# Do engineers still design for volume using discrete through-hole bipolar transistors?

As a hobbyist who likes plugging BJTs and discrete components into solderless breadboard for entertainment and education, I recognise that people like me are an insufficient market to make it worthwhile for manufacturers to continue production.

I've noticed that on this website a lot of questions about BJTs are answered with recommendations to use a purpose specific IC or a PIC instead.

I also enjoyed reading 555 best IC ever or obsolete anachronism

I note 555s and through-hole BJTs are available and pushed by hobbyist shops online.

I have some cheap solar powered LED lamps to mark the edge of a path, recently one was broken so I opened it up and found this board with three through-hole BJTs and a bunch of other discrete through-hole components on a single-layer board.

So why do some large-volume manufacturers still use this sort of technology?

The only reasons I can come up with are

• Its a really old design and not worth redesigning?
• The educational market is bigger than I imagine?
• Maybe old BJT production machinery can be shipped to low-wage economies and kept running for years at minimal cost?
• Wages are still low enough in China to sustain this kind of technology?
• There are a lot of old engineers around leveraging accumulated wisdom with 555s & BJTs?
• Discrete through-hole BJTs are still the best solution for something (what?)

Do engineers still design for volume production using through-hole BJTs?

• I've seen the first stage of audio amplifiers designed with discrete BJTs in respectable brands. The reasoning is that even an ordinary discrete BJT has lower noise than integrated transistors. I don't know how true this is, especially in an industry known for snake oil products, but it is common practice anyway. – Phil Frost May 6 '13 at 11:59

Low low low cost is the aim.

A small volume maker with no expertise in anything other than manual soldering (if that) can make these. They can be even made at home by workers if desired.

Single side phenolic board. Cheap.

Not only manual assembly but component size and hole spacings not well matched and leads are hand bent and nobody cares.

A design like this does not have to be efficient or long life - it just has to work well enough to persuade a western company's "buyer" that they can sell it to the eg US public at an adequate profit. The parts on that board plus the board cost probably 10 to 20 cents. (Junk resistors under 0.1C, inductor 1 or 2 cents maybe, transistors about 1.3 cents say, LED a few cents, PCB under 5 cents, wire 1 cent. Labour costs need to be low to match. That's for quality components that I am accustomed to :-) - presumably even less if quality matters nowt. Even 10 cents plus labour is getting costly as you have to add battery, housing, lenses, box etc and sell it out the factory door at about 20% to 25% of eg US selling price.

Chinese minimum official labour rate is now about 2000 RMB/month AFAIK (maybe more by now) but you may work 6 days/ week and 10 hour days for that. Say 6 x 10 x 4.333 = 260 hours.

An RMB is worth $US0.16 at present so per hour that's 2000/260 *0.16 ~~=$US1.23/hour.
Some workers will get that for a 5 day, 40 hour week = 50% more per hour. Others will work longer and get half that.
And be pleased enough in many cases ! :-(. Worst case I saw was 10 hours per day x 365 days/year except leap years. Really. (US owned/managed. NOT mainstream. No names). At many such places you get free lunch. That added 25% to the labour cost at the 3650 hours pa place.

Which may explain why your PCBA looks as it does :-(.

5 year old photo. Tasteful precision blanking by me. Factory has long passed on to other things. This is typical enough "cottage industry" level manufacture. Not much equipment needed. Entirely reasonable quality can be attained when done well. Or you can get things like in your picture if not.

• Which may explain why your PCBA looks as it does -- I'd be happy to produce something that clean myself! – Chris Gregg May 6 '13 at 12:17
• @ChrisGregg: you have to train just 10 hours each day for a year. :-/ – David Schmitt May 6 '13 at 16:07
• Argh what did you blank out? It's going to drive me nuts. – Bryan Boettcher May 6 '13 at 18:33
• @insta The nature of the product is obvious enough from what you can see BUT it was specifically identifiable in the blanked areas. While this was long ago the owner of the manufactured goods may not wish me to show them. – Russell McMahon May 7 '13 at 1:48

Yes they do, particularly if the margins are very tight. That is part of engineering is to find the right solution, which for certain markets may be only about cost. That circuit above has 3 transistors and 3 diodes and with the pierce and blank phenolic board, might be only a few cents fully stuffed in volume production, for example the resistors in total may only be ~ $0.01. In fact there may not be an IC solution that can implement this function, and spinning your own you may not ever be able to recoup the NRE (Non Recurring Engineering - i.e up front) costs or the talent base may not be there. They can easily change the design and get it back into production quickly in it's current form. The performance may not be any better in an IC. The product lifetime may not be long enough to justify spending too much tie on the design cycle. The often hidden presumption is that an IC solution is always superior. That is not the case. You must evaluate on a case by case basis. You will see this kind of design in toys and high volume simple consumer goods. As to your " ... recommendations to use a purpose specific IC or a PIC instead." thats simply because that is what those people know best. If your only tool is a hammer - every problem begins to lok like a nail: is one saying that might apply. Or it could simply be that being a smallish site there isn't the diversity present. • Info only: It happens in this case - and this is of course NOT true in some other cases - that there is an excellent IC solution available at low cost and it's possible that the PCBA shown costs more to make overall. The IC's cost under$US0.10 and need a total of the 4 pin TO92 size IC + inductor, an LED, a PV (solar) panel and a 1 cell battery to produce a "lawn light". No diode, resistor, capacitor. Works well enough. Turn off in sunlight. Light at dusk. Efficiency of up conversion around 70-80% (more claimed) which is good for a 1 cell unit without sychronous rectification. – Russell McMahon May 7 '13 at 1:23

Looking at that board, it wasn't even assembled with an ancient piece of pick-and-place machine - it was assembled by hand. In this case, I guess, the manual assembly was done for extremely low labor cost.

However, some other new designs use through-hole parts as well, most prominently switch-mode power supplies, even expensive ones. The reason is that you have some big through-hole components anyway (terminal blocks, transformers, large electrolytic capacitors), so you need the assembly and soldering process already. Using through-hole components for the rest may be a bit more expensice as far as the stuffing process for just these parts goes, but the overall cost may still be smaller.

As you guess, there may be plenty of reasons.

Remember that on this site, most people asking questions are building circuits in small quantity. If you only need one of something, then it can make sense to spend \$20 on parts so you get it done and working in a reasonable time. If you need to build tens of thousands, then you'll want to spend a lot more time redesigning to get the unit cost down. That redesign may use much lower cost parts and manufacturing techniques.

Also, as rawbrawb alluded to, some of it is education and what you know. I very often see hobbyists using expensive parts and complex designs to do simple tasks because they don't have the engineering background to design something efficient. Like the saying goes: anyone can build a bridge; It takes an engineer to build one that barely works.

• "Like the saying goes: anyone can build a bridge; It takes an engineer to build one that barely works." I can only assume you've never played any bridge builder games. Once past the obligatory intro levels the only way to continue is to come up with the most preposterous, would get a real engineer fired for even joking about, on the verge of collapse design possible. – Dan Is Fiddling By Firelight May 6 '13 at 13:42