, and you can program in-circuit. Another nice thing to use is a chip clip. This it a plastic spring clip that goes over a DIP chip. It is quicker than any socket, and it can do ICSP even on chips that are soldered onto a PCB.
so, would it be possible to program all types of PIC chips with this?
Nope, not all. Theoretically it could program pretty much all chips that can operate at 5V. But in reality, it depends on what the software you are using supports. Back when I made this, I used PicPGM software. Some of the older, more "obsolete" chips weren't supported, such as the 10F and many of the 12F series chips. And 18F series chips were also not supported. But most all the recent 12F and 16F series flash chips were supported. This includes a good variety of 8 pin, 14 pin, 18 pin, 20, 28, and 40 pin chips. But if you used a different software, you might be able to program different chips. I think the other one I used was called WinPic, or something. Using that software, you could program a good bit of the 18F series. Nowadays I use Microchip's PicKit2 programmer, so I'm not really up on the latest free programming softwares.. :)
Just as an FYI, even some of the nonsupported chips can be programmed, sometimes. What you do is set the software for a similar chip, with the same instruction word length and memory size, but then you have to read the datasheets to figure out how to set the configuration registers, manually, in hexadecimel. So it's a pain, but it can be done.
is there any way to extract code from a microprocessor?
i cant see where to put the pic here.. where do i put the chip to be programmed??
From the header, you can plug in whatever you want. Sure, you can permanently wire a socket if you want. But I'd rather plug in a socket when I need it, because I rarely program a chip using a socket. And even if I did, there are many different chips you can program with this and many different pinouts. I've seen clever setups where a large ZIF socket is wired with multiple connections, so you can program several types of chips by placing them in different spots/orientations. But I'd rather not need to refer to a manual just to figure out how to use my own homemade programmer. :) When I get need to program a new chip, I look at its manual and hard-wire and/or breadboard a new programming adapter.
Haha. Your cat is so cute. I remember when my cat was that small.
if i just follow the schematic and not the instructions, do i still need the resistor network? i find the instructions a little confusing. also, has anyone tried this with a laptop? thanks
Nope. Just follow the schematics. I also find the instructions very confusing. I no longer own a laptop with a parallel port, but I had one at one point BEFORE I made this particular version of the programmer. On other variations of the TAIT I had made, most did not work with the laptop's lower port voltage. I managed to make a version that worked with the laptop, though. What I did ended up doing was using a quad comparator to buffer all the signals so I could manually set the logic levels to whatever I wanted. In this case, you just need to reference the comparator inputs to a low voltage, like 1.5V, and you're all set. Now any voltage that is higher than 1.5V will switch the comparator, and there's your new laptop compatible logic level. So basically, you can replace all the 4 npn transistors with a quad comparator IC, if necessary and you know what you're doing. You don't even need the pnp to control the programming line, because the outputs of an LM339 are open-collector. They can be pulled up to whatever voltage you want. So you end up providing programming signals with pullups.. which means the output impedance is fairly high, but it should work ok as long as you're not doing ICSP with a lot of circuitry loading down those pins on your micro.
sorry, i'm just learning, so you lost me. i looked at the datasheet for the 339, but couldn't quite figure out how to replace the four npn transistoer with the chip. thanks for the reply though.
Yeah, np. Maybe it would help to learn the theory behind it. Basically a comparator has 2 input pins and one output pin for each comparator. One of the input pins is usually set to a reference voltage from an external source. Then a signal is applied to the other input pin. When the voltage on the signal goes higher than the reference voltage, it makes the output pin change state. The way a comparator works it basically the same as an opamp, except it has a couple extra transistors on the output to make it go "all-or-nothing," which makes it work well for this app. For this particular app, you would make a voltage reference of 1.5V-2V anyway you want... with a resistor divider or using a voltage reference IC, or whatever. Then connect that reference to all four of the chips I(+) pins. Then anywhere you have a transistor, you connect the part that goes to it's base to one of the comparator IC's I(-) pins. The output of the comparator is what's called an "open-collector" output. So it's really just a transistor collector, itself. That goes where the transistor's collector would go. Pulled up to the positive rail and becomes the output signal. Well, I know that confused you even more. But I'm sure you'll figure it out if needed. Necessity is the mother of invention.. or of learning stuff. :)
maybe i understood better than i thought. i started drawing it out with the reference going to all four + pins of the lm339, but wasn't sure and tossed it. i'll give it another try tomorrow. thanks!!
More Comments