Hello, everyone.
I’ve designed a pretty simple board, and it’s working fine on the breadboard, but when I try to export the file to a Gerber, Fritzing crashes and closes, and the folder contains only the .pnp file.
I’ve tried reinstalling Fritzing, but it didn’t help. Is there an error in my file that’s stopping it from working?
Thanks.
Thermoprobe.fzz (15.1 KB)
Welcome to the forum.
We need a few more details to try to help (much). When I tried it, gerber export worked with your sketch. What operating system are you using, and which version of Fritzing? Maybe how much memory too.
It worked for me on Fedora 31 (64 bit) and Fritzing 0.9.4b on a system with plenty of memory.
You say the circuit is working on breadboard. Looking at the circuit and breadboard view though, I see something suspicious. both ends of the potentiometer are connected to 5V. I think that is supposed to be used as a voltage divider, to provide a reference voltage to control the brightness of the display. In that mode, one end should go to 5V, the other to ground. The centre goes to the LCD. I assume you copied this from a project online someplace. A hint that the connection is wrong, is that the one connection to 5V is a black wire on breadboard view. The rest of the black wires are going to ground, and the red to 5V (or other +/power). Which is the convention.
On the PCB view, a design rules check (on the routing menu) complains that one wire is too close to some of the arduino pins. The trace coming off of the variable resistor. From the look of that, it was auto-routed. The trace routing could be cleaned up and simplified a lot. Especially if the parts can be moved around on the PCB.
Windows 10, 15 Gigs of RAM, Fritzing 0.9.3
You say you have 0.9.4b, but a Fritzing “Check for updates” says there’s no newer version than my 0.9.3 ?
EDIT: >>> I looked directly on the Fritzing site, and found a link to 0.9.4.64. I installed it, and both the Export to Gerber and Design Rules Check now work. My only problem now is the ‘extra’ traces Fritzing has added to the schematic and PCB.
[quote]
I see something suspicious. both ends of the potentiometer are connected to 5V. I think that is supposed to be used as a voltage divider, to provide a reference voltage to control the brightness of the display. In that mode, one end should go to 5V, the other to ground.[/quote]
Good catch… Its right on the ‘real’ breadboard, but I’d done it wrong on the ‘virtual’ one. I’ve corrected it.
Hmmm, that brings up another bug. When I try to run the design rules check, the progress bar gets to 2%, and Fritzing crashes and closes.
I’ve been studying the traces, and I see a bunch of traces on the PCB, and wires on the schematic, that don’t correspond to actual wires on the breadboard.
For example, Fritzing added a wire connecting LCD 15 & 16 - the +5V & Gnd for the LED, and one connecting the Arduino +5V & Gnd. Another connects J3 pins 1 & 3, which ar +5V & Gnd for the sensor.
If I delete them, and run autoroute again, it puts them back,
Thermoprobe.fzz (14.0 KB)
Hello Rickinnocal,
I’ve taken a look and I think the same thing happened to be me some time back, basically my guess is that you once had it shorted on the breadboard side, and autorouted the PCB at the time, then since corrected the mistake on the breadboard, but didn’t re-autoroute the traces on the PCB.
You can see here that this dotted wire indicates the short on the breadboard: 
A tick I learned from Peter is that if you go into PCB view then ‘Routing --> Select all traces’ then hit delete it removes all traces, hit autoroute again and now there’s no short.
James.
I make techy YouTube videos you can find here: CuriousJames.net
With the voltage divider connection confirmed, here are a couple of alternate pcb layouts and trace routings that demonstrate what I meant about simplifying. These could be tuned a bit further, depending what your constraints and goals are. I also shrunk the pcb to fit the layouts, since usually you are paying by the board area.
I also cleaned up the schematic view, to make it easier to read, but did not do the full routing there. Just enough to link to the added power and ground symbols.
Thermoprobe.fzz (15.2 KB)
Thermoprobe2.fzz
Thank you, I’ll remember that trick.
Thank you very much. These layouts are really easy for me to read and follow with the naked eye, too, unlike the autorouted one Fritzing generated.
May I ask, did you start by running autoroute then manually tweaking it, or did you just lay it all out manually?
I think, since I’m trying to learn this, I should maybe just stick to doing it manually. Then at least if there’s a MCF I can see where and how I did it.
After looking at the ratsnest lines, I “played” with the part positions a bit to get what looked like reasonably simple routes. Then I manually placed the clean groups of parallel traces, then the majority of the power and ground traces. Then I let auto-route finish up, then manually deleted and rerouted what autoroute did, one trace at a time, using what it did as a (very rough) template. With additional tweaks of the part positions as I got a better look at where the traces were going. It is an iterative process. Once the layout is mostly finalized, go back back to clean up exactly where the bend points go, and the spacing between pads and traces. This layout is very simple, with lots of space for the traces. For tighter situations, tweaking the trace positions to maintain spacing while minimizing board usage outside of the parts, squeezing the parts closer together can be time consuming. There are a lot of tradeoffs that only experience will teach.
The goal is usually to use minimal board area, short traces, minimal vias, 45 degree (and less) bends on traces. While keeping track of physical constraints, like where the off board connections need to go. I am not an expert, but have a good eye for the patterns.
There can be other electronic considerations too, like wanting filtering/bypass capacitors close to the power connections. Avoiding signal cross talk between traces. Lots more when starting to get into high frequency/speed circuits. With the circuits typically seen here, those tend to be less important (have less impact on how well the circuit works).
