THT versions of SMD components

[glowredman]

Hi there!

I want to build a weather station, for which I found this guide: https://www.instructables.com/Solar-Powered-WiFi-Weather-Station-V40/

I intend to do it a bit different though:
- I won't use the wind/rain sensor nor the air quality sensor (PMS5003)
- I'll replace the SI1145 UV-index sensor with an LTR390 (SI1145 is disconntinued and only approximates the UV-index)
- I want to use a little SMD components as possible because I have zero experience with it and THT components make it easier to test things first

This is what is left:



Replacing the capacitors, resistors, LEDs and switches/buttons is easy but the rest is not:
- SMF5.0CA TVS diode(I think this one can be substituted with SA5.0CA)
- SS24 Schottky-diode
- MMBT3904 transistor
- SI2301S MOSFET
- LP4060S integrated circuit
- AP6685 integrated circuit
- XC6220B331MR integrated circuit
- Ra-01SC-I LoRa module
- ESP32-WROOM32-U (I want to use a devkit to get around the SMD requirement, I hope the power consumption won't increase significantly by doing so. This is also the reason why I crossed out some capacitors and resistors in the ESP32 box - I hope it's correct)

Maybe these components are just general purpose components or super specialized and their specific behaviour is important or a mix of both, I can't tell. So this is where any help would be much appreciated!

[Vovk_Z]

- SMF5.0CA TVS diode(I think this one can be substituted with SA5.0CA)
- SS24 Schottky-diode
- MMBT3904 transistor
- SI2301S MOSFET
*****
- LP4060S integrated circuit
- AP6685 integrated circuit
- XC6220B331MR integrated circuit

The first portion of your list really can be substituted with any jellybean or not jellybean part.
(2 A 40 V Schottky diode, 2N3904, 2 A n-type low voltage Mosfet).
As for the last portion - integrated circuits - I'm not sure they exist in different case types. Either you use them 'as is' or redesign it totally.

[Kleinstein]

SMD resistors and capacitors are relatively easy to solder. For the start one may want 0805 or larger. After a few parts one usually does not want THT again.  In the layout it helps if there is a little extra space around the parts to make soldering easier.  The MMBF3904 is still only a small challange.

For the LDO there are alternatives in TO92 case - though likely not identical. Depending on the regulator C3 may need to be different (e.g. ceramic vs electrolytic or maybe both combined). The battery charger and protection likely don't have a THT equivalent. The protection part in the schematics looks suspicious - as shown there is very little function.
The power LED should get away with way less current.

[Solder_Junkie]

Trying to stick with through hole parts is getting increasingly difficult as the manufacturers have long since migrated to SMD. Some components are only made in SMD.

Hand soldering of SMD is straightforward and surprisingly easy, even down to 0603 resistors and capacitors. You need fine solder and at least an “Optivisor” type head magnifier.

This YouTube video shows how straightforward hand soldering of SMD parts is.
https://youtu.be/eZtPR8_x0nc

SJ

[tychob]

This YouTube video shows how straightforward hand soldering of SMD parts is.
https://youtu.be/eZtPR8_x0nc

SJ

And how does that apply to 'test things out' which is likely before the design is complete and ready for fabrication? That video assumes a PCB.

[tooki]

Replacing the capacitors, resistors, LEDs and switches/buttons is easy but the rest is not:
- SMF5.0CA TVS diode(I think this one can be substituted with SA5.0CA)
- SS24 Schottky-diode
- MMBT3904 transistor
- SI2301S MOSFET
- LP4060S integrated circuit
- AP6685 integrated circuit
- XC6220B331MR integrated circuit
- Ra-01SC-I LoRa module
- ESP32-WROOM32-U (I want to use a devkit to get around the SMD requirement, I hope the power consumption won't increase significantly by doing so. This is also the reason why I crossed out some capacitors and resistors in the ESP32 box - I hope it's correct)

Maybe these components are just general purpose components or super specialized and their specific behaviour is important or a mix of both, I can't tell. So this is where any help would be much appreciated!

Instead of a basic ESP32 dev module, why not buy one of the ones that has the battery charger built right in? For example, one of the several models available in the Adafruit Feather series, e.g. https://www.digikey.ch/en/products/detail/adafruit-industries-llc/5400/16129667 or https://www.digikey.ch/en/products/detail/adafruit-industries-llc/5303/15786745 (which even has a BME280 temperature/humidity/pressure sensor built in, but it's an ESP32-S2 so you'd want to make sure that will work for your application). Both of those have USB-C and a battery charger, so they basically comprise everything on the first schematic sheet.

[Solder_Junkie]

And how does that apply to 'test things out' which is likely before the design is complete and ready for fabrication? That video assumes a PCB.

Exactly the same as everyone else using parts that are SMD only. You can have a prototype board manufactured very quickly if you make it yourself using heat transfer paper, plain copper PCB and a ferric chloride bath. I can make one in an hour. Alternatively PCB prototype boards are cheap enough, if you can wait a few days, JLCPCB cost as little as 4 or 5 GBP inc delivery and tax for 5 boards, or a bit more for 24 hour service and courier delivery by DHL.

SJ

[tychob]

And how does that apply to 'test things out' which is likely before the design is complete and ready for fabrication? That video assumes a PCB.

Exactly the same as everyone else using parts that are SMD only. You can have a prototype board manufactured very quickly if you make it yourself using heat transfer paper, plain copper PCB and a ferric chloride bath. I can make one in an hour. Alternatively PCB prototype boards are cheap enough, if you can wait a few days, JLCPCB cost as little as 4 or 5 GBP inc delivery and tax for 5 boards, or a bit more for 24 hour service and courier delivery by DHL.

SJ

And thats as 'easy to test things first' as laying out components on protoboard once you have the components?

[Doctorandus_P]

I also find SMT parts easier to work with on a PCB than THT parts.

* You don't have to cut leads to length, deal with bent leads, etc.
* You don't have to turn the PCB around all the time and work with the (confusing) mirror image from the bottom view.
* You don't have to turn the PCB around all the time so part's won't fall of while you are attempting to solder them.
* SMT parts are easier to source. there is much more choice.

But you do have to choose your battles.
0805 resistors are easy to do by hand, they also fit quite nicely on matrix boards with a 2.54mm pitch.
0603 is considered doable by a lot of people without extra tools such as stereo microscopes.

I did buy a stereo microscope myself a few years back though.
I should have done that 10+ years earlier.

Even when soldering big parts such as the old DIP IC's, you see so much more details that I can't do without anymore.
You see the solder meniscus wetting the pad and the pin during soldering.
You see the flux working on the edge between the solder and the solid.
You see the flux boil of during soldering.

That microscope has improved my soldering quality a lot, and it also gives much more confidence in the reliability of your work.

For testing and verifying properties of SMT parts, I sometimes make little adapters to make them breadboard friendly, or solder them directly to 2.54mm headers. You can for example quite easily solder two of the pins of an SOT23 to pins of a header, and use a short wire for the third pin.
One of the remaining problems is part recognition. I recommend you keep a list of the SMT codes of the parts you use.

[tychob]

I also find SMT parts easier to work with on a PCB than THT parts.

And if you don't have a custom PCB? Is 1.27mm matrix board with SMT as easy to use as 2.54mm perf board? Or is there a form of matrix/perf board with a larger pitch than 1.27mm that can solder 1.27mm pitch ICs to as easily as DIPs to 2.54mm perfboard (ie not to a breakout board and then also do the 2.54mm pins)? Theres no point switching to SMT just to use passives. ICs are essential.

Surface mount soldering is not the challenge. Its what do you solder to that's the biggest shift in surface mount for prototyping, and that needs to be addressed. Not wanting to jump straight to a custom PCB is a valid desire for beginning hobbyists. So long as through hole remains the preferred technology for the interconnect fabric for these low end one-offs/prototypes, its understandable to want to stick with through hole parts rather than have to do both surface mount and through hole soldering for every part.

[donlisms]

Little boards that adapt SMD devices to 2.54 mm pins that will fit a through-hole prototyping system are readily available and inexpensive. You will have to solder once.

Buy lots and lots of the cheapest SMD parts you can find, and practice practice practice. It might be cheaper and better experience to design your own soldering practice boards than buying the ones that are out there.  There's an advantage to soldering hundreds of a part (e.g. 0805, SOT 23) rather than, say, 2 or 12 of them.

Then populating the THT adapter boards will be super easy, and making an SMD board will be super easy.  Eliminate the obstacles with simple one-step-at-time mini-projects. And "practice" is an extremely valuable concept too often ignored.

[tychob]

Little boards that adapt SMD devices to 2.54 mm pins that will fit a through-hole prototyping system are readily available and inexpensive. You will have to solder once.

With a DIP I only have to solder 0 times. 0 is always easier than 1.

Then populating the THT adapter boards will be super easy, and making an SMD board will be super easy.  Eliminate the obstacles with simple one-step-at-time mini-projects. And "practice" is an extremely valuable concept too often ignored.

Why though? What benefit other than a very marginal cost difference (~30c on a TL072 opamp for example, which can be very easily eaten up by the adapter if you don't buy bargin bucket adapters and header pins), do you get if there is an equivalent DIP available that provides the through hole capability in a single pre-packaged offering? Why would you want to use a SOT23 on an adapter over a TO-92? Or an 0805 vs an axial resistor?

[donlisms]

Little boards that adapt SMD devices to 2.54 mm pins that will fit a through-hole prototyping system are readily available and inexpensive. You will have to solder once.

With a DIP I only have to solder 0 times. 0 is always easier than 1.

Then populating the THT adapter boards will be super easy, and making an SMD board will be super easy.  Eliminate the obstacles with simple one-step-at-time mini-projects. And "practice" is an extremely valuable concept too often ignored.

Why though? What benefit other than a very marginal cost difference (~30c on a TL072 opamp for example, which can be very easily eaten up by the adapter if you don't buy bargin bucket adapters and header pins), do you get if there is an equivalent DIP available that provides the through hole capability in a single pre-packaged offering? Why would you want to use a SOT23 on an adapter over a TO-92? Or an 0805 vs an axial resistor?

I might prefer THT parts - if they are available.  Maybe for this one, we can swap in any OPAMP that's handy, but as has been said, it's more and more likely that a THT package is not available these days.  When you work through a design with mixed... hole-osity, the reality of it stares you in the face, and it's pretty clear what you'll need I do with SMD vs THT.

What I was trying to say, in case I wasn't clear (which wouldn't surprise me!), is that you can fairly easily get comfortable with both technologies, and mix and match, or convert, as necessary.  I'm not advocating one or the other; me, I need both, depending.

[Yuu]

I recently bought one of those SMD soldering practice kits on amazon. It was great and actually not hard at all. I feel very comfortable with SMD soldering now. I definitely recommend.

[schmitt trigger]

I hear you loud and clear. With THT components, a breadboard and some solid jumper wire, one could test a concept almost immediately. Been using them myself for 45 years.

Having said … the fact is that the World has been moving away from THT for a while now, and during the Great Chip Shortage, this trend accelerated -enormously, as chip manufacturers consolidated products and processes.
Most if not all new and interesting ICs are not even released in THT format anymore. Even some legacy ICs are now becoming only available in SMT format.

So the choice is: adapt to the brave new world or quietly fade into the sunset. 

[sparkydog]

I'm surprised no one has pointed out that there are plenty of components that are only available in THT form.

The projects I've been working on, I've been using almost exclusively THT (in fact, I have exactly one exception). Many relays (power relays, anyway) are only available as THT. Certain capacitors (especially electrolytic) are only available as THT, and in some cases you want THT because the (larger!) physical size is actually critical to their design. Many headers are THT, and when there's a choice, personally I would prefer THT for the sake of mechanical stability. Also, I've seen an awful lot of "SMD" components that are clearly just THT components with the pins bent sideways and clipped; anything of this nature is likely to be around in THT form for a while, yet.

That said, I've also seen THT capacitors that are literally potted SMD components with leads soldered to them. 🙂

[Watth]

- I want to use a little SMD components as possible because I have zero experience with it and THT components make it easier to test things first

It is true that THT allows using a breadboard.
Nevertheless, SMD components can be used on breadboards: smallest (resistors, capacitors, LED, ...)can be soldered to pin headers. Bigger capacitors can be soldered to wires or leads. For ICs, there are cheap adapters.
Hope that helps.

[JustMeHere]

Don't be afraid of SMD.  Get a rework station.  Pratice a little. Once you get used to it, you'll find SMD much easier than through hole.  The boards are much easier to route because you don't have the holes going through all of your layers.  Since you can 4 and 8 layer boards for cheap, it's the way to go.

I find prototyping with PCBs much easier than breadboards.  There may be a delay in getting them from China, but they only cost $2 and will save you hours of frustration due to loose connections.

[tychob]

I find prototyping with PCBs much easier than breadboards.  There may be a delay in getting them from China, but they only cost $2 and will save you hours of frustration due to loose connections.

What do you use to reliably fix design mistakes and or if you realize you need an extra chip? As opposed to perfboard where you can just solder more jumpers to the board.

[schmitt trigger]

Easy; you analyze the circuit a little more, prior to laying out the board.

One recent instance: I built an amplifier for an AWG. I required it to have very low offset. The traditional and sure-fire approach is utilize a trim pot. But I wanted an auto zero solution, with an ultra low offset opamp to null the offset, but I was concerned about unstability issues. What I did is include in the board both the trimpot and the untested auto zero circuit, selectable with a jumper.

I built the board first the traditional way. After it worked to my satisfaction, I soldered the components for the auto zero and changed the jumper position. It worked! Although I had to modify some of the RC values, but as others have mentioned, desoldering and replacing SMT components is very simple. 

If for some reason the auto zero circuit didn’t work and I could not fix it, I would simply replace the jumper connection to the trimpot and would have called it a day.
Another trick: if in a particular circuit an opamp section is not used, I still lay in front of it several 0805 pads, to connect them in whichever configuration could be required later.
I also place in unused spaces the footprints for SOT23 transistors, again with pads to tack components to.

[tychob]

Easy; you analyze the circuit a little more, prior to laying out the board.

If what I wanted to do was just math and analysis in a computer program, I'd be writing software, not trying to do circuit design. If I wanted a system I knew would work, I'd use Arduino shields. As a hobbyist, there are only really 3 reasons to try and build a circuit yourself these days 1) You have already validated your design of sticking together dev boards and want something product like, 2) You are an advanced hobbyist working with high bandwidth signals that aren't easily routed between dev boards, 3) you actually want to play around with physically building circuits, rather than only working with abstract math (and simulators are abstract math). Its that 3rd option where through hole parts and non-committed prototyping techniques come into there own. The value may not be in the end product itself, but the process of trying things and seeing what does and doesn't work along the way. After all, usually a better end product is available cheaper commercially.

[schmitt trigger]

Tychob: as I mentioned previously I totally understand you. You are preaching to a wholly converted individual. 

I have come long, long ways. My first projects that I built used vacuum tubes. Then there were germanium transistors, expensive and fragile as heck. Then silicon transistors.  Up to then one could still use the point to point wiring, utilizing solder lug terminals, as shown in the photo.

Then came ICs in DIP form. Those definitely one could not build a project, without one learning how to do your own boards, which in the pre-CAD days, required a lot of messy and slow manual work. Thank goodness for protoboards. I still have one that I purchased over 30 years ago.

Later came the digital circuits and microcontrollers, which required one to learn assembly language. 
Nowadays is the time for SMT components, and worse, for the fine pitch or leadless components, some of which I believe are the work of the devil.

The point I am attempting to drive is that, thorough these past 50 years I have had to adapt.  Do I always like it? Do I enjoy that any DSP can provide a much better filtering function, than anything I could ever accomplish with opamps? No, many timesthe answer is not really. But I recognize that the world has moved on.

I also miss the ruggedness and the warm glow of tubes.

[JustMeHere]

I find prototyping with PCBs much easier than breadboards.  There may be a delay in getting them from China, but they only cost $2 and will save you hours of frustration due to loose connections.

What do you use to reliably fix design mistakes and or if you realize you need an extra chip? As opposed to perfboard where you can just solder more jumpers to the board.

Usually the PCB design tool like Eagle or Kicad make it easy to discover your mistakes in the design phase.  If you can give the deign tool the schematic the OP provided, the tool will show you where to connect things.  If you make a big error, you can use a bodge wire.  If that fails, just order a new board.  They are $2 for set of boards. 

The stability you get in a 4 layer board is well worth the effort of learning how to use a PCB tool.  Connections to power and gnd are literally everywhere. The ground plane gives you excellent EMI protection.  ADC circuits aren't full of noise, and power is very stable.

[tychob]

Later came the digital circuits and microcontrollers, which required one to learn assembly language. 
Nowadays is the time for SMT components, and worse, for the fine pitch or leadless components, some of which I believe are the work of the devil.

The point I am attempting to drive is that, thorough these past 50 years I have had to adapt.  Do I always like it? Do I enjoy that any DSP can provide a much better filtering function, than anything I could ever accomplish with opamps? No, many timesthe answer is not really. But I recognize that the world has moved on.

In which case, is there any point for a new beginner to even try to progress beyond what circuits can be achieved on a breadboard with passives and a 555? It feels impossible for anything you can do with circuitry to be better than the software domain. So a dev board furfills all the needs.

At least for me, I'm not going to adapt to embedded engineering. If that is the future than I don't see a way to enjoy that as a hobby. But thats me as a burned out professional software developer who doesn't want to touch another line of code when I leave work.

Usually the PCB design tool like Eagle or Kicad make it easy to discover your mistakes in the design phase.  If you can give the deign tool the schematic the OP provided, the tool will show you where to connect things.  If you make a big error, you can use a bodge wire.  If that fails, just order a new board.  They are $2 for set of boards. 

For 50x50mm. If your design isn't just a microcontroller + sensors, thats a serious constraint. And if you want to use more than one board, the cost of interconnects that don't lose you all the benefit of the ground plane adds up quickly.

[abeyer]

Later came the digital circuits and microcontrollers, which required one to learn assembly language. 
Nowadays is the time for SMT components, and worse, for the fine pitch or leadless components, some of which I believe are the work of the devil.

The point I am attempting to drive is that, thorough these past 50 years I have had to adapt.  Do I always like it? Do I enjoy that any DSP can provide a much better filtering function, than anything I could ever accomplish with opamps? No, many timesthe answer is not really. But I recognize that the world has moved on.

In which case, is there any point for a new beginner to even try to progress beyond what circuits can be achieved on a breadboard with passives and a 555? It feels impossible for anything you can do with circuitry to be better than the software domain. So a dev board furfills all the needs.

At least for me, I'm not going to adapt to embedded engineering. If that is the future than I don't see a way to enjoy that as a hobby. But thats me as a burned out professional software developer who doesn't want to touch another line of code when I leave work.

Usually the PCB design tool like Eagle or Kicad make it easy to discover your mistakes in the design phase.  If you can give the deign tool the schematic the OP provided, the tool will show you where to connect things.  If you make a big error, you can use a bodge wire.  If that fails, just order a new board.  They are $2 for set of boards. 

For 50x50mm. If your design isn't just a microcontroller + sensors, thats a serious constraint. And if you want to use more than one board, the cost of interconnects that don't lose you all the benefit of the ground plane adds up quickly.

You seem to have an arbitrary set of constraints you've decided to impose on yourself, are unhappy with the results of those constraints, and yet won't consider any alternatives that aren't some imaginary perfect solution without any tradeoffs. 

Your hobby is yours to enjoy as you wish... but I'm not sure you're going to get the validation you seem to want arguing the point here.