Selling retro computer products

[Steve Cousins]

Bill suggested it would be interesting to discuss the selling of retro computer products, so here's a place to do it.

[Steve Cousins]

Here's my story...

Most of you will know that I sell a range of retro computer products on Tindie. I'm happy to discuss all aspects of that "business".

I didn't start off with the intention of selling kits. 

I had been thinking about getting back into 8-bit computers for some years. One day I found a website showcasing people's homemade CPUs. Some were made from logic chips, others from transistors, and some just used relays. That was the final inspiration for me. 

I decided I wanted to start from scratch but was willing to use a modern PC as a development tool to develop my own software tools! My first target was to write an integrated development environment so I had the software tools to write code for a home-made CPU. To prove the IDE, I first got it to assemble Z80 code and simulate a fictitious Z80 system. I got this working adequately for my purposes but it is quirky and full of bugs. You may know this as the Small Computer Workshop (SCW). I consider it to be poor quality and incomplete software so I would not normally share it. However, I've made SCW available as a download so people can modify SCM.

At that point, I guess I should have started my homemade CPU but instead this is what actually happened:

I wrote a Machine Code Monitor for the simulated Z80 system. This eventually became the Small Computer Monitor (SCM).

My grand plan was now well off course. 

I can't remember if I started my SC101 SBC design before or after ordering the RC2014 Pro, but whichever it was, I definitely decided it would be sensible to have a known working modular system (the RC2014 Pro). 

Once I got the RC2014 Pro and built my SC101 SBC I found myself drawn into the RC2014 community. This lead me to improve my monitor and create SCM. I then got involved with porting SCM to the LiNC80 SBC1.

At some point during all this, I realised that PCBs were no-longer so expensive as to be out of the reach of hobbyists. Now an active member of the RC2014 community the obvious first board was a module for the RC2014 bus. Looking at the available modules I spotted several obvious holes and set about filling them. Thus was born SC102 a Z80 CTC module, SC103 a Z80 PIO module, etc.

I had 9 spare boards of each design. Hm! I gave a few away and then thought perhaps I should put them of Tindie as a way to cover my costs. Once I had done that the natural progression was to order more boards to replace those I had sold. Once I was in the cycle of ordering PCBs specifically to sell on Tindie I was hooked.

Whilst selling the PCBs I became aware that the volume I was selling was less than I'd have expected given the size of the RC2014 community and the number of systems Spencer had sold. The reason, I decided, was that most people wanted kits. I don't really remember why I thought it was a good idea, but that's when my kit "business" started.

Actually, I genuinely think this hobby needs people to offer kits and other forms of support to attract more people.

As for the finances, I'm happy to give away my designs (they're in the public domain) but the cost of inventory, costs incurred sorting out problems, and the hours of work it all takes means it is very unattractive to try to offer kits at cost. While money may not be a prime motivator it is about the only way to justify offering kits. So, yes, I aim to make a profit.

I think it would be very good for our community if more people offered kits.

Steve

[Nils-Arne Dahlberg]

I think one reason people want kits is that nowadays there are few of these over-the-counter one-of-each electronics shops. You need to buy the parts on-line and the low number of parts needed makes p&p a big part of the cost. After making this mistake with my RC2014 kit, I realised that it is not worth it to just buy the PCB and then the parts separately. Not having access to a prom programmer makes kits even more interesting...

[Mark T]

I think one of the problems with offering bare boards through tindie is that with the cost of shipping its almost better just to order direct from jlcpcb, although you also end up with a number of spares that may seem a waste. Also those confident enough to source components and build a board in this way probably also enjoy the process of designing their own board with some slight variation of design to keep it interesting.

Supplying boards also requires a certain level of documentation. I think the documentation you generate is very good, and probably helps to encourage those that haven’t had any previous experience of building retro computers. I don’t have the patience or interest in trying to match that and tend to lose interest other than including notes in schematics intended for myself for future reference. The extra work in documentation is probably not going to be covered by the price of bare boards but maybe could be supported by a slight markup on the kit prices.

Kits are probably better for those with less experience and looking for more confidence that what they solder is likely to work, either first time or after some simple corrections. This is why kits should probably be through hole components only.

I can see how making kits available could make the hobby more popular, but the time spent collecting parts into kits and shipping in small quantities is not my idea of fun. I’m glad Spencer was able to do it successfully to create this community but as more people try and do the same with variations of the same design it might saturate the market.

Mark

[Richard Lewis]

Same for me, i.e. I will just buy the bare PCB or if the gerbers are available send it off to have it made. I  order the parts on Mouser if not in my inventory and time it so I get both by Friday. I have a lot of extra boards now so I thought about populating them then doing a sale of pre-made boards on Tindie until I run-out then stop. Most of my designs (the ones that work at least) are all SMD with 0402 and 0603 passives and TSSOP or DSN chip packages. I'm not willing to offer these as a kit and then subsequently get slammed because the person buying doesn't have the right tools to make it successful.  

Steve set the bar really high for documentation. It's on the level I used to see from Velleman or the old Radio Shack P-box kits. For me, I have a 10hr/day job and a very active 9 year old at home so I have about 2-3 hrs per week max I can spend on my projects. No way I would have time to do proper documentation. That's why all my projects are on GitHub with a basic README and free for anyone to do whatever they want with them. 

What irks me about some kits that I've purchased is that the documentation is either none-existent, sparse or wrong. Especially annoying is when the kit costs $100-200 and I find that the silkscreen/schematic and supplied part do not match. 

No way would I quite my day job and go into the kit business. Maybe after I retire and need the extra income. 

-Richard

[Bill Shen]

I think we all like to see our designs being used and have some values to others.  There is another reason why we should make some efforts to sell:  it is a way to convince our spouse that we are not just puttering around wasting money (actually we are, but shush!).  I'm fortunately enough to have a very indulgent wife who tolerated me having a messy lab and a garage full of junks, but when some of my items fetched good money on eBay, she suddenly developed real interests in my projects.  I think women are practical and nothing spell "practical" like prospective buyers willing to pay good money for the otherwise "junks" that we've designed.  Selling elevates "nerdy junks" to "practical goods".
 
  Bill

[Francis Pierot]

It's probably also relevant to discuss about *buying* them so I add my point of view which parallels yours, but as a buyer.

My own experience in hardware is still low. I know how to solder through-hole stuff pretty well, which is sufficient to buy full computer kits without fearing of making a mess of it. Like lots of people, I also buy kits on chinese shops like aliexpress or banggood, I also buy components like resistors or whatever to try to learn more about electronics using arduinos and such. Frommy experience, these chinese components are low in price but also in quality and I probably wouldn't bear using them in a computer kit. I'd probably turn to more professional and specialized component sellers, but this also means a higher cost. From what's I've seen with RC2014 and SC126 kits, the quality of component is good. It means some serious work from the seller, and a higher components cost but it also means the buyer is less prone to have a problem with quality.

I'm perfectly ok with the retro-computing seller making a profit, because from my experience with you (Stephen) and Spencer Owens, the sellers are working hard to give birth to great designs, but also to insure a high quality level for kit components, good support and communication on forums, good or excellent documentation (your SC kits have fantastic documentation!).

And as you point out, the seller also have additional work to handle its stock, manage packaging and shipping. This is how buyers receive the kits quite fast, in a good protective packaging.

Paying only for cost would basically mean that service has no value. When I buy a retro-computer kit I expect a lot of things to compose the price, of which most are services:

* components costs

* service: kit quality assurance (including documentation)

* service: packaging quality

* service: shipping quality

* service: support

I'm also ok with the idea of giving even a little money more, so the seller stays motivated and continues his fantastic work :-)

[Francis Pierot]

Nils, in 1982 as I was still a student I wanted to add a very nice graphic board to my TRS-80. One of my  friend at ease with electronics would solder it for me. As the full kit price was out of reach for me I bought the PCB and thought I'd go to an electronics shop to buy parts. Although the shop had a really good reputation, that proved to be hell: some parts they replaced by others, some were not available, etc. Eventually my friend and I had all the parts or equivalents for a much higher cost than the original full kit, but after soldering the thing never displayed a pixel.

What I've learned from this is that if you don't have reliable components sources and/or don't really know what you're doing, buying full kits is the best if not the only option, because you have an implied quality insurance and you're assured to have the right components.

Sill, no kit protects the buyer from messing with it if he/she doesn't have the knowledge and soldering level for it, so sellers could include some king of easy to read label for the difficulty level, like pepper sauces with none, one, two or three red peppers:

* the RC2014 Pro fulll kit is one or two peppers: maybe it doesn't look that difficult to seasoned peoples but it's about 1000 soldering points, some tricks to know about bus connectors, choices to do about powering etc. Beginners, they could give up even before they end soldering the 600 points for bus connectors alone

* the SC126 is probably one pepper: it has reasonnable number of soldering points and very good documentation but you have a lot of choices to do before soldering and this is not beginners level

* any SMD kit would be three peppers, as soldering itself is not for beginners, but also any soldering mistake is hard to fix, let alone fatal

Like Richard Lewis is saying SMD soldering needs proper tools and some practice, so offering boards with already mounted SMD can be moneyed.

[Tim Hardy]

I think sellers are perfectly justified in charging for the service they
provide on top of the costs. There’s no reason to be ashamed of making
a modest profit if you are providing something that people want.

For most of the last year, I’ve found myself with almost no free time
at all and on the very rare occasion I have a couple of hours for my
hobby, it’s frustrating if I discover that I am missing a component
for the board I want to solder up. It’s also expensive and slow to put
in an order for just one or two components. For this reason, a kit is
fantastic. I can be confident that everything I need will be there when
I am finally ready to sit down and make it.

Steve, I honestly think you could charge more. Your margins must be very
low - I can only assume that you’re getting deep discounts by buying
components in large quantities. You are taking a risk that you’ll end
up with components you cannot sell and you are right to charge a price
for this. You are also giving up time to source components and package
and send items. You deserve to be compensated for that. I know I am not
the only who is very grateful for what you give so generously to this
community and I hope you take care not to burn yourself out by giving so
much. Thank you.

I honestly think Spencer’s kits are reasonably priced too, even though
they are at the expensive end of the range. His documentation is very
poor especially for the price but he does go out of his way to help
people who are having trouble putting their kits together and the
components are of high quality.

Is the market saturated? I’m not sure. Is it confusing to a beginner
to know where to start? Definitely. I think the establishment of a
different bus may have interesting consequences for this especially if
there are boards available using the new bus that have decent
documentation. (I haven’t contributed to that thread because I know
enough to know that I don’t know enough to make an intelligent
contribution!)

[Bill Shen]

These are great feedback.  It is a good idea to mark the difficult level of a kit; sellers should also state how long it takes for an experienced assembler to put together the kit.  A simple rule-of-thumb for pricing small lot of assembled/tested boards is 3 times the cost of parts.  Sometimes it is cheaper to sell assembled boards than kits because the lower level of support, documentation and ease of packaging & shipping.  Fine-lead SMT parts are particularly expensive to ship, so partially assembled board with SMT parts already soldered is a cheaper option than full SMT kit.  Having said all that, I just offer assembled/tested board on eBay auction starting with 99 cents and let the market determines the selling price.

  Bill

[Steve Cousins]

This discussion is very interesting. As a supplier, it is great to get this type of feedback.

I like the idea of marking kits with some sort of difficulty rating. Any ideas on a definition that can be applied to give consistent ratings?

Another aspect of pricing that is "fun" is setting the shipping rates on Tindie. I've tried several schemes to generate fair shipping totals. Currently, I have quite accurate shipping charges generated if you order just one product but if you order several products together the total is often not very representative of the actual cost. Tindie's system has an odd way of handling products from different shipping bands. When I buy products I really dislike it when the shipping costs seem unreasonably high, so I'm keen to make my shipping charges reasonable and fair. Unfortunately, there seems no way to achieve this.

Thanks for all the encouragement guys.

Steve

[Richard Lewis]

If the following is too long, the short version is that there should be a progression of of kit difficulty so that people who want to aggressively learn won't get stuck in a rut of paint-by-numbers kits.To me the RC2014 kits are all just variations of Grant's original design so I quickly lost interest. 

Now for the TL;DR part

My progression started this way... 

Two to three years ago I started off with a basic level of electronics knowledge: I knew the fundamental differences and function of the parts,I knew Ohm's law and Ampere's law,  I could read a resistor color code, knew roughly how they  are used but wasn't competent enough come up with something on my own and didn't know how to interpret a datasheet.  Also, could do some basic debugging with a multimeter if things went wrong. The first time I was building the RC2014 kits (the mini about a 2 years ago) I was very dependent on the documentation because to me this was still a "paint-by-numbers" hobby. The only practical knowledge I had if I ran into trouble and the documentation was not good was that resistor direction didn't matter (unless one had an OCD about that) and that electrolytics, diodes, transistors and IC's had to be put in the right-way around or blue smoke would ensue. 

My first kit (the RC2014 mini) was a massive fail... There were problems with the ACIA serial board and the instructions were abysmal. I finally gave up and it went to the bottom of the closet. I moved onto the "pro" but was dismayed that I had soldered the wrong headers to the wrong boards or otherwise made mistakes because there were no part values on the silkscreen and the schematics didn't match. I eventually got it to work but soured on wanting to invest in more RC2014 style kits. Took several days to get CP/M functioning correctly. Finding Grant Searle's original work was a big help in getting me to understand how the various components interacted and functioned. I started to learn about how to properly read a schematic. 

Anyway was getting bored playing around with Star Trek/Zork etc (was fun 35 years ago). Ran into Steve's boards on Tindie and because the documentation was so good I felt confident enough by this point (10 months ago) to order just the board from Tindie and then source the parts from Mouser. When boxes from Mouser started showing up my wife was wondering why I was taking a sudden interest in buying stuff for the cat... but I digress... 

At first I was often buying the wrong parts because I didn't carefully understand the variations in digital logic families. However, whilst building Steve's boards something "clicked" and I started to understand how all of this fits together. So was time to move onto the next stage which was sending off (someone else's) board designs to be fabricated. 

I was interested in the BusRaider board for the TRS-80 emulation. For this I decided send the Gerbers off to JLPCB rather than buy the premade boards. After I had done so, I realized to my dismay that there were a few SMD parts (albeit in hindsight some rather large parts) that needed fitting. I had been scared of SMD since my eyes are getting older and they were getting harder to see. I grabbed the BOM, ordered the parts from Mouser then bought 3 or 4 cheap SMD practice boards from Amazon. The first practice board was a disaster and realized I had the wrong equipment. Fast forward to today and I have a Hakko soldering station, a medical grade dissection microscope, a variety of liquid fluxes and a reflow oven (that I built from a kit and a $40 toaster oven). I'm pretty comfortable now with SMD at least down to 0402 parts. The bus raider was a success and I was beginning to get confident enough to design my own boards. 

I took Steve Ciarcia's original design for a single-step circuit with digital display of the address and databus, with some ideas from Scott Baker's bus monitor as well as the YAZ180 and decided to design my own board. I installed KiCAD and started to move my way up the giant learning curve. At first I breadboarded the circuit and it more or less worked. When it came time to map the schematic on the the RC2014 form-factor PCB I quickly ran out of space. So I started substituting the passives for SMD 0805 parts and even some of the smaller actives. At the time I had no idea how to properly route a board nor did I understand the purpose of ground planes. When I got the boards back from the fab house I excitedly solder it together and nothing worked. I realized I had made a ton of mistakes in routing the traces. It had also taken me about a week to route the thing since I didn't know about vias yet so the traces were a mess.  Fortunately I was able to patch it to the point that it worked. But then that's when I finally soured on THT. Easy for a beginner but takes up too much real estate. Also was having trouble finding TH parts that would cater to the ideas I was starting to have. 

Recently I've been focused on glue logic using CPLDs and replacing various component peripherals with an FPGA. 

Since then I've gone on to design 6 boards which 4 of the 6 working as intended. I now pretty much instinctively know what parts I need for my ideas and how they should fit together. I guess it's like moving from Legos to Lego Mindstorms. I have a rather sizeable inventory of parts but I 2-3hr window on the weekends to work on this stuff. 

Richard

[Richard Lewis]

I think Velleman at one point had up to 5 different levels. My rough take on it would be:

Level 1 - a one-sided board with a 555 timer blinking an LED, a schematic with a simplified explanation of parts and function. Can be put together with stuff found in the garage like plumber's solder and a propane torch (just kidding... maybe...)

Level 2 - The "standard" RC2014 boards, lot of soldering but still easy enough to build with simple tools

Level 3 - Densely populated boards, backplanes with 100's of solder joints or bare board kits. Hybrid boards with a mix of SMD and TH (maybe the Bus Raider would qualify). Kits that involve mains wiring. 

Level 4 - All SMD like 0603/0402 passives and TSOP or QFP chips, can still be hand soldered but requires precision tools and a microscope. 

Level 5 - BGA, DSN etc, parts that cannot be hand soldered and require reflow equipment: oven, hotplate, hot air gun. 

[Francis Pierot]

Taking views of Richard Lewis in count helps.

There are 2 ways of defining a difficulty level: one is fro the beginner's view, the other is from the experienced guy.

I can only help fr the first way, yet.

As a beginner, my view is that hole-through soldering difficulty is mainly proportionnal to the number of soldering points and the number of different types of parts.

If a kit only has resistors,capacitors, LEDs, switches, then it is levek 0, anybody will eventually get it with even a basic iron and poor solder wire.

Level 1 would mean more points, or more not so easy to sodler arts: say some ICs, or more than a few hundredssoldering points, a few choices of powering or options you have to set on dip-swtches or jmpers. This implies a minimal knowledge of how the electronic world works, and you cannot simply draw by numbers to get the kit working, you have a minimal thinking required.

Level 2 would be as level 1 but with a quantity level, like say, 10 or more different kind of components, more than 200 soldering points, different powering ways, some I/O (pots, or switches, or digital/analog inputs connectors, hdmi or leds outputs...). really, more like a higher level 1.

Level 3 would imply SMD, or something implying a lot more knowledge and tools, like EEPROM or Flash programming, you get the idea.

From a seasonned maker, I bet these levels would all look like level 1 except the 3rd. I think difficulty level is mostly usefull for the beginners like I was 2 years ago (and still am, given I stopped any related activity for about 8 months.)

A number of work hours to build the kit  could be given to the seller by the buyers themselves so that he could tell what would be a reasonnable amount of hours on his selling page. It would not be mandatory, just an indicaition of how much time it generaly takes for peoples to finish the kit. Some buyers will go faster, others slower, depending on the time they want or can spend. But it would at least give an indication. If you want an RC2014-Pro and the estimated time is about two weeks (shared with normal life) then some people would rather start with a less demanding kit (RC2014 mini for example) and not be discouraged or worse, put their family life in danger. (Just exagerating things, of course :-)

[Phillip Stevens]

I guess as the seller of the "most expensive" z80/z180 system on Tindie, at $499, I should say a little about why I think that price is sensible, compared to Steve's SC130 kit which is almost exactly 1/10 the price, at $49.

Well, frankly, I can't see how Steve can sell his kit so cheaply. The BOM must be close to his selling price and, unless he's buying in bulk, I'm sure that it would be higher than the kit price. Not to mention the work involved in assembling (collecting, making safe, and packaging) of the kit contents for shipping.

When I'm pricing up what I produce, I try to look at three levels.

1. Designs. I do designs and software for enjoyment. I don't want to feel like I "have" to do anything for anyone. Interest guides the work. That is the main reason why all the design and code I write is all open source, and well documented. Take it, use it, or not.
2. PCBs. Specifically for the YAZ180, the PCB is very expensive in small quantities. 160x100x2mm, 4 layer, 2oz copper, ENIG. An order of 5x costs me around US$170 shipped to my door. So, I've recently ordered 40x (at US$620) to make available on Tindie. That is a cost of US$15 each, which I sell at US$25 each. Not a commercial margin, but enough to make holding the stock worthwhile. And, for the buyer it is still cheaper than ordering an individual PCB.
3. Finished. As a pre-built working system, the US$499 selling price is based on a PCB costing US$34, a short BOM costing US$135, as well as Am9511A (US$10+) and 9x TIL-311 devices (US$60+), making the input cost total US$270 in parts. So then I have to look at how much time I have available, and how to spend it. Building a YAZ180 takes me about 8 to 10 hours, based on my experience last Christmas where it took a full week to get 5x successfully built.  I'm valuing my labour at around US$20 per hour. So, although I'm happy to do that occasionally, it is not something that I want to do every day. Hence, although the selling price is not completely commercial, I set it high enough to encourage me to build more (and to justify to my CFO / wife that it is worth doing), when they get sold.

I would love to sell kits, but the effort to assemble the components is not justified where a complete BOM can be obtained from Digikey (or similar), which is packaged and priced with only one set of shipping and tax to pay. So, I'd rather just go with quoting the BOM shortlink, and leave it there for anyone to order as they see fit.

Another issue that I have is that postage from Australia for bare PCBs is more expensive than the PCBs themselves. I have set the Tindie Shipping at $25, though the actual price at the Post Office usually it is slightly more than that. It varies depending on EU, UK, or USA/Canada. But I figure that I'd rather have the business than worry about the price of a beer amongst friends.

And that is why the YAZ180 fully built costs 10x the SC130 kit.

Cheers, Phillip

[Bill Shen]

If a board is designed to be built in stages so it can be powered up and tested at each stage, does that reduce the difficulty level?  For single board computer, such assembly approach is quite reasonable: a minimal configuration can be assembled and tested, then more peripherals and connectors added later.  Staged assembly is easier to test because at first stage the board only has a few parts; it also simplifies the problem so if a board is working at stage 1 but failed in stage 2, the problem is most likely due to the parts added in stage 2.  I think it is also more educational in the sense that different aspects of the hardware is explored at different stages.  The downside is the documentation take quite a bit more efforts.  I have a couple examples of staged assembly published on Hackaday.  This one is about Z80SBC64: 
https://hackaday.io/project/163786-building-a-22mhz-z80-computer-in-4-stages

  Bill

[Sergey Kiselev]

Interesting discussion.

I've been designing my own boards for years. And as Richard and others have mentioned I would order 10 or more PCBs, and then I'll use one or two and sell the rest for the price that would cover the PCB manufacturing costs.

I was considering selling kits, and every so often people would ask me about a kit. But so far I was avoiding that. My concerns are:

• Kits require more support. If I sell PCBs buyers are pretty much on their own. It also guaranties some level of expertise on the buyer's part. My assumption is that if people by a kit, they'd expect me to provide more support and troubleshooting, and possibly returns if something goes wrong.
• Inventory management. I'd need to invest in purchasing a large stock of various parts, with the hopes that I'll eventually I'll sell enough kits to pay off my investment.
• Packaging: It takes more time to package a kit properly... Bags/boxes/anti-static packaging for ICs. 
• Especially for the discontinued parts normally obtained from various Chinese parts recyclers, there's always a concern that parts will be damaged. I had a not very pleasant experience, where with one of my PCBs I've decided to offer such a part (all other parts could be ordered from Mouser). I purchased several tens of these discontinued parts from China. I've spent an evening testing them on my own board (and found a few defective parts), and then I shipped parts that tested OK with the PCBs. One buyer contacted me and claimed that part didn't work. It is possible that I didn't test all the aspects of the part functionality, or that he damaged the part, or it was just a dishonest person wanting to get yet another spare part for free (hope not). In any case, I've shipped the new part, losing $10 for shipping and part cost in the process. Maybe in a large scale operation it wouldn't be a concern (one can simply plan for 10% returns, and include that in the kit costs).
• For my (usually relatively large) designs, the parts cost is fairly high. About $100-$200. That would make kits pretty expensive... adding the PCB cost, a bit of markup (packaging, my time, a bit of profit), etc. $150-$300 a kit. And I was skeptical that hobbyists would be willing to pay that much for a project.

I am wondering how the people that sell kits are dealing with these concerns?

How much time do you spend supporting the kits? Do you allow returns/refunds? How do you deal generally with parts ordering (buying in bulk, what distributors, how about "vintage" components)? Any other recommendations?

Thanks,

Sergey

[Steve Cousins]

In answer to Sergey's questions, the following has been my experience.

Kits do require more support but so far it hasn't been a big problem for me. My designs are not very complex and with detailed documentation, there have been very few buyers who have had to ask for help. When I do get an issue I generally revise my documentation to reduce the chance of the same thing causing problems for others, 

I've not had any returns due to problems with the kit. I had one where the buyer could not get it working despite his photos showing good workmanship. I sent another kit to solve that problem. I have had a few kits that never made it to the buyer. I just replace these.

Inventory management is a real problem. I don't have the space to organise stock in a way that would be easiest to manage. Despite reasonable effort, I do run out of things. This can be a problem with items that take weeks to get here (I don't price in the cost of fast component delivery). I have ended up with quite a bit of money tied up in stock. At some point, I will end up with stock that hasn't sold and ends up being worthless. I'm just ignoring that at the moment!

I could reduce my inventory issues by buying everything from the big distributors, such as Mouser, RS, and Farnell, but this would significantly increase my costs and thus my kit prices. I accept the hassle of more stock management to keep my prices down. I find I can save money by buying simple mechanical components, such as bus connectors, from China. There can be significant savings buying ICs from China but not in most cases - especially once you include delivery and risk.

Packing does take quite a while. Storage of all the packaging materials is a space issue for me. I have been careful to cost in the packaging materials. Individually the bags, boxes, tape, etc. doesn't look much but it all adds up. If I was better set up for this kind of business I'd use a more appropriate printer and self-adhesive address labels. It takes a lot of time and tape to turn a batch of cheap inkjet printing, on standard photocopy paper, into robust and water-resistant labels.

I've had bad experiences buying ICs from China. Where I have resorted to such sources I try a few and if I'm happy I order a lot of them (and hope I get the same batch as the one I tested). I really dislike using this method but sometimes it is necessary. I think I've been lucky so far. I've not experienced problems where some of the chips fail. If a sample seems to work then the batch has so far always been good (as far as I know). If I start to get problems like random chip failure then my kits would be unreliable, and that is unaccptable to me. 

I deliberately design most of my products to be at the low end of the price range. I think this is the area where a kit supplier can be most useful to the community, especially to new users, which is my main focus. More niche products, like my CTC module, sell in low volume and probably only appeal to people who are quite experienced - and would quite likely be happy with just a PCB. 

One problem I get here in the UK is customs charges. I often get hit with charges for items I import. This usually amounts to 20% of the purchase price, which is what we are lucky enough to pay on most goods purchased in the UK. Our government thinks that is good for us! The real sting in the tail is that the companies processing these imports add a handling charge each time there is a customs charge. This charge is between 8 GBP and 13 GBP (about 10 USB and 16 USD) depending on the company involved. This handling charge can wreck the costings on all but the largest of orders.

One reason I don't get many problems is that you guys who buy the kits are very reasonable, tolerant, and take responsibility for your purchasing decisions. With typical consumer goods, there will be people who order on a whim and expect to open the box, change their minds, and send back the product. Retro computer kits attract a much more reasonable buyer.

I'd like to see more people offering kits but be warned: It is a huge commitment. It takes much more time and effort than you might expect. Stocking the number of components needed requires a different approach to storage than your personal R&D stock - the boxes are bigger and you need more space. Of course, if you only sell a few niche products these problems are much reduced.

I think I've been lucky so far but good documentation improves your luck :)

Steve

[Richard Lewis]

I totally agree with this. The best example I have ever encountered of this philosophy is the Nixie Transistor Clock which I built 2 years ago and has over 700 components. I was a beginner then so at first it was a bit daunting but as one can see in the manual, the board is developed in sections where each is tested sequentially as it's being built. I spoke to the designer about build failure rates for his kits and he said: "at worst there maybe a bad transistor or meta-instability from forgetting to place a bypass cap on a flipflop". 

I think at minimum for beginners documentation should get one to the point where the power distribution can be tested sans IC's (which Steve does in his docs), followed by clock (assuming one has a way to measure frequency such as a multimeter, oscilloscope etc) and a section on troubleshooting. 

One thing I've started to do is to add test points to my designs. It's often hard to get probes attached if I have to debug something on the inside of the board

Richard

[Bill Shen]

I do buy parts from China; they are mostly problem-free, but I am concerned about parts functionalities.  I tried to keep my design simple to keep the part count down to save cost as well as reducing inventory.  CPLD has been very useful in keeping the part count down.  I reuse parts in my designs so I can buy large quantity of few parts.  Sometimes you can get great deal in large quantity purchases and that drives down the cost of boards.  Because of the concerns about parts functionalities, I sell mostly assembled & tested boards so I know the parts are all working correctly.  With low part count, parts from China, and lucky purchases of large quantity, $50 for assembled & tested SBC is possible but probably not sustainable long term.  Recall I've said previously that assembled & tested board should sell for 3 times the part cost. 
  Bill

[Peter Willard]

I would like to point out --->  "detailed documentation"  It makes a BIG difference.

[Nils-Arne Dahlberg]

And detailed documentation that can be checked before I order is even better. I have not seen any problem with hidden documentation from the people in this forum, but I have run into it on ebay some times.

[Karl Albert Brokstad]

It’s a very interesting discussion, with many constructive views

 

My first experience with the RC2014, was the Classic.  I assembled the kit and it didn’t work. After some extensive troubleshooting which failed, I ordered the CP/M upgrade kit. I hoped this would help me identify any problems with the memory modules, but it still didn’t work. I was then sitting with nearly $200 worth useless junk.

 

At this point, I was considering building my own modules to help get “the thing” up and running. I had some initial trials with KiCAD and Eagle but found the EasyEDA to be the best fit for me. Then I designed my own classic modules and sent them off to Jlcpcb. After assembling my own kit (“NeoClassic”), the system was up and running within minutes. I was also then able to test all the other modules and locate the fault of the original serial module. This was the start of designing modules, and it becomes a bit of an addiction. Within a few months I had designed Z180 and 1m memory modules, SIO, KIO, and RC80 bus backplanes (S. M. Baker was the inspiration), and suddenly I had over 50 designs. Many of these were variations in the layout of components or a combination of functions. 

 

I have never been very keen on selling my kits, and I gave away lots of modules (often working prototypes I had laying around) to people who needed help to get their RC2014 running. I hope I managed to help a few this way. After some consideration, I did start offering some kits, basically, repair and upgrade kits. I think some find it difficult to make contact on the forum and ask for help, and it is easier to just buy a repair kit. I have sold a few of these kits and I hope it has been helpful. My biggest worry is having lots of sales, that would involve getting large inventory, a lot of work with securing the supply chain, making the kits, packaging, and posting and so on. So, I have settled on offering niche products, and with the current sales of a few kits per week is fine with me.

 

Keeping up with supplies is demanding. I tried the China method, but it doesn’t work very well for me, so I source most of my parts from reputable vendors like Mouser and other European suppliers, but it comes at a cost. There are a few parts I am dependent on China/eBay and that is the MC68B50 chip, the 22pin IDE to CF adapter and CF memory cards. Of these only the CF memory cards are troublesome. 

 

When pricing my products, I make a little profit but not much, and I am not able to compete with Steve on price. When I sold my “Black adder kit”, the components cost was over $50, and I sold the kit for $99. I have not had any complaints on my kits yet, touch wood. If the kit doesn’t work after assembly, the customer can contact me. If we are not able to easily remedy the problem, I will either send new parts or a complete kit, regardless of whose fault it is.

 

One reason I think Steve’s products are a very good buy regardless of price is the added value to the products. You receive good product design, documentation, support, expandability, and functionality. The effort Steve has put into his products has been an inspiration to me, but there are also many others not mentioned here contributing to the community.

 

[Bill Shen]

Karl,
You mentioned that you are having difficulties getting CF memory cards.  If you don't mind large quantity of CF cards, you can try this eBay seller:
https://www.ebay.com/itm/Lot-of-100-Unbranded-64MB-CF-Compact-Flash-Memory-Cards-Tested-Formatted/392463554883
I've tried these cards; they are used Toshiba 64M CF cards.  I believe I've test all 100 cards and they all worked, so I basically have acquired 64MB cards for 60 cents each.  Shipping internationally will be more expensive, but it is still a bargain.  This allows you to include CF disk already loaded with software in your kits.  I've found transferring software to CF disk to be one of the hurdle for the new users.
  Bill

[Phillip Stevens]

Steve Cousins wrote:

While money may not be a prime motivator it is about the only way to justify offering kits.

I forgot one additional key motivation for me to sell things on Tindie.

Selling stuff provides a positive PayPal balance, making the acquisition of other kits seem completely free... ;-)

Phillip

[Karl Albert Brokstad]

Thank you Bill

It's a tempting offer.

I do add software to the CF-disks for the "missing module". 

That's how I find and isolate bad cards.

I found that the cleaner the CF cards the higher the failure rates. The really dirty ones with finger marks are always fine.

Karl

[Bill Shen]

Karl,
Randomly picked 9 CF disks and took a picture.  They are clearly used disks without labels, but the electronic ID says it is Toshiba 64MB.  It is thin enough that I think I can send you one in regular international letter mail.  If it got damaged or kicked back, oh well, it is no big deal.  Would you like to try one out?  I believe I have your address already.
  Bill

[Bill Shen]

Several designers, myself included, are not interested in offering kits for their designs for many reasons.  I do agree with many reasons stated in this discussion topic, but I also like having economical kits for the hobbyists without too much work on my part.  Mouser and Digikey offer Bill-of-Material service, but they are really quite expensive and a number of retro parts are not available.  If you search for "Z80 kit" on eBay, you can find a large number of Z80 kits offered.  For examples, there are Z80 Zeta kit for $22 which 35 were sold and Simple80 kit for $6 which 40 were sold.  I don't know about the origin of the Z80 Zeta kit, but I can tell the story of the Simple80 kit:  Originally it wasn't called Simple80--it was just Z80, SIO, RAM, flash, and few logic for $6.99.  I thought of a design (Simple80) that can use most of these parts so I ordered a kit to try it out.  It turned out the kit description was not correct and the SIO turned out to be a PIO, so I wrote to the seller whose English is quite good and resolved the mistake satisfactorily.  In subsequent conversations I mentioned the Simple80 bill of material and asked whether he can build a kit for it.  He can't supply the oscillator, but was able to supply rest of the ICs for $6 (plus $6 shipping).  So I ordered 2 kits and verified they do work.  I ordered two more kits in the past few months just to spot check the quality of parts and they were all good.

Based on this experience, I think the designers can work with the parts sellers to provide cheap kits for their designs, but I also believe part sellers are aware of hobbyist computer market and have put together corresponding kits on their own initiative.  The easiest approach is publishing the bill of material and letting part sellers discover the BOM and putting together kit independently; there are no connection between the designers and kit supplier so if the parts are not right or defective, the designers are not responsible in anyway.  However, because there are no communication between the designers and kit suppliers, the kits are more likely to have problems.  Ideally the designers should monitor the quality of their kits and while not taking responsibility for quality of every kit sold, can confirm a particular sample of kit is correct and working.  If we do this right, we can dramatically lower the price of retro computers, making them available to more hobbyists without the designers getting involved with the tedious kitting process.  I like to hear people's thoughts.
  Bill

[Richard Lewis]

I'm having really bad luck with FPGAs and CPLDs bought on the cheap from eBay. I order 2 MAX7000S CPLDs from a Chinese reseller on eBay and by accident received 20 (I'm thinking win for me). So far 7 out of the 20 will not respond to JTAG ISP. I created a simple test bed, dropped in 2 chips I got from a domestic source who insisted they were NOS and they both worked right away. It's possible the chips I purchased had the JTAG pins permanently disabled, where repurposed and sold as NOS. Noticed on some of the chips that the logo and labeling are faded suggestion they've been cleaned by solvent then repackaged in a pick-and-place strip.  

Since I'm not really interested in building kits for the masses, sourcing cheap parts is not a prerequisite for me (although my wife wishes otherwise). Nonetheless, I've decided if I can't source my BOM  from Mouser or Digikey then I probably won't build it unless I know the pedigree of of the parts that I'm purchasing.

-Richard

[Steve Cousins]

One observation I can make about life as a kit builder is this:

If the volume gets above a few kits a week, the kit building uses up all your playtime :-(

I'm having trouble finding time for any new projects, which I think is the biggest reason not to embark on that path.

Steve

[Bill Shen]

The lost of play time is probably the main reason I don't sell much.  I sell enough to keep my PayPal account in the black.  However, looking at the evolution of your SC126, I find a new reason that I should consider selling: prior to SC126, Z180 is not well supported, but the success of SC126 brings out enthusiasms for Z180 and supports from z88dk and there are several discussions about new hardware for SC126.   I've been interested in Z280 for quite a while and have designed several Z280 SBC.  Other than "lowen" in retrobrewcompter.org, I'm the only other person actively working with Z280.  My designs are all published online, but probably because I'm not actively selling them as kits or assembled board, the Z280 hardware/software base remain weak.  Perhaps it is time for me to sell Z280 products cheaply to create enthusiasm and support base for it.   This should be an interesting social experiment.  What is a good price that may generate enthusiasm?  Obvious below $50. $40, $30?
  Bill

[Karl Albert Brokstad]

Hi Bill

I think it's a good idea you start selling your stuff e.g. at Tindie. Start small, and your Z280 design is a good start.

I found complete kits to be easier to sell than just empty boards, so focus on complete kits.

I would recommend asking a reasonable price for your kits, eg. in the range $50 -$100 for a kit containing Z280 with RAM and ROM.

For comparison (starting price)

Z80 SBC

SC114             $38

https://www.tindie.com/products/tindiescx/sc114-z80-motherboard-kit-for-rc2014/

Z180 SBC

SC126             $58

https://www.tindie.com/products/tindiescx/sc126-z180-sbc-motherboard-kit/

SC130             $49

https://www.tindie.com/products/tindiescx/sc130-romwbw-cpm-computer-kit/

Z180 modular systems

SC203             $95

https://www.tindie.com/products/tindiescx/sc203-modular-z180-computer-kit-for-rc2014/

Black Adder     $95

https://www.tindie.com/products/karlab/full-z180-kit-black-adder/

Z180 Micro      $59 --> $95

https://www.tindie.com/products/karlab/70-z180-micro-kit/

Karl

[Bill Shen]

I think an EPROM programmer is a cost hurdle for people new to retrocomputing.  my Z280 SBC are all ROM-less so no EPROM programmer is needed.  I'm thinking of a Z280 design that can also program EPROM and serve as an EPROM programmer.  Z280 is inexpensive, about $3 from UTSource, so 100mmX100mm ROM-less Z280 SBC is in the range of $20-$25 part cost.  Giving the fee charged for Tindie and PayPal and overhead expense, $30 is probably the lower limit of a CP/M-ready kit with a CF disk included.  Shipping would be extra.
  Bill

[Steve Cousins]

Hi Bill,

It would be good to see more kits available, especially those that offer something not currently available. eg. Z280

A few things worth noting about pricing (mostly obvious, but still worth mentioning I think):

Once you sell in any significant volume you have to start buying in packing materials (you won't have enough 'free' recycled materials). Some parts will cost more than you normally pay as you won't be able to rely on picking up occasional good deals. You will spend more on stock than you may expect. Allow some 'profit' to cover problems that you will probably want to solve at your expense, such as items lost in the post and kits that don't work when assembled.

It takes a lot more time than you may expect, especially if you have a lot of different kits with a lot of different components to stock.

Steve

[Karl Albert Brokstad]

Steve, you have some very good points, the cost increases.

Selling stuff regularly requires an improved supply chain and quality components.

I would appreciate having a boot ROM, for the sake of ease of use.

If it cost $5 or $10 more, its worth it. The buyers don't need a ROM programmer if you supply it with ready programmed PROMs.

Also, I think it would be a bonus if the Z280 system could be supplied with something like SCMonitor or ROMWBW.

Bill, don't let this discourage you, as we have seen there is a marked for your designs, and don't sell to cheap.

Karl

[Bill Shen]

Steve,
What to do about non-working kits is a particularly good point. I have no experiences with that because I've always sold assembled & tested boards for $50, which is actually quite profitable. My concern is kits are no cheaper than assembled & tested boards when additional documentation, defects and end user mistakes are taken into consideration. I would be curios to know which will sell more if I offer kits and assembled/tested boards both at the same price? I'm incline to think I'd sleep better if I offer functional, but partially assembled Z280 motherboard at say, $30. It won't have expansion connectors so the end users need to buy and assemble their own expansion connectors. In parallel I can also offer $50 for fully assembled boards like I've always done.

I already have to buy new boxes and packaging materials. I'm on my third batches of 50x shipping boxes. I thought I had a large surplus of ESD bags, but I'm now looking for more bags.

Karl,
The ROM-less designs do not mean they need to be reloaded every power cycle. The battery-backed bootstrap RAM code is banked out of way and only visible during reset. It needs to be reloaded when the battery ran down which is a couple years, according to my calculation. SCMonitor is easy to port so is always included. Quite often I also offer a version of SCMonitor with StarTrek already loaded, so uses can do a BASIC warm boot and run StarTrek right away. I have not worked with ROMWBW, but felt the ROM-less SBC can easily be reloaded with application-specific boot code that ROMWBW may not be needed. Then again, I really haven't work with it to speak knowledgeably.
Bill

[Karl Albert Brokstad]

Hi Bill 

I have built one of your systems, the zz80rc and it works well, well it used to until I started modifying it.

I think I first mounted a battery pack, but this only lasted a few days, and then I mounted a supercapacitor and then I never got it working again.

I struggled a lot to get the bootstrap up and running, I could load it with TerraTerm after some trials, but not from a mac. 

I think this was due to some windows specific code in the bootstrap binary file.

Karl

[Bill Shen]

I pulled out my ZZ80RC which has a tag "Battery @ 9/20/18" and measure its voltage.  It reads 2.91V, and still power up fine.  Having the battery ran down in a few days doesn't sound good.  Are you using AS6C4008 for RAM?  It has low standby current.  Another possibility is DS1210 may not be working.  One temporary solution is removing DS1210 and jumper across pins 1&8, and pins 5&6 which bypass the battery-back feature and power up again.  You should be able to bootstrap with serial port setting of 115200, odd parity, 8 data, 1 stop.  Instruction here:
https://www.retrobrewcomputers.org/doku.php?id=builderpages:plasmo:zz80rc:getting_started

ZZ80RC and its enhanced version with compact flash were designed with the constraints of 100mm X 50mm form factor and through-hole technology.  With a 102mm X 102mm PC board format, the design can go back to the traditional EPROM & RAM approach.  Serial bootstrap mode is still desirable because it allows the board to be powered up without any program in EPROM, loads the programming software in RAM to program the EPROM.  It is like having a build-in EPROM programmer. 

Hmmm, this is a textbook microprocessor design so easier to understand for people new to retro computers; it also bypass the cost and hassle of EPROM programmer.  Gee, I guess I've just talked myself into it!

  Bill

[Alan Cox]

>You should be able to bootstrap with serial port setting of 115200, odd parity, 8 data, 1 stop.

I had a lot of trouble with that until I realised that some of the USB-serial adapters I have (or their drivers at least) don't actually support the wacky 8bit with parity (9 bit total) serial that Z280 uses for download. It reported it did, but the output was not 8O1.. 8) I guess it's not a case anyone tests any more!

Alan

[Bill Shen]

Thanks for the feedback about possible problem with odd parity.  The other problem I've struggled with was the lack of hardware handshakes on some of the USB-serial adapters even though the handshake signals are clearly labelled on their connector.  Of all the problems I've had with the first hardware I sold, Tiny68K, vast majority of that were the 5 wires connecting to the serial port.  It has gotten so bad, that I now always include the USB-serial adapter with everything I sell.  CP2102 is $1.25 each in bulk quantity but well worth the trouble it saved.
  Bill

[Francis Pierot]

A question to all of you who crossed the step to sell kits, but also to peoples who regularly build circuits.

I'm starting to build a stock of components and am already confronted to the inventory problem.

What do you use to keep track of your inventory?

Some specialized inventory software, a simple Excel sheet, a (less likely) text file ... ?

[Steve Cousins]

Eyeballs!

I do have a spreadsheet of inventory but it is too much hassle to keep it up to date.

I struggled for quite a while with the spreadsheet but I kept running out of things. I was relying on the spreadsheet as my storage solutions were very poor and I had stuff stacked up so I couldn't see what I had.

To fix this I purchased storage drawers large enough for my modest production needs. My previous "hobby" sized drawers were too small. Now I can see what I have and have got used to the volume of parts I get through it is relatively easy to keep on top of stock.

Demand spikes and occasional incompetence still catch me out from time to time. I could still do with making improvements to my storage solutions but space is a real problem for me.

I keep a history of purchases (suppliers, prices, and notes) to help manage buying.

Steve

[Sergey Kiselev]

I've tried it all :-)

• Text files
• Spreadsheets
• A special inventory application
• Writing and using my own web-based application

Any inventory tracking method requires constant updates. To make things worse, when working on a project, components leave the storage, and then some components return back. And that might happen many times during a few hours of project testing and debugging session.

So I kind of gave up. Instead I try to keep the components more or less organized in the storage:

• I use a couple of hardware storage cabinets with drawers (e.g. https://www.amazon.com/Akro-Mils-10124-Plastic-Hardware-6-5-Inch/dp/B003TV3NL0 and https://www.amazon.com/Akro-Mils-10144-20-Inch-16-Inch-Hardware/dp/B003P2UOCO). The nice thing about the bigger drawers is that they fit Mouser's transparent plastic bags nicely. So I keep most components in these bags.
• The resistors are sorted by value. I have a drawer for each one of the following: R < 100 ohm; 1 ohm <= R < 1 k, 1 k <= R < 10 k, 10 k <= R < 100 k, R >= 100 k)
• The capacitors are sorted by type. I have separate drawers for ceramic MLCC capacitors, for electrolytic capacitors, polymer capacitors, and various others capacitors.
• Several drawers for connectors: A drawer for D-sub connectors, a drawer for pin headers, etc.
• Several drawers for small hardware (M3 screws, nuts, standoffs, etc.)
• Several drawers for individual ICs in DIP packages, sorted by series/type (e.g. a drawer with 74LS, with 74HCT/AHCT, etc)
• Several drawers for LEDs (by size/type and color)
• A drawer for transistors and another for diodes... I don't use too many of these...
• A couple of bigger IKEA storage bins for ICs and sockets packed in tubes and bundled by series/type
• Several smaller bins for SMD components, switches, small displays, crystals and oscillators.
• Several bins for the current and upcoming projects
• When working on a project, I collect all the components it might require in a bin, moving bags from other storage. This makes it easier to ensure I have all the components for a project. And since the components are organized, I know where to look, and I can quickly identify if I am missing and need to order any components.

Generally, I try not to stock up on components, and try to order the exact number of components I need. Although it is a difficult balance, as sometimes there is a minimal purchase quantity or good discount for higher quantity orders. In addition it is nice to have commonly used components in stock.

Thanks,

Sergey

[Karl Albert Brokstad]

I basically keep tract of components by visually inspect my inventory.

Most used components are kept in 2 racks with 3x6 drawers. 

I have also made shopping (project) list at Mouser, so I get all the parts I need to make a certain number of kits, that helps a lot.

I do run out of parts sometimes, but thats mainly due to the unpredictable supply of China parts (very few).

Karl

[Denis Hennessy]

I don't have any affiliation with the author but can recommend http://partsbox.io, the free plan is enough for most makers AFAIK.

The inventory management isn't too rigid, so it's possible to store different parts in single locations if that's what you do in real life (e.g. a box labelled ZIF Sockets with various sizes).

Denis

On Thursday, 12 December 2019 16:58:47 UTC, Francis Pierot wrote:

[Bill Shen]

I'm an "opportunistic" designer.  I buy unnecessarily large amount of components off eBay (cheaper by the dozens!), and try to design something with them.  The best example is EPM7128S; I used it in most designs because I have hundreds upon hundreds of them; 68000 with 16Meg SIMMs are another example.  I didn't get into Z80 until I bought 10 of them for $9.90, and so on.  I do try to keep my components the same, CPLD really helps reducing the inventory of random logic.  I buy the common components (resistors, capacitor, headers, power jack, connectors) in the hundreds or even thousands because they are generally not expensive.  I design with CF disk only because I have big lots of them.  I don't worry about the items I can get from Jameco or Mouser because they deliver very quickly.  I like SMT resistors & capacitors because they are cheap and easy to store, and they are useful in prototyping and engineering changes because it is easy to sneak a few somewhere on board.

So no, I don't have much of an inventory system, just use few common types and buy a whole bunch of them and make up the rest on the fly--the Design-to-What-I-Have methodology, my mentor is spinning in his grave right now...
  Bill

[Alan Cox]

On Friday, 13 December 2019 14:17:11 UTC, Bill Shen wrote:

I'm an "opportunistic" designer.

Ditto sometimes, and then I design around the chips I have somewhat. eg I tend to use 74HCT688s for all the address decode because I once bought a couple of thousand of them and of a few other chips I needed a few of for less than mouser wanted for about 20 new ones.

Other than that I try and keep a rough tab on what I have a few spare of and order what I need as I do builds or new board designs to match the boards,

Alan

[Tom Szolyga]

Hi Denis,

Thank you for mentioning PartsBox.  I had forgotten that I created a free account 3 years ago.  I logged on and my part list was still there!

Best regards

Tom

[Francis Pierot]

Thanks to all of you who replied. I now have a much better view of how to keep track of my inventory.

First I bought two storage units with drawers for my most common ICs and passive components and pin headers etc. It already helps me a lot!

Second, I bought another kind of more little storage unit to keep less common components, like a few Parallax Propellers, EF9367, EEPROMs etc.

Now, about keeping stock for kits, I'm not there yet so I have no need for this so there will be no excel sheet or specialized software, not even BOM lists on components sellers like LCSC or Mouser.

I'll make this evolve if/when needed.

Thanks again for all the very different and appropriate advices!

[Francis Pierot]

Hello everybody,

I'm working on a possible kit and have a number of questions related to selling this.

Where do you find your parts? I have three levels of quality suppliers right now: low-cost with variable quality and slow shipping (AliExpress and Banggood), reasonnable quality and shipping (LCSC) and high quality and fast shipping (Mouser).

At LCSC and Mouser I find components I won't find at AE and BG, but at these last too I'll find mechanical parts for low prices that neither LCSC nor Mouser are selling.

Any other suggestion?

Another question is where do you find the little boxes in which you send your kits?

What shipping options did you consider or adopt for your kits for:

• overseas (in my case, europe -> US)
• uk/europe (europe->uk)
• europe->europe

Also, how do you evaluate the right price for your kit? I have a base price which is made from:

• the cost of each individual component, computed from how they cost me in lots
• a minimal price for specific cables building
• the cost price of each PCB (how much it actually cost me, although I don't count the first versions with mistakes in that price)
• a little cost for all the work it implied
  

I'm also afraid of the price I reach right know, and am hesitating between building a stock already so people won't have to wait one month while I order the parts, or asking for pre-orders so people are advancing the money for me to buy the components.

The cost of the kit is not something I can advance too much, so I'd like to get an idea of how many people would be interested, but how can you reliably do that?

Thank you for any advice!

[Steve Cousins]

Hi Francis

I get my PCBs from JLCPCB (no surprises there!). I get most of the ICs from Mouser, RS, or Farnell. I get most of the simple passive components from UK based distributors. Most of the simple mechanical parts and connectors I get from China.

As my volume of sales increased, time became the limiting factor, so I now get more parts from 'proper' distributors. Searching for good prices takes a lot of time. Often the prices were only significantly better if I avoided customs charges (particularly the handling charges which are daylight robbery). To avoid customs charges here in the UK the threshold is about 20 USD. This means buying lots of small batches which takes much more time to organise and keep track of. Often I found that when I went to reorder the deal had changed or was no longer available so more time needed to be spent. As the volume increases the price from proper distributors comes down, closing the price gap on the shopping around approach.

I get the padded bags and cardboard boxes from UK sellers. For example:

https://www.ebay.co.uk/itm/C4-C5-C6-C7-SIZE-POSTAL-BOX-ROYAL-MAIL-LARGE-LETTER-POSTAL-CARDBOARD-MAILING-BOX/112064402396?ssPageName=STRK%3AMEBIDX%3AIT&var=414177448000&_trksid=p2057872.m2749.l2649

The cardboard boxes I use are made specifically to match the sizes and price points our national mail service (Royal Mail) offer.

I have tried various shipping options on Tindie. I tried complex systems to try to give a fair price depending on what was ordered (eg. total weight), but due to the complexity and the limitations of Tindie's system it didn't work very well. I now offer a simple fixed shipping price regardless of how many items are ordered. I set the shipping charge so the most common package size and weight is almost exactly what it costs me. On average I think I slightly subsidise the cost of shipping. I have three prices: UK, Europe, and everywhere else.

I price my kits to make a profit, not just cover my costs. Remember all those slightly less obvious costs I listed in a previous post. They have to be covered too. The actual price I end up with does not follow a formula. I price different things in different ways. If I want and expect higher volume I tend to keep the price low but if I expect the volume to be low I go for a higher price as I end up holding stock for longer and fixed costs are harder to cover. At the end of the day the price is whatever I feel is right. When pricing don't forget to cover Tindie's charges (or eBay etc).

Trying to determine in advance what the demand will be is difficult. I've got products I think should sell well, like SC114 which I think is suitable from a large number of new users, but does not sell well. Similarly things like SIO modules don't sell well. As mentioned previously I think this is because Spencer's officially branded products are the obvious purchase for most users when they start out. That means core modules have already been purchased when users start looking at the broader options.

Niche products will by definition be low volume and most designs that are likley to be popular have already been done. The volume sales are likely to be from things that are sufficiently different to attract second purchases. For me that is my Z180 based SBCs. My modular backplanes also sell quite well.

I don't think you can reliably determine demand until you have a product. I think the best thing to do when launch a new product is this:

• buy a large stock of the items that are relatively cheap  (eg. resistors)
• buy a large stock of the items that take a long while to arrive (eg. connectors  and pcbs from China)
• buy just a few of the relatively expensive items that are in stock at major distributors and will be delivered in a few days (eg. CPU)

Steve

[Karl Albert Brokstad]

Hi Francis

There are many good sources

Some are better then others in your region

Mouser 

https://mouser.com

Farnell  

https://www.farnell.com

RS 

https://uk.rs-online.com/web/ 

elfadistrelec

https://www.elfadistrelec.com

Kessler electronic

http://www.kessler-electronic.de

Utsource

https://www.utsource.net/

(for discontinued/odd parts)

and Ebay

There is a misconception that ebay/china is much cheaper then e.g. mouser. This is not always true.

For More established vendors on larger orders and you get a good price.

Ebay is particular good on small orders and the cost is not inflated by the shipping cost.

Look forward to see your kit for sale.

Karl

[Tom Storey]

Some people will tell you that the retail price of your product should be 2.5x the cost of the raw materials/manufacturing. EEVBlog has a video about pricing your product and I believe he follows this rule.

If you don't buy in large enough quantities to bring the cost of your parts down enough then this could make your product quite expensive.

On the other hand, if you're selling something similar to an existing product then you might not have much choice and you'll have to match the price of the competing product as best you can.

On Sun, Jun 7, 2020, 13:07 Francis Pierot <fpi...@gmail.com> wrote:

Hello everybody,

I'm working on a possible kit and have a number of questions related to selling this.

Where do you find your parts? I have three levels of quality suppliers right now: low-cost with variable quality and slow shipping (AliExpress and Banggood), reasonnable quality and shipping (LCSC) and high quality and fast shipping (Mouser).

At LCSC and Mouser I find components I won't find at AE and BG, but at these last too I'll find mechanical parts for low prices that neither LCSC nor Mouser are selling.

Any other suggestion?

Another question is where do you find the little boxes in which you send you kits?

What shipping options did you consider or adopt for your kits for:

• overseas (in my case, europe -> US)
• uk/europe (europe->uk)
• europe->europe

Also, how do you evaluate the right price for your kit? I have a base price which is made from:

• the cost of each individual component, computed from how they cost me in lots
• a minimal price for specific cables building
• the cost price of each PCB (how much it actually cost me, although I don't count the first versions with mistakes in that price)
• a little cost for all the work it implied
  

I'm also afraid of the price I reach right know, and am hesitating between building a stock already so people won't have to wait one month while I order the parts, or asking for pre-orders so people are advancing the money for me to buy the components.

The cost of the kit is not something I can advance too much, so I'd like to get an idea of how many people would be interested, but how can you reliably do that?

Thank you for any advice!

--
You received this message because you are subscribed to the Google Groups "retro-comp" group.
To unsubscribe from this group and stop receiving emails from it, send an email to retro-comp+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/retro-comp/84397584-ea3e-4184-a359-e98ea318c6a8o%40googlegroups.com.

[J.B. Langston]

All of this these considerations reinforce for me that not selling my own kits was the right decision.  I'd like to see more people using my designs but it's not worth the effort to me. Luckily Michael Kamprath volunteered to put together kits based on my design and sell them on Tindie. I was happy for him to do it and I don't ask for a cut since I don't need the money, and I think he is selling for close to cost.  Since all my designs are MIT-licensed, in theory anybody who wanted to could sell kits without my permission.

[Francis Pierot]

I see what you mean but I feel a kind of excitment preparing this kit.

I'm pretty sure there are people who will love what I'll show, but from this to a product I can sell, there's a distance and this is what I try to evaluate.

I can spend some money for tries and fails while I reach the state where the kit has a true existence.

From this future starting situation, my main points will be 1) not loosing money 2) not disappointing buyers.

In my work I always take care of point 2 and my boss takes care of point 1. For this kit, I'm alone on both points.

I already discovered industrial design is a job by itself. I have 4 unuseable PCBs that have reversed pins on headers. I have 4 next version unuseable PCBs because I put 2 screw holes too close from a component and the nut won't fit. And another 4 panels where holes are in alignment with bus connectors and you'd have to choose between the panel connector and a board on the bus.

I spend an unbelievable amount of time preparing a documentation I hope to be the same quality level as Steve's SC126 excellent build doc. For ease I write it in my (french) language and will translate it to english when finished.

Aligning screws holes and writing doc are much more work than designing the product and PCB themselves.

I already know I won't price this time and money spent in prototyping. But the idea that somebody could buy and build my kit is really exciting.

[J.B. Langston]

Oh I didn't mean you shouldn't do it. It is a personal decision. I just read all the Steve had to deal with and thought, gee I'm glad I'm not doing that.  I am very happy that other people are willing to make and sell kits, it's just not something I want to do.

[Steve Cousins]

Hi Francis. 

If it was easy, everyone would be doing it :-)

Steve

[Bill Shen]

Has anyone tried auction on the European eBay?  Since I have to assembled & test some number of boards to validate my design, I would auction them on the USA eBay starting from 99 cents.  Quite often new designs would receive very high bid and price settled down to something more reasonable after half a dozen auctions.  It may be a good way to recover development cost and get an idea of the fair market value for the item.
  Bill

[Steve Cousins]

I've just hit 1000 orders on Tindie.

Steve

[J.B. Langston]

Congrats!

[Karl Albert Brokstad]

Very well done!

Karl

[Bill Shen]

Excellent!
  Bill

[Francis Pierot]

Grats for your success Steve, you know how I appreciate your work and I think without any hesitation you deserve this success. I know it also has a cost, as it takes time from your other projects and probably personal life so I hope you'll find how to arrange things so you can still give us new ideas, projects and software in the future.

I will look the other components sources from Karl to see if I can reduce my base costs. And check the points Steve puts in light.

Right now, I choosed to take some risks: I ordered 10 of my last version of pcbs from JLCPCB and will build a stock for 10 kits. It's quite an investment but I can't see where I'm heading if I don't give it a try.

I still have work on documentation and prototyping but I think I'll be ready for when the PCBs will arrive.

Not sure yet if it's the right time to show pictures or videos. Annoucing before being ready is so Sinclair-QL like ....

Le vendredi 4 octobre 2019 19:36:51 UTC+2, Steve Cousins a écrit :

Bill suggested it would be interesting to discuss the selling of retro computer products, so here's a place to do it.

[Xander Soldaat]

Hi Francis,

What kind of boards will you be selling? What do they do?

= Xander

[TonyD]

Well done Steve,

1000 and still going strong !!!

Tony