This one sounds very promising, and they claim it will be the same price as a Raspberry Pi.
Go to www.krtkl.com for more info.
If something sounds too good to be true, it usually is.
Ask yourself, why does the RPi / Beaglebone cost as much as it does? Huge retailer markup? Apple tax? No! They cost that much because itâs the only way to produce a product like that in the quantities that they do. So you tell me youâre going to take a Beaglebone, add an FPGA, wifi, and bluetooth chipset, and somehow make it cost less?
2 Likes
It isnât impossible. Looking at the rendering, they have stripped off most of the connectors of the Raspberry Pi. That cuts out of a lot of the costs, right there. I suspect they are using a chip aimed at the lowend Android tablet market. Dual core, integrated onchip Bluetooth and WiFi. Integrated onto the package the RAM and NOR flash. The FPGA isnât a particularly large one and probably doesnât have that much I/O. Such devices are not expensive, think close order of a buck. Maybe 2.
Most of the I/O is on the header. That means one or more boards that stack with it to provide that functionality. Which is OK, you have a couple of them for development. When you embed it, you use one with the connectors you need, ditching the HDMI, probably the USB connectors and maybe the ethernet connector which are more useful for development.
Now yes, they need enough volume to hit those numbers. But that is what Kickstarter is for. Well, maybe not originally, Kickstarter seems to have been orginally aimed at the arts, but not everything goes as planned. However, it has been successfully used to launch companies in a similar chicken and egg bind where they need enough volume to make their product viable. If they hit their numbers, success! If not, wellâŚ
So this product is feasible. The only joker in the deck is convincing one of the Chinese chip companies like Allwinner to sell you the chips you need. It helps to understand Chinese because documentation in English is scarce. So is support, but if the RAM and flash is integrated on package, the rest of the stuff is pretty slow speed and a lot easier to design without support.
2 Likes
Beware of wishful thinking. Itâs easy to convince yourself that thereâs some magical way that by cutting connectors here and integrating RAM there, youâll magically make something twice as cheap as the nearest competitor. I, for one, am not convinced.
I didnât say it wasnât feasible, I said itâll never hold that magical âsame as the RPiâ price point. Same with the CHIP and the dozens of other ARM dev boards clamoring for attention.
2 Likes
Look at it this way.You can source from China, quantity one, a tablet with a 7" or so LCD screen with touch for $60 or less. That incudes battery, wall wart, box and all.
How much do you think the processor costs for something like this? It had a quad core, 1 gig. of RAM and 8 gig. of NAND flash. And a camera with LED flash.
As to the connectors, open a Mouser catalog and add up the cost of connectors in quantity. Be sure to check that the connector is surface mount, they tend to run quite a bit more. Placing connectors by hand on a mass produced board is a non-starter. It is quite easy for the cost of connectors to dominate your BOM. And the cost of placing them on a board only increases that cost. They are high mass and bulky and therefore cannot be placed at the same speed as smaller components. And time is money.
There is no magic required. Just knowledge of what it actually costs to manufacture things like this.
Um, yeah. Your blithe assumption that I am engaging in magical thinking annoys me. Especially when you show no real understanding of the issues involved.
2 Likes
Part of the value in products like the Raspberry Pi and the Beaglebone is the wealth of information and support that is available. Most of which is created by users of the products, but at the core that support infrastructure was provided by the companies that created these products. It is the same phenomena that has made the Arduino so ubiquitous.
I have yet to see a Chinese company that has produced (or made much of an attempt) such a wealth of support material for their products. It happens that others will produce it despite the originating companies efforts, but unlike the aforementioned products it isnât produced with the encouragement and support of the company.
The result is that many of these products simply languish. It is very easy to underestimate just how much work it takes to do something useful when you canât rely on a multitude of documentation, libraries, etc⌠that others have created and shared.
All that said, it is certainly possible for this device to be as successful as the others; but the odds are against them. And for most people, waiting for that success to be demonstrated is the better choice.
Also, in terms of price/performance it is easy to find incredibly powerful development boards priced for far less then cost⌠It is how the manufacturers gain market share.
Here is a perfect example:
$4 Dev board with ARM, FPGA, and programmable analog circuitry!
2 Likes
That is a different issue.
Yes, the ecosystem of the board is very important. The Snickerdoodle runs Linux, so it can leverage that with the proper drivers. What is critical is library support of the FPGA. Some nice examples of the FPGA, say some crypto functionality. Or possibly some image analysis on the FPGA to help with your cat herder. Some reason why anyone should care about having a FPGA on the board.
There are some versions of OpenCV that uses FPGA acceleration for some of the functions. I donât have much in the way of FPGA experience, so I donât know if the promised âup to 1.3M ASIC gates + 222 GMACsâ is enough to do much.
2 Likes
Ok, I have dug into this a bit more.
By Googleing around with the mentioned specs, it looks like they are promising a Xylinx Zynq Z-7020.
If this is the case, then they have to add the RAM, NOR flash and a Bluetooth/WiFi chip. That changes the economics quite a bit. Those Chinese chips aimed at the lowend tablet market are very cheap, well under $10 and even $5. Xylinx SOCs are not in that class at all. While they talk about low-cost, they are also aiming their chips at the medical devices market.
This canât be priced with the Raspberry Pi. The problem is that Xylinx likes to package these things in high ball count BGAs. And that costs. In single quantity, Digi-Key lists this SOC at prices from $58 to $2300, depending on the ball count.
I suppose if the quantity was high enough and they ran with thin enough margins, they could possibly sell a board in the $75 to $100 range. I suppose quantity 1000 could drop the cost of the cheapest SOC into the $30 to $35 range. The PCB wonât be cheap, they would probably need at least an 8 layer board to route those balls. 10 or 12 would be safer.
And it would be a steal at that price. Even at $150. Because Xylinx has a lot of quality support software. Especially for machine vision.
2 Likes
Howdy folks! One of the founders of krtkl hereâŚ
Just came across this post and figured Iâd take some time to address the questions, concerns, doubts, comments, suggestions, etc. people have expressed.
Iâll just start by saying: I love the passion. Stoked to see people getting excited for the product, even if they doubt our ability to effectively pull it offâŚ
There are a couple details weâre keeping closer to the chest leading up to the launch but I can personally assure you everything I say here is honest and accurate. If you still donât believe me after reading the responses, please let me know publicly/privately and I promise Iâll follow up.
So here we goâŚ
[apparently only two users can be mentioned per post so I have to split this into multiple pieces]
@David_Walker - Thanks for bringing this to the communityâs attention, David! Shoot us an email at hello at krtkl.com and weâll hook you up with something special for going out of your way to post this.
@zootboy - John is kind of on the right trackâŚplease see below. As far as the âsame price as a Raspberry Piâ goes, keep in mind that this board has Wi-Fi and Bluetooth 4.0 pre-integrated so thatâs factored into the price comparison. Youâll also be able to order a version without wireless if you chooseâŚwhich will cost the same as the R Pi or less. We havenât released official pricing but will definitely do so ahead of the launch so you can determine ahead of time if you want to buy one - which we hope you will!!
âcontâdâ
1 Like
@JohnKocurek - Some excellent observations, John! The rendering is a little outdated now (that was actually done before we changed the SoC weâre using to up the I/O count). Yes, eliminating the connectors does allow us to lower the BOM cost. It also allows for a much smaller footprint thatâs much more friendly for use in ârealâ applications while still being able to quickly bang out prototypes. Since different projects call for different I/O types, weâve addressed that by providing a couple different baseboard options. More details to come on that shortlyâŚ
The parts isnât quite as integrated as you describe but thatâs the right idea. And yes, the pricing only becomes sustainable in high volumes - which is why we need your help 
No need for âconvincing Chinese chip companies like Allwinner to sell the chips we needâ though⌠We wonât stand for poor documentation and support. Weâre (unofficially) going with guys like Xilinx, STMicro, and other vendors who have hoards of documentation and specs freely available. No NDAs, none of that.
@zootboy - Weâre still baffled by what the Chip guys are doing. Some of the stuff - like memory - is sold on exchanges and itâs simply impossible to get the parts for less than the Ask price (itâd be like buying Apple stock for less than itâs listed for). No integrating RAM or developing an ASIC or anything like that - just 12 months of cost engineering, negotiations, and no sleep.
âcontâdâ
1 Like
@JohnKocurek - yup!
@wandrson - Couldnât agree more with every one of your points - including that the odds are against us 
Documentation, community, and user experience - those are the primary drivers to the majority of success with any of these platforms.
Weâre setting up the backbone of the community as we speak and the robotic, drone, and computer vision apps will be fully support by ROS, OpenCV, ArduPilot, FreeRTOS, etc. We HATE poorly documented products - software or hardware. I promise that wonât be a problem.
And as far as user experience goes, weâre gong to blow away all of these other boards. Canât go into a ton of detail on that front at the moment but just think about how much better things would be if these boards came with Wi-Fi and Bluetooth on boardâŚand actually implemented in a user-friendly way.
And the Discovery Board is a good example of a super low-cost dev board to get their foot in the door. That is totally the right idea and itâs honestly pretty shocking that more companies donât do this. Thatâs not our play though.
All that said, we certainly hope we can convince you that we will be able to pull this off and that youâre willing to take a small gamble on us during our campaign to give us a running start!
@JohnKocurek - You just hit the nail on the friggin head! Without going into a ton of detail (which I promise we will share soon enough), the nice thing about the SoC architecture is that you can pretty much pretend the FPGA isnât even there and just treat the chip like a uP with a TON of I/O / peripherals. But weâve got the FPGA hesitancy totally under controlâŚand depending on how a couple of our upcoming meetings go we might pull off something on the software side that will blow your mind.
@JohnKocurek - Youâre good⌠I canât go that into depth re: specific part pricing but I can say that we will not market, build, or sell a product that we cannot deliver and sustain. I promise you that.
I really hope I didnât miss anything⌠If you have any other questions, comments, etc. please donât hesitate to fire away! The only favor I ask is you guys sign up for our pre-launch mailing list and follow our updates as we near the launch later this year. If we canât convince you by the time we launch that youâre backing a winner, then so be it. But this will not fail - we will die before we let that happen.
Best regards,
Ryan
1 Like
My suggestion would be to spend less time trying to market people and more time actually getting your product to market. Once it is commercially available we can decide ourselves if your product meets your marketing hype and if it is cost effective.
My understanding from Davids OP was that youâll were the ones who compared your pricing to the Raspberry Pi. Admittedly, I couldnât find that on your current web page so I am not sure where @David_Walker got that from. That said, it is a bit of a âbait and switchâ to list a fully loaded products specs and compare its price to a machine that doesnât share those specs, then say in the fine print that the price is only comparable when many of those âoptionsâ are removed from your product. People donât care about anything other then the bottom line cost of what you are advertising. For instance, I donât know many people who care that they can get a crippled raspberry pi for $15 less then the current flag shop model.
As I said in my earlier post, what will matter is whether you have a community grow up around your product to provide most of the software library support most hobbyists need. And that will only have a chance of happening after your product actually ships.
I suggest practising a modern version of Teddyâs âWalk softly and carry a big stickâ. IMO, its best to let the actual users of your products produce such hyperbole. Many of us have been dealing with such tech industry promises for decades, mostly for products that never made it to marketâbut they would have changed the world if they had.
@wandrson - Iâm just trying to help answer peopleâs questions; apologies if youâre put off by my responsiveness (or lack thereof?) or wording. As you said, a huge part of the potential success of this product is dependent on the health of the community so Iâm just trying to get involved early on as opposed to leaving you all in the dark with all questions and no answers.
Sorry if this wasnât clear, just giving an apples-to-apples comparison. Put another way: if wireless isnât a need for you, you can get a version of our product without wireless for the price of a Raspberry Pi without wireless. If you do want wireless, you can get a version of our product with wireless for price of a Raspberry Pi with wireless. Does that make more sense?
Yessir, I hear ya! I admit I can get a little carried away when excitedly talking about this stuff. I guess a better way to put this is: you wonât be disappointed. If youâre more comfortable âwaiting and seeingâ if we can actually do what weâre saying weâre going to do before pre-ordering a board, I totally understand. Although in this case - as is the case with any crowdfunding campaign - we are indeed relying on people to have a little faith⌠If everyone wanted to just âwait and see,â crowdfunding wouldnât exist and these projects would be dead anyway - which is what we love about it! Everyone (users and creators alike) is in it together so either everyone wins or everyone loses.
Best,
Ryan
1 Like
You have my best wishes that your product makes it to market and is successful. And while I understand how easy it is to become really passionate about a project your working on, it is equally easy to over sell your product (âweâre gong to blow away all of these other boards.â) and produce nothing but disappointment no matter how well done your product is.
There is also the fact that many individuals/companies before you have made promises that they never delivered on, even after taking money from people who funded such campaigns. Your sites emphasis on marketing and âcreating a communityâ seems very premature to me. Getting the engineering and business aspects done (or nearly so) before such efforts almost always equate to actually making a successful delivery for a crowd funding (or any other new product release).
As I said, I couldnât find the comparison text that @David_Walker referenced in his OP on your site. Do you know where such a comparison was actually made and by whom?
(post withdrawn by author, will be automatically deleted in 24 hours unless flagged)
Thanks @wandrson!
I wholeheartedly agree. We actually would have launched months ago if we werenât so obsessed with developing an actual, working design prior to launch (something that I can guarantee most companies with crowdfunding campaigns are not doing).
The most obvious place re: the âprice of a R Piâ was probably one of our landing pages, however if can also be found under the âwhat can I don with snickerdoodle?â dropdown on the homepage.
-Ryan
1 Like
@jast - not entirely sure whoâs âtrollingâ who? Just here to answer questions. If Iâm bothering people, please let me know and I will cease all posting to the thread.
-Ryan
[quote=âkrtkl, post:17, topic:4150â]
I wholeheartedly agree. We actually would have launched months ago if we werenât so obsessed with developing an actual, working design prior to launch (something that I can guarantee most companies with crowdfunding campaigns are not doing).[/quote]
Well apparently you are also spending a fair bit of time creating a marketing plan and hyrperbole for a product that doesnât exist yet. And most reputable crowd funding sites require a working design prior to launch. They allow you to figure out how youâll manufacture after however, which is where most of those have failed. Well except for the blatant frauds who sold imaginery products that âwould change the world.â
Thank you for pointing that out. I have included a screen copy with annotation so as to make it clear in perpetuity. I find your web page to be an extremely misleading piece of marketing that has confirmed by overall impression that youâll appear to be putting more effort into âsellingâ rather then building a good product.
Thanks @wandrson. Fixed to help clarify⌠[screenshot from http://krtkl.com ]
2 Likes