Kickstarter 2 Ramblings

This was supposed to be easy, a rinse and repeat (mostly) of the first Kickstarter. A global chip shortage conspired against us but a new manufacturing partner opened up some fresh possibilities.

We had always intended on improving the board. We needed to get the CE marking (or UKCA depending on how long this process was going to take) and this compliance required a large number of modifications in order to reduce those dreaded electromagnetic fields (EMF) and such. The board was pretty noisy.

Cleaning up the board and designing it more around the desired end result, rather than basing it on the original design and “upgrading it” was a better option. It gives us more wiggle room. Memory on a single 2MB chip close to the FPGA avoids many issues that could arise from a four RAM chip configuration when you’re pushing the limits. A more solid and shielded Wi-fi module helps us to a sweet spot when combined with the many other modifications to remove noise. Moving it to a location where it is less likely to interfere with impressionable components gives us even less noise. Little by little, resistor arrays were added along with a lot of circuit protection to protect your valuable board from spikes, or end user mistakes.

Some PSU plugs look the same right? One word of warning, whilst the new board is much more tolerant to accidental abuse and incorrect voltages and polarity, it is not bomb proof. A 21V power source was a step too far and possibly lower than this would be too. Only one component was fried and the board returned to working order once this was replaced. Don’t test you luck though, your mileage will vary so always be careful.

The inner workings of Kickstarter 2 are not that remarkable. Gone now is the huge bulk of work which previously not only included doing stuff, but figuring out what needed to be done, how and where. We had to find people to work with on the many threads. We had to deal with multiple companies for each item and convey our ideas and requirements and work out who offered the best quality at a reasonable price. We had to deal with companies backing out either due to the toxicity of the Vega+ (nothing to do with us mate), or a sudden reappraisal of their priorities (read profit). There are a lot of moving parts on a project such as this and when each is being handled by separate entities it is difficult logistically. Different countries and therefore timezones, different methods of working, and lots of cross dependencies.

Breaking this down for the first Kickstarter, I can count 12 different companies/entities involved (excluding SpecNext Ltd and the team of highly motivated and skilled people who helped make this happen). This goes from something as small and insubstantial as screws and threaded inserts, right up to cases, keyboards, boards, manuals, boxes and final assembly.

In the new world we have had a number of hefty kicks in the trousers but the silver lining in all this is a relationship with a new manufacturer in China. In fantasy terms they are akin to “One ring to rule them all” or Highlander’s “There can be only one”. They provide an extraordinary range of capabilities which align very well with our requirements. We have Jim Bagley to thank for establishing this connection, this one-stop shop which is effectively a holy grail of sorts. They have over time earned our trust and our thanks. Importantly, they get it, what we’re doing.

For the first Kickstarter, we endeavoured to do as much as we feasibly could in the United Kingdom. We wanted to continue the tradition first established with the original ZX Spectrum computers back in the day. Moving forwards we understood the challenges better, there was uncertainty in the world and we knew we had to cut back on costs where we could. The Kickstarter 2 pricing reflects an assumption that things will go wrong at some point and also that our original costings for the first Kickstarter were way off the mark. Pushing the Spectrum Next more into China would allow us to deliver a product for less, but would also enable us to do some things better than before. The main issue is finding a manufacturer who is consistent and trustworthy. Without the funds to absorb significant mistakes, we always had to ride the path of risk aversion.

In the early stages and evolving a design from Victor Trucco, we employed an electronics design consultancy to oversee the intricacies of designing the new boards for the new Kickstarter. It often wise to spend money in these essential areas even though we have good amount of electronics knowledge within the team. We were moving into the realm of standards, compliance and conformity. Things like CE and UKCA have to be done in very specific ways. They also provide a sounding board for members of our team and give us their ideas and recommendations.

Back to China and the “Wonka” factory as I will now refer to them, took on our initial requirement for making prototype boards which were designated as Issue 3. They used a similar design to the Issue 2B which shipped with the first Kickstarter units but with a good number of improvements. This would prove to us that they could deliver a fully functional Spectrum Next board with great quality components, consistency and robustness. They passed this test.

Jim had extensive conversations with them around all aspects of the Next and what we needed to fulfil the 5500+ pledges we received. In pretty much all cases, the Wonka factory could supply what was required. If they couldn’t then they knew somewhere that could. Power supplies, SD cards, inline switches, boards, boxes, manuals and a partridge in a pear tree. We could slim down our list of partners significantly and in some areas this meant less cost for us. Less cost helps us greatly in being able to absorb some of the items which increased in cost due to global circumstances.

Sometimes things just drag on and on and sometimes a month can feel like a very long time where bad news is concerned. The Raspberry Pi Zero situation was fairly dire. We needed over 5000 of them and availability was as the product suggests, zero. It was March 2020 and in the midst of Covid when we reached out to the Raspberry Pi sales manager with our requirement. They expressed their concern to us that the lead time would be around 10-12 weeks. Apparently this was a long time back then but we have gotten accustomed to waiting much longer for crucial components.

Given the time scales, we held off on placing our order. This would avoid us incurring excessive costs for part storage and typically we try to aim for everything to arrive a few months ahead of assembly at most. In October 2021 we finally bit the bullet though. Were they successfully navigating this new world of parts shortages? Wellllll no. Due to global chip shortages there were no firm production dates for the Pi Zero line. There was nothing out there.

They obviously have their own plans and requirements, given the quantity of Raspberry Pi variants they produce. We just hoped that they would be able to fit us in somewhere, if indeed they were able to get any kind of production started again. So no accelerated Spectrum Next’s until such time as we can either source the part, or adapt our own approach and use a different flavour of Pi. Unfortunately though, it seemed that the whole Zero range was impacted and we couldn’t get anything. No Pi Zero, no Pi Zero with Wi-fi, no Pi Zero 2.

Fast forwards a number of weeks (which felt like many months), and the landscape started to change colour. Out of the blue we received an email with the question “If we can supply these, when would you need them delivering?”. It wasn’t a 100% guarantee at that point but it was something. That something built and eventually we were given a green light and a delivery date. It is quite destabilising when something so fundamental suddenly becomes an issue, we’re more focused on the items that have caused us issues in the past, and of course the sourcing of suitable FPGA’s.

In the many conversations on social media platforms, people often ask questions about other external endeavours and how they are able to source the FPGA’s that we seemingly cannot. The N-Go is a prime example whereby they have been producing a constrained but fairly steady quantity over the past few years. They order in low numbers as low numbers are pretty much all you can get. There is an element of risk in these transactions as dealing with random factories in China can be a lottery. You don’t necessary know the history of the items you are purchasing. Are they reconditioned? Do they need re-balling? Are they even genuine? They will be paying over the odds in most cases and will have to pass this cost on. Those seas are rough and when you have the hopes (and money) of thousands of people, and where one wrong move could have a significant financial impact, you don’t go aiming your ship through the Drake Passage.

As mentioned, the China move has enabled us to save money in some areas but as we’re not out to actually make profit, we tend to invest these savings to enhance parts of the Next experience where we can. Of course we are also walking the line of caution and not counting avian beings until they are hatched so we are mindful to keep some savings in the bank of “Just In Case”. As work streams are marked as completed, we gradually eliminate possible rogue elements, reducing our risk factor significantly. If we budget for the packaging element of the Next based on the first Kickstarter costings and see what we could do in China for the same amount, we can do a lot better. Do we do the same as before or do we push the materials and design to new heights, incurring more cost but still living within our budgetry constraints?

There are two factors at play here. There is an issue to fix with the first Kickstarter packaging which involved the internal platforms. There is a desire to make the whole packaging element an even more enjoyable experience. We are for the most part very discerning and focused individuals when it comes to the world of vintage computers. We like our original boxes to be as mint as possible, both inside and out. We as consumers have experienced the fruits of the very best packaging minds in the world (I think Apple has to be my own personal favourite). We cannot match Apple though, their budget for packaging research and design on a single line exceeds our Kickstarter total. We can do a pretty damned good job though. We can do the whole lifting by the lid and watching as the base slides out infinitesimally slowly and evenly. We can give your eyes a feast when the lid is off and the internal components are laid bare in all their enhanced glory. Hell we can even give you a more beautiful manual with a table of contents!

Back in the first Kickstarter days, we arrived at two internal box designs. One used foam/cardboard inserts and the other used cardboard inserts.

Foam version

Dreaded card version (what we used)

Hindsight is a wonderful thing and given that, people may presume that we made a mistake and should have gone with the foam upper version (yes and no). This would have prevented the damage to the insert (well technically it wouldn’t have existed) and prevented popping the odd key off the Next when the insert lifted and the Spectrum Next slid underneath. The fact is that the foam element was too expensive to be produced, it blew our budget for packaging. We also thought it looked a little ugly but that’s just our opinion.

Foam comes in a number of different flavours. Ester, Polyethylene (PE) and Pro Cell foam (there are probably more out there). The more you pay, the more compact and textureless it becomes. Jim had spoken to our manufacturer about the packaging element and they looked at our original packaging and came up with their own approach. Reducing the height of the overall package and coming up with an internal arrangement using PE foam. It wasn’t exactly pretty looking (PE is notoriously textured and cheap looking) but it would do a far better job of protecting the valuable innards. The costings for this package were hugely encouraging, so much in fact that we decided to explore the various foam options, pushing things as far as we could. At the same time I called upon Phil Candy, our industrial designer to create a CAD model that would present the components in a more favourable way. Everything is now head and shoulders above the first Kickstarter packaging.

Alfredo Tato had created the artwork for the original Kickstarter box and it was time to start thinking about the second coming. A familiar but different approach was required. Prior to this, Alfredo had imported original Next CAD STP assets into Cinema 4D and had begun the laborious task of correcting a large number of glitches that had arisen in the conversion. Vertexes, textures, he spent a lot of his spare time getting it into a perfect state so that he would be able to render a new shot of the Spectrum Next for the box artwork amongst other things such as Kickstarter 2 promotional materials.

It was at the end of 2020 when focus turned again to the box artwork. A new colour was suggested, it was tweaked then agreed. We also agreed that the overall structure should be preserved and we would increase the number of screenshots on the back from 8 to 10. We tried 12 but it broke up the flow and gave us alignment issues. Over the months, on and off, we refined the look, the words, improving it as best as we could. We also reached out to a certain member of the community for their input which was hugely beneficial to us. All names of course will be credited in the manual.

For the screenshots on the back of the box, there is a wealth of software out there now. Back in the first Kickstarter days, there was virtually nothing available and most titles were in development. It became almost a curse as few titles featured on the Kickstarter 1 box actually made it to completion. There is still time though! Sometimes things just take a lot longer than you expect. I selected the screenshots for the box, so if something you think should be there isn’t, you can blame me. It started out as a rough idea of capturing the stretch goals first and featuring those (Head Over Heels, Baggers Detour, Night Knight), then looking at a very broad range of other software by other people, trying not to give too much to one developer or genre. It is what it is and at some point it has to be submitted for production. In each case I ask for permission to feature a particular title on the box however this does not mean it will make it to the finished product. Some changes were made along the road, and with only seven non stretch goal screenshots to fill, it’s not a lot. Apologies if you feel let down. If there is another Kickstarter, we will have another ten empty slots to fill so who knows?

In summary then, the second Kickstarter is not exciting and dangerous like the first Kickstarter was. We know what we know now thanks to all the people we worked with the first time around. They educated us about working in the real world of manufacturing. We learned a lot about electronics production, mould flow and warp analysis, injection points, ejector pins and cooling and a whole raft of other things. We know the right questions to ask as we have already lived it. If there is a third time it will hopefully be a charm! With no constraints in the world of components we could literally make a fresh batch of Next’s in a few months.

The main cloud looming at the moment is the logistics of shipping. The world has changed since we shipped the first batch. This is one for Henrique.