Shift Register LED board

Hello all,
I am working on designing a board which will be able to drive 8 channels at up to 1.5 amps and 30ish volts per channel. It will have a shift register controlling the gates on 8 mosfets which are controlling the Vout from 8 LM317 voltage regulators. I will have 8 trimpots on board so the voltage for each channel can be set. I've not designed a board to be manufactured before so it will be a bit of a learning process with eagle for me.

The idea is to have a flexible, easily expandable (modular) board which I can use in many different projects thats not too expensive. I plan to use the PWM 595 library to get pwm for each channel so that each one can control 2 RGB LEDs. I should be able to control tons of LEDs with them because its only limited by the libraries max number of shift registers. The first thing I want to use it for is to replace the hopelessly inadequate circuitry in my LED blinder that I made a while ago when I wasn't very experienced with hardware. One thing I want to include in my design is a way to easily plug everything in and take it apart, so I plan to have terminal blocks so that I can just run a ribbon cable for the signal and maybe two larger wires for the high power bit and keep everything mice and organized.

First of all I'd love it if you guys would just check over my schematics and board design so I don't make some obvious (and expensive) error.

Second I would like to order way more than I need and send a good number of them to people who are interested for the cost of the board (hopefully 2ish dollars a piece) so that the per board cost is lower, a sort of early adopter scenario like what happened with the very small arduinos.

Thirdly I would like to make it work for common anode LEDs which are so wonderfully cheap from dealextreme (like 3 dollars) so that I don't have to get the common cathode ones which are about twice as much. I think I know how to do but I might be totally mistaken. I'll post the schematics I think will work sometime tomorrow.

Thanks for your help, I'll post all of my work as I go along :slight_smile:

P.S. The diagram I'm posting only has 6 channels because I was planning on 6 at first but decided to change to 8 because its pretty cheap and easy, also when I was increasing contrast and everything to make it easier to read it decided to become red, probably because of my bad lighting situation but whatever haha. The second picture is just my planning and messing around before I drew to schematic

I can't follow those. The red on black is hard to see, and both are too big to scroll around and make sense of.
You can draw up a quick schematic with expresspcb for discussion here before you get bogged down in learning eagle.

I don't have pcbexpress but I'll get it, eagle is... counter intuitive to say the least

I've been trying to make it on there for a while

http://www.expresspcb.com/ExpressPCBHtm/Free_schematic_software.htm

eagle can take a while to become ... proficient.

Here are some symbols I created if you need.
Put them in your expresspcb/SchComponents_Custom folder

(right-click, save target as, rename if needed)
http://www.crossroadsfencing.com/Arduino_Mini_Pro.s
http://www.crossroadsfencing.com/ATMega328P.s
http://www.crossroadsfencing.com/Generic 328 Arduino.s
http://www.crossroadsfencing.com/Generic_328_Arduino.s

Can use to play with sizing on a shield if you go that route.
http://www.crossroadsfencing.com/arduino_shield.pcb
http://www.crossroadsfencing.com/arduino_shield.sch

Sadly theres only download available for PC users on PCBexpress so ill have to figure out eagle. In the meantime (sometime tomorrow during school) I'll use a sharpie on some graph paper so its easier to read and post a picture of that.

Thanks for the links, ill add those. I don't think I'm going to make it a shield because I plan to daisy chain about 10 of them, and I've never really liked shields, I like to lay them next to each other and run ribbon cable. But in some cases shields are nice, if enough people want maybe I could try to make it arduino shield compatible

HI,
I don't think that will work.
I can't read your writing as to the FET number but I will assume it is a P-channel one
You have a regulator going into each P-channel FET with an adjustable pot in the ground. The problem is that to turn the FET off you have to supply a voltage within about a volt of the output from the voltage regulator. However your shift register is only 5V so there is little point in making the voltage variable.

If it is an N-channel one then the matter is even worst. In that case the maximum voltage you can get out of the FET is the gate voltage, again this is 5V, the extra voltage gets burned in the FET.

Ok so how do i get around that? Is there a type of mosfet that will make this setup work? Or what if i put a signal transistor between the Vout of the regulator and the gate of the mosfet, wouldn't that supply the gate with voltage very close to that of the Vout? Sorry I'm new to mosfets, I've only ever used bjt transistors which are pretty easy to use as switches.

I'll draw a schematic of what I mean

The part number on the mosfets I originally designed with is ndp6030pl. It looked like it would work for my purposes.

Is there a type of mosfet that will make this setup work?

No.

Or what if i put a signal transistor between the Vout of the regulator and the gate of the mosfet, wouldn't that supply the gate with voltage very close to that of the Vout

That's the idea, base to the shift register (via a resistor) collector to the gate of your FET and a pull up from the gate to the output of the voltage regulator. Emitter to ground of course. This will then allow you to trim the voltage from the regulator and have the FET switch that voltage.

Here's the schematic I drew, I'm not sure if it will work or exactly which mosfet I need in it. I'm open the suggestions

Ps yeah for uploading pictures from an Android

This is a full schematic i drew up. I changed a few things. First of all the transistors have been added. Also I changed it and i hope I dis it right so it will work with common anode leds. I source the current straight from the offboard v+ terminal block and then sink it through the lm317s, hopefully that is do able. The boxes with dots in the middle are terminal blocks. On the bottom is a general layout of the board with terminal blocks around the outside. On the top part is the input stuff, and the bottom has the output stuff that would go to the input of another board, and on the side are the 8 sink ports and 2 source ports that are wired directly to the high power in. To the right of the schematic are the ports that don't interface with any of the components directly if that makes sense

Well you have shown the FETs going to ground on the latest one instead of going to an LED. I don't have to tell you what that will do. :slight_smile:
And the arrows are missing off the emitters.
Otherwise it looks OK electrically, esthetically it leaves a lot to be desired. Grounds should point down, positives should point up. It would look a lot better if the transistors were rotated 90 degrees clockwise. The same goes for the FETs.

Yeah i know the fet is going to ground instead of the led, I instead put the led before the lm317 so it would be sinking current instead of sourcing it so that I could use common anode leds. It looks like it should work to me but can you do that?

I instead put the led before the lm317 so it would be sinking current instead of sourcing it

You had it like that before when you connected the LED between the ground and the FET, that was sourcing current.

The way you have it is not going to work, in effect you are shorting out the regulator and relying on the voltage it drops in total to control the LED voltage. No sorry I can't see it working.

Ok so I'm gonna change it up a little. I'm going to cut it sown to 3 channels to simplify the board design, replace the shift register with the ws2801 driver so i can daisy change more of them, and I'm going to reverse the circuit with a lm337 negative regulator so that it will work with common anode leds. Sorry for all the changes guys, I'm not great with hardware design, hopefully college will help with that haha

I'm going to reverse the circuit with a lm337 negative regulator

Not sure what you mean but sounds wrong. A negative regulator is for negative voltages, where do they fit in.

Anyway post your new stuff and we will see.

I thought that i would use the lm337 on the ground side or the power source, which would be negative relative to the v+ of the power supply, and then I would have the mosfet, and then the common anode led with the anode going straight to the v+ of the power supply. Is there a way I could use the common anode led with 3 lm317s?

Yes I think you have misunderstood what an LM337 actually does. I don't think it will work the way you think.

Ok so I just finished making the eagle schematic on what I hope is near the last revision of the board. I've decided just to go with the common cathode design I had originally. Sorry but its late here and I'm to tired to label all of the various resistor/capacitor values so I'll do that when I have time. I'm going to lay out the board as soon as the schematic is all good. I've attached both a pdf printout and the actual eagle file

Oh yeah there are two different positive nets. The regular 5v one goes to the logic level stuff, microcontroller ws2801... The 12v is the high power that the board is controlling, it doesn't actually have to be 12v I just wanted a different symbol. There are two identical power only terminals, one is an input and the other is an output just for ease of wiring.

The LM317 I used has a heat sink in the footprint, I think I'll get rid of that but set it up on the board with plenty or room for heat sinks so that its optional, depending on the kind of power each one has to dissipate

Ws2801 Board.s#1 (87.5 KB)

Ws2801 Board.pdf (15.8 KB)

When posting eagle, need to post the .sch files.

Sorry, its too late haha

Ws2801 Board.sch (88.5 KB)