Author Topic: Correct way to show screw holes for manu  (Read 1710 times)

0 Members and 1 Guest are viewing this topic.

Online rs20

  • Super Contributor
  • ***
  • Posts: 2013
  • Country: au
Re: Correct way to show screw holes for manu
« Reply #25 on: March 04, 2018, 02:11:42 pm »
Just a side note; I received some third-hand advice a while ago which goes like this: if the through-hole version of a part is actually no larger than the surface mount version, then just use the through-hole version. For example, electrolytic capacitors and big DC jacks, you want to use the through hole versions. Reasons include:
  • Greater mechanical strength: the part ends up practically riveted in place as opposed to relying on the much weaker bond between copper traces and the underlying FR4.
  • Easier SMD reflowing: big parts catch the air and get blown around, plastic can start degrading, you can give the electrolyte in a cap a harsh introduction to the world, etc. Just popping the through-hole part in and soldering it by hand is much faster, easier, and less stressful to all the components of the board.
  • (Often) lesser parasitics: Now of course your 0805 and 0603 SMD components have way better ESL than pretty much any through-hole components, but we're not talking about those. We're talking about SMD components that are the same size as their through-hole counterparts; electrolytic SMD caps of this type are basically just their through-hole counterparts with their legs splayed out and a plastic base added. This makes the leads longer, worsening the ESL, and you lose the lovely property of through-hole electrolytics that they dive directly down to the power and ground planes (where present) with no intermediate vias required.
Exceptions to this rule:
  • May be more difficult for automated assembly (irrelevant to the home hobbyist)
  • If leads sticking out the other side of the board are a problem (bit of a stretch)

On this basis, if I were you, I'd be using the through-hole barrel jacks and electrolytic caps instead of the SMD ones you're using there.
 

Offline paulca

  • Super Contributor
  • ***
  • Posts: 1700
  • Country: gb
Re: Correct way to show screw holes for manu
« Reply #26 on: March 09, 2018, 06:48:09 am »
On this basis, if I were you, I'd be using the through-hole barrel jacks and electrolytic caps instead of the SMD ones you're using there.

Thanks.  I agree, but a lot of this is about exercise and learning through playing.  I just wanted to make my first SMD board as SMD as I could make it.  It still has THT pin headers and switch wiring.

Anyway, the board arrived!  Oddly it arrived before the one I ordered 2 weeks before it. (The binary clock).  Note to self, DHL is worth it, EMS not so much.  The Binary Clock is stuck with Parcel Force because customs were a dick about it, so hopefully I only have to pay the £1.80 or so VAT and the extortionate admin fee to parcel force to release it.  Unless customs have valued it higher or consider it a commercial import.

Anyway... "Piper Lamp".

* The DC jack custom foot print aligns perfectly, the two drill holes for it's mechanical stays align perfectly. 
* It fits in the case.
* The screw holes align!
* The lid closes with the DC barrel jack in place, although it's tight, probably 0.5mm too high, but that will not both how it looks.  I could sand the top of the barrel jack or bend the SMD feet up a fraction, but I don't think I'll bother.

All looks great.  I might try and make one up tonight to test it.  I have 5 sets of components and 10 boards, I can make a few of them and give the best one to my daughter.

For completeness, the rest of the device has come along.

A section of 2" white PVC waste pipe from the hardware store fits absolutely perfectly into the ring where the lamp holder would go in the lamp shade, perfectly in that it actually holds the pipe in place. 

Around this bit of pipe a 1m length of WS2811 LEDs (12V 3xseries sets) wraps in a spiral.  The glue, although it claims "3M" of course is pants and it has started to peel at the ends which are under more stress.  A dob of thick CA glue will hold it.

The chaining wires on the top I will remove by pulling or melting the hot glue away and desoldering them, they won't be needed.  The other end the tail + - wires will be cut right back, hopefully in such a way that they can never short againt each other and the remaining + - and Data go into the 3 pin heading in the device.

I will drill small holes to add cable ties for durability and keep the pipe in place, then add hot snot around the rim and cable tie it in place.  It has to survive a few knocks as and table lamp will get.

Of course the final piece is a 12V DC wall wart.

The lamp is free standing as it doesn't require air flow for cooling.

A lot is obviously involved in programming the LED patterns and the modes selected by the switch.  Having an ISP header is all but essential of course.

Fingers crossed that everything works out.
"What could possibly go wrong?"
Current Open Projects:  3 Channel Audio mixer with DAC, BT, pre-amps and h/phone amp, WS281x LED controller Version 2 5V/5A w/Atmega328, FY6600 Power supply, 5A DC Load (Still!)
 

Offline paulca

  • Super Contributor
  • ***
  • Posts: 1700
  • Country: gb
Re: Correct way to show screw holes for manu
« Reply #27 on: March 09, 2018, 08:27:18 am »
So it all soldered together, sort of okay.  I had issues with the paste melting and ended up with the hot air at 380C to get the capacitor to reflow.  I think I will pre-heat the board, next time and leave the solder paste out of the fridge longer.  I will also do the DC Jack and Capacitor with the iron.

It powered on, in that it took 12V and didn't heat up, melt, let smoke out or consume a lot of current.  All voltages registered okay.

So I connected a programmer to it and suddenly it started consuming current, 200mA of it and the PSU went into current limit.  I disconnected it and tried powering it from the USBAsp alone.  No smoke.  So I checked the two ICs for heat.

I have a blister on my finger now.  The ATTiny was about to burst into flames, hot enough to badly burn my finger.

So I checked again the programmer header and found... GND and 5V are backwards.  Damn.

Well wired a quick adaptor, connecting back up and ... no response from the ATTiny, but at least it doesn't burn me now and it's wired right.

It's getting late, so I am going to leave it for tonight.  Either I have cooked the ATTiny, I have 4 others... or it survived but there is a problem programming them.  It's happened before that I couldn't program them due to weird factory fuse settings.  Tomorrow I will solder one of them onto an adapter and check it programs on a breadboard or if I need to refuse it.  If I need to refuse it, it will be annoying as how can I HV program it without soldering it to something!

I'm actually hoping I cooked the ATTiny or the 5V regulator and don't need to HV refuse the MCUs.

I can live with the backward header power pins it will only be programmed by me and probably only once.
"What could possibly go wrong?"
Current Open Projects:  3 Channel Audio mixer with DAC, BT, pre-amps and h/phone amp, WS281x LED controller Version 2 5V/5A w/Atmega328, FY6600 Power supply, 5A DC Load (Still!)
 

Online rs20

  • Super Contributor
  • ***
  • Posts: 2013
  • Country: au
Re: Correct way to show screw holes for manu
« Reply #28 on: March 09, 2018, 09:17:31 am »
You also need to consider whether you've killed your programmer -- that's a definite possibility (given that you effectively provided a strong -5V to its Vtg line).

Btw, this is why you slowly ramp up voltage and current on your power supply when connecting your board for the first time -- you wouldn't have done any damage if you had a 10mA current limit in place.
 

Offline paulca

  • Super Contributor
  • ***
  • Posts: 1700
  • Country: gb
Re: Correct way to show screw holes for manu
« Reply #29 on: March 09, 2018, 10:43:28 am »
It's okay I tested the programmer still functions by programming a Nano with it, it does, though I was probably lucky.
"What could possibly go wrong?"
Current Open Projects:  3 Channel Audio mixer with DAC, BT, pre-amps and h/phone amp, WS281x LED controller Version 2 5V/5A w/Atmega328, FY6600 Power supply, 5A DC Load (Still!)
 

Offline paulca

  • Super Contributor
  • ***
  • Posts: 1700
  • Country: gb
Re: Correct way to show screw holes for manu
« Reply #30 on: March 12, 2018, 09:50:14 pm »
Well it took me quite a while to get it working.  I had issues with the fall time on the data waveform.

I fixed it with a pull down botch resistor.  However as was suggested in the MCU sub forum I should double check that FastLED library did indeed set the port as output.  if it didn't then (I might be wrong), setting the port high turns on the pullup resistor, setting it low turns the pull up off, which might explain the slow fall time, ie the port is not sinking current to GND but just disabling the pull up which would show as a ski slope fall time.

Other than that the board works fine.  Fits perfectly in the case and I was able to solder the switch onto the wire pads instead of fitting the switch to the case.  The switches I got just high enough that the top is flush with the case and as I drilled the hole in the lid a millimeter wider it is easy to press without protruding a lot.  It's a little "botchy" but you wouldn't notice at first glance.  Switch is secure.

Just the programming to do now.  Spent Saturday evening devising about 4 different rainbow patterns, but suddenly they all started animating at the same speed, so I expect I have variable scope leakage or have done something blonde.

Soak tested it on bright white 100% for 4 hours and the wall wart PSU became warm, but not hot.  The bench supply reported this as 1.1A, the PSU is rated for 2A.  100% cold hard white is the worst case usage.  So all good.  Even survives being switched off and on repeatedly, although I couldn't switch it off and on again faster than the cap discharges to try and lock it up.

The only minor issue, which is being pedantic.  On the bench PSU the 100uF cap pulls a lot of current at first power on.  The PSU flicks to CC mode for half a second.  I'm sure this won't stress the wallwart that much.

« Last Edit: March 12, 2018, 09:54:41 pm by paulca »
"What could possibly go wrong?"
Current Open Projects:  3 Channel Audio mixer with DAC, BT, pre-amps and h/phone amp, WS281x LED controller Version 2 5V/5A w/Atmega328, FY6600 Power supply, 5A DC Load (Still!)
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf