Jump to content

Photo

F18A MK2

F18A VDP 9918A

259 replies to this topic

#251 --- Ω --- OFFLINE  

--- Ω ---

    --- Ω ---

  • 12,487 posts
  • Location:워싱턴 주

Posted Mon Jul 30, 2018 4:35 PM

What if you are actively 'developing a product', but decide to not bring it to market. 

Is there a limit on how many 'beta testers' a product in development can have?  icon_wink.gif

 

If you pay to join the 'development team'... but withdraw a few weeks after receiving your 'test unit'...



#252 Tursi ONLINE  

Tursi

    Quadrunner

  • 5,170 posts
  • HarmlessLion
  • Location:BUR

Posted Mon Jul 30, 2018 7:22 PM

In the 3-5 years of F18A MK1, Matthew only netprofited maybe $5K to $10K total for all the units. If he had to pay an annual fee of $5K per year to use VGA, he wouldn't have been able to do this.


Dude, you're only doing Matt a disservice by making numbers up here... ;)

#253 OLD CS1 OFFLINE  

OLD CS1

    Quadrunner

  • 5,361 posts
  • Technology Samurai
  • Location:Tallahassee, FL

Posted Mon Jul 30, 2018 7:59 PM

Dude, you're only doing Matt a disservice by making numbers up here... ;)

 

I get Wildstar's idea, but, yeah, we kind-a need to work with real figures not hypotheticals to plead our case.



#254 Wildstar OFFLINE  

Wildstar

    Chopper Commander

  • 125 posts

Posted Mon Jul 30, 2018 10:23 PM

Dude, you're only doing Matt a disservice by making numbers up here... ;)

 

I'm figuring off the top of my head the price of each item and the BOM (bill of material) which he illuminated. There isn't a while lot of margin between BOM cost per unit and product price of the original F18A.

 

Developing hardware 'products' for the retro-computing market isn't and never been about making it rich. $5K to $10K or so spread out over 5 years is only $1K to $2K a year. Not even close to a viable sales run for the HDMI license.

 

The numbers are ball park figures. Even if he made a little more, it really doesn't change the issue. It is apparent that Matthew isn't making enough annual sales from the F18A MK1 to begin with. The F18A MK2 would have similar sales. If there was enough demand like 1500+ units sold a year, it wouldn't be a big issue.  Without being able to subsidize the annual fee on other products, it will require a sales volume of 1000+ units to pay the fee. Lets not forget that there is more than just having an HDMI Adopter license. The product has to be tested at an ATC (Authorized Testing Center) to comply with HDMI specifications. 

 

Current sales volume (F18A MK1)  is a variable range of 100-250 a year. This is based on what matthew had already said about the MK1 sales. There is no reason to expect sales to jump to 1000+ sales. We would have to be looking at sales upwards of a magnitude of an order higher volume than what was of the original F18A.  We are a small niche community of an already small niche community in retro-computing. Only a subset of the retro-computing users buys hardware upgrades and replacements for their original. The rest of the retro-computing community only work from the emulation environment. This is simply the reality of retro-computing business. You're not going to be the next Bill Gates from the sales of products for these 30+ year old computers. The TMS99xx technology is roughly 40 year old technology which started around 1977 according to Karl Guttag (one of the TEAM OF SEVEN "miracle workers"). 

 

If we think about it, there are only so many of us from that generation that still has the passion for these systems. There is even less of us that uses the physical hardware. There is a lot more of us as "emulator" users than those of us with the real hardware at our fingertips. Even I wouldn't spin up the process of making runs of 288 F18A MK# without being able to sell them let alone running a production run of 1000+ to 2000+ units unless I stand to recoup the hard costs and a little for my time. 



#255 Wildstar OFFLINE  

Wildstar

    Chopper Commander

  • 125 posts

Posted Mon Jul 30, 2018 10:32 PM

 

I get Wildstar's idea, but, yeah, we kind-a need to work with real figures not hypotheticals to plead our case.

 

I agree but there is some real figures to extrapolate from. Price per unit. Matthew already mentioned how many F18A MK1 were sold. He already given some figure on his BOM costs. 

 

matthew isn't doing this as a full fledge retro-computing business.



#256 --- Ω --- OFFLINE  

--- Ω ---

    --- Ω ---

  • 12,487 posts
  • Location:워싱턴 주

Posted Tue Jul 31, 2018 8:27 AM

Those sales are across all the different retro platforms,  Many of those platforms will not need a MK2, other people who already have an original F18A might not feel a driving need to open up their legacy hardware just to replace a perfectly fine working piece of hardware.  While the device WILL be an improvement for many, especially newbies, the projected sales curve may not be as steep as some envision.



#257 Wildstar OFFLINE  

Wildstar

    Chopper Commander

  • 125 posts

Posted Tue Jul 31, 2018 1:06 PM

Those sales are across all the different retro platforms,  Many of those platforms will not need a MK2, other people who already have an original F18A might not feel a driving need to open up their legacy hardware just to replace a perfectly fine working piece of hardware.  While the device WILL be an improvement for many, especially newbies, the projected sales curve may not be as steep as some envision.

 

I expect sales in a stand-alone drop in replacement to be of limited sales. While I can envision more sales by making it a cartridge that a game, app, etc. would use and be a required hardware and be part of the price of the game, app, etc. but as a drop in replacement, sales would be limited especially to those willing to go through the effort of installing it. There is a number of people uncomfortable in removing and installing ICs when soldering is involved. So, a plug-in board on PEB or Apple II, or a cartridge on the C64 and some other systems would apply to a different level of comfort zone. I couldn't predict sales on that front but I can reasonably project that F18A MK2 direct sales to customers would likely be equal to or less than the original MK1 sales because some who already has the MK1 might not feel the need to buy the MK2 just because it plugs into HDMI or Displayport. 

 

I can see the value of the F18A MK2 used in various forms of embedded computing as well. However, sales on that front is uncharted territory. Matthew isn't planning to produce this for that and anything like that would have to come from a person licensing F18A MK2 from Matthew to explore that. 

 

If we need to reach annual 1000+ unit production runs, we'll have to really find a way to sell that many but we'll have a very difficult time trying to do that by only targeting the TI/MSX/Colecovision users.

 

Please note: It is not ALL the different retro-platforms. I would say all the different retro-platforms that originally has the TMS9918.


Edited by Wildstar, Tue Jul 31, 2018 1:09 PM.


#258 MrPix OFFLINE  

MrPix

    Space Invader

  • 27 posts

Posted Tue Jul 31, 2018 2:08 PM

One exciting aspect of this specific hardware is that it's really, REALLY generic. It will go in almost any socket where a ULA was used and can be configured by FPGA programing to be almost any device. The limits are basically the power and ground pins, and if any pins are not GPIO but just GPI or GPO. This could replace the 8301 or 8302 in a Sinclair QL, a lot of video processing in the earlier Ataris or Commodores - all it takes is the right FPGA code to map pins and to decode the video correctly.

Also, while the MSX market may be quite small here, it is quite large in Japan and other EMEA markets. Also, the UK/EU retro scenes are much more active and those guys (and gals) eat this stuff up.

I think he would sell far more by also having an "unprogrammed" option and people can just read the datasheet, create and install the JEDEC of their choice, etc. He'd find a whole bunch of aftermarket uses for this device.



#259 Wildstar OFFLINE  

Wildstar

    Chopper Commander

  • 125 posts

Posted Wed Aug 1, 2018 3:25 AM

One exciting aspect of this specific hardware is that it's really, REALLY generic. It will go in almost any socket where a ULA was used and can be configured by FPGA programing to be almost any device. The limits are basically the power and ground pins, and if any pins are not GPIO but just GPI or GPO. This could replace the 8301 or 8302 in a Sinclair QL, a lot of video processing in the earlier Ataris or Commodores - all it takes is the right FPGA code to map pins and to decode the video correctly.

Also, while the MSX market may be quite small here, it is quite large in Japan and other EMEA markets. Also, the UK/EU retro scenes are much more active and those guys (and gals) eat this stuff up.

I think he would sell far more by also having an "unprogrammed" option and people can just read the datasheet, create and install the JEDEC of their choice, etc. He'd find a whole bunch of aftermarket uses for this device.

 

While it could conceivably be mapped to work in place of the VIC-I or VIC-II, but the reality is, we would more likely install it on a cartridge and map the address and data lines and other lines as needed. The TMS9918 works differently enough to the VIC-I & VIC-II to pose some problems. You have an interesting idea there since it is an FPGA, after all. While initially, I didn't understand the ULA part but after a brief 'look it up', I can see how a FPGA can be useful that way. The key trick is how we print or layout the circuits from the underside pins to the FPGA. With minor modifications of the circuit designs, we can arrange for any known locations for power and ground  and route it. This design profile is awesome BECAUSE it would be functional mechanically for most if not all 40-pin DIP based microcontrollers and in some cases, the protector IC and HDMI connector is not needed. Therefore, the brilliance of coming up with drop in replacements chips for our beloved computers components would be awesome. Our limit is ultimately in the FPGA's available logic elements/cells. In the future, I am confident FPGAs 5mm x 5mm will have the amount of logic cells/elements and logic gate resources to make any chip from 1975 to 1995 with enough contacts for any chip up to 208 pins such as having more contacts on the FPGA than a Pentium in a 15mm x 15mm chip for a drop in replacement for Pentium processor. it is all a matter of time that they can even run processors as complex as a Pentium on an FPGA as fast if not faster than even a Pentium processor with less heat dissipation without even needing a heat sink. Of course, that is years down the road but from what I see, we can see our beloved computers having drop in replacements before that in the next 5 years as I can see the Pentium drop in replacement on FPGA in 10 to 15 years.

 

I think we can see a FAT AGNUS on FPGA on an 8mm x 8mm to 10mm x 10mm FPGA on a PCB that plugs into an 84 pin QFP in the next few years. I can see all kinds of stuff made for our beloved systems whether it is a Sinclair, C64, TI, Apple II, Amiga, or so forth.

 

The trick is getting the FPGAs inexpensive enough so people can afford to buy all those components without costing a fortune (in the perspective of discretionary funds budget people tend to have on retro-computing).


Edited by Wildstar, Wed Aug 1, 2018 4:07 AM.


#260 towmater OFFLINE  

towmater

    Dragonstomper

  • 537 posts

Posted Wed Aug 8, 2018 9:18 AM

Shut up and take my monkey?







Also tagged with one or more of these keywords: F18A, VDP, 9918A

0 user(s) are browsing this forum

0 members, 0 guests, 0 anonymous users