-
Scam Alert. Members are reminded to NOT send money to buy anything. Don't buy things remote and have it shipped - go get it yourself, pay in person, and take your equipment with you. Scammers have burned people on this forum. Urgency, secrecy, excuses, selling for friend, newish members, FUD, are RED FLAGS. A video conference call is not adequate assurance. Face to face interactions are required. Please report suspicions to the forum admins. Stay Safe - anyone can get scammed.
Schematics and Circuit Board layout with Fusion 360
- Thread starter Tomc938
- Start date
Well hold onto yer socks. SPICE is included with Fusion360. (Its fabulous GIGO software!)
Fusion360 Simulation <-- Autodesk Help Link
Tomc938
Ultra Member
I will look into this and make the changes before I send you the schematic.
I will be running wires up from these connectors to a DB9 socket. I did this just because the DB9s are too large to mount right on the board.
There are so many types of plug in connectors, is there a particular one you would recommend? Something like this for example(this is the 3 pin, not the 4 pin):
Connections will be to DB9 plugs mounted above the board, so not leaving the enclosure. Thought this woudl be easier than wires coming between wall and socket. This will change when I change to plug in terminals.
The nice thing about through mount is as long as the holes are in line and don't cross other components, I can mount them in holes that are between about 8 mm and 40 mm apart.
Had to look up this acronym. I went with the Fusion default settings. They know more about this than I do. (The janitor at Autodesk likely knows more than I do). I"ll do more digging.
Will send to you. It is the same schematic Yuriy designed and posted at https://www.touchdro.com/resources/adapters/diy/esp32-diy-dro.html but I added the 5V, 3.3V, LED, socket for switch and 12V plug.
You bet I will! Thanks!
If I get these two working, that would be a great next step.
I assumed Yuiy knew where to put them. I have added 100uF to the +5v and +3V3 outs and to ground. When you say they are not in the right place, do you mean in the schematic, or the physical placement on the board? I think I got the schematic right on the scale inputs, but I certainly could be wrong, or Yuriy could have made a mistake.
Will be trying to figure out how to do a ground plane. I soul also look at the traces to the scales, as they
I thought I might be breadboarding it and no further. Also easier to take up spacing errors. SMT is within my skill level for soldering, so maybe next time.
I agree., In this case the circuit is easy, and there are only 3 resistor values (two really, and one outlier) and they are grouped by type, so no searching needed on this one.
Great idea! Once I learn how to label in Fusion.
The chips are notched, which I think will print, and the ESP32 is orientated with the antenna to the outside of the board. (seemed like a good idea to me)
I appreciate you taking the time to give these suggestions! I have a feeling I couldn't afford you.= normally.
I will be looking at adding a ground plane and beefing up the traces in the 5V section. The scales draw a maximum of 200mA, so I'll have to see what I need and how to do it in Fusion.
I haver learned so much already, and I know I have like 8 university courses before I know enough to do much of anything. I may have to look at online offerings...
slow-poke
Ultra Member
0) Cap should be as physically close as possible to the regulator terminals. This is where SMT makes your life a lot easier,
1) Plugable connector, perfect that's exactly the type I was thinking of. There are two types; the ones with the rising cage clamp are more user friendly than the cheap leaf spring type
2) Yes it will, although placing the DB's directly on the board, saves time, lowers cost and reduces failure points
3) The common error is having the hole size the same size as the pin and then after the PCB house "plates up" the hole, the hole is now too small.
4) A good DRC engine is priceless, this may sound like I'm exaggerating, but I'm not.... when I first complete a layout I run the DRC and it's very common to have a hundred or so errors flagged. If I did a good job these are mostly cosmetic, like silkscreen too close to pad, or some generic spacing error, that when you fix the offending component, 20 errors go away at once. The big one to be aware of is when you use nets to describe signals and then call one say 3v3 and the other 3V3, these are treated as different nets and so if not connected will not flag an error.
9) When I do a layout the first step is to layout the ground and power traces. For complex boards muti-layer boards this is easy because you dedicate an entire layer to say ground, and another for 3v3 etc, but with a two layer board, its best to dedicate one layer to (for the most part) to be used for ground and Vcc so a solid plane for the ground with the VCC traces within that ground plane, and then use the other layer for your signals. I will show an example below. This is a very simple low speed design so decoupling and ground planes are not super critical, however this circuit will be working in a potentially noisy (electrically) environment. I have fixed countless designs (some in production) that work fine on the bench (everything is connected correctly), however the moment the device is asked to work in the real world it becomes flaky.
10) SMT is your friend
I'm retired, more than happy to help, everyone can afford free;-)
Two examples showing ground plane and regulator decoupling:
Top example is a two layer board, U5 is a 3 terminal regulator. The bottom blue plane is (for the most part used for ground) and is a solid plane. That same ground is repeated on the top layer where possible. Note C4 and C25 and how they have very short connections to both ground and their voltage terminals
Bottom example is a 4 layer board, the gold plane (mid layer 1) is ground, and the (not shown 3v3 Midlayer 2 is everywhere. Note C7 and C8 and their close proximity to U3
The important point that is often overlooked by electronic hobyists is that they think in "DC" terms, I need to connect pin 1 to ground, I have a trace to ground so I'm good. Now appreciate that the digital signals are not DC they are AC and can change very rapidly << a micro-second, and long narrow long traces have inductance, that inductance will not allow those signals to change quickly, will effectively act as a resistor (if you think in DC terms), so will change the signal level and generate noise during the transition. With very high speed GHz designes (for example the QorIQ above), nothing would work without proper ground planes, even the countless signal traces to things like memory have to have matched lengths so the signals arrive at the same time, it makes for a tricky layout.
1) Plugable connector, perfect that's exactly the type I was thinking of. There are two types; the ones with the rising cage clamp are more user friendly than the cheap leaf spring type
2) Yes it will, although placing the DB's directly on the board, saves time, lowers cost and reduces failure points
3) The common error is having the hole size the same size as the pin and then after the PCB house "plates up" the hole, the hole is now too small.
4) A good DRC engine is priceless, this may sound like I'm exaggerating, but I'm not.... when I first complete a layout I run the DRC and it's very common to have a hundred or so errors flagged. If I did a good job these are mostly cosmetic, like silkscreen too close to pad, or some generic spacing error, that when you fix the offending component, 20 errors go away at once. The big one to be aware of is when you use nets to describe signals and then call one say 3v3 and the other 3V3, these are treated as different nets and so if not connected will not flag an error.
9) When I do a layout the first step is to layout the ground and power traces. For complex boards muti-layer boards this is easy because you dedicate an entire layer to say ground, and another for 3v3 etc, but with a two layer board, its best to dedicate one layer to (for the most part) to be used for ground and Vcc so a solid plane for the ground with the VCC traces within that ground plane, and then use the other layer for your signals. I will show an example below. This is a very simple low speed design so decoupling and ground planes are not super critical, however this circuit will be working in a potentially noisy (electrically) environment. I have fixed countless designs (some in production) that work fine on the bench (everything is connected correctly), however the moment the device is asked to work in the real world it becomes flaky.
10) SMT is your friend
I'm retired, more than happy to help, everyone can afford free;-)
Two examples showing ground plane and regulator decoupling:
Top example is a two layer board, U5 is a 3 terminal regulator. The bottom blue plane is (for the most part used for ground) and is a solid plane. That same ground is repeated on the top layer where possible. Note C4 and C25 and how they have very short connections to both ground and their voltage terminals
Bottom example is a 4 layer board, the gold plane (mid layer 1) is ground, and the (not shown 3v3 Midlayer 2 is everywhere. Note C7 and C8 and their close proximity to U3
The important point that is often overlooked by electronic hobyists is that they think in "DC" terms, I need to connect pin 1 to ground, I have a trace to ground so I'm good. Now appreciate that the digital signals are not DC they are AC and can change very rapidly << a micro-second, and long narrow long traces have inductance, that inductance will not allow those signals to change quickly, will effectively act as a resistor (if you think in DC terms), so will change the signal level and generate noise during the transition. With very high speed GHz designes (for example the QorIQ above), nothing would work without proper ground planes, even the countless signal traces to things like memory have to have matched lengths so the signals arrive at the same time, it makes for a tricky layout.
Last edited:
Tom make sure that your ground plane does not extend under the esp32 antenna section if your using blue tooth or wifi. It can diminish your wireless distance. I usually try to make my ground plane as large as possible. In easyeda you can set it to just that. Not sure with fusion. I didn’t see if you socketed the esp32, so you can swap out the esp32 if you have issues.
Looks like it is coming together great. excited to see how your design works out. I put mine to the back burner some time ago.
Looks like it is coming together great. excited to see how your design works out. I put mine to the back burner some time ago.
Tomc938
Ultra Member
Excellent information! I'll work on the things you suggested. I'm enjoying the challenge and learning involved.
Tomc938
Ultra Member
Socket is a great idea. I didn't do that. And thanks fro the tip on the ground plane.
Now ti figure out how to do it in Fusion. Been doing looks of google searches and video watching. Just like being in school again. Except I think a classroom setting might work better than the scatter-gun approach of Youtube.
I'll take a good teacher over youtube any day. Few youtube artists are really experts. Most of them are just guys trying to get popular or trying to make money the easy way instead of earning it.
IMHO, the absolute best way to learn from an expert is to read a book he wrote. Few real experts will bother with YouTube but most will take the time to write a text book.
Of course, the best teacher of all is experience.
Just my opinion.
The Magic Smoke is known by many names. Experience is just one of them.......Many a lesson has been learned by the release of the magic smoke
Also the aroma of popping electroylic capacitors
When I majored in electronics in highschool a few of the guys made sport of it, lol
Also the aroma of popping electroylic capacitors
When I majored in electronics in highschool a few of the guys made sport of it, lol
Who says learning can't be fun?
140mower
Don
Yup, I duct tape all vent holes on anything I initially wire up these days, it prevents the magic smoke from escaping.......Many a lesson has been learned by the release of the magic smoke
Although I have not used it the fusion electronics module it is supposed to be quite good. I have a friend who makes custom power supplies. He does it all with fusion. He resents that he has to pay for the CAD/CAM part of fusion because he considers it to be useless.
slow-poke
Ultra Member
The fun part of electronic design is that there are so many ways to accomplish any particular design.
Here are a couple of possibilities that will only require about half your available board space and use less components.
I would actually replace TB6 with the DB and eliminate all the hoopla associated with the terminal block to DB wiring, you have enough room.
I'm not sure if Fusion imports Altium files, if it does your welcome to the files for these examples.
BTW with your original schematic. if the intent was to power the board with 5V, your circuit won't work.
Here are a couple of possibilities that will only require about half your available board space and use less components.
I would actually replace TB6 with the DB and eliminate all the hoopla associated with the terminal block to DB wiring, you have enough room.
I'm not sure if Fusion imports Altium files, if it does your welcome to the files for these examples.
BTW with your original schematic. if the intent was to power the board with 5V, your circuit won't work.
slow-poke
Ultra Member
Okay, I have not done a layout for a while and I was curious so I spent an hour to see what all six inputs would look like space wise. Answer we only took 15% of the available board space so the board can be a lot smaller if desired. Requires some clean-up and re-annotation, but at least it has a ground plane and proper decoupling now. I'm off to the pool for my daily swim, will check back later.
Last edited:
Tomc938
Ultra Member
Man, what a difference it makes when you know what you are doing! SMT also offers a lot more flexibility on a smaller footprint.
I think my basic design compares favourably to yours - like a Kindergarten drawing compares to a Rembrandt!
The design and layout I used is from Yuriy, so I can take no credit (or blame). His webpage explains that the board can take inputs from scales that use either +5V or +3V3. The ESP32 has inputs for both values. My assumption is the board functions at 3V3, and the +5V in is in case you want to send 5V out of an IO pin.
Having said all this, I'm going to look seriously at your files. If I can steal them, you can bet I will! One big question in this who process is would what I design end up working in the real world. Up to this point, I think the answer is a resounding "no". But it's been a fun trip!
Last edited:
Tomc938
Ultra Member
Thought I would give a quick update on this thread as the discussion has been fast a furious via email with @slow-poke, who has been helping me figure out just about everything about this schematic and PCB design, and myself. We have now moved beyond that and he has so very graciously stepped in and is designing a board that will be different from what I was working on in one significant detail - it will actually work!
I thought if I connected the components in the fashion indicated in Yuriy's schematic and placed them on a board that things would work. I learned that a circuit like this needs a proper ground plane, not just traces that go from A to B. They just wouldn't handle the current. But they would have made cool black lines on the board. And running power to the various parts at 5V and 200mA with traces is also not a good idea. (refer to comment on ground traces)
I also learned the SMT is definitely the way to go. Much easier to layout and reduces the footprint. Most can be soldered by hand with practice. I ordered 3 practice boards from Aliexpress. This will give me about 900 practice connections. Should be able to be doing a good job after that.
And finally, I have learned that a circuit designed by someone who knows what they are doing is a beautiful thing.
Stay tuned , there is light at the end of the tunnel.
I thought if I connected the components in the fashion indicated in Yuriy's schematic and placed them on a board that things would work. I learned that a circuit like this needs a proper ground plane, not just traces that go from A to B. They just wouldn't handle the current. But they would have made cool black lines on the board. And running power to the various parts at 5V and 200mA with traces is also not a good idea. (refer to comment on ground traces)
I also learned the SMT is definitely the way to go. Much easier to layout and reduces the footprint. Most can be soldered by hand with practice. I ordered 3 practice boards from Aliexpress. This will give me about 900 practice connections. Should be able to be doing a good job after that.
And finally, I have learned that a circuit designed by someone who knows what they are doing is a beautiful thing.
Stay tuned , there is light at the end of the tunnel.