RE: New project...PCI based servo control board

["Craig Edwards" <cedwards-at-ceinetworks.com>]

Understand and agree John....it is difficult to sell product to folks like
ourselves (yea... I'm a cheapskate too...) at a profit.

Honestly, I do have aspirations of SELLING something someday, but recognize
it's unlikely to be at a profit for the DIY / hobbyist segment of the
market.  IF this project does result in PCI board, I will sell the first
proto batch "cheap" to members of this list, but since my family is way too
attached to eating....:), I think the only way forward for me is to either
treat this strictly as a "business hobby" (which many do....), or get
serious and market products to OEM customers in the various segments of the
small machine industry.  I've not done enough work to convince myself that
the latter is feasible in my case, but this is certainly a topic I would
like to discuss with other "like minded" individuals who are working with
EMC.   If it stays a hobby, things will have to move forward on a "as time
allows" basis.... if I put together a business plan, then things better move
pretty fast.

With regard to "open sourcing" hardware designs, I'm not entirely opposed to
the concept, but I'm also not sure it would work very well for most
hobbyists.  I respect where you're coming from with regard to a willingness
to build your own... But with current IC packaging technology it's just
getting real difficult for even the average engineer to build their own
boards using up to date technology.  Depends on the design of course, and
just how "niche" you're willing to be. 

One thought that I did have in the spirit of being "open" is that with the
right definition of hardware interfaces and software APIs, it should be
possible to offer a "base" board with all the harder stuff done (PCI PnP,
pulse generators, encoders), but still allow for easy hacking (both software
and hardware) for the inevitable "if only".  I wasn't planning on including
much "machine interface" stuff on the PCI board anyway...just need to make
it really expandable.

BTW, I apologize to the EMC list members if this has gotten WAY off topic
from the normal type of traffic here.... I'll try to refrain from further
posts on "business" issues.... but I guess it was inevitable once the issue
of "price" came up.

Regards,
Craig
       



Jon Elson wrote:

> I have had the PPMC on the market at US $650 for the 4 axis setup with 
> 16 opto-isolated digital inputs and place for 8 SSRs to mount right on 
> the board, so you don't need an Opto-22 board, too.  This has 4 16-bit 
> DACs, 4 24-bit encoder counters with index channel.  For $30 extra, 
> you get differential receivers for the encoder inputs.  And, it has 
> all the E-stop logic, with a watchdog timer, built in, too.  This has 
> been on the market for over a year, and I've sold SIX units!  I 
> haven't even managed to use up the run of 10 prototype boards!

> I have had the Universal Stepper Controller out for nearly a year (I 
> demoed it at the last NAMES show) at $200, and Ive sold FIVE!  I'm 
> basically giving these away at this price, at this volume.  The FPGA 
> on this product costs $30.  But, it has 4 24-bit encoder counters 
> (with index) and 4 digital step rate generators with 24-bit rate 
> selection,

> So, here I have developed these products, no one wants them, and yet 
> you guys are STILL arguing over what would be the all-time BEST CNC 
> interface.

> All I can tell you, is unless you want to invest several thousand $ a 
> MONTH into advertising, you won't sell more than 10 of them a year.

I can't speak for Craig or anyone else, but I don't intend to SELL anything.
I am well aware of how hard it is to make money selling to hobbyists.  Face
it - we're cheapskates.

As I see it, there are two ways to build things at a profit.  One, you build
something that you can sell in quantity, and you invest in the machinery,
etc., needed to make it cheap in volume.  That's basically what Gecko has
done with their drives
- there's
no way they would be as cheap as they are without automation and quantity.
The other way to make money is to build something that is customized or
otherwise unique, and sell only to those with deep enough pockets to pay for
that unique quality.  Jon, I think you are well aware of that approach with
with your nuclear instrumentation.  I have done something similar selling a
specialized test jig to
tile testing labs.   Those deep pocket people are NOT the ones who build
their
own CNC machines for "fun and profit".  I just don't think it will ever be
profitable selling CNC hardware to people like us. I'm not  even gonna try.

Whatever designs I come up with, I intend to publish as completely open
projects, and let anyone who wants to build them do so.  If there is enough
interest to warrent buying a batch of bare boards to get the cost down,
great. If not, that's OK too. I'll buy one board and build my own system.  I
know it will cost as much if not more than your $200 USC, but I certainly
hope it will be under the $650 pricetag of the PPMC system.  (Not counting
the value of my time, of course!)

If the USC had everything I want/need for my project, I might well be buying
one instead of getting involved in this thing.  But is doesn't - I need a
couple of additional encoder counters, and an analog output.  So being an
engineer, I'm gonna build my own.  If someone else is doing much the same
thing, then I'll collaborate with them.  If my needs divirge too much from
what everybody else wants to do, then I'll go it alone again.  Regardless of
how it works out, I'll post my schematics and code.

Because I don't expect to make any money doing this, I've been keeping the
design as simple and straightforward as I can.  I've avoided PCI like the
plague, because I have neither the information, skills, or time to design a
PCI interface. The single board computers I intend to use for my own project
have an ISA bus, and I know that an ISA design is reasonable for me to do on
my own.  That, or an EPP version, is the path I was planning.  If Craig,
combined with the guys at opencores.org, can make a PCI version, that's
great.  Perhaps I can contribute to the control hardware design, and others
can contribute to the bus interface... In any case, the discussions are a
good thing that everyone can benefit from.

John Kasunich