8
\$\begingroup\$

I have built this circuit on a solderless breadboard, and now I want to build a permanent version. I have all the parts, I want to build it today.

How should I put this on a perfboard?

enter image description here

Note that I am answering my own question to create a reference for the future. Seeing several very poorly made perfboards recently on this site motivated me to create this Q/A. I realize that some of this will be opinion based, but some concepts are universal and have been used since the 1970s or earlier. Feel free to add your own answer describing your build methods. An example with an IC would be great. I won't accept any answers to encourage more answers.

Here a poorly constructed perfboard caused a lot of frustration: Astable Multivibrator Not Oscillating

Discussion on meta: Reference Question for Soldered Breadboard/Perfboard

\$\endgroup\$
2

4 Answers 4

7
\$\begingroup\$

I am answering my own question to create a reference for the future. Seeing several very poorly made perfboards recently on this site motivated me to create this Q/A. I was still in high-school when a master build technician taught me most of these techniques.

This is a general-purpose method good up to about 100 kHz or so. Although you can potentially go higher, I have an old processor perfboard that runs at 1 MHz. The intended audience is a beginner who shouldn't be building high-speed digital or RF circuits yet.

In general, you want to emulate the layout and traces of a single layer PWB. I have been using this method for almost 50 years, and it will continue to be viable as long as through-hole parts are available.

Start with an appropriately sized board. Unless you are building something that must be small, use a board large enough so you don't waste time trying to get the components to fit.

In the past, I used perfboards without copper pads. Now perfboards with copper pads are cheap enough that there is no reason not to use them. You can buy a large bag of assorted sizes for 14 USD.

enter image description here

enter image description here https://www.amazon.com/Prototype-Soldering-Universal-Printed-Electronic/dp/B079DN31SW

Most through-hole integrated circuits are still using 0.1 inch (2.54 mm) spacing. This is the most common type of perfboard and this is what you should use.

For most breadboarding, I don't recommend perfboards that have traces connecting pads. They are more of a hindrence than a benefit, they will force you to use odd component layouts that will make troubleshooting more difficult.

enter image description here https://www.amazon.com/SchmalzTech-Protoboard-Solderable-Breadboard-Electronic/dp/B0C4HW97V7

Plan your component layout. If you are starting with a well done schematic, follow the general layout of the schematic.

Many circuits have a power bus and a ground bus. I normally put these at the top end and bottom end of the board. I use 20 AWG for these because the wire is stiffer and will stay in place better. I loop the ends to the front side and back to the bottom. The loops are used to make the connection to the power supply, test equipment, and other modules.

This example doesn't have any connections besides power. When I do have I/O, I may use more wire loops if this is a temporary setup. Or I may use male header pins since I am often connecting to an MCU. Or, I may use screw terminals.

For higher current circuits, I will use 16 AWG for the power buses. But this size won't fit in the holes, I need to drill them larger.

Decoupling capacitors should be close to any ICs. An old guideline is one 100 uF for the board, plus a 0.1 uF ceramic for each IC. The wires to the IC for the ceramics should be as close as possible. I sometimes put the decouplers on the back of the board in between the IC pins. The decouplers here are not critical, I put 2 out of habit.

Insert the components into the board. For those that will tend to fall out, apply a small amount of solder to hold them in place.

For many of the connections, the component leads will be long enough to make the connection. You should always start with a good mechanical connection, loop the lead around the next component lead. (these loops should be a little tighter)

enter image description here

When the component leads won't reach the destination, I use 24 AWG bare solid wire. You should only use tinned copper wire. Untinned wire oxidizes quickly. Don't cut the protruding leads until after you solder.

enter image description here

For very long connections you may need to use insulated wire. I like teflon insulated wire, but it is expensive, so I usually use silicone insulated solid wire. PVC insulation melts easily, not recommended. You can put the insulated wire connections on the top or bottom of the board. (note that this example doesn't have any)

Solder the connections. 60/40 or 63/37 tin/lead is easier if you are a beginner. I use a solder with a rosin core, then I don't need to use flux. Don't use anything meant for plumbing.

A beginner will tend to use too much solder. You want the solder at the joint to be concave, a concave layer of solder will have good wetting to the component lead. A blob won't always indicate a bad joint, but you can't tell, the blob is hiding the evidence.

You should use a temperature-controlled soldering iron. I use very small conical tip for 95% of my soldering. I use an old Weller that uses the Curie Effect to control the temperature. The temperature control is in the tip, so it isn't convenient to change the temperature. They don't sell this version anymore, but stations with a digital readout are about the same price that I paid 40 years ago. I use 700 F tips for most work (assuming tin/lead solder). This is 370 C, or about 350 C, a common recommended temperature. Some people use a lower temperature, it is a personal preference. What is important is not to keep the iron on the joint for more than a few seconds so you don't overheat the component. You want the solder to flow within roughly 2 seconds. Then hold the iron for about 1 more second to assure good flow.

The small boards often don't have space to label the connections. I often use fingernail polish to color code the connections.

enter image description here

enter image description here

enter image description here

Edit: Here is an MCU perfboard with some of the concepts missing on the first perfboard.

Insulated wires. Use solid wire. If you are emulating a single layer PWB, then put them on the front side; but I sometimes put them on the back.

Male header pins. The bottom leads on some are a little too short to attach wires to, I buy those with slightly longer bottom pins.

Sockets for ICs. If the board may see condensation you might want to omit sockets, but you generally want them. Sockets with machined contacts are more reliable. The pins will be round when viewed from the top. A common beginner's mistake is to bend a pin underneath the IC. The issue won't be easy to see, and it may be barely touching and work for a while.

enter image description here

enter image description here

enter image description here

Machined Socket: enter image description here

\$\endgroup\$
10
  • \$\begingroup\$ You mention this is only suitable to 100kHz bandwidth. For the purposes of defining what you mean by 'high speed digital', it would be useful to also add what rise time is associated with that. Is a 10ns rise time greater or less than 100kHz? \$\endgroup\$ Nov 27, 2023 at 14:21
  • \$\begingroup\$ A rule of thumb for rise time is (0.35 / f_BW). So 0.35/100kHz corresponds roughly to a rise time of 3.5 µs. The other way around you can calculate that 10ns rise time needs a bandwith of ~35MHz. \$\endgroup\$
    – jusaca
    Nov 27, 2023 at 14:40
  • 2
    \$\begingroup\$ Placing components on top of each other is bad practice and not allowed by standards like the IPC ones. We shouldn't teach people to do that. \$\endgroup\$
    – Lundin
    Nov 27, 2023 at 15:17
  • 1
    \$\begingroup\$ Oh, I would be more than happy to use this construction method for 10s ns range signals -- more to the point though, there is increasingly much crosstalk between signals, due to the lack of ground plane; this dominates decisions about how to prototype circuits in this speed range. \$\endgroup\$ Nov 27, 2023 at 15:37
  • 2
    \$\begingroup\$ I would never ever recommend that a beginner rely on rosin-core solder as their only source of flux. Sure, it's not technically required to add additional flux when soldering; but covering everything in flux makes it much easier. I consider underuse of flux to have been my biggest mistake when learning to solder, as I perceived flux as an "advanced technique", when nothing could be further from the truth. Soldering without flux is an advanced technique. \$\endgroup\$ Nov 28, 2023 at 4:13
3
\$\begingroup\$

Additional remarks:

  • Start by considering connectors and mechanics. How large can the board be, where will it fit, can it be attached with screws, do we need to insulate the solder side etc.

  • Drill holes etc before doing anything else. Mainly because you don't want to stress the boards/joints by such things the last thing you do.

    But also if you have 4 holes in a PCB, you can use that to temporarily attach spacers etc pointing upwards and get a board which you can easily place upside down on 4 legs while soldering without having it resting/wobbling on the tallest component.

  • Consider currents and EMC. Since there is no ground plane, you'll want to build everything so that grounds are routed as directly to the supply ground pin of the connector as possible. It's common that these boards end up with a "line" of supply connections and another "line" with grounds - ideally place these for the shortest path to the ground pin of the connector.

  • Place as much of the components out in advance before soldering. By the book, component legs should be cut before soldering.

    Ideally (not a must but good practice), also:

    • Place components so that all component texts face the same direction.
    • Have all polarity of electrolyte caps etc point in the same direction. Ceramic or polymer caps without polarity makes life easier, if cap characteristics aren't that important.
  • Wires should be affixed tightly to the board, never to leave it in big loops etc. That's horrible for EMC and mechanics both.

  • Don't place any insulated wires on the solder side. Never have two wires cross each other, insulated or not! This is a very common cause of faults, either because of crosstalk/EMI or because if you manage to melt the cable insulation, you could get a short between the wires.

  • Similarly, do not place components on the solder side either. Components shouldn't cross each other either. Ideally they shouldn't even touch each other. Make sure that the components are fixed to the board and that their legs are cut before soldering. (Tape can be used for this during soldering, but it's not to everyone's liking.)

  • Even if you have a flux core solder (you should), apply external flux where necessary. If you are like me and have a bunch of perfboards lying around unprotected, they do gather oxidation over time.

  • For through-hole boards, there's probably no good reason to use anything but a chisel solder tip. These are easiest to use since they transfer the most heat to the surfaces fastest.

  • Keep a multi-meter close as you solder. Double check polarity of diodes and connectivity of traces/connectors etc if you aren't sure.

  • Proper soldering of through-hole components in through-plated holes means that the hole is filled and there is a cone-shaped wetting a bit up along the leg on each side of the component. This isn't always possible with perfboards since the hole diameter doesn't necessarily fit one particular component leg size, but we should strive for that none-the-less. Otherwise the joint may be cold and is also exposed to oxidation over time.

\$\endgroup\$
6
  • \$\begingroup\$ Keep a tweezer LCR meter on hand and check each decoupling capacitor as you solder it in to make sure your power rails aren't shorted. Hunting for the one shorted decoupling cap on the board is such a PITA. \$\endgroup\$
    – DKNguyen
    Nov 28, 2023 at 2:26
  • \$\begingroup\$ @DKNguyen There's really no reason to use polarized caps for standard 100nF decoupling these days. \$\endgroup\$
    – Lundin
    Nov 28, 2023 at 7:22
  • \$\begingroup\$ When did I ever talk about polarized caps? \$\endgroup\$
    – DKNguyen
    Nov 28, 2023 at 15:20
  • \$\begingroup\$ @DKNguyen I misunderstood you. Though I don't quite get exactly how you would short a cap in particular, assuming we're still talking big through-hole mounted components here. You could just measure resistance between Vdd and GND with an ohm meter as you solder, and that's the very same thing...? \$\endgroup\$
    – Lundin
    Nov 28, 2023 at 15:24
  • \$\begingroup\$ I suppose you could also measure resistance. I'm talking mainly about smaller, more heat sensitive ceramics where you might have lots and lots. \$\endgroup\$
    – DKNguyen
    Nov 28, 2023 at 15:26
2
\$\begingroup\$

This will be a somewhat unconventional answer: rather than a direct approach, I'll essentially review a board I made, including what I remember of building it, what's good, what's bad, and what could be done better. In the process, branching off to various topics, which, probably aren't very focused, but hey, comments are welcome.

The Project

Today's throwback comes courtesy of my old [personal] website, archived here:
LED Strings | Seven Transistor Labs

Just to set the stage, this was ca. 2008 or so, freshman college age, and I had been using perfboard for a while. Since the 2010s or so, I transitioned more to using copper clad (hand-carved, Manhattan-style, or deadbug-style) construction, better suited to most of my projects since then (power switching, high speed, and SMTs). So I don't really have more recent projects to share. (These techniques might make a good basis for another question, however..!)

For what it's worth, I wouldn't mind -- aside from the live mains voltage that is -- building a circuit like this on solderless breadboard. (I probably did initially.) The currents are low, the fastest thing is the comparator (~10s mA / ~100s ns), impedances are average. The biggest risk to such a build would be loose/unreliable contacts; stray capacitance and inductance aren't a problem.

Schematic:

LED Strings schematic

I won't go into detail here, but the basic explanation is: an off-line power supply for strings of LEDs, PWM switched, and an audio (envelope / peak detector) input to modulate them. PWM is convenient for two reasons: one, saving power (otherwise the IRF624 needs a medium to large heatsink); two, the control section is easily isolated from the mains. Thus you can plug in an audio source and get flashing lights to the music -- simple enough.

Top View:

Top View

The power supply section is on a separate board that I didn't feel like removing and photographing is simple enough not to worry about; it holds the transformer, rectifiers, filter caps and current-limiting resistors. (Not shown on the schematic is a power switch, and not shown in the photo is a fuse holder at the rear of the enclosure.) There is also assorted chassis wiring, including an old school terminal strip (pictured).

Bottom view:

enter image description here

So let's see here. Labeling this will be a confusing mess, so I'll resort to a mix of vague description, and highlighting when description would be impossibly vague. (Also doesn't help that I wasn't in the habit of using schematic designators back then...)

Techniques

  • Overall board: probably scissor (shear) cut from a larger piece. Classic RadioShack brand paper-phenolic per-per-hole perfboard. ('Member RadioShack?..) Sawing is a better method, causing less stress and cracking in the material, but either will get you there. (These days, I'm especially fond of cutting FR-4 with a hacksaw or abrasive blade, then sanding the edges smooth and straight with SiC sandpaper over a flat plate.)
  • Mounting: a hole has been drilled in the end, to screw onto a mounting boss / pillar. The front end of the board is constrained by the potentiometer shafts sitting in their holes.
  • Thru-hole components: all electrical components are thru-hole, inserted from the top side, soldered to the corresponding pad on the bottom, and the lead is either trimmed short, or bent onto adjoining pads. Entirely radial lead dress was used, with X or Y (horizontal or vertical in the photo) alignment, spanning one or two holes' distance, except for one placed diagonally (I think, the 10k below the differential pair?).
  • Soldering techniques: Mostly lap joints, with leads or buses routed between pads (mostly adjacent, occasionally at angles), solder dragged along the route to secure the lead to all pads along the way.
  • Wiring and connectors: lap joints, top and bottom side placement, no strain relief. Solid core wire, 20-24 AWG size by the looks of it.

One combined blessing-curse of perfboard is, the holes serve as pilots for larger drills, but if you happen to need a hole slightly off grid, you're kinda screwed (or rather...the board isn't, hah, get it?). Likewise, if you need a hole bigger than the edge-to-edge span, the drill breaks into neighboring holes and makes a mess (maybe munging up the hole, maybe shattering the board). Usually you'll end up wallowing out a hole to make an elliptical or slot shape, to move it a bit off-grid; this is rough on drills (both the machine and the bit), and can be dangerous, so be careful.

Pros

  • Reasonably dense layout; probably >50% of holes are used in active wiring areas.
  • Responsible layout: notice the red and black wires arching over the board, these are the rectified 12V supply, forming parallel buses along the circuit.
  • Insulated jumpers for longer connections (these could also be vertical, in dedicated holes).

Cons

Just to preface, it's not that I have many complaints about this -- few bullet points follow -- but it's interesting to explore the reasons behind them. And I've never been one to shy away from excruciating levels of detail...

  • Unreasonably dense layout. (Ha!) This pair of contradictory judgements is probably to say they cancel out on the whole. At this point in time, I would've been pretty experienced with placement and routing of this sort of design; the dense layout, of a known (solderlessly-)breadboarded circuit, was probably fairly straightforward. That's fine, given the experience, and given that the circuit doesn't need to change -- but it leaves very little room for subsequent mods, or servicing in case anything blows up.

    Especially as a beginner, do give yourself some space to work in -- you're much more likely to need to make changes later, not just replacing components, but adding new ones as well, and there's simply no place for new components to go in a dense layout like this one!

    It can also help to diagram out the circuit as you would place it. Draw circles, ovals, lines, etc. between grid points, see where everything needs to go, jumper wires as needed, etc. Label circuit nodes, and component pins, and make sure you're wiring them the right way around. I might not be using perfboard these days, but I do plan as much as ever -- often, my prototypes include rough sketches of SMT components, where I'm going to cut pads and traces into the board, or add pads on top, etc. Even if it's just sketching a local area to get a feel for where things need to go, it helps to avoid awkward congestion and stuff. Do recommend.

  • Lack of strain relief on wires. This isn't a huge down-side, as the circuit sits in a box on a shelf -- it's not subject to vibration, it's not flexing a lot (I put more stress on the wires today than they saw in the last 13 years, I guarantee!), nothing's tugging on them, it's fine, it's not like it's going to spontaneously crumble into dust. But it would be just that little bit better with stranded wire (the strands slide over each other, flexing much more easily than solid, and fatiguing less), and better still with strain relief.

    For an example of strain relief, see the bottom of this answer: https://electronics.stackexchange.com/a/77526/311631

  • Poor mechanical constraint (mounting). The single screw holding in the board is very "how-ya-doin". The front end of the board rests against the potentiometer shafts in their holes. There's even a spare hole in the board in the front corner; this could be secured with an angle bracket to a screw through the front panel, perhaps.

  • Isolation could be better:

    enter image description here

    The isolation gap is highlighted in green here. Isolation is quite poor on pad-per-hole PCB, because only the total distance between pads counts. It also doesn't help that I ran +V up inside the optoisolator's footprint -- putting traces here is generally not recommended, I mean obviously it doesn't matter specifically here as it's not much closer in this region than elsewhere along the gap, but more to say, it would be easy enough to remove the pads in this area to greatly increase the separation distance.

    Isolation is easy to overlook; it's something you can get away with quite flagrantly, in the average case, until months or years later, that one surge finally hits your device and damages it (or worse..). To be fair it's not something well known to beginners, nor all that simple of a topic to introduce.

    The real underlying take-away though, is: don't attempt circuits with mains voltage until you have the experience to handle it safely. There are rules and regulations that must be met for commercial equipment; granted, you're probably not selling whatever kits and bobs you're making for yourself, but there is still some responsibility to consider here, such as whether your insurance (if any) will cover such equipment -- even if such equipment was not causative, or involved, or perhaps even plugged in, in event of a fire.

    Which, looking back... this was a very "eeeeh" project in this regard, and would've been better done with a low-voltage transformer, and series-parallel LED strings instead of all-series high voltage ones. Be safe!


Some comments on soldering, more generally:

Just to preface; I think it's worth building a mechanical, microscopic understanding of these processes -- at least if you're interested enough to explore the ideas.

Soldering is a metallurgical joint. The base metal, solder, and the other metal, are in intimate contact with each other. Usually there is a microscopic interfacial (diffusion) layer, which together with the solder itself, dictates strength of the joint.

Metals, in air, oxidize. Basically all of them. (Gold being a famous exception, hence its great value as contact plating for connectors.) Solder doesn't stick to oxide, and for that matter, the oxide on that dry blob of solder sticks even less! To get intimate contact, pristine -- chemically clean -- metal surfaces are required. To do this, we clean them with active chemicals: fluxes. These are usually mildly acidic, and disrupt and dissolve the oxides, moving them off the metal surfaces, allowing pure metallic solder to contact pure metallic base metal, and then they wet readily, solder spreads out across the surface, and you get a joint.

An interface layer occurs when certain metals or alloys are mixed and partially melted. Copper base metal, and tin in the solder alloy, are a typical case. Iron not so much (hence why it's an excellent metal for making soldering iron tips out of, it doesn't react and barely dissolves into solder), nor lead (but, it's rare that pure lead is used as a solder, especially these days). The layer forms and grows by diffusion, and diffusion accelerates exponentially with temperature. This is why you don't want to overheat and cook joints, particularly those using copper and tin (there are some quite brittle Cu-Sn compounds that can form, given enough time and heat). Therefore, you should solder with the lowest iron temperature that is feasible, and heat only as long as needed to make a secure joint.

Notice that doesn't mean setting the lowest temperature period that can do soldering at all! Too low, and it just takes an eternity for the joint to come up to temperature, you're cooking the thing, insulation is melting off the wires, rosin is unreactive, it's a whole mess! Rather, temperature is a balance between getting it done quickly, and making an adequate joint. The hotter the iron, the faster it heats a joint, the more aggressive the flux is -- but the faster the tip itself burns up, the faster the joint cooks, etc. Good soldering is a practiced skill!

You want a good clean, shiny tip to make ready contact to the parts being soldered, and flux helps wet the joint -- both by spreading heat into the joint (a layer of solder on the iron, might make close, but not intimate, contact with the parts), and by cleaning them.

It is possible to do plenty of work with just a roll of rosin-core solder; but it is much easier to do so with some kind of flux handy. Especially if dirty and dry joints are involved (such as in repair).

As a metallurgical joint, it's not going to just spontaneously come loose. I think there's common a misconception that solder joints shouldn't bear stress; it would be more accurate to say, they shouldn't bear stress above some critical point. It's just that solder is so much softer (~5k PSI) than the base metals (15~60k PSI), so it takes very little leverage from something bulky like a wire, to peel apart a thin layer of solder. (There is also creep and fatigue to worry about, and even esoterica like tin whiskers; creep is, in part, solved by setting a lower stress limit, or by using a non-creep alloy i.e. lead-free; and fatigue likewise is solved by reducing stress.)

Given all this, there is definitely some work we can do with joint design. Basic lap joints are easy, but they aren't very strong; they can peel or tear apart. The more we can twist, knot and crimp the base metals/leads around each other before soldering, the stronger the result will be; ideally, in one sense, the solder merely serves as a sealant to keep the joint frozen in place, free from oxidation. Such is the basis of the famous Western Union splice.

If we use thru-holes, and get PTH (Plated Thru Hole) PCB material, we can have not just the ring pad on bottom, but indeed the whole "barrel" of the hole, filled with solder and anchoring the wire or lead all the way through. This is one of the manifest successes of 2+ layer PCBs. It's very rare that a cold solder joint develops on these (but altogether too common on non-plated single-layer boards!).

So for the pictured case here, it's not a problem per se that the solid-core wires are merely lapped onto the circuit board; but it's definitely not something that will handle much flex or stress. (Just digging out the board here, flexing it around for some pictures, is probably like 5% of the total lifetime flex I'd want to put on this thing... Shove it around a few more times, and wires will definitely start popping loose!) If this were something that gets handled or serviced regularly, then stranded wiring, and some manner of strain relief (looping through holes, using cables ties, or even gluing down the wires) would be welcome. Or using connectors. Which...

Connectors

Case in point: I didn't have any connector systems back then, and actually soldered the LED strings directly into the board(!). I didn't feel like taking down the strings to take these photos... so, I took the opportunity to cut them off entirely, then "splice" with proper connectors.

Specifically, I used Molex SL, a 0.1" pitch, 0.025" square pin header family, with shrouding and locking tab. Pins: 0016020103, male housing 0050579402, female header 0705430001. The header is thru-hole, but I'm making an inline connection, so I stripped, tinned and lap-soldered the wire ends onto the pins, and covered them with heat-shrink tubing. (A proper inline housing like 0701070001 would be preferred, but I don't keep them in stock.) Obviously, I said inline, but preferably such connectors would be on PCB of course, and being 0.1" pitch like most perfboard, they're excellent candidates.

The official crimp tool is quite expensive unfortunately; it's well worth for professional use, and is easy to use, but it's hard to justify the cost otherwise. Numerous cheaper and variously-universal crimp tools exist, but require more finesse to use -- good crimping is itself an art form and skill, worthy of another question/thread perhaps.

Molex SL, C-Grid, KK, "Dupont" (actually Amphenol ICC, and widely cloned), TE/AMP MTA, and many others, are good candidates for use with perfboard. Even if you don't want to do crimping, you can find premade cable assemblies in these families, and make a quite professional looking build from off-the-shelf components!


Minor electromagnetic sidetrack:

Interestingly, there's no supply bypass cap on the control board itself; the power supply section is a good, oh, 20cm away, in terms of wire length. Take it or leave it I suppose, but it at least doesn't seem to be a functional problem.

This is easy to understand, actually, if we consider the maximum load rate of the circuit: 20cm should provide on the order of 0.2µH stray inductance (probably less), and at a maximum peak load change of say 50mA in 200ns, the peak supply voltage change should be \$V = L \frac{dI}{dt}\$ or 50mV -- negligible out of a 15-17V supply.

Such calculations were not in my repertoire at the time (or, more than a rough feel, not justified by calculation), but that shouldn't stop us from taking a critical look backwards with the benefit of such tools!

\$\endgroup\$
4
  • 1
    \$\begingroup\$ I think we should probably post pictures of good examples rather than bad ones? :) Personally, over the years I mastered the art of how to do quick & dirty lab boards in non-recommended ways, for the purpose of testing things quickly... so I can probably post a whole gallery of warning examples. \$\endgroup\$
    – Lundin
    Nov 30, 2023 at 10:22
  • \$\begingroup\$ I mean -- do's and don'ts aren't the worst way to give instruction :) I was hoping for a more comprehensive example, but didn't find any surviving good examples in my junk box. It serves here more as a starting point for discussion, which maybe makes up for that, maybe not. \$\endgroup\$ Nov 30, 2023 at 10:27
  • \$\begingroup\$ Well, to comment further on your example beyond what you have already done, this is a good example of why insulated wires shouldn't be placed on the solder side. This would be one of the more common not recommended quick & dirties I think. The black and red ones upper left corner that cross component legs and joints could easily get torn during use or melted during soldering then cause an accidental, hard-to-find short. Also it is quite hard not to melt the insulation and cause it to wander up the wire here. -> \$\endgroup\$
    – Lundin
    Nov 30, 2023 at 10:34
  • \$\begingroup\$ Proper mounting would have been to place the entire wire on the component side, tightly fixed to the board and measured + peeled in advance. (This does require more board space than here though, speaking of dense/cramped layouts.) If one wants to get extra pedantic one can also pre-solder the wire at the peeled parts before slipping it on the board - this helps against melted insulation since it will wet easier. \$\endgroup\$
    – Lundin
    Nov 30, 2023 at 10:36
2
\$\begingroup\$

Quick answer with a few pics. This is for prototyping, not permanent builds.

SOT-23 transistors and 0805-1206 passives are just the right size for perfboard over double sided sticky tape over ground plane.

enter image description here

I actually prefer having the perfboard upside down with SOT23 and SMD passives because it only takes a few seconds to desolder and change a value when prototyping. If there's no ground plane on the back, then it can be used for wires. With the new cheap plated through hole perfboards, it's possible to put SMDs on one side and thru hole parts on the other side, which is convenient, but unfortunately there is no third side for wires.

Copper tape over Kapton tape makes low inductance connections (50MHz oscillator and feedthrough filter).

enter image description here

A random shunt voltage regulator:

enter image description here

The nice thing about kapton tape is that it's pretty thin, so the bit of copper tape on top is well coupled to the ground plane and has very low inductance.

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.