New Art / Study Project

datasmith

Well-known member
After finishing my Electric Rainbow Sculpture project I have taken an interest in playing a bit more with shift registers. There's a lot of stuff out there about the SN74HC595, so I picked up a bunch to play with. I've got a a bin full of these rectangular red LEDs (image), so I thought I would try and animate them using the the shift registers fed by an 08M2. Well, the first thing I discovered is that the SN74HC595 can only sink about 6MA per output pin. While that's fine for directly attaching LEDs while experimenting, I'm gonna want to drive the LEDs at their optimum current (20MA+) if I intend to put them into some kind of art piece. So I have ordered a few of these ULN2803A Darlington array chips. A bit of an overkill I guess as these things can drive motors and relays, but they have input resistors built in so I can attach them directly to the shift register outputs.

So.. my art concept will be a circle of 48 LEDs driven by six shift-register-Darlington-array pairs dasy-chained together. While I am waiting for the ULN2903s to arrive I will start playing with some Picaxe code and figuring out how I will put it all together.

I had thought about using discrete transistors to drive the LEDs, but that's a lot of transistors and bias input resistors. Does anyone know of anything better to drive these LEDs, or am I going in the right direction with the ULN2903?
 

Attachments

Aries

New Member
In an earlier thread https://picaxeforum.co.uk/threads/multiplexing-using-a-40x2.31145/#post-322003 I showed my underfloor heating controller with an array of LEDs, driven by 74HC595 shift registers. The rows have the shift registers on the high side, the columns use one ULN2803A Darlington on the low side. The whole array is multiplexed. In my case, the 74HC595s sourced enough current to light one LED indicator per pin to an adequate brightness - series resistors between 330R and 1K. So, from my own experience, I agree with using a Darlington if you need the higher current. Your last reference (to 2903) is probably a typo, because the 2903 has only seven outputs, whereas the 2803 has eight.
 

datasmith

Well-known member
Thanks Aries. Nice work on that heating controller panel. Can I ask how you did your text/labels?
Yup, I fat fingered the '8'.
 

hippy

Technical Support
Staff member
my art concept will be a circle of 48 LEDs driven by six shift-register-Darlington-array pairs dasy-chained together.
Wouldn't an APA102 RGB LED strip not be the easiest path to walk, possibly cheaper, easier to put together, and offering a lot more flexibility when it comes to 'the artwork' ?

Most strips can have each LED separated or you could potentially add additional lengths of strip to match the distance. It may be a fair amount of wiring, but likely no worse than wiring discrete LED's, and it will all be daisy-chaining; more a long slog than anything requiring concentration or figuring out.
 

datasmith

Well-known member
I certainly agree with you Hippy, if I were designing this thing from scratch. But, I've got a box of these LEDs and already have the shift registers. So, I'm going with the "it's the journey" mindset here, playing with shift registers, liking the shape of the LEDs and all. Also, I think I am going to go down another new prototyping path for me as well... etching my own circuit boards on this one. Looking at all my "found parts" I have come up with this concept. Maybe I'll call it "Ring of Fire" or something like that. Below are some concept sketches and circuit board workups I've put together this weekend. I'm thinking on building 8-LED modules that are pie-shaped, that when 6 are put together create the "ring of fire" of 48 LEDs. But, I could also do 8, 16 and 24 LED builds as well.
 

Attachments

hippy

Technical Support
Staff member
So, I'm going with the "it's the journey" mindset here
That's understandable. I like the 'pizza wedge' idea. It could be worth routing your ins and outs on each wedge to the edges of the boards so you can simply connect one to the next, but I guess going to the centre means they can be hidden behind a hub.

And might be worth considering mounting holes if you haven't.

Art-wise I guess there's the choice of exposing the IC's or the tracking. You could go to quite ornate lengths with tracking, even adding tracks which aren't used and are purely decorative.
 

datasmith

Well-known member
I thought about doing ins and outs from edge to edge, but single-sided PCB real estate made synchronizing ins on one side with outs on the other nearly impossible, and I will also have to have connections to the master source on each board anyway. So, I thought I would just do the daisy-chaining on a master board. The row of pins at the bottom of the pie will plug into headers on the master board that sits behind. Yes, I will have some mounting holes of some kind at mid-point on either side of the pie slice. My son has a laserjet printer that I plan to use to make the resist stencils for the circuit board. It a color printer, so I thought I might be able to transfer a color stencil for the component side. A stylized circuit design like I have on the concept, or maybe something completely different. Now as I am even writing this I am realizing a couple of issues with attaching a connector header to the back side of a single sided PCB. 1) Where/how do I solder-I'm mounting on the solder side. 2)Reverse pull stress - pulling the connector from the master board socket could easily rip the pads from the PCB board. I guess I could modify the header...push the black plastic header block all the way to one end of the pins and use it as a stop on the component side, then carefully solder around the pins on the copper side such that the pins still have clearance to go into the master board socket. I might even add a little epoxy under the header block before pressing the pins through before soldering. I put together a diagram... Comments? Suggestions?
 

Attachments

AllyCat

Senior Member
Hi,
2)Reverse pull stress - pulling the connector from the master board socket could easily rip the pads from the (single-sided) PCB board. I guess I could modify the header...push the black plastic header block all the way to one end of the pins and use it as a stop on the component side, then carefully solder around the pins on the copper side such that the pins still have clearance to go into the master board socket. I might even add a little epoxy under the header block before pressing the pins through before soldering. I put together a diagram... Comments? Suggestions?
Yes, that's a problem I encountered just today, although it's usually caused by the heat during soldering. However, once the plastic block is pushed to the end, the adhesion to the pins won't be great, so definitely use "superglue" or maybe remove the plastic block and smother the pins with epoxy after soldering. Then you can have pushed the block right to the end of the pins, to leave as much free pin length above the solder as possible. Also consider right-angled header pins which may be longer and reduce the direction of pull away from the PCB.

Cheers, Alan.
 

Hemi345

Senior Member
Art-wise I guess there's the choice of exposing the IC's or the tracking. You could go to quite ornate lengths with tracking, even adding tracks which aren't used and are purely decorative.
I like Hippy's idea of adding tracks even where you don't need them to add some additional bling to the boards.

Etching my first PCB was rewarding, but I quickly lost interest with all the drilling (broken bits), having tracks lift while soldering, etc, etc. I'd love to see what your design would look like using OshPark's "After dark" PCBs (a blackish-purple FR4 with clear soldermask so the traces can be seen). I had the little PCB that's plugged into the breadboard made with the "After dark" option:



Judging the size of the each wedge from your render, you could probably have two sets of three made for about $30 and you could do ins and outs at the edge since you'd have both sides of the board to work with.
 

tmfkam

Senior Member
After finishing my Electric Rainbow Sculpture project I have taken an interest in playing a bit more with shift registers. There's a lot of stuff out there about the SN74HC595, so I picked up a bunch to play with. I've got a a bin full of these rectangular red LEDs (image), so I thought I would try and animate them using the the shift registers fed by an 08M2. Well, the first thing I discovered is that the SN74HC595 can only sink about 6MA per output pin. While that's fine for directly attaching LEDs while experimenting, I'm gonna want to drive the LEDs at their optimum current (20MA+) if I intend to put them into some kind of art piece. So I have ordered a few of these ULN2803A Darlington array chips. A bit of an overkill I guess as these things can drive motors and relays, but they have input resistors built in so I can attach them directly to the shift register outputs.

So.. my art concept will be a circle of 48 LEDs driven by six shift-register-Darlington-array pairs dasy-chained together. While I am waiting for the ULN2903s to arrive I will start playing with some Picaxe code and figuring out how I will put it all together.

I had thought about using discrete transistors to drive the LEDs, but that's a lot of transistors and bias input resistors. Does anyone know of anything better to drive these LEDs, or am I going in the right direction with the ULN2903?
I've used a lot of '595 devices to sink current through some common anode 7 segment LED displays. I'm running these at around 10-15mA with a single resistor in the anode supply. The 595's have been totally reliable connected this way, and they have been in use for a number of years in a commercial device which runs for most of the day, every day. I'd say you should be able to use them if you are not running every LED at once as (according to the TI datasheet) they can handle up to 35mA per output, though a maximum of 70mA through either the VCC or Gnd pins. I did use some ULN2803 for the four inch displays, as they needed 12V to drive them. Of course with the ULN2803 connected, the output logic was reversed requiring different code to drive them. I got around this by including some inverters for the Clk/Data/Latch lines, double inverting the Clk and Latch, which allowed me to use the same code for any of the displays (half inch, one inch or four inch) making them fully interchangeable.
 
Last edited:

westaust55

Moderator
You can consider alternatives to the ULN2803A using DMOS technology.
See my past post here:
 

datasmith

Well-known member
Thanks for all the driver info guys. Starting my breadboarding this weekend. I'm hanging in there with our social distancing and self sequestration. I figure I'll be fine as long as FedEx keeps delivering. :)
 

datasmith

Well-known member
Hey Hemi,
Farming out the production of circuit boards would indeed give me a more professional looking double-sided product. But, this is an art project and making my own boards will be part of that art. And for that matter, just tooling up to make my own boards is probably going to cost me four times as much as it would to farm them out. I know, but artists must struggle for their art. ;)
 

premelec

Senior Member
I used to make PCBs with Speedball pens and evaporated marking pen ink - big trick was to layout on graph paper and use that to drill stack of boards BEFORE drawing lines with pens - the holes also provided two side registration if needed. [I glued graph paper to PCB with rubber cement]. Cleaned with copper pot cleaner before drawing and then etching. Only for largish through hole designs... ;-0 No chance of mistaking my boards for professional design... [ok later I had prof boards made...]
 

datasmith

Well-known member
Got my prototype circuit breadboarded. A Picaxe 08M2 feeding an SN74HC595 shift register feeding a ULN2803A Darlington array driving 8 LEDs.
I'm sing a 9-volt supply; 9-Volts directly to the LEDs, a little 5-Volt voltage regulator to feed the Picaxe and shift register.
I've actually got an NPN transistor feeding a PNP darlington sourcing the 9 Volts to LEDs tied back to the PWM output of the 08M2. This should provide a fade in-out effect on the bank of LEDs, eventually. I just have the PWM output pin set to High right now to keep voltage to all the LEDs on. I have just programmed a simple racer routine right now to test the circuit.

RoF-Prototype.gif
 

datasmith

Well-known member
OK, so I have been staring down this whole "make my own PCB" thing.
I went ahead and got some copper clad board, so now I'm invested. (Well, I won't REALLY be invested until I drop $70 US on a table-top scroll saw to cut out my pie-shaped circuit boards, and another $90 US for a mini drill press for making component holes. "But honey, its just $160. You can pay that much at the salon!" Let's see how that goes.

My son has a laser printer which I will be able to use to make my resist stencils. One article I read suggested printing on pages from a magazine and using an alcohol/acetone mix to rub/melt the toner onto the the board from the paper. But my son says "You are not going to run magazine pages through my laser printer!", so I'm gonna have to find some other "approved" transfer/glossy paper he will let me print on.

Videos of rubbing with solvent looked like it worked well, while other articles suggest ironing the toner onto the copper. I'm not sure what would be more reliable, heat or solvent?

Then there's the etching solution. The first thing that pops up is ferric chloride. It's ugly looking stuff.
Then there were two different articles I saw about making your own etchant, one with hydrogen peroxide and muriatic/hydrochloric acid and another with hydrogen peroxide, salt and vinegar mix.

Alas, I am still going to do this. But this project is slowing into a chemical/physics research phase and equipment funding phase.
Suggestions? Opinions?

20200325_222025.jpg
 

premelec

Senior Member
Don't know if you have looked at plethora of youtube stuff on PCB toner transfer... ferric chloride is hard to see through while etching - H2O2 + HCl works... they all end up with poisonous copper solutions... There are also mechanical methods just grinding away the unwanted copper - you've got copper rasa [a la tabula...] to try out now... go to it... I made a 'scroll saw' by laying a saber saw on it's back and using a Dremel blade to an overhead spring leaf... time to be creative... ;-0
 

Electronics Learner 123

Well-known member
OK, so I have been staring down this whole "make my own PCB" thing.
I went ahead and got some copper clad board, so now I'm invested. (Well, I won't REALLY be invested until I drop $70 US on a table-top scroll saw to cut out my pie-shaped circuit boards, and another $90 US for a mini drill press for making component holes. "But honey, its just $160. You can pay that much at the salon!" Let's see how that goes.

My son has a laser printer which I will be able to use to make my resist stencils. One article I read suggested printing on pages from a magazine and using an alcohol/acetone mix to rub/melt the toner onto the the board from the paper. But my son says "You are not going to run magazine pages through my laser printer!", so I'm gonna have to find some other "approved" transfer/glossy paper he will let me print on.

Videos of rubbing with solvent looked like it worked well, while other articles suggest ironing the toner onto the copper. I'm not sure what would be more reliable, heat or solvent?

Then there's the etching solution. The first thing that pops up is ferric chloride. It's ugly looking stuff.
Then there were two different articles I saw about making your own etchant, one with hydrogen peroxide and muriatic/hydrochloric acid and another with hydrogen peroxide, salt and vinegar mix.

Alas, I am still going to do this. But this project is slowing into a chemical/physics research phase and equipment funding phase.
Suggestions? Opinions?

View attachment 23741
If you have a CNC machine or a local maker space (if it is still open) you could mill your PCB like this , if you explore design packages some fo them allow you to auto convert a Gerber file into a CNC file (I think that you may have to flip it over if you are going to mill it), you can also do 2 sided pcbs using copper and in my experience they are very efficient and work well.

Electronics Learner 123
 

Buzby

Senior Member
$160 will buy you a hell of a lot of professionally made PCBs from a fabricator.

Rolling your own is not easy. You will have a lot of scrap during your learning phase, not to mention chemical stains everywhere, and plenty broken drill bits.

I suppose it boils down to 'do you want to make an art project, or do you want to make PCBs ?'.

Cheers,

Buzby
 

Hemi345

Senior Member
I don't recall the etching solution I used to make my home-made PCBs. A friend gave me the oxidizing powder and I simply mixed it with water. It was clear with a slightly blue tint when mixed. I used my Brother laser printer to print on pages cut out of a magazine, then taped the paper to the PCB and ran it a few times through a device specifically for doing the toner transfer that looked like a Seal-A-meal machine (also borrowed from the friend). This part of the process was actually kinda fun.

Then came the drilling and soldering (I love to solder). Tiny drill bits that break if you look at them wrong. Don't forget to budget for a miniature drill press and many spare drill bits. Drill registration was an issue unless I made oversized annular rings which then limited my component choices to mostly 2.54mm spaced devices. Single sided designs which meant either very creative track routing or using wire links (more drilling!). Soldering was aggravating without a solder mask to keep solder where it should be and then dealing with pads that would lift because the drilling process peeled up some of the copper from the substrate. The last board I did was almost entirely surface mount to avoid drilling and when it came down to just drilling some holes to mount headers for power, sensors and programming, I nearly chucked it across the shop in frustration. I can't imagine trying to build 6 identical boards that you're planning to do with all those holes.

Untitled.pngUntitled.png


Am I glad that I tried it and made a few boards successfully with this process? Yes. Would I recommend it to others if they don't have any of the tools already or could borrow them? H*ll no.
 

premelec

Senior Member
I'd have to agree that commercially made boards are easiest and best looking - BUT there is the time spent learning a PCB program to produce Gerber files... not trivial.... many decades ago I could take a drawing in to a manufacturer at 2X scale and they would produce undrilled boards photographically...
 

Hemi345

Senior Member
@premelec Most manufacturers will now take the board files directly, generate the gerbers from those and show you what your board should look like. If it looks good, then no additional learning or steps required. But some manufacturers will include layers like names and values in the silkscreen which can sometimes look like crap so the benefit to generating your own is you get to choose what is included in each layer. Most reputable manufacturers will offer cam files for download that you import so cam settings, design rules, etc match their production specifications. Oshpark does a good job of offering config files for most software: https://docs.oshpark.com/
 

premelec

Senior Member
@Hemi345 you are way ahead of me - Last boards I had made with Expresspcb - proprietary files though now for a fee they can be converted to Gerber... currently just trying to clean up some of the mess I've made over the years... PCBs have come a long way since I first made some [1956!]. :oops:
 

tmfkam

Senior Member
To offer a more useful opinion than just spotting old Windows wallpaper...

If you are set on making your own PCBs, I'd look at a milling machine. I used to make PCBs for the college I worked at, on behalf of the students. Printing the PCB artwork on transparent paper then exposing this to photo sensitive copper clad board. Attempting to align both sides of the image with one another with enough accuracy was really difficult. Putting it in the 'dunk tank' to etch it, ensuring both sides got an even coverage of etchant (without over-etching one side over the other and etching away the wanted copper) was also difficult. I always thought milling would be a better option. Sadly I couldn't persuade my Line Manager of this before I left.

As most fabrication plants can produce boards at lower cost than you can buy copper clad board, it would make sense to have the boards made.

We use JLCPCB at work, I've used them myself. Never had any problem with their boards that wasn't my own fault. Had a batch of 500 only last week. A smaller batch arrived this week. All are great, and value for money too, even with the tax paid up front.
 

datasmith

Well-known member
OK, OK. You've all scared me away from making my own boards. Of particular note is the hassle of drilling. Also, along with the comment above I read a comment elsewhere that Ferric Chloride should have a brand name called "StainsAlot", so there's that. I'm gonna hang onto my copper clad boards, though. Thinking on doing some art etching on them, like logos to attach to my art pieces. Maybe use a home-made etchant. Much easier to cut rectangles, too. That will come later though.

So I downloaded and installed CircuitMaker, and went though some design efforts using the software over the last couple of days. I think I have a good two-layer configuration. Imported my odd-shaped board outline from my cad program. Used community library elements for the resistors, rectangular LEDs, header pins, shift register and Darlington array. Created an artful silk screen for the top layer. Got my through-hole drills and vias set up. Generating Gerber files. Went to the JLCPCB website. They say they can use Gerber files generated from CircuitMaker. But, I am struggling with the file names for the board outline and drill information. I get those layers, but the JLCPCB recommended names don't match up with what CircuitMaker generates. I might submit them for Quotes anyway, just to see what happens.

Since JLPCB is in China there is this statement on their website;
"Due to the Novel Coronavirus (COVID-19) outbreak, restrictions on international transport, traffic in individual regions and border controls may affect the delivery time or even prevent delivery. Restrictions and delays for some geographies began to appear. We will monitor the situation day-by-day together with our logistic partners. "

So, I'm not sure about the turn-around time right now. I thought about finding a US company to get a quote. But, right now I haven't a clue as to what "cheap" and "expensive" should be for my range of cost. Also, could anyone suggest a US company?
 

lbenson

Senior Member
Also, could anyone suggest a US company?
Oshpark.com is in the U.S., but their boards are made in China, so may suffer the same supply issues as JLCPCB. Also, they're considerably more expensive, though 10 years ago I would have considered them quite cheap. From JLCPCB it's hard to beat 5 boards for $7.24, shipping included. Give 'em a try--there's not much to lose.
 

Hemi345

Senior Member
Oshpark's FR4 *might* come from China, but the production process is entirely in the USA, one of the reasons I support them. I usually get my boards in less than 10 days from the time I order using the free shipping option. One of their latest blog posts said production is normal with COVID-19.

Someday I'll try JLCPCB for a bigger design or for something I need more than 3 as it's more cost effective. Plus they have other color options than purple or "after dark".
 

lbenson

Senior Member
but the production process is entirely in the USA
I didn't realize that. I thought when they said the latest batch was sent off to production, that meant to China. I'll have to check them out again. I always liked that they take Eaglecad brd files with no Gerbers needed.
 

datasmith

Well-known member
Well I placed my order with JLCPCB. Ordered twenty boards. With a first order coupon the cost was just $10.70 shipped. Don't know the arrival time. I asked for a blue solder mask. That adds 2 days to the delivery date, whatever that is. But, I thought the red LEDs against a blue board would make an artful contrast. My wife said my orange-ish/yellow concept drawing above reminded her of tacos. :LOL: So I figure the blue mask will give it more of a night theme. Well, now I wait. Let's see how long this takes to get these during a pandemic.

If I read their website right Oshpark's deal is three copies at $5 per square inch. Not sure how to measure that on my odd-shaped circuit board, but it fits in a 3x2.7-inch rectangle. That's 8.1 square inches at $5/sq-in which equals $40.50 shipped, for just three boards. I'll get 20 from China for one fourth the cost. How in the heck are we ever going to beat China on price? How can they be so low?
 

Hemi345

Senior Member
Oshpark's standard specs are upgrades on JLCPCB. Did you upgrade the gold ENIG finish? The FR4 at OSHPark has a higher temp rating too. That'll probably close the gap in price difference a little bit, but Oshpark will still be more expensive, no denying that. I don't want to know what I've spent with them over the years. LOL
20200329_085431.jpg
 

datasmith

Well-known member
Just an update... I just got word that my first set of circuit boards have shipped from JLCPCB. I submitted the specs on the 29th, they shipped on the 1st. Can't beat that! Now I'll find out how long it will take for DHL to get them here.

As a side note here... I simply cannot believe in the sophistication of the available free development software that's out there. First, I found Blender. It's like the '747' of 3D modeling and animation software, professional movie quality capable... and it's free. So when it came to generating Gerber files to use for making PC boards, I found this program called CircuitMaker. And it is cool as can be. Even cooler 'cause it's free. Takes you from schematic to commercial parts picking to complete 3D rendering of your board, let alone making the Gerber files. It's pretty complicated too, like Blender. I fell off the bike a couple of times, but once I got peddling... Well, here's a couple of screen shots. The top half is my LED display module. The bottom half is my master board that will control six of the LED modules.
CircuitMaker.png
 

datasmith

Well-known member
Hi lbenson,
For my previous Picaxe art project I was looking for an effective way to graphically visualize what I was going to make. You can get a glimpse of the Blender software in action and the output product about halfway through my video of that project: Electric Rainbow Project. Turns out that Blender is more of a graphic artist's tool. It's extremely complex with everything you need to make photo-realistic scenes and animations, realistic, movie quality. It was a bit of an overkill for my project, like flying a 747 to get to the grocery store. But, I am a bit of a graphic artist myself and wanted to see if I could learn the tool. If I were looking to get just a spacial/visual representation of an idea now, I would consider TinkerCAD or FreeCAD. Those are much easier to use and can give you a good 3D representation, but certainly not a movie quality representation. My short Blender sequence is near the lowest-end of photo quality that Blender is capable of. Of course to use Blender's maximum capabilities you need to have one or more high-power GPU's in your computer to do all of the hi-rez frame rendering. Took my NUC computer some hours just to crank out my little video sequence, without having a GPU.

Now, when it comes to CircuitMaker, the 3D visualization I think serves two purposes; 1) to give designers the ability to see component height and width clearances, 2) to provide mapping for automated pick-and-place circuit board fabrication. CircuitMaker allows you to generate primitive shapes; spheres, cylinders, polygons, that you can use to represent the 3D space that a component will take on your board. There is a small electrolytic capacitor (black cylinder) and multi-layer ceramic capacitor (yellow box) on either side of the voltage regulator in the second PCB above where I used CircuitMaker's primitives. The red LEDs on the first circuit board are also CircuitMaker primitives. CircuitMaker is also tied to a big community library of component parts. Many, but not all, include either a 3D model or CircuitMaker primitive representation, which you can grab and place right on your schematic. The resistors and transistors are examples of complete component parts from the library. You can also "steal" 3D "footprints" from other parts and use them to represent custom components. I used a DIP-8 OP-AMP footprint to create my 08M2 component on the second board. Then there are some components that I created from scratch. While I could have just used some primitives to represent their space, I thought I would try and see how difficult it would be to create my own 3D models. The voltage regulator laying on its back and the two right-angled header connectors are my own creations. While the community library had the voltage regulator and header pins. They were all standing straight up.

So, what is used in this environment is an object file called a .STEP file. (Standard for the Exchange of Product Data). I used two programs to make these components, both TinkerCAD and FreeCAD. I used both because I am more familiar with TinkerCAD interface than FreeCAD, but FreeCAD can generate the .STEP file. I built the objects in TinkerCAD and exported them as .STL files (Stereo-Lithograph) typically used for 3D printing. .STL files have no color component , so I mported the .STL files into FreeCAD, where I could color the objects, and then exported them into .STEP files, which could be used in CircuitMaker. I guess I should learn to use FreeCAD and save a step.

FreeCAD.png
 

julianE

Senior Member
Thanks datasmith for sharing all the info. Have you tried the easyEDA the program that JLCPCB suggests? I gave it a whirl but did not do too well with it. It's probably me, it you have tried easyEDA how does it compare to CircuitMaker as far as usability. Thanks again for your generous sharing.
 

lbenson

Senior Member
Thanks indeed, datasmith, Since I am nothing in the way of a graphics artist it may fall into the category of information which is interesting, but for me, not useful.

You do good work.
 

datasmith

Well-known member
Hi julianE,
So, I installed easyEDA and gave it a test drive. Here's my (limited, new to EDA in general) comparison of the two programs. For the most part easyEDA and CircuitMaker are very much the same. Both are online systems. This provides team collaboration, version control and access to community component part libraries. This also means you can't work offline with them. But, that's the price you pay for the tools being rubust, and free.

The only real differences that I noticed between the two programs were in the available component 3D footprints from each community library. CircuitMaker seemed to have more 3D objects in their libraries than easyEDA. Most of the components in the easyEDA libraries just had 2D footprints (ie. those that define component pin placements and an outline of the part on the silkscreen layer). So right off the bat when you visualize your easyEDA circuit board in 3D, you only see the board. I created a simple circuit (3 pin connector, a TO-92 transistor, a couple of 1/4 watt resistors and a .5 mm LED. When I did a 3D view, only the transistor had a 3D footprint on the board. If I wanted to see the resistors, connectors and LED, I would have to find or build my own 3D files for them. Now, I didn't do a deep dive into all the community library components for easyEDA, so there may be 3D models out there which I didn't find with a quick scan. Additionally, easyEDA uses an .OBJ or .WRL 3D object file format. These are general purpose, kind of universal 3D object file formats, where CircuitMaker uses the .STEP (Standard for the Exchange of Product Data) file format, which is kind of an emerging industry standard for component part 3D objects.

Except for some syntax and naming convention differences they both otherwise operate exactly the same.

Both of these programs require strict adherence to a development process that starts with first laying out a schematic diagram of the circuit that you want on your PCB board. You need to have EVERYTHING that you want on your circuit board represented on your schematic diagram, board edge pads, header pins and connectors, resistors, capacitors, diodes, LED, transistors, regulators, ICs, etc. etc.). These objects must come from the community libraries, or your own creations (that get submitted to a community library). You need to pay attention to each component footprint to make sure it fits your needs (surface mount, through-hole, small, medium, large, etc.) There will be several versions of a particular component in the libraries to pick from. Then you need to connect everything together in the schematic with a "wire" or "connection" tool. This is a drafting-like process, but you need to make sure that the "wire" gets connected (snapped) to component leads so the software knows the paths between components. The software will indicate an open connection as an error. Once you are pretty sure you have your circuit properly laid out on the schematic you can then transfer that info over to a PCB layout window. There all the part footprints get piled into a lump with rubber-band like connections that represent your schematic. Over there in that window you define your board dimensions and start dragging your component footprints around to create your PCB layout. You'll see those rubber bands stretch around as you place your devices. The rubber bands help you visualize the circuits. Then you have to create the copper paths between component pins. The rubber-bands will snap to your copper paths if you lay it out right. The software has rules that warn you when you have a crossed connection or you are too close to another path or pin. For a double-sided board you will be on the "top layer" and "bottom layer" when creating your copper paths. The minimum layers needed for Gerber files are Top Overlay (the top silkscreen layer), Top Solder Mask, Top Layer, Bottom Layer and Bottom Solder Mask. The Solder Mask layers are automatically generated, as is the Top Overlay for the most part. You will also need a drill map, which is often created separately. The drill map is not a layer file, but the map file must be included in the Gerber files to get your holes drilled.

Whew! And this is actually just a short overview of what I discovered in creating a circuit board to be manufactured. It's no simple task to start with a hand drawn schematic and end up with a circuit board design laid out in Gerber files.

Every time I get my hands on a new piece of development software (from Photoshop to Vegas Studio to Visual Studio to Blender, now EDA) I see how whole professional careers can be had using any one of these tools.
 
Top