I should get around to learning a full fat version of PCB layout, like EAGLE or DipTrace. Everytime I mean to I just revert to the FreePCB, and do a quick and dirty board, they are plenty good enough for me and 100mil stuff.
I try to use 0.6mm drills for as many holes as possible, it doesn't disrupt the tracks as much and they seem to stay sharper longer, they can be opened up later, if too small. The small drill finds the 20mil pilot points better.
To Jean-Marie, Schematic looks fine - I can't wait to see this as a real board! Oval pad should be ok.
To Toby, Eagle has proper sized pads. Big enough for even a 1mm drill on the DIP pads (resistor pads can also be bigger than the ones shown). I'm hooked on Eagle!
For my home-made PCBs I use 100x55 mil oval pads with a 0.8 mm drill for DIL chips, I can get a track between them without any problems. For resistors and capacitors I use 60 mil pads with a 0.8 mm drill.
I should get around to learning a full fat version of PCB layout, like EAGLE or DipTrace. Everytime I mean to I just revert to the FreePCB, and do a quick and dirty board, they are plenty good enough for me and 100mil stuff
It's about the same for me. My soft is so easy that I am too lazy to learn Eagle.
Anyway, Dr_Acula, I wouldn't like to be hooked on an eagle. Rather ride it
I have checked about everything.
One cap was lacking near the EEPROM in the schematic.
I redraw the serial part to make it clearer.
I hope this is the last attempt before etching.
This time, I went through etching. It was not to bad... but also not completely successful. As you can see on the picture, there is a bit of copper left in the upper corner. I know why. It is a bit stupid.The PCB artwork was printed at the corner of the page. So I covered the page with the sunlamp so that the left upper corner of the PCB was at the corner of the lamp. The rays are too poor in the corners.
But I think I can repair the PCB by scratching it with a knife.
Now I have to test for short-circuits and break points (I am not sure about this English word ).
Also, I found an error : there are no pads for the crystal ! Not too difficult to repair.
Jean-Marie, I think the word you want is "opens" as in open circuits, the opposite of closed circuits. Since we all know you're talking about electronics you could just say "test for shorts and opens" and we'd all know what you're up to.
The copper at the top corner should not matter electrically.
This has been great watching the board go from an idea to a design to all the helpful modifications from people all over the world, to a real board. I can't wait to see it all soldered up!
I have always had a problem with edges, and corners are just two edges, with the toner ironing. I have to cut the blank board at least 1cm bigger than the end layout. This is probably due to the movement of the ironing.
Here's a case in point. It is not a thin track (40mil) but it is up there in a corner. It will be perfectl ok for the 9V feed to a sub-board, but it isn't as perfect as we would like it to be.
Hey-Ho.
(It's a Prop Radio experiment, is anybody wonders)
@ Localroger and everybody
I am very pleased you gave me the right term. I am sure I use a lot of strange or odd words but if you don't tell me, I endlessly continue to use them. So, please help me to improve my language.
@ Dr_Acula
That is a very nice looking board!
That was ! I checked the board and I found 2 shorts (near the corner with copper), no opens, 4 lacking tracks and 1 error. I would have better checked it before etching but fortunately everything is repairable.
Anyway, it is true that without the advise of people around the word, the board would not exist or would not work. Many thanks to all of you.
@ Toby
I don't use a laser printer, so I am not aware of the problems of that method. Quite often, I read that the toner ironing method gives better results than inkjet + UV. I am satisfied with the UV but I should have put the board at the centre of the light.
UV photo-etch with transparencies printed on an inkjet gives much better results than TT, and is just as quick. I can do 8/8 mils without any problems, although I usually use 10/10, and have managed 5/5 as a test.
My board has been for a swim, no opens, no shorts ... just a bit rough on a few bits. TT is a bit unpredictable, I agree, but is the best way, I have found to try things out quickly ( before I lose interest ).
Sometimes the boards, often the larger ones, turn out so well that I allow myself to feel more smugg than usual.
I was away for a few days, so it took me a while to have everything soldered and unbelievable ! it worked (almost) at the first try !
A few new changes have been made.
I couldn't find any DC power socket of the size of the layout. The socket I found was larger, so I had to put it at another place, with short wires going to the previous holes.
I realized that the 3.3V voltage regulator didn't warm up, so I removed the heat sink (remembering the demo board and protoboard don't have any heat sink).
By the way I saw the voltage regulator and the capacitor were in the way if one wanted to remove the propeller. Thus I changed the layout and put the regulator flat on the board.
The main capacitor (2700 µF) was in my drawers. It should be possible to find a much smaller size with the same value. Anyway, I don't think I would reduce the size of the board because all the boards have to have the same size since they are piling up.
I have the same problems, because 90% of the components are what I have, and all of those are scrounged off of other scrap boards, everything becomes a "best fit possible".
It makes some of the pad layouts tricky as there are no definate drawings to use. I often push the pins of connectors through a piece of paper, with foam underneath, and measure the holes.
It would be just too easy to use brand new components on commecially produced, through hole plated PCB, wouldn't it ?
I have a couple questions that maybe someone could answer, because I know nothing about the how and why of most electronic circuits.
First, what is the purpose of the large capacitor connected to the power supply connector? The last time I saw something like that, Gunther Daubach used a 3500 micro faraed cap in between the regulator and SX chip to prevent a brownout while driving R/C servos.
Second, the traces that go between the pins on your 2x20 header, are they less capable of drving something with their thinner traces? (I tried to clarify my question with the attachment below). Or is that the same as a water pipe analogy?
Third, is Ferric Chloride the same as using pool acid (muriatic acid) mixed with hydrogen peroxide? My first attempt at making a homemade pcb, I used a laser printer on photo grade paper and etched it with the muriatic acid & hydrogen peroxide. I thought it came out quite nice, is there disadvantages to this method?
Jean-Marie, your project has made good reading, and is an excellent example of what these forums are capable of. Thank You.
I will try to answer your questions but electronics is not my professional activity.
About the capacitor(s), they appear between + and - in about every schematic of voltage regulator. They play the role of reservoir of electric charge to avoid as much as possible voltage variations. I think large capacitors mainly prevent voltage drop in case of sudden need while small caps prevent high frequency but slight variations.
About the size of the tracks, I use thin tracks where I can't use larger ones without having shorts. The intensity of the current is limited by the cross section of the copper track but all the thin tracks go to an I/O port of the Prop, so the maximum intensity shouldn't be higher than 40 mA. On the other hand, the power lines and the ground are all quite large tracks because the intensity can reach much higher values.
For the etching method, I used Ferric Chloride in the past, but it was very dirty. Then I tried Sodium Persulfate. It isn't dirty and is efficient but it was quite expensive (at least at my electronic store). Since a few months I use Hydrochloric acid (Muriatic acid = HCl) + Oxygenated Water (H2O2). It is efficient, clean and cheap (but a drop on your clothes leaves a hole after a few days). When acting on copper, cupric chloride is formed which is itself a etching agent regenerated by the oxygen of the air. So you keep the solution and just add a few ml of H2O2 before etching. http://www.instructables.com/id/Stop-using-Ferric-Chloride-etchant!--A-better-etc/
Ah, the anchient art of a low resolution picture; no detail, no shame.
Anyway, you can show Dr_A the bottom side, just keep your neck hidden !!!
Capt Quirk
The "decoupling" capacitors that go across the rails are there to absorb the variations in curent demands, just like a shock absorber. Large caps for large and slow (ish) changes, and ever smaller ones to acount for smaller yet faster changes. A relay or a motor would be a heavy but slow change going right down to the near instantanious spikes that the Prop's edges give switching at 80MHz.
A large electrolytic just wouldn't handle the edges, and a 100nF wouldn't cope with the motor, that is why you see a mixture in designs.
This Basic Prop Board is working fine but, as is, it does nothing else than giving access to all I/O pins and power through the 40 pin connector. In the near future, I plan to try Z80 emulation, PropDos, etc
So the next stage should be a module with some In/Out devices like keyboard, mouse, VGA / TV connection, SD Card reader. But also the opportunity to add SRAM.
Should I split all these in several modules ?
Should I have permanent tracks between (for examples) the resistors of the VGA output and the 40 pins connector or use a flat ribbon cable from VGA connector to pin headers (representing the I/O ports of the Prop) ?
The forum is much more oriented towards software than hardware so I am hesitating about the organization of the nest module.
The homemade PCB I mentioned earlier in this thread has a prototyping area enabling me to add components for any peripheral circuit, or connectors for external boards.
I noticed there is no on-off switch; does the large capacitor control the power spikes created from connecting and disconnecting the power supply to your board?
The homemade PCB I mentioned earlier in this thread has a prototyping area enabling me to add components for any peripheral circuit, or connectors for external boards.
As a matter of fact, I've never used a prototyping board. When I look at the Protoboard I see pads on the upper side but what about the other side. Are there pads ? tracks ? Do you have to solder components or just plug them in. Would you have a photo of the bottom side ?
However, my aim is not to make a prototyping board (I use a breadboard combined with the basic board I have made) but I am interested in pursuing a modular system with piling up boards. After having keyboard and VGA facilities, I plan to have a module with an Atmel Atmega communicating with the Prop by SPI or other quick serial connection.
That's right, to switch off the board I just disconnect the power supply. I use 2 different sources of power : an old power supply from a PC or a small 5V 2A AC/DC adapter. I don't know if I have more spikes by removing the DC plug or using a switch. When I disconnect the plug, the 5V LED continue to light for a few seconds but the 3V LED switch off much more quickly. When I plug the power in, both LED light almost instantly. I don't know if the capacitors are able to absorb all the spikes because I don't have any oscilloscope but the Prop is still alive :yeah: That's all I can say.
The good thing about putting a regulator on the PCB is that it prevents my trick of picking up the wrong PSU lead. I had a 3.3V PSU, for dirrect powering and my usual 9V one, for PCBs with regs.
One day I just picked up the 9V lead and that was the end of a Prop (RIP) and an EEPROM, the SD card got very hot but worked again on cooling down.
Beware the PC PSUs, they just don't care about fragiles, "lots of grunt, will travel" is their motto.
My board has only one 3.3V regulator. For the moment, there is no danger because all the components use 3.3V
But later, there could be a problem with the keyboard, mouse and Atmega: they need 5V so if I plug 9V or 12V by error there will be some fireworks :freaked:
You are right Toby, I should have put a 5V regulator before the 3.3V and provide 12V or preferably 9V. I will have a look if it is still possible to modify the existing board
I modified the schematic and the layout in order to add a 5V regulator (LM7805). It was easy on the paper but I am not sure I will implement that on the board itself. I would have to grind a lot of copper , make new holes, move some existing components, add 2 new ones (the regulator and one more cap) and make connections with wire. Besides, the components wouldn't be soldered to pads, so they would wobble a lot.
Edit: I almost forgot to tell you that I am leaving for holidays till August 29 (South of France)
In the near future, I plan to try Z80 emulation, PropDos, etc…
Excellent!
Add a ram chip, and you have lots of space for games, big programs in C (Catalina) and other languages.
Since you are building these boards so quickly, you can do this in stages. VGA is just some resistors and the VGA socket. Keyboard is just a few resistors and the 6 pin mini-din socket. SD card looks complicated, but it is really just a few pins and pullup resistors. 5V is just another regulator. All those could be quite easy on a single sided board.
Ram might be a little more complex on a single sided board. A 138, 3 latch chips and a ram chip. But Toby has done it.
South of France for holidays? That sounds a very good idea, and the weather should be perfect at this time of year (4 degrees C and hailing outside here in Adelaide at the moment). Please drink some wonderful French red wine for me!
I looked tor the gerber files of the "spreadout DracBlade" version. It looks as though those files were on the sick HD.
No I didn't have it backed up, dont start ....
That was laid out with the different areas as just that, on the end was the connectors, in the middle was the Prop with all its pins brought out (so that it could be recycled, on failure) and the memory/LCD bits. I use an external rs232 interface so that I do not have to build that bit again, and again ...
All was done on a single sided PCB with wire links. The last pic is double sided and is a 64KB restricted version (only two latches) and runs Nascom EMU perfectly.
Of course the sensible, if lazy, way forward is to just buy a real DracBlade PCB from Dr_A.
Ram might be a little more complex on a single sided board. A 138, 3 latch chips and a ram chip
I am not sure to understand what you are saying.
Does "a 138" mean a 74HC138 (3-to-8 line decoder/demultiplexer) ?
Does "3 latch chips" mean 3 x 74HC573 or 74HC374 (Octal latch or flip flop 3-state) ?
Could "a ram chip" be A56C4008 (512K x 8 3V 55ns CMOS SRAM SOP-32) ?
Does this all correspond to your schematic here ? http://www.smarthome.viviti.com/files/documents/Propeller_v5.pdf
Is it that board which is called "DracBlade" ? The price is very tantalizing but you know I like making my own boards and I feel free using bits and pieces salvaged from everywhere.
I suppose the purpose of all these chips is to address a large memory with the minimum of I/O pins, but I need some more explanations to understand what's going on. I will ask you a bit more details when I come back from holidays.
By the way, you dream of France and the French wines. I dream of crossing Australia from Perth to Sydney and from Adelaide to Alice Springs with a Land Rover (but probably I would no more dream of it if I had made it )
___________________________________________________________
Hello Toby
What is the "spreadout DracBlade version" ? Did you make a "DracBlade" on a homemade one side board ?
Comments
I try to use 0.6mm drills for as many holes as possible, it doesn't disrupt the tracks as much and they seem to stay sharper longer, they can be opened up later, if too small. The small drill finds the 20mil pilot points better.
To Toby, Eagle has proper sized pads. Big enough for even a 1mm drill on the DIP pads (resistor pads can also be bigger than the ones shown). I'm hooked on Eagle!
It's about the same for me. My soft is so easy that I am too lazy to learn Eagle.
Anyway, Dr_Acula, I wouldn't like to be hooked on an eagle. Rather ride it
I have checked about everything.
One cap was lacking near the EEPROM in the schematic.
I redraw the serial part to make it clearer.
I hope this is the last attempt before etching.
This time, I went through etching. It was not to bad... but also not completely successful. As you can see on the picture, there is a bit of copper left in the upper corner. I know why. It is a bit stupid.The PCB artwork was printed at the corner of the page. So I covered the page with the sunlamp so that the left upper corner of the PCB was at the corner of the lamp. The rays are too poor in the corners.
But I think I can repair the PCB by scratching it with a knife.
Now I have to test for short-circuits and break points (I am not sure about this English word
Also, I found an error : there are no pads for the crystal ! Not too difficult to repair.
The copper at the top corner should not matter electrically.
This has been great watching the board go from an idea to a design to all the helpful modifications from people all over the world, to a real board. I can't wait to see it all soldered up!
Hey-Ho.
(It's a Prop Radio experiment, is anybody wonders)
I am very pleased you gave me the right term. I am sure I use a lot of strange or odd words but if you don't tell me, I endlessly continue to use them. So, please help me to improve my language.
@ Dr_Acula That was !
I checked the board and I found 2 shorts (near the corner with copper), no opens, 4 lacking tracks and 1 error. I would have better checked it before etching but fortunately everything is repairable.
Anyway, it is true that without the advise of people around the word, the board would not exist or would not work. Many thanks to all of you.
@ Toby
I don't use a laser printer, so I am not aware of the problems of that method. Quite often, I read that the toner ironing method gives better results than inkjet + UV. I am satisfied with the UV but I should have put the board at the centre of the light.
_________________________________________________________
I give you the corrected pictures of the layout, hoping these are the last ones.
Sometimes the boards, often the larger ones, turn out so well that I allow myself to feel more smugg than usual.
I was away for a few days, so it took me a while to have everything soldered and unbelievable ! it worked (almost) at the first try !
A few new changes have been made.
It makes some of the pad layouts tricky as there are no definate drawings to use. I often push the pins of connectors through a piece of paper, with foam underneath, and measure the holes.
It would be just too easy to use brand new components on commecially produced, through hole plated PCB, wouldn't it ?
First, what is the purpose of the large capacitor connected to the power supply connector? The last time I saw something like that, Gunther Daubach used a 3500 micro faraed cap in between the regulator and SX chip to prevent a brownout while driving R/C servos.
Second, the traces that go between the pins on your 2x20 header, are they less capable of drving something with their thinner traces? (I tried to clarify my question with the attachment below). Or is that the same as a water pipe analogy?
Third, is Ferric Chloride the same as using pool acid (muriatic acid) mixed with hydrogen peroxide? My first attempt at making a homemade pcb, I used a laser printer on photo grade paper and etched it with the muriatic acid & hydrogen peroxide. I thought it came out quite nice, is there disadvantages to this method?
Jean-Marie, your project has made good reading, and is an excellent example of what these forums are capable of. Thank You.
Bill M.
Thank you for your comment.
I will try to answer your questions but electronics is not my professional activity.
About the capacitor(s), they appear between + and - in about every schematic of voltage regulator. They play the role of reservoir of electric charge to avoid as much as possible voltage variations. I think large capacitors mainly prevent voltage drop in case of sudden need while small caps prevent high frequency but slight variations.
About the size of the tracks, I use thin tracks where I can't use larger ones without having shorts. The intensity of the current is limited by the cross section of the copper track but all the thin tracks go to an I/O port of the Prop, so the maximum intensity shouldn't be higher than 40 mA. On the other hand, the power lines and the ground are all quite large tracks because the intensity can reach much higher values.
For the etching method, I used Ferric Chloride in the past, but it was very dirty. Then I tried Sodium Persulfate. It isn't dirty and is efficient but it was quite expensive (at least at my electronic store). Since a few months I use Hydrochloric acid (Muriatic acid = HCl) + Oxygenated Water (H2O2). It is efficient, clean and cheap (but a drop on your clothes leaves a hole after a few days). When acting on copper, cupric chloride is formed which is itself a etching agent regenerated by the oxygen of the air. So you keep the solution and just add a few ml of H2O2 before etching.
http://www.instructables.com/id/Stop-using-Ferric-Chloride-etchant!--A-better-etc/
Hello Dr_Acula. Don't be too enthusiastic. I didn't show you the bottom side!
Anyway, you can show Dr_A the bottom side, just keep your neck hidden !!!
Capt Quirk
The "decoupling" capacitors that go across the rails are there to absorb the variations in curent demands, just like a shock absorber. Large caps for large and slow (ish) changes, and ever smaller ones to acount for smaller yet faster changes. A relay or a motor would be a heavy but slow change going right down to the near instantanious spikes that the Prop's edges give switching at 80MHz.
A large electrolytic just wouldn't handle the edges, and a 100nF wouldn't cope with the motor, that is why you see a mixture in designs.
Anyway, there is no shame to show my bottom side to a Dr.
This Basic Prop Board is working fine but, as is, it does nothing else than giving access to all I/O pins and power through the 40 pin connector. In the near future, I plan to try Z80 emulation, PropDos, etc
So the next stage should be a module with some In/Out devices like keyboard, mouse, VGA / TV connection, SD Card reader. But also the opportunity to add SRAM.
Should I split all these in several modules ?
Should I have permanent tracks between (for examples) the resistors of the VGA output and the 40 pins connector or use a flat ribbon cable from VGA connector to pin headers (representing the I/O ports of the Prop) ?
The forum is much more oriented towards software than hardware so I am hesitating about the organization of the nest module.
However, my aim is not to make a prototyping board (I use a breadboard combined with the basic board I have made) but I am interested in pursuing a modular system with piling up boards. After having keyboard and VGA facilities, I plan to have a module with an Atmel Atmega communicating with the Prop by SPI or other quick serial connection.
____________________________________________________________
That's right, to switch off the board I just disconnect the power supply. I use 2 different sources of power : an old power supply from a PC or a small 5V 2A AC/DC adapter. I don't know if I have more spikes by removing the DC plug or using a switch. When I disconnect the plug, the 5V LED continue to light for a few seconds but the 3V LED switch off much more quickly. When I plug the power in, both LED light almost instantly. I don't know if the capacitors are able to absorb all the spikes because I don't have any oscilloscope but the Prop is still alive :yeah: That's all I can say.
One day I just picked up the 9V lead and that was the end of a Prop (RIP) and an EEPROM, the SD card got very hot but worked again on cooling down.
Beware the PC PSUs, they just don't care about fragiles, "lots of grunt, will travel" is their motto.
But later, there could be a problem with the keyboard, mouse and Atmega: they need 5V so if I plug 9V or 12V by error there will be some fireworks :freaked:
You are right Toby, I should have put a 5V regulator before the 3.3V and provide 12V or preferably 9V. I will have a look if it is still possible to modify the existing board
Edit: I almost forgot to tell you that I am leaving for holidays till August 29 (South of France)
Excellent!
Add a ram chip, and you have lots of space for games, big programs in C (Catalina) and other languages.
Since you are building these boards so quickly, you can do this in stages. VGA is just some resistors and the VGA socket. Keyboard is just a few resistors and the 6 pin mini-din socket. SD card looks complicated, but it is really just a few pins and pullup resistors. 5V is just another regulator. All those could be quite easy on a single sided board.
Ram might be a little more complex on a single sided board. A 138, 3 latch chips and a ram chip. But Toby has done it.
South of France for holidays? That sounds a very good idea, and the weather should be perfect at this time of year (4 degrees C and hailing outside here in Adelaide at the moment). Please drink some wonderful French red wine for me!
No I didn't have it backed up, dont start ....
That was laid out with the different areas as just that, on the end was the connectors, in the middle was the Prop with all its pins brought out (so that it could be recycled, on failure) and the memory/LCD bits. I use an external rs232 interface so that I do not have to build that bit again, and again ...
All was done on a single sided PCB with wire links. The last pic is double sided and is a 64KB restricted version (only two latches) and runs Nascom EMU perfectly.
Of course the sensible, if lazy, way forward is to just buy a real DracBlade PCB from Dr_A.
I am not sure to understand what you are saying.
Does "a 138" mean a 74HC138 (3-to-8 line decoder/demultiplexer) ?
Does "3 latch chips" mean 3 x 74HC573 or 74HC374 (Octal latch or flip flop 3-state) ?
Could "a ram chip" be A56C4008 (512K x 8 3V 55ns CMOS SRAM SOP-32) ?
Does this all correspond to your schematic here ? http://www.smarthome.viviti.com/files/documents/Propeller_v5.pdf
Is it that board which is called "DracBlade" ? The price is very tantalizing but you know I like making my own boards and I feel free using bits and pieces salvaged from everywhere.
I suppose the purpose of all these chips is to address a large memory with the minimum of I/O pins, but I need some more explanations to understand what's going on. I will ask you a bit more details when I come back from holidays.
By the way, you dream of France and the French wines. I dream of crossing Australia from Perth to Sydney and from Adelaide to Alice Springs with a Land Rover (but probably I would no more dream of it if I had made it
___________________________________________________________
Hello Toby
What is the "spreadout DracBlade version" ? Did you make a "DracBlade" on a homemade one side board ?