CaptKirk wrote:I think the SOT-23 and TO-92 footprints could co-exist if plate-thrus were used instead of the land pads. Same trick can be used for the smd resistors(?) where a 1/8th watt on end fits in the space of a SMD resistor.However, I do not mind SMD (0805 or bigger) as I actually have taken the time to "learn a word processor program rather than sticking to a typewriter"... 😉 I've invested my $10 in a mag hood, good tweezers, solder paste and tiny tip on my good iron. It is amazing what you can learn to do if you HAVE to do so!!
Keep in mind that a SMD only takes up space on ONE side of the board and a thru-hole takes up space on both sides so that could be where the space issue comes in.
Tried it on 805 it just dosen't work well the solder solder lieks to wick to the reverse side when doing smd . Also unless i am doing modwire on smd parts i always use my bigest chisel tip.
kingofkya wrote:Tried it on 805 it just dosen't work well the solder solder lieks to wick to the reverse side when doing smd . Also unless i am doing modwire on smd parts i always use my bigest chisel tip.
The trick is to pre-solder the plate-thrus then use THAT solder to tack on the SMD. I've done a bunch of that putting the led + resistors in place of the 680 ohm resistor to have the LED channel indicators on Renards...
G'day, G'day... Yes to the "Hammond Enclosure". I like the size, and having a translucent box means I don't have to drill holes for the LEDs.
The original design had TO92 for the 3.3v and a TO220 lying down for the 5v. However, based on where I think the antenna will be connecting (top right corner of the image), I wanted to reduce the vertical height of the components. That led to the SMD 3.3v regulator, and caps. The slippery slope kicked in.. and I figured.. if I'm doing the ONE.. I may as well do the other (oh.. and I also still have three of the 5v SMD regulators sitting in a box in the basement. 🙂 ).
The board has undergone a further revision, as I realized I forgot to include the necessary resistors and pull-down control to have the PIC generate DMX signally. This means one less status LED (not really a big deal).
Then last night I started to rethink the need for electrical isolation, so I'm at it again seeing if I can squeeze all the parts into that tiny enclosure. The NRF24L01+ boards arrived today, so I can do a dry fit test later this evening. I actually enjoy the SMD side as I can try and fit more functionality into a smaller space. I realize it's not for everyone.
Thx for the feedback.
So I got it into my head that it would be a good idea to drop a couple of extra bucks in order to isolate the RS485 driver, from the rest of the circuit. I've run into ground loop issues in the past, and for the prices of a hand-full of components, it seemed wise. The challenge, however, was to try and make it all fit onto the PCB.
Here's a copy of the new schematic (it's still a bit rough to look at), along with the latest 3D render of the board. Yes.. I've pretty much gone completely to SMD. I might try and move some of the component back to through hole, now that I have a layout. There was quite a few more resistors introduced, so space is going to be tight if I go back to everything being through hole.
[EDIT: There are two issues that were discovered in the schematic. Isolated Ground - identified in this thread, and a lack of pull-ups on the output of the 6n137 opto-isolators. Just noting this here in case someone tries to incorporate part of this design into their own work]
... and the rendered model ...
The new NRF modules arrived, and a dry fit say I will have the necessary room. So I'm almost ready to send this out to get a batch of 10 boards made up.
On the schematic, the power pins on the 75179 look like they will auto-connect to Vcc and Gnd, not to VCC_ISO and GND_ISO. Likewise, the RJ45s should connect to GND_ISO.
Also, something is wrong with the named nets in the power supply; it looks like you have Vcc on both sides of IC2 (actually, since the connection dot on the positive end of C4 is missing, you have no named net on the output of IC2, and C4 is connected from Vcc to GND)
/mike


