The 3 letters of doom: SMT

Let’s not hijack smrl’s discussion on his new project and discuss this in another place:

Electro druid wrote:

I have to say that before that post, the idea of doing kits with SMT boards didn’t sound like something I’d be interested in – I think a big part of the beauty of your kits is the feeling of accomplishment from building them yourselves. SMT seemed like a step towards fully pre-manufactured synths, which is less interesting. That said, if you think you can do what you say can be done with SMT and get smarter DSP and polyphony then it could be a really interesting next generation of instruments.

Now I’m sort of in two minds.

Out of interest, what’s the limiting factor with through-hole? Is it the processing power available to through-hole-sized processors (I noticed you went for a more powerful processor with the Shruthi-1 – is that as powerful as through-hole chips get?) Or is it more about physical PCB size, or the number of connection pins for things?

Only very basic chips are available as DIP packages. The limited number of I/O ports is not a big problem (though not having to do everything through shift registers or I/O expanders would make things easier) ; the hard limit is processing power for rendering samples and pushing them to the DACs…

  • On the AVR side, the top of the line are the 644p and 1284p (the 1284p has more flash). They can’t be clocked at more than 20 MHz, and a MAC with 8 bits operands to a 16 bits register takes 5 cycles.
  • On the PIC side, there’s the dsPIC series. Instruction set seems more suitable for DSP (with a 1 cycle MAC), and stripped of some of the PIC awfulness. But what I’ve seen of the PIC development tools/environment keeps me away.
  • Maybe some antiquated DSP chips from TI or freescale? It would be ironic to be locked with outdated parts for something as trivial and available as a microcontroller!

On the other hand, there’s plenty of cheap and advanced ARM-based chip with well-supported toolchains: ST’s STM32, NXP’s LPC17xx… all in SMT packages. Putting such chips in through-hole packages would make little sense from a business perspective. There are a couple of options for integrating them with a through-hole design (schmartboard, breakout boards) but I am not a fan of those - typically they are the kind of things you can get only from one supplier, in limited supplies, and at a horrible price - because they target the amateur that will buy one or two to tinker with a chip, not a larger scale project. Lesson learned from the Shruti-1 LCD: never integrate in your project a part dedicated to prototyping/tinkering - it doesn’t scale. Indeed, I don’t know if there are still professional designing for through-hole anymore - and if there are, it might be for low-tech projects for which massive computing power is not needed.

But it’s good to hear your feedback about this. Would a board with only one pre-soldered SMT part (the MCU) be ok? Or is it an “all or nothing” thing?

Personally, I don’t see a problem with having a single SMT part pre-soldered to the board. Given the comprehensive build instructions provided with the Shruti-1 plus the fact that the Shrutis were generally available as kits, you essentially created Lego sets ( a compliment, btw; I loved legos as a kid) that beginners with very limited knowledge of electronics could assemble. In comparison, the GSSL mix bus compressor that I built was much more involved than the Shruti with respect to parts sourcing, assembly, calibration, testing, troubleshooting and general difficulty. For the Shruti, the hardest part I encountered was waiting for the components to arrive from the various suppliers that I used (still waiting for the damn rack enclosure, btw…grrr).

I don’t believe that providing the end user with a single SMT component pre-soldered is going to be that much of a departure from what you provided in the Shruti-1 kits. To continue the Lego analogy, It’s sort of like those cool little Lego electric motors that were available. The motor was preassembled and ready to go, but housed in a lego case that allowed you to build around it and incorporate it into your own constructions.

It’s certainly not an “all or nothing” thing for me, and I should state upfront that you can design whatever the hell you want to design :slight_smile:

If the limiting factor is processing power, and the choices available are between chaining a bunch of less powerful chips (which I can imagine will be a big headache), using obsure and/or rare chips, or going with an SMT part, I can see why you’d favour SMT.

As far as I can tell, the founding principles of Mutable Instruments seem to be to make stuff that is:

  • As simple as it can be whilst still being useful and musical
  • As open, moddable and hackable as it can be
  • As cheap as it can be (with some profit for you to make it worth your time, of course)

Personally speaking, I happen to also like the following selling point:

  • To be able to build as much of it as possible myself.

My initial hesitation about SMT was that I thought it might lead down a road where kits come increasingly pre-made until eventually there’s nothing to build, which would make Mutable Instruments the same as any other synth manufacturer, which would be boring. But if what we’re talking about is a future where the MCU is presoldered but you build everything else yourself, and the result is something with a lot more power and flexibility than if we got to solder that one part, then of course - that would be brilliant :slight_smile: So long as we still get to hack the rest of the hardware, and can hack the firmware on the MCU, then I’d have no problem with that at all.

EDIT: Konketsu’s analogy about the Lego motors is pretty much spot-on. Yes, that’s one pre-built part, but it’s one which can open up a whole bunch of possibilities depending on what you build around it.

SMT doesn’t put me off at all. I haven’t done a lot of it (in fact just one thing so far, my WTPA), but having that under my belt gave me a lot of confidence that I could do more. And if you added the option of having the SMT component(s) pre-soldered (at an extra cost perhaps) gives those with less confidence the option to still partake…

@Electrodruid:
You’re absolutely spot on about my goals. I am not interested in selling finished products ; and the farther I will go will be pre-assembled boards for everything that requires SMT.

I’m with dnigrin.
After the WTPA build, I’m not scared of anything. (Actually, the MB SID was a big part of that: “build your own power supply”, “find your own chips”, “don’t ask for help”…)

The main issue for me is cost: If the kit is cheapest with an SMT part included, but not pre-soldered, then give me that.

I really agree with the direction this thread went.

All I have to add is that the hackability is interesting to me only in the bits that make the sound. I see this as the code for the digital bits and the components for the analog bits.

I would not be put off even by a pre assembled shruthi interface board. But I like the idea of building that filter, for some reason.

Adam.

SMT is no biggie for me either, the RAM on the WTPA is actually more more of a pain to solder than most SMT parts. I say offer SMT parts soldered in for a fee. People that can do it themselves makes kits easier for you and people that cant earns you money so it’s a win/win from where I am sitting

i got a bit used to smt soldering after gm5, wtpa sram, aout_ng (not tested) and opl3 and now after all this i even own flux, which should make things even easier in the future. i actually like it, it’s a bit more challenging than other parts. so i’m ok with smt parts if used as mcu, or ram etc.
but i’d rather like to avoid smt in parts of circuits that carry analog sound signals, eg resistors and capacitors. i’m not sure how much this really affects the sound compared to bigger through hole parts (like there’s also a big difference between transistor and tube technolgy), for me it begins with a smt diode sticking to my soldering iron when i try to bend a circuit (like happened with one of the new stylophones).
but for processors, memory etc, why not?

i wussied out and bought the wtpa board with the sram pre-soldered. i’d be up for giving it a go though. i reckon my solder skills are pretty ninja at this stage :slight_smile: