Component choice.

[NMNeil]

A very broad general interest question.
I'm putting together a design for a CDI ignition circuit based on an old 1970's Wireless World article. Many of the components are obsolete and have been superseded with either improved versions or with IC's.
This leads me to ask the question.
When the members of the forum design a circuit, do you design it with an open mind then pick and order the perfect component for the job, or do you design it specifically to use the big box of components that you ordered sometime in the past but can't remember why you ordered them in the first place?
Just wondering.

[tpowell1830]

For hobbyests, it is usually the latter due to time and budget constraints, however if this will end up as a professional commercial device for sale, then the former.

[JRosario]

For hobbyests, it is usually the latter due to time and budget constraints, however if this will end up as a professional commercial device for sale, then the former.

Is the difference noticeable to the quality and functionality when you do this?

Because I would think for the issue of time, it is always best to get the perfect component rather than trying to make just anything work. But in terms of functionality (perhaps not longevity), if it works it works, right?

This is me as a super beginner, but I am so worried about my skills that I don't want to also have to worry about the components not working and sabotaging success.

[aylons]

For most low-volume (professional or not) projects, price will be dominated by a few expensive central components to the project. I prefer to focus my energy in making sure these are the "perfect" choice, and get from the "big box", or simply from previous projects, whatever else.

Also, most of the time, there's no perfect component. A range of them will do, and nitpicking some detail will most likely be only a waste of time. So, anytime I can get away with it, I prefer to use tried and well known (by me) components. It reduces the risk of silly mistakes, speed up the design, and allows me to have a library of circuit designs ready for use.

Meeting the specs in the least development time is usually the name of the game.

[james_s]

If I have parts on hand that I'm trying to use then I'll often try to incorporate those into a project. Otherwise I just buy the right part. The process of designing around parts I already have is not much different than designing around new parts that are within my budget.

[rstofer]

The 'perfect' part may actually exist, somewhere.  But I'll never find it wading through a half-dozen web sites and spending hours looking at parametric data only to find out that nobody stocks it and lead time is 26 weeks.

I try to find something that will do the job and that's about it.  OTOH, all of my projects are hobby grade.  If I was doing this stuff for a living, I would try to find the 'perfect' part and let Purchasing worry about getting it.  Of course, they would also want to downgrade my design because the 'perfect' part is too expensive.  And so it goes... Round and round...

[tpowell1830]

For hobbyests, it is usually the latter due to time and budget constraints, however if this will end up as a professional commercial device for sale, then the former.

Is the difference noticeable to the quality and functionality when you do this?

Because I would think for the issue of time, it is always best to get the perfect component rather than trying to make just anything work. But in terms of functionality (perhaps not longevity), if it works it works, right?

This is me as a super beginner, but I am so worried about my skills that I don't want to also have to worry about the components not working and sabotaging success.

It depends on how much of a perfectionist that you are  if you are a hobbyest. If the idea is to build an astable oscillator from a few BJTs and capacitors and try and hit a specific frequency with a perfect sinusoidal output then you may try with the grab box devices on hand and prove the concept. In other words, get it as close to working as expected as you can and then move on to your next experimental project. Since this is only something that you want to prove that you can do, then you have done it.

On the other hand, if you are building a larger project that you will use in your lab, like a power supply, you may want to order better suited BJTs or even FETs for this to work well in your power supply.  You may try the ones you have and not be completely happy with the performance and order a different part number for one of the components in order to make it more perfect.

I will use the grab box component for proving an oscillator and get it close and then move on, being satisfied that I could get it better and put some thought into what would improve the device.  The next time, could be months, that I place an order, I will order the different part numbers to try my oscillator experiment again with the new parts.

As a hobbyest, I run my experiments to enjoy building a specific type of specialty circuit and getting it to work.  That is how I view my hobby, as enjoyment.  I have been involved in designing circuits at my work and the parts list for those designs are all ordered as a BOM and the parts are very carefully specified to do the job. The circuits are put together as a prototype and then experimented with and tested in nine ways to Sunday. As a hobbyest, I don't have that need unless I am building a larger device using that circuit and I need it to be reliable. I will probably never design and build a professional device for sale at home, so I will never have to go through all of the extensive prototyping and testing required for a commercial product. (But never say never.)

The bottom line, to answer you more specifically, is that it depends on you. 

[radiogeek381]

Personally, when working around the shop, "perfect" means "in a place where I can find it."  For lots of common parts, I've settled on a go to list. 2n3904 2n3906, some surface mount versions of the 3904, a surface mount diode that I bought a bunch of, and a basket of standard resistor values.  For lots of experiments and even a few useful widgets, good enough is good enough, and getting the thing built while the ambition is hot beats waiting for a parts delivery.

Beyond the pedestrian, I've collected lots of RF parts that get used from time to time in one-off designs and widgets.  Even here, there are favorites that I've accumulated.  If you can't be with the part you love, love the one you're with.


To help with "supply chain management" many companies have qualified parts lists and/or qualified vendor lists that limit the choice of components, often for very good reasons. Back when dinosaurs roamed the earth, I worked for a company that had a "qualified parts list."  If you needed a part, it had to be on the list.  If you couldn't do the job with the parts on the list, then you had to make a case, and be prepared to sweat it out.

There were FOUR chips on the QPL.  A quad NOR, a four bit counter, a dual D flop, and a 741 op amp.  I added a 40 pin USART to the QPL.

The company built modems(!) and lots of other odd chunks of gear that was aimed at hooking terminals, computers, machines, and valves together.  By keeping the QPL small, they bought bazillions of each part and got volume economies that overcame the occasional sub-optimal design choice.