Thinking of starting your own Crowdfunded project? Ask me anything!

[jmole]

Hey All,

I raised $50k on Kickstarter last year for a project called freeSoC.

I've been meaning to write a big blog post about the experience and the process, but still haven't managed to get around to it, so I figure a Q&A is the next best thing. It was the first time I'd done anything on this scale, as far as delivering real products to customers goes. I was late by about 3 months on delivery, but I naively assumed that it would only take me a month to get things together. Then, backers started requesting design changes, and I acquiesced. Anyway, it was a great experience, and I'd love to share any information you're interested in, so ask away!

[amspire]

Congratulations! I have been thinking about a project that needs a micro + hard logic and I was seriously thinking about the Cypress PSoC 5. I will have to take a good look at your board.

If you were doing a new electronic design project, what work would you attempt to complete before starting the fund-raising?

Also do you think you can get the money if you are totally honest and conservative in your estimated delivery. In your case, if you had quoted, say, 4 months until the first deliveries, do you think you would have got the backers? Is quoting delivery based on the most optimistic assessment of the project a required part of the game in order to raise decent money?

[EEVblog]

Also do you think you can get the money if you are totally honest and conservative in your estimated delivery. In your case, if you had quoted, say, 4 months until the first deliveries, do you think you would have got the backers? Is quoting delivery based on the most optimistic assessment of the project a required part of the game in order to raise decent money?

That's a very interesting question!
I'd love to see some real studies on this sort of thing, do any exist?
Or is that sort of stuff covered in Marketing 101 class?

[jmole]

If you were doing a new electronic design project, what work would you attempt to complete before starting the fund-raising?

In terms of planning the reward structure and project costs:

• Prototype (obviously)
• Cost Estimates -- Pick some random quantity of boards you'd like to sell. Let's say 100. So get a complete BOM estimate, production estimate, assembly estimate, etc, for 110% of the quantities you're building.
• Packaging -- This means, go to uline.com, pick out your packaging materials, add to cart, and write down the total with shipping
• Shipping -- Are you going to ship this stuff yourself? For 100 orders, it's feasible. Above that, you're gonna be screwed unless you have some help.  I had ~600, so I used a fulfillment service called Shipwire.
• Shipping Costs -- For domestic, you can figure about $5, for a small product like mine. International, about $30.
• Total per unit costs -- Now add up all the totals above, divide by your number of units, and figure out what it will cost you.
• Markup -- Figure out what kind of markup you want on the product. This should be around 2.5X, at a minimum, or you will probably end up losing money. It's funny how fast unexpected costs can creep up.

Also, you're gonna need a bunch of copy for the Kickstarter page, a good video, yada yada. Don't underestimate these things. You've gotta pitch your product well. Emulate the most funded projects. Look at their videos and how they sell the idea, and do that for yours.

Also do you think you can get the money if you are totally honest and conservative in your estimated delivery. In your case, if you had quoted, say, 4 months until the first deliveries, do you think you would have got the backers? Is quoting delivery based on the most optimistic assessment of the project a required part of the game in order to raise decent money?

This, I can't say for certain.

In my case, I don't think I had any real complaints that the ship date was so delayed. It was clear that I was iterating on the design, and making progress and that kept the confidence of my buyers.

I've backed a few other projects, and in a few, this has not been the case.

CruxSkunk, a laptop-type case for an iPad, has been seriously overdue (project finished about when mine started), and there are tons of people complaining about it. I couldn't care either way, because I can see that he's making progress and I will get it eventually.

Memoto, a lifelogging camera, had an estimated delivery date of March 2013, which has come and gone, but the creators have been very good about keeping us up to date, and I'm certain they will deliver a great project.

That said, I think time overruns are part of the game. It's pretty much expected by anyone who's not a first time kickstarter backer.

Now, will someone back a project if the delivery date is "realistic" (assuming that you can actually hit the mark)? I think so. Take a look at all the video game projects. Millions of dollars for a release date 1-2 years down the line. The key is to be honest. I honestly thought I would be able to deliver the product in a month. We had just finished a first run of 100, and I was ready to pull the trigger on the next 1000. Then the design change requests came.... Then I realized I could make a better board by listening to my backers. And they supported me through it. So be honest. If you have a great project, the good people of the internet will help you make it work.

[jmole]

BTW, this is a great article to read if you're interested in why most Kickstarter projects are overdue:

Coding, Fast and Slow: Developers and the Psychology of Overconfidence

[amspire]

By setting the short deadline, was that a big self-motivation pressure for yourself? If you set a 1 month deadline, I would guess you would start out working day and night to push the project.

With a 4 month deadline, you may feel you have to pace yourself, and you could still end up missing your deadline by the same amount if you just are not pushing yourself as hard.

It is very interesting to hear how co-operative the contributors are as long as you are honest and you keep the communications going. I am really interested to hear how much contributors actually help the project and in fact feel that they have become part of the project. Is that something that really helps in the motivation?  I just know how easy it is to loose motivation in a project when I am working on it by my self.

[Corporate666]

Where did you publicize your Kickstarter campaign?  Or did you just launch it and sit back and wait for people to see it?

I was one of your backers by the way - been using the PSoC chips for 3-4 years.  I think if more people used them and understood them, they would be so much more popular.  They are amazingly flexible and powerful, but priced in line with other MCU's. 

[jmole]

By setting the short deadline, was that a big self-motivation pressure for yourself? If you set a 1 month deadline, I would guess you would start out working day and night to push the project.

I was (and still am) doing research and grad school when the Kickstarter project was going on. So my attention was very divided. But believe me, when there's other people's money on the line, you're going to spend every spare second thinking/worrying about and building the project.

It is very interesting to hear how co-operative the contributors are as long as you are honest and you keep the communications going. I am really interested to hear how much contributors actually help the project and in fact feel that they have become part of the project. Is that something that really helps in the motivation?  I just know how easy it is to loose motivation in a project when I am working on it by my self.

The backers helped a ton. If you're familiar with the idea of customer development, it's the extreme version of that. You've got an excited group of early adopters, and you get real feedback, because they're actually invested in the project.

[jmole]

Where did you publicize your Kickstarter campaign?  Or did you just launch it and sit back and wait for people to see it?

I sent out some emails to a few bloggers, but got no coverage there. Some Hackaday reader submitted the project, and it got front page coverage. Viral is key. I also got organic coverage on Hacker News, which was great. A lot of the traffic comes from Kickstarter itself, and Kicktraq, which i now follow religiously.

I was one of your backers by the way - been using the PSoC chips for 3-4 years.  I think if more people used them and understood them, they would be so much more popular.  They are amazingly flexible and powerful, but priced in line with other MCU's.

Ditto. I've used PSoC for the past 5 years, and I will never look at another microcontroller again, unless I need something a lot more powerful, like an FPGA or a Cortex A8/A9. I think the new PSoC Pioneer board Cypress put out recently will help encourage the cost-sensitive hobbyist market to give it a try, and I think my board is a great option for the hobbyists who actually know something about MCU development.

Still, there's a lot of people who think looking up register address in a TRM is super fun, and it seems like these people are the hardest to convince, because they have a lot of time and experience on their favorite platform. Believe me, I enjoy a good TRM as well as the next guy, but honestly, I'd rather have an IDE and a MCU that will let me get a project going in a matter of minutes, not hours.

[Corporate666]

Ditto. I've used PSoC for the past 5 years, and I will never look at another microcontroller again, unless I need something a lot more powerful, like an FPGA or a Cortex A8/A9. I think the new PSoC Pioneer board Cypress put out recently will help encourage the cost-sensitive hobbyist market to give it a try, and I think my board is a great option for the hobbyists who actually know something about MCU development.

Still, there's a lot of people who think looking up register address in a TRM is super fun, and it seems like these people are the hardest to convince, because they have a lot of time and experience on their favorite platform. Believe me, I enjoy a good TRM as well as the next guy, but honestly, I'd rather have an IDE and a MCU that will let me get a project going in a matter of minutes, not hours.

There is some kind of opportunity there between Cypress and the hobbyist market.  I often see the Arduinos and, no disrespect to them, but if the average person knew what they were missing out on in PSoC, they would be shocked.  Arduino is such junk comparatively.  I did a project a few months back that had 27 separate PWM channels - all on a single PSoC3 that cost < $4.  Same chip went into another project that has USB and MicroSD cards.  Same chip went into another device with several timers, counters, PWM's and some analog logic.  I see people talking about which Atmel chip to buy because they need 2 timers but also 2 interrupt pins and I wonder how on earth Cypress hasn't absolutely dominated these other guys yet.  Perhaps with PSoC4 they will.

Do you work with the Cypress company very much?  I have a few contacts there and they have been absolutely phenomenal.  We use maybe 20,000 MCU's a year and Atmel couldn't care less about us.  They outright lied to us on a few occasions and left us in the lurch a few times - including us designing the X-Mega into an important product only to be left hanging for more than a year on getting parts, then having a bug in the chip that rendered it useless in our design and requiring a complete do-over.  They also told us to F-off when we experienced a very high chip failure rate on Tiny13's.  On the other hand, Cypress gives me so much design support that they have actually offered to design and layout boards for me to win the design, and they sold me tons of parts and dev kits basically at cost.

The only downsides are the lack of community size...  you can do a search on any task and "PIC" or "Atmel" and find dozens of example projects - but not so much with the PSoC. 

Someone will crack the PSoC <--> Hobbyist nut perfectly and make lots of money.

[jmole]

Do you work with the Cypress company very much?  I have a few contacts there and they have been absolutely phenomenal.  We use maybe 20,000 MCU's a year and Atmel couldn't care less about us.  They outright lied to us on a few occasions and left us in the lurch a few times - including us designing the X-Mega into an important product only to be left hanging for more than a year on getting parts, then having a bug in the chip that rendered it useless in our design and requiring a complete do-over.  They also told us to F-off when we experienced a very high chip failure rate on Tiny13's.  On the other hand, Cypress gives me so much design support that they have actually offered to design and layout boards for me to win the design, and they sold me tons of parts and dev kits basically at cost.

I've got a couple connections there, and they've featured me in their Tech Spotlight video series.

But, yeah, their tech support is phenomenal. On a different product of theirs, the FX2 USB MCU, I was trying to find a crystal with a lower footprint than the typical HP49 through hole beast. I found a couple, but the max power was significantly lower than what the FX2 specd. Submitted a case to their team, and they sent me back the exact circuit I needed to use, with load crystal specs and damping resistor. Never had a case go unresolved with Cypress, on any of their product lines.

The only downsides are the lack of community size...  you can do a search on any task and "PIC" or "Atmel" and find dozens of example projects - but not so much with the PSoC. 

Someone will crack the PSoC <--> Hobbyist nut perfectly and make lots of money.

This is one thing I'm trying to do now, with freesoc.net. I think the dev kit market has potential, but building the community is the key. I'm focusing on educational videos now, and hope to have a full set of videos showing how to use each major component in PSoC Creator.

[Corporate666]

This is one thing I'm trying to do now, with freesoc.net. I think the dev kit market has potential, but building the community is the key. I'm focusing on educational videos now, and hope to have a full set of videos showing how to use each major component in PSoC Creator.

I remember when I was starting with Atmel, there was a guy who had a video about using the AVR Butterfly to develop with.  I had gotten a Butterfly at some Atmel event, so it was easy to follow along, and I felt I was making quick progress getting the LCD to stay stuff, etc. 

Cypress has some very cheap dev kits (or intro kits - whatever you'd call them) like the FirstTouch and your FreeSoC.   Actually, one of Cypress' biggest assets is they have a dev kit for just about everything - so as you progress to more complex stuff, it's easy to buy pre-made examples.  The biggest problem is they need a PSoC for dummies book, or at least a section of their website for people who know nothing about SoC's and PSoC in specific.  Just something that starts out with the assumption that a person knows maybe a little bit about programming, but virtually nothing about electronics or embedded design - and that lets them quickly and cheaply make significant progress, like getting an LCD to say stuff, getting LED's to react to accelerometers, etc. 

I got into PSoC from an evangelical technical salesman @ Future Electronics who kept telling me how great it was.  But there was no dummies guide... so I had to take the time to sit down and spend a few days playing with the software and a dev kit to have that "a-ha" moment.  I think if Cypress or you can find a way to make that a-ha moment come sooner or easier for the average hacker, it's a potential goldmine.   I find Creator very easy to use with lots of API's for most stuff, so the needed components are all there - it just needs to be assembled into an easy to use package for the layman.

[amspire]

I know this is not quite the right thread, but I can't pass up the chance to ask about the PSoC.

I need a micro clock speed of 10 - 20MHz (I will probably have it variable driven from an external high stability synthesizer) and I need to generate a PWM-type output accurate to the nearest clock cycle. I need to have about 20  x 18 bit hardware registers presettable by the micro, and when a 18 bit hardware clock counter equals the value in a register, the PWM output flips.

It may be possible to have a single 18 bit register instead that is reloaded live by the micro code after each clock counter counter match if I can generate an interrupt back to the micro each time there is a match between the register and the hardware clock counter.

I actually need two banks of these registers/PWM outputs, and if I can extend it to 4 banks, that is even better.  If I can add an extra 4 x 8 bit conventional  PWM outputs as well, then that is even better again. That would definitely be enough to convince me on the spot.   %-B

Does that sound like something the FPGA in the PSoC can do? Hope it makes sense.

Richard.

[jmole]

I know this is not quite the right thread, but I can't pass up the chance to ask about the PSoC.

I need a micro clock speed of 10 - 20MHz (I will probably have it variable driven from an external high stability synthesizer) and I need to generate a PWM-type output accurate to the nearest clock cycle. I need to have about 20  x 18 bit hardware registers presettable by the micro, and when a 18 bit hardware clock counter equals the value in a register, the PWM output flips.

It may be possible to have a single 18 bit register instead that is reloaded live by the micro code after each clock counter counter match if I can generate an interrupt back to the micro each time there is a match between the register and the hardware clock counter.

I actually need two banks of these registers/PWM outputs, and if I can extend it to 4 banks, that is even better.  If I can add an extra 4 x 8 bit conventional  PWM outputs as well, then that is even better again. That would definitely be enough to convince me on the spot.   %-B

Does that sound like something the FPGA in the PSoC can do? Hope it makes sense.

Richard.

Sounds feasible, but you could probably just download Creator for yourself (it's free), and see if you can synthesize the verilog you're looking for. If the constants are hard-coded, a simple counter state machine with combinatorial logic for the specific PWM transitions will do. If you want to be able to change the PWM transition states, you may run into issues with resource allocation if you attempt a naive implementation (which would probably work fine on a typical FPGA).

The digital blocks in the PSoC are part PLD, part datapath, so you can use some tricks to extend their functionality considerably. This document is a great resource on the architecture: http://www.cypress.com/?docID=43512

One approach might be to store your coefficients in a table in RAM, and after each PWM transition, DMA loads the next coefficient into the verilog register. If the PWM coefficients aren't sequential (i.e., you've got like 10 or more clock cycles between each PWM output), the DMA engine will have enough time to transfer the new coefficient to the register. (Just looked back, and that's more or less what you said.  )

[Keef Wivanef]

Also do you think you can get the money if you are totally honest and conservative in your estimated delivery. In your case, if you had quoted, say, 4 months until the first deliveries, do you think you would have got the backers? Is quoting delivery based on the most optimistic assessment of the project a required part of the game in order to raise decent money?

That's a very interesting question!
I'd love to see some real studies on this sort of thing, do any exist?
Or is that sort of stuff covered in Marketing 101 class?

Probably not.
It IS covered in Ethics 101.
Dishonesty in sales and marketing is SOoooooo 70's 

[amspire]

This document is a great resource on the architecture: http://www.cypress.com/?docID=43512

Thanks - that is very useful.

But I can feel a headache coming on from the look of the document. You don't have the version written for 3 year-olds?

[jmole]

This document is a great resource on the architecture: http://www.cypress.com/?docID=43512

Thanks - that is very useful.

But I can feel a headache coming on from the look of the document. You don't have the version written for 3 year-olds?

Haha, that's what the drag-and-drop logic components in PSoC Creator are for!

[BravoV]

There is some kind of opportunity there between Cypress and the hobbyist market.  I often see the Arduinos and, no disrespect to them, but if the average person knew what they were missing out on in PSoC, they would be shocked.  Arduino is such junk comparatively.

.. <snip>..

I see people talking about which Atmel chip to buy because they need 2 timers but also 2 interrupt pins and I wonder how on earth Cypress hasn't absolutely dominated these other guys yet.  Perhaps with PSoC4 they will.

.. <snip> ..

The only downsides are the lack of community size...  you can do a search on any task and "PIC" or "Atmel" and find dozens of example projects - but not so much with the PSoC. 

Someone will crack the PSoC <--> Hobbyist nut perfectly and make lots of money.

I had the chance to play with PSoC since it was born maybe in early 2002, started with the brain damaged early release CY8C25xxx or C26xxx that had to use a "precise" timing just to program the flash, which is very cumbersome for enthusiasts.   
Up to newer and much-much better C27xxx and C29xxx series which fixed that really pesky problem. In those days, supporting software tools that available were just PSoC Designer, heck, even PSoC Express (PSoC Creator's predecessor), that GUI based click & drag programing thingy  didn't exist yet.

It was quite a phenomenal at least to me as enthusiast especially with those highly configurable blocks, a really flexible peripherals configurations. Related to enthusiast market, I believed Cypress made a huge mistake at one period by not providing any free compiler at all, even the crippled one, and was totally abandoned the enthusiast communities. At that time PSoC users had to buy from 3rd parties just for the compiler which was way too much for hobbyist. Public compiler didn't exist like today GCC, and that period practically wiped off enthusiast/hobbyist interest, until the crapduino Arduino wave arrived. 

[FrankBuss]

Still, there's a lot of people who think looking up register address in a TRM is super fun, and it seems like these people are the hardest to convince, because they have a lot of time and experience on their favorite platform. Believe me, I enjoy a good TRM as well as the next guy, but honestly, I'd rather have an IDE and a MCU that will let me get a project going in a matter of minutes, not hours.

You can still do this with PSoCs, setting all registers by hand, needs just a bit more time. I've tried it once:
http://www.frank-buss.de/rs232toi2c/
Of course, doesn't make much sense. The IDE and visual placing of the hardware blocks and connections is really nice.

[Corporate666]

There is some kind of opportunity there between Cypress and the hobbyist market.  I often see the Arduinos and, no disrespect to them, but if the average person knew what they were missing out on in PSoC, they would be shocked.  Arduino is such junk comparatively.

.. <snip>..

I see people talking about which Atmel chip to buy because they need 2 timers but also 2 interrupt pins and I wonder how on earth Cypress hasn't absolutely dominated these other guys yet.  Perhaps with PSoC4 they will.

.. <snip> ..

The only downsides are the lack of community size...  you can do a search on any task and "PIC" or "Atmel" and find dozens of example projects - but not so much with the PSoC. 

Someone will crack the PSoC <--> Hobbyist nut perfectly and make lots of money.

I had the chance to play with PSoC since it was born maybe in early 2002, started with the brain damaged early release CY8C25xxx or C26xxx that had to use a "precise" timing just to program the flash, which is very cumbersome for enthusiasts.   
Up to newer and much-much better C27xxx and C29xxx series which fixed that really pesky problem. In those days, supporting software tools that available were just PSoC Designer, heck, even PSoC Express (PSoC Creator's predecessor), that GUI based click & drag programing thingy  didn't exist yet.

It was quite a phenomenal at least to me as enthusiast especially with those highly configurable blocks, a really flexible peripherals configurations. Related to enthusiast market, I believed Cypress made a huge mistake at one period by not providing any free compiler at all, even the crippled one, and was totally abandoned the enthusiast communities. At that time PSoC users had to buy from 3rd parties just for the compiler which was way too much for hobbyist. Public compiler didn't exist like today GCC, and that period practically wiped off enthusiast/hobbyist interest, until the crapduino Arduino wave arrived. 

I remember PSoC Express!   What a stupid program - every time some company thinks they can make software so easy that programming can be done by a layman - they are fooling themselves   I think PSoC Express is gone now, and the functionality is moved into PSoC Designer.

If you have not tried the PSoC3 or PSoC5 chips, you should try them for sure.  I thought the PSoC1 is "ok", but the 3 and 5 are just a dream.  All of the annoyances of the PSoC1 were addressed - like the restrictive pin routing rules, difficulty with interrupts (like only one GPIO interrupt for all pins and all state changes), the limitations on resources like needing "analog" and "digital" blocks, etc.  All of that is gone and there are wonderful things like pin API's and you can add any components you like.  The cheaper PSoC3 ships are in the $3-4 range right from Cypress and for that you can get 32k or 64k of flash and 25-62 IO pins and 2-8k of SRAM.  Low power modes on the PSoC3 and 5 are very good, and they can run on 1.7 volts. 

Can you tell I like the PSoC a lot?

The PSoC3 and 5 are so much better than PSoC1.  As I understand, Cypress bought another company that was making the SoC chips and the 3 is their first clean sheet design.

The new 4 that is coming out (awkwardly named) is actually below the PSoC3 in price and features, but I believe the entry level unit will be <$1 per chip which is amazing considering the features. 

[jmole]

Lets take the PSoC debate to the thread in the MCU forum. I want to talk about crowd funding here.

[Rasz]

Hi.
I saw your project on Hackaday when it was running, but didnt follow it so Im sorry if Im asking about something covered on Kickstarter page/comments.

Where did you manufacture? Did factory do any tests/inspection you specified or did you just wing it? Were chips flashed in the factory? What was the failure rate?

Im also interested about shipping cost. I cant understand how Chinese ebay sellers are able to ship free of change items they sell for $1- its like Chinese government sponsors China Post to subsidize local economy/bankrupt World retail market . Have you considered making a deal with someone in China (seed studio?) to handle packaging and shipping directly?

ps: Congratulations on successful project.

[Corporate666]

Im also interested about shipping cost. I cant understand how Chinese ebay sellers are able to ship free of change items they sell for $1- its like Chinese government sponsors China Post to subsidize local economy/bankrupt World retail market . Have you considered making a deal with someone in China (seed studio?) to handle packaging and shipping directly?

The Chinese government most definitely does heavily subsidize shipping costs in an effort to promote Chinese exports.  That goes for China post, freight shipping and other shipping methods as well.  Heavily subsidized.

[JoannaK]

Lets take the PSoC debate to the thread in the MCU forum. I want to talk about crowd funding here.

You forgot to mention you apparently have got quite substantial sponsorship and support from the Cypress.

[BravoV]

You forgot to mention you apparently have got quite substantial sponsorship and support from the Cypress.

A Cypress marketing subsidized project ?