Rayman: Thanks and yes, this module is ideal for utilization of PropBASIC.
David: The connectors I am using are Samtec TS-120-T-G-1, Link to the TS series on Samtec's site is here: http://www.samtec.com/ProductInformation/TechnicalSpecifications/Overview.aspx?series=TS
Other than the Pinout diagram on post #1, what would you like to see? I will upload an improved diagram that shows the pinout options based upon header loading later this evening.
I've been down this road with the PropKey (which isn't dead, I just have so much time to devote to all of my extracurricular activities). The manufacturing cost for a double sided SMT board was $32.
I have serious doubts you will be able to produce and sell it at your proposed price point, and do anything other than barely break even.
Prof_Braino: I am currently working on a site for taking pre-orders. Life has gotten the best of me the past week and I have not been able to touch anything, but will be getting back on track tomorrow. I will be actively keep this thread up to date now.
Paul: Yes, a "carrier" or "breakout board" is fairly imminent so that the module can be easily used in it's entirety. Once the first orders of the module are rolling out, this will be revisited. Thanks for the compliments, order form is coming....
pedward: Yes, I am very aware of the challenges with manufacturing costs of lower volume PCBs, but I am confident at being above break even at the $35 price point. I have no concerns with the double sided SMT design and the cost impact of the additional process will be kept at a minimum. There are startup costs to factor for the initial 50-100 which is part of the delay with getting these out to customers. This is why I am setting up a process for pre-orders.
It might be helpful for other makers/hackers/etc if you share any of the tips for minimizing short run assembly/fab costs.
I've got a few boards in the pipeline that I'd love to make available (likely through Tindie) but I realize my only option is either hand assemble as many as I can stand (a couple hundred), or to make several thousand.
Instead of making your own site for pre-orders, it might be better to use Indiegogo or Kickstarter...
UPDATE: 1/21/13 Unfortunately, this project took a back seat to some "matters of day-to-day life", but I am now working on it again. I will be adding information on the first post about a pre-order site. It looks like I will be targeting for commitments for 60 modules @ $35 to kickoff a production run, with a 6-8 week turn around. I am leaning towards starting a Tindie Fundraiser if interest shows that 60 will be possible. We are also considering a breakout board to be offered upfront with the kickoff of the module.
UPDATE: 1/21/13 Unfortunately, this project took a back seat to some "matters of day-to-day life", but I am now working on it again. I will be adding information on the first post about a pre-order site. It looks like I will be targeting for commitments for 60 modules @ $35 to kickoff a production run, with a 6-8 week turn around. I am leaning towards starting a Tindie Fundraiser if interest shows that 60 will be possible. We are also considering a breakout board to be offered upfront with the kickoff of the module.
Nice little module, two comments
a) Can you bring out the handshake lines of the FTDI device to solder-lands. I've seen too many boards have no means at all to add wires, and the smarter ones bring them to larger vias, or smd sized solder lands. It does not need to be a full header, but it does need to be more than the edge of a QFN package
b) 50mil headers would seem restrictive. a dual row 0.1" has the same connection count, and is a lot more universal.
Or, is there some magic dimension you are desperate to keep inside ?
a) Since the FT230X does not have DTR, RTS is used for resetting the propeller during programming and as such is tied to the reset circuit. RTS is brought out to the header via the ATN pin (see the first questions in post #2). CTS is ignored in this design and left unconnected.
b) 50 mil headers may pose a few limitations for some people, but if you notice the pinout pattern of the "EVEN" pins in the first image attached in post #2, you can see that you can easily gain access to Propeller IO pins P16-P29, 5V, 3.3V, GND, and all 4 CBUS pins of the FT230X using 0.1" pitch headers that have machined pins. (machined pins=smaller diameter to fit the 50 mil pitch holes)
Really nice job on this board. You have a truly impressive collection of features.
There's a lot to be said for a good buck/boost onboard supply. Sure simplifies prototyping and project design not having to worry about that.
I really hope you are successful in your fundraiser. I also hope I have enough in my hobby fund to support it when the time comes
The HaD publicity I got for my Prop breakout on Tindie is probably the only reason the fundraiser didn't totally tank like a few others have. Truth be told I didn't get quite as many orders as I'd hoped but just enough to get started. Propellers seem niche. Easily understood differentiation surely helps given the number of Prop board designs. I didn't do enough to get the word out, I think.
Anyway I'd be happy to blog/tweet/g+ about it when the time comes if you want, just let me know.
a) Since the FT230X does not have DTR, RTS is used for resetting the propeller during programming and as such is tied to the reset circuit. RTS is brought out to the header via the ATN pin (see the first questions in post #2). CTS is ignored in this design and left unconnected.
Yes, but my point was, could someone connect CTS, if they wanted to ?
It is there in hardware, as a pathway back to the PC, and can be used as a simple activity indicator, or in a more complete design as a confirmation signal.
I'm always interested in things like this, even if I don't see an immediate need for a board, I like to collect them, because I'm a packrat, and you never know when something will suddenly be exactly what you need.
I'd get one of these anyhow, because the whole bs/prop conversion problem I have is the whole 5v/3.3v thing. I've pretty much targeted all of my parts for 5v, and when trying to port to the propeller, I run into the whole 3.3v isn't enough to power my component parts, so a board with 5v is great.
Of course, Parallax solved this problem with their human interface board (gotta get me more of those) so it's not as critical for me as it was) but it's still good to have anyway.
I also see this as a useful transition for educators who are currently using the bs2, who want to switch to the prop, but don't want to retool their lessons to accomodate the prop directly. With this little bugger, they'd not have to change a thing.
Very cool.
Put me down for one when they're available.
It's been a slow process of getting other projects off my plate to finish the pre-order site for the "$35 Mini Propeller Module". It is now being called the PropBSC or "Propeller: BASIC Stamp Compatible". Got the site and pre-order email loaded up here: http://bsc.powertwig.com (click on the module)
If you are interested, send an email per the instructions on the main PropBSC page and you will be added to the list. An official pre-order style fundraiser through Tindie will be started once interest proves to be supportive of an initial run. This looks to be somewhere between 60 and 70 units. The lead time for production will be about 6 weeks.
The fact that you offer BOTH 3.3v and 5.0v logic compatibility, and a USB interface make this a better value at that $35 USD than the Propeller Mini.
Even if you needed to increase the price a bit, it is still in many ways a more complete product that users might prefer. I just suspect that the economy and other factors have slowed interest in this. There may come a day when user interest will pick up.
I was also very pleased to see that more than one Ground was provided and several +5 outputs. If a mini device really is going to be fully exploited, quite a few devices will need to at least share grounding. The Propeller Mini has only one Ground (if you exclude the programing port interface).
Loopy, thanks for the comments. Yes, I was rather surprised at the lack of interest honestly. It features many things still not available with Propeller boards currently offered from anyone. This should have been a great way for BOE-BOT owners to migrate to the Propeller with very little expense. Maybe that is why the interest was low, only Propeller users saw the thread? If I had time, I would toss it over onto a Kickstarter, but I have other priorities that show more promise for now.
Sorry to hear this. It would appear that the market for Propeller boards (and Basic Stamp boards?) is either much smaller than we expected or people just tend to go with the Parallax gear and/or don't know other options exist. It doesn't help that the Prop chips are so expensive and require so many expensive support components. Unless you're making 1000's of them (and maybe even if), margins have to be set painfully low to sell at a price anyone actually wants to pay. Bummer.
Sorry to hear this. It would appear that the market for Propeller boards (and Basic Stamp boards?) is either much smaller than we expected or people just tend to go with the Parallax gear and/or don't know other options exist. It doesn't help that the Prop chips are so expensive and require so many expensive support components. Unless you're making 1000's of them (and maybe even if), margins have to be set painfully low to sell at a price anyone actually wants to pay. Bummer.
I wonder if in addition to creating a ~$15 powerful P2 chip it might be possible for Parallax to make a cost-reduced, <$4 P1 chip?
It would appear that the market for Propeller boards (and Basic Stamp boards?) is either much smaller than we expected or people just tend to go with the Parallax gear and/or don't know other options exist.
This has been my biggest question. I think if I had 100 in stock and available through a webstore, things might be different. There are only 2 design changes to be made to this current prototype to make it rock solid:
One is only a safety circuit since the current design would allow Vin and Vusb to meet head on if the device were locally powered and plugged into a USB port. Not a big deal as long as you remove local power while programming.
The second is designing in some way to handle the design flaw that exists in all of Parallax's BASIC Stamp BOE boards. The BOE boards have the 5 volt output of the onboard regulator tied to the 5 volt output of the BS2 module's regulator. That may not cause issues with two low current linear regulators, but with my 2.5 amp capable 5 volt switcher, there could be some problems.
It would be nice to tap into the market of BOE-Bot owners though. Last year, the number I heard from Parallax was over 300,000 had been sold. Even 1% of that market is 3,000 of my modules.
To tell the truth I had forgotten all about it between announcement and pre-ordering. I'm also notoriously impatient. I'd rather not order when there is a long wait for something. If the wait is high then I tend to move on to the next best thing that is available now. There is definitely an impressive array of features on that board. Any consideration to selling just the design files and BOM and I can send off to SeeedStudio and stuff my own components later on? I like it more than the QuickStart due to the breadboarding ability.
I wonder if in addition to creating a ~$15 powerful P2 chip it might be possible for Parallax to make a cost-reduced, <$4 P1 chip?
Anything is possible, and I wonder how many die XMOS actually have, on their 'family'.
The problem is you need to have some way to disable some cores/threads, and smaller packages certainly help.
However, with P1, a new die would be needed to disable options, and the die is already package-bound, so smaller options are off the table.
On the topic of smaller-packages I did see a Cypress road map, of an Automotive PSoC4, CY8C40xx
(SO16, QFN24) - but showing only an i2c port + PWM + IDAC (no uart, no SPI, no ADC)
I cannot fathom how someone can plan a tiny Automotive part and not include LIN ? Is LIN losing favour to i2c ?
Maybe they have a single large customer for this part.
Anything is possible, and I wonder how many die XMOS actually have, on their 'family'.
The problem is you need to have some way to disable some cores/threads, and smaller packages certainly help.
However, with P1, a new die would be needed to disable options, and the die is already package-bound, so smaller options are off the table.
On the topic of smaller-packages I did see a Cypress road map, of an Automotive PSoC4, CY8C40xx
(SO16, QFN24) - but showing only an i2c port + PWM + IDAC (no uart, no SPI, no ADC)
I cannot fathom how someone can plan a tiny Automotive part and not include LIN ? Is LIN losing favour to i2c ?
Maybe they have a single large customer for this part.
I wasn't thinking that this new P1 would be a P2 die with some pins unconnected. After all, P1 and P2 are not binary compatible and not really source compatible either. I was thinking it could be a resynthesis of the old P1 RTL converted to Verilog. At one point Chip said that would be an easy job. Of course, easy for Chip isn't necessarily easy for anyone else! :-)
I wasn't thinking that this new P1 would be a P2 die with some pins unconnected. After all, P1 and P2 are not binary compatible and not really source compatible either. I was thinking it could be a resynthesis of the old P1 RTL converted to Verilog. At one point Chip said that would be an easy job. Of course, easy for Chip isn't necessarily easy for anyone else! :-)
Problem is not the Verilog, but the full custom stuff still in P1.
Either way, the important words here are 'new die'. = a lot of cost, in $ and man-hours.
Do you try to re-do P1 custom cells, or cut/paste from what will be die-proven P2 cells ?
Then you can choose a Verilog P1 or P2, but P2 is far ahead of P1, so if it was me, I'd look at what P2 can fit into a TQFP32 package, with an option for 64 pin bonding. ie One new die fits either side of P1..
Any cost-down package ideally needs an on-chip regulator, but that could be an 'external transistor' design, to avoid using die area just to cool a regulator. (depends a little on just how much mA/MHz a P2 comes in at )
Problem is not the Verilog, but the full custom stuff still in P1.
Either way, the important words here are 'new die'. = a lot of cost, in $ and man-hours.
Do you try to re-do P1 custom cells, or cut/paste from what will be die-proven P2 cells ?
Then you can choose a Verilog P1 or P2, but P2 is far ahead of P1, so if it was me, I'd look at what P2 can fit into a TQFP32 package, with an option for 64 pin bonding. ie One new die fits either side of P1..
Any cost-down package ideally needs an on-chip regulator, but that could be an 'external transistor' design, to avoid using die area just to cool a regulator. (depends a little on just how much mA/MHz a P2 comes in at )
Isn't the P1 all digital? I would have thought it would be possible to synthesize the whole thing and get rid of any custom layout. Sorry if I misunderstood.
Comments
David: The connectors I am using are Samtec TS-120-T-G-1, Link to the TS series on Samtec's site is here: http://www.samtec.com/ProductInformation/TechnicalSpecifications/Overview.aspx?series=TS
Other than the Pinout diagram on post #1, what would you like to see? I will upload an improved diagram that shows the pinout options based upon header loading later this evening.
Excellent design by the way. Awaiting an order form.
I have serious doubts you will be able to produce and sell it at your proposed price point, and do anything other than barely break even.
Paul: Yes, a "carrier" or "breakout board" is fairly imminent so that the module can be easily used in it's entirety. Once the first orders of the module are rolling out, this will be revisited. Thanks for the compliments, order form is coming....
pedward: Yes, I am very aware of the challenges with manufacturing costs of lower volume PCBs, but I am confident at being above break even at the $35 price point. I have no concerns with the double sided SMT design and the cost impact of the additional process will be kept at a minimum. There are startup costs to factor for the initial 50-100 which is part of the delay with getting these out to customers. This is why I am setting up a process for pre-orders.
I've got a few boards in the pipeline that I'd love to make available (likely through Tindie) but I realize my only option is either hand assemble as many as I can stand (a couple hundred), or to make several thousand.
Instead of making your own site for pre-orders, it might be better to use Indiegogo or Kickstarter...
Nice little module, two comments
a) Can you bring out the handshake lines of the FTDI device to solder-lands. I've seen too many boards have no means at all to add wires, and the smarter ones bring them to larger vias, or smd sized solder lands. It does not need to be a full header, but it does need to be more than the edge of a QFN package
b) 50mil headers would seem restrictive. a dual row 0.1" has the same connection count, and is a lot more universal.
Or, is there some magic dimension you are desperate to keep inside ?
a) Since the FT230X does not have DTR, RTS is used for resetting the propeller during programming and as such is tied to the reset circuit. RTS is brought out to the header via the ATN pin (see the first questions in post #2). CTS is ignored in this design and left unconnected.
b) 50 mil headers may pose a few limitations for some people, but if you notice the pinout pattern of the "EVEN" pins in the first image attached in post #2, you can see that you can easily gain access to Propeller IO pins P16-P29, 5V, 3.3V, GND, and all 4 CBUS pins of the FT230X using 0.1" pitch headers that have machined pins. (machined pins=smaller diameter to fit the 50 mil pitch holes)
There's a lot to be said for a good buck/boost onboard supply. Sure simplifies prototyping and project design not having to worry about that.
I really hope you are successful in your fundraiser. I also hope I have enough in my hobby fund to support it when the time comes
The HaD publicity I got for my Prop breakout on Tindie is probably the only reason the fundraiser didn't totally tank like a few others have. Truth be told I didn't get quite as many orders as I'd hoped but just enough to get started. Propellers seem niche. Easily understood differentiation surely helps given the number of Prop board designs. I didn't do enough to get the word out, I think.
Anyway I'd be happy to blog/tweet/g+ about it when the time comes if you want, just let me know.
Yes, but my point was, could someone connect CTS, if they wanted to ?
It is there in hardware, as a pathway back to the PC, and can be used as a simple activity indicator, or in a more complete design as a confirmation signal.
I'd get one of these anyhow, because the whole bs/prop conversion problem I have is the whole 5v/3.3v thing. I've pretty much targeted all of my parts for 5v, and when trying to port to the propeller, I run into the whole 3.3v isn't enough to power my component parts, so a board with 5v is great.
Of course, Parallax solved this problem with their human interface board (gotta get me more of those) so it's not as critical for me as it was) but it's still good to have anyway.
I also see this as a useful transition for educators who are currently using the bs2, who want to switch to the prop, but don't want to retool their lessons to accomodate the prop directly. With this little bugger, they'd not have to change a thing.
Very cool.
Put me down for one when they're available.
If you are interested, send an email per the instructions on the main PropBSC page and you will be added to the list. An official pre-order style fundraiser through Tindie will be started once interest proves to be supportive of an initial run. This looks to be somewhere between 60 and 70 units. The lead time for production will be about 6 weeks.
The fact that you offer BOTH 3.3v and 5.0v logic compatibility, and a USB interface make this a better value at that $35 USD than the Propeller Mini.
Even if you needed to increase the price a bit, it is still in many ways a more complete product that users might prefer. I just suspect that the economy and other factors have slowed interest in this. There may come a day when user interest will pick up.
I was also very pleased to see that more than one Ground was provided and several +5 outputs. If a mini device really is going to be fully exploited, quite a few devices will need to at least share grounding. The Propeller Mini has only one Ground (if you exclude the programing port interface).
This has been my biggest question. I think if I had 100 in stock and available through a webstore, things might be different. There are only 2 design changes to be made to this current prototype to make it rock solid:
One is only a safety circuit since the current design would allow Vin and Vusb to meet head on if the device were locally powered and plugged into a USB port. Not a big deal as long as you remove local power while programming.
The second is designing in some way to handle the design flaw that exists in all of Parallax's BASIC Stamp BOE boards. The BOE boards have the 5 volt output of the onboard regulator tied to the 5 volt output of the BS2 module's regulator. That may not cause issues with two low current linear regulators, but with my 2.5 amp capable 5 volt switcher, there could be some problems.
It would be nice to tap into the market of BOE-Bot owners though. Last year, the number I heard from Parallax was over 300,000 had been sold. Even 1% of that market is 3,000 of my modules.
Anything is possible, and I wonder how many die XMOS actually have, on their 'family'.
The problem is you need to have some way to disable some cores/threads, and smaller packages certainly help.
However, with P1, a new die would be needed to disable options, and the die is already package-bound, so smaller options are off the table.
On the topic of smaller-packages I did see a Cypress road map, of an Automotive PSoC4, CY8C40xx
(SO16, QFN24) - but showing only an i2c port + PWM + IDAC (no uart, no SPI, no ADC)
I cannot fathom how someone can plan a tiny Automotive part and not include LIN ? Is LIN losing favour to i2c ?
Maybe they have a single large customer for this part.
Problem is not the Verilog, but the full custom stuff still in P1.
Either way, the important words here are 'new die'. = a lot of cost, in $ and man-hours.
Do you try to re-do P1 custom cells, or cut/paste from what will be die-proven P2 cells ?
Then you can choose a Verilog P1 or P2, but P2 is far ahead of P1, so if it was me, I'd look at what P2 can fit into a TQFP32 package, with an option for 64 pin bonding. ie One new die fits either side of P1..
Any cost-down package ideally needs an on-chip regulator, but that could be an 'external transistor' design, to avoid using die area just to cool a regulator. (depends a little on just how much mA/MHz a P2 comes in at )
http://www.criticalvelocity.com/item.php?itemid=ard4