@Cluso99 Wouldn't it be easier and less expensive to just make the P2_MaxiBlade? How much smaller would it be?
Now it's only a little wider and doesn't have the expensive switcher, and with more I/O brought out, it makes much more sense.
Almost finished the routing, with a little tweeking to do, and then just the power and ground planes to do
Have you thought about using a USB-C connector, USB-C cable, and DB9 connector like Peter talked about? Seems like a great idea to me.
Then also someone could possibly use the same connector for HDMI if they wanted.
How about using the unused pins for a pair of HyperRAM chips on the bottom on another version?
This is a small simple board with basic features. I’m not trying to be everything for everyone. Peter’s board covers way more features than mine.
I am not into HyperRAM
There are standard sized pin headers for user expansion to add other features if required.
This is a small simple board with basic features. I’m not trying to be everything for everyone. Peter’s board covers way more features than mine.
I am not into HyperRAM
There are standard sized pin headers for user expansion to add other features if required.
Whatever you decide, I want one. I'm guessing yours will be the fastest way to get up and running with the P2 without shelling out the big bucks for the P2 Eval & accessories.
It looks like Chip is designing a pair of HyperRAMs into a version of the P2Edge anyways.
The (maxi) board is progressing. I need to move the bulk caps for the P2 1V8 and 3V3 supply around as they are blocking the P2 ground plane. It’s just the usual minor tweaking that goes on in the final ground plane pouring. All the bypass caps are routed.
Some good news.....
I have added an optional SOIC8 FLASH chip on the underside of the pcb under the microSD.
I am going to make an alternative version that replaces the VGA connector with a 2x10 pin 0.1” header to bring out P16-23 and P24-31 plus 3V3 and GND. This means P0-47 will be brought out on 0.1” headers. P48-51 go to the microUSB connectors, P52-57 are not brought out, then P58-61 go to the microSD and optional FLASH (onboard), and P62-63 are brought out to the 5/6 pin 0.1” headers (with Reset, GND and 5V) which can connect to the fleabay $1.50 CP2102 USB-TTL 6pin board. The transistor reset circuit is on my P2 pcb (solder linkable).
On the VGA version, P16-20 goes to the VGA connector and P21-31 are not brought out.
I’ll post a new 3D pic when I get home later this morning, although it’s not much different to what I posted a few posts back.
I bought a cheap fleabay microUSB to USB female 6” cable (connects to a mouse or keyboard with standard USB-A connector). Cost ~$1.50 shipped. Just need to verify it connects the 4 pins correctly.
* PCB 1.5"x2.0" 2 layer ENIG (probably I'll go for 1.2mm pcb rather than 1.6mm)
* P2X8C4M64P chip of course
* 20MHz xtal
* Requires 5V regulated input from one of headers, either microUSB, or USB-TTL CP2102 (not supplied - fleabay ~$1.50)
* 3V3 1A LDO SOT-89 regulator
* 1V8 generated by series 3V3 1A LDO SOT-223 to 1V8 1A LDO SOT-223
* Proper bulk and bypass capacitors
* Transistor reset circuit (as per P1) configured with solderable link
* microSD socket (microSD card not supplied) - can boot direct from microSD without Flash
* Two microUSB sockets - for power / keyboard / mouse - cables available from fleabay for ~$1.50
* VGA socket for video screen / LCD
* P00-P15 and P32-47 brought out on 2 * 1x10pin 0.1" headers
* 1x6pin 0.1" header for Reset/SerialOut/SerialIn/5Vin/Gnd/nc compatible with CP2102 USB/TTL 6pin
* Optional SOIC-8 footprint on underside for Flash W25Q32JVSIQ/W25Q64JVSIQ/W25Q128JVSIQ (4/8/16MB Flash) with 0.050" header/link to enable Flash boot via 10K pullup resistor (fitted as standard)
Notes:
* The regulators will likely self-limit to somewhere in excess of 500mA. Switchers are not used.
* The 3d image shows the wrong microSD and microUSB connectors. Peter showed the ones I'm using on his P2D2 thread the other day.
* I have all the parts except for the P2 and optional Flash, and pcbs of course.
* I might do an alternative version (depends on interest) with P16-23 on a 2x10 0.1" header instead of the VGA connector (shown a few posts back).
KiCad finally passes all the design rule checks. It's all laid out now, so I only need to generate and verify the gerbers. Then I can panelise and send off for pcb manufacture.
Your regulators might thermally limit before 500ma as there is not much there to heatsink it. If you use JLCPCB you can panelize them just by selecting "panel by JLCPCB" and enter the x/y and edge rails etc. If you need some P2 chips I can send you some since I have to order more anyway. I have plenty of the 1MB Flash chips that will fit that smaller footprint you have too.
Looks good Ray. Looks like you're really getting the hang of Kicad. How have you found it? I haven't used it all that much, yet, but surrounded by many who are using it
Looks good Ray. Looks like you're really getting the hang of Kicad. How have you found it? I haven't used it all that much, yet, but surrounded by many who are using it
I've been using it too and now that I've sorted out all those V4 vs V5 library problems. Most of it is fairly easy, some a little easier than Protel, others a bit more awkward but there seems to be an ongoing bug with selecting existing fill zones when other footprints and text etc keep getting selected.
I’m finding KiCad fairly easy. Some thing are easier than Protel (99SE so it’s old) and others more complex. Routing tracks is easier except when you need to force the angles the right way as sometimes it has a mind of its own.
There is no global changes but since the files are straight text it’s easy to edit with notepad++. I needed to turn off all the reference numbers for the caps - there are ~50 on my pcb - setup a modified footprint and then changed the footprint name with notepad.
The fills are both good and bad. You can have multiple levels, but you have to wire all the pads (eg ground visas) with a track first as otherwise the DRC gives errors even tho they are connected. Also there’s a problem with multiple levels if a subsequent fill is over an existing one and you have reliefs set - there is a pixel fragment that gets missed so they don’t join properly. You have to place tracks over the join
My VGA has the mounting pads set to GND but only one gets connected to the GND fill so I had to put in a track to GND.
I’m running 5.1.7 )September release.
All up I’m happy as I can do slots easily. Also bitmaps. 3D is good. Not used to the keyboard shortcuts yet. Nicely integrated with schematics.
Looks good Ray. Looks like you're really getting the hang of Kicad. How have you found it? I haven't used it all that much, yet, but surrounded by many who are using it
I've been using it too and now that I've sorted out all those V4 vs V5 library problems. Most of it is fairly easy, some a little easier than Protel, others a bit more awkward but there seems to be an ongoing bug with selecting existing fill zones when other footprints and text etc keep getting selected.
Click on the edge of the fill zone and when selected hit E. Turn off other layers helps. Have the layer the fill is on selected as the working layer. To get under a fill, I select and then move the centre dot to move one edge off the pcb and move it back later.
Your regulators might thermally limit before 500ma as there is not much there to heatsink it. If you use JLCPCB you can panelize them just by selecting "panel by JLCPCB" and enter the x/y and edge rails etc. If you need some P2 chips I can send you some since I have to order more anyway. I have plenty of the 1MB Flash chips that will fit that smaller footprint you have too.
Thanks Peter. If you need any more microSD, microUSB, or 0.050” m or f just let me know, or anything else you know/think I might have.
I have planes under the regulators top and bottom. I tested them on a bit of breadboard with virtually no pads at all and I could get above 500mA so I expect more. TI did a nice app note on the pad sizes and results for sot223 regulators. Works much better using copper both sides.
Rather than polute my P2-RetroBlade thread, I thought I would ask here...
As an alternative to the VGA connector version, and the other version of 2x10 pin 0.1” pitch header for P16-23, what about if I fitted 2x RJ45 sockets?
Each RJ45 would connect 8 P2 pins so we would get the whole P16-23 set. Any ground would need to be provided by the P2 sacrificing a pin.
Now we can have two VGA’s if required. We also have the full P2 pins on the connector/cable for some other uses (remember, you may need to provide GND on one of those pins, or else you will need to provide an extra GND wire from one of the headers).
Both VGA and RJ45 connectors are so bulky in this smd world. Whichever way you go it wouldn't hurt to allow for a pin header on the board so you can adapt it very easily with a cable or another tiny pcb as need be.
@Tubular,
Using the shield would require both shielded RJ45 socket ($$) and shielded CAT5 cables (not readily on hand, and $$).
My main objective is to get a cheap interface to VGA. Anything else is a bonus. The CAT5 cable and VGA adapter are readily available and cheap and the CAT5 is more flexible than the VGA cables and doesn't put the big VGA connector directly on the pcb.
@Peter,
Don't think I will have the space for 0.1" headers and RJ45. I want to keep the pcb to 2" in length and I am not in the mood to redo the layout even if I could squeeze it in - you know the problems with this - I remember someone (who shall remain nameless) asking you to squeeze the 0.050" pitch to 1.0"
Hey Ray, you owe me. Remember the "just move it in a little"? I figure you could fit a single row in surely but I would allow for dual row under the socket. No socket? Use the header.
Bad news...
The dual RJ45 requires more pcb area (depth) than the VGA so the pcb must grow in length by 0.1" to 2.1"x1.5" providing I can squeeze the tracks in in this space
PCB size now 2.2"x1.6" and has 1x16 0.1" P16-31 across the bottom with an RJ45 option (for VGA) that overlays the P24-31 headers.
I had to move the existing headers out 0.050" to get them to match the 0.1" grid of the new 1x16 across the bottom edge.
Comments
Almost finished the routing, with a little tweeking to do, and then just the power and ground planes to do
Note the 3D connectors shown are not the real parts being used (microSD, microUSB, VGA)
Then also someone could possibly use the same connector for HDMI if they wanted.
How about using the unused pins for a pair of HyperRAM chips on the bottom on another version?
I am not into HyperRAM
There are standard sized pin headers for user expansion to add other features if required.
Whatever you decide, I want one. I'm guessing yours will be the fastest way to get up and running with the P2 without shelling out the big bucks for the P2 Eval & accessories.
It looks like Chip is designing a pair of HyperRAMs into a version of the P2Edge anyways.
The (maxi) board is progressing. I need to move the bulk caps for the P2 1V8 and 3V3 supply around as they are blocking the P2 ground plane. It’s just the usual minor tweaking that goes on in the final ground plane pouring. All the bypass caps are routed.
Some good news.....
I have added an optional SOIC8 FLASH chip on the underside of the pcb under the microSD.
I am going to make an alternative version that replaces the VGA connector with a 2x10 pin 0.1” header to bring out P16-23 and P24-31 plus 3V3 and GND. This means P0-47 will be brought out on 0.1” headers. P48-51 go to the microUSB connectors, P52-57 are not brought out, then P58-61 go to the microSD and optional FLASH (onboard), and P62-63 are brought out to the 5/6 pin 0.1” headers (with Reset, GND and 5V) which can connect to the fleabay $1.50 CP2102 USB-TTL 6pin board. The transistor reset circuit is on my P2 pcb (solder linkable).
On the VGA version, P16-20 goes to the VGA connector and P21-31 are not brought out.
I’ll post a new 3D pic when I get home later this morning, although it’s not much different to what I posted a few posts back.
* PCB 1.5"x2.0" 2 layer ENIG (probably I'll go for 1.2mm pcb rather than 1.6mm)
* P2X8C4M64P chip of course
* 20MHz xtal
* Requires 5V regulated input from one of headers, either microUSB, or USB-TTL CP2102 (not supplied - fleabay ~$1.50)
* 3V3 1A LDO SOT-89 regulator
* 1V8 generated by series 3V3 1A LDO SOT-223 to 1V8 1A LDO SOT-223
* Proper bulk and bypass capacitors
* Transistor reset circuit (as per P1) configured with solderable link
* microSD socket (microSD card not supplied) - can boot direct from microSD without Flash
* Two microUSB sockets - for power / keyboard / mouse - cables available from fleabay for ~$1.50
* VGA socket for video screen / LCD
* P00-P15 and P32-47 brought out on 2 * 1x10pin 0.1" headers
* 1x6pin 0.1" header for Reset/SerialOut/SerialIn/5Vin/Gnd/nc compatible with CP2102 USB/TTL 6pin
* Optional SOIC-8 footprint on underside for Flash W25Q32JVSIQ/W25Q64JVSIQ/W25Q128JVSIQ (4/8/16MB Flash) with 0.050" header/link to enable Flash boot via 10K pullup resistor (fitted as standard)
Notes:
* The regulators will likely self-limit to somewhere in excess of 500mA. Switchers are not used.
* The 3d image shows the wrong microSD and microUSB connectors. Peter showed the ones I'm using on his P2D2 thread the other day.
* I have all the parts except for the P2 and optional Flash, and pcbs of course.
* I might do an alternative version (depends on interest) with P16-23 on a 2x10 0.1" header instead of the VGA connector (shown a few posts back).
KiCad finally passes all the design rule checks. It's all laid out now, so I only need to generate and verify the gerbers. Then I can panelise and send off for pcb manufacture.
I've been using it too and now that I've sorted out all those V4 vs V5 library problems. Most of it is fairly easy, some a little easier than Protel, others a bit more awkward but there seems to be an ongoing bug with selecting existing fill zones when other footprints and text etc keep getting selected.
There is no global changes but since the files are straight text it’s easy to edit with notepad++. I needed to turn off all the reference numbers for the caps - there are ~50 on my pcb - setup a modified footprint and then changed the footprint name with notepad.
The fills are both good and bad. You can have multiple levels, but you have to wire all the pads (eg ground visas) with a track first as otherwise the DRC gives errors even tho they are connected. Also there’s a problem with multiple levels if a subsequent fill is over an existing one and you have reliefs set - there is a pixel fragment that gets missed so they don’t join properly. You have to place tracks over the join
My VGA has the mounting pads set to GND but only one gets connected to the GND fill so I had to put in a track to GND.
I’m running 5.1.7 )September release.
All up I’m happy as I can do slots easily. Also bitmaps. 3D is good. Not used to the keyboard shortcuts yet. Nicely integrated with schematics.
Click on the edge of the fill zone and when selected hit E. Turn off other layers helps. Have the layer the fill is on selected as the working layer. To get under a fill, I select and then move the centre dot to move one edge off the pcb and move it back later.
I have planes under the regulators top and bottom. I tested them on a bit of breadboard with virtually no pads at all and I could get above 500mA so I expect more. TI did a nice app note on the pad sizes and results for sot223 regulators. Works much better using copper both sides.
forums.parallax.com/discussion/172262/clusos-p2retroblade-and-p2maxiblade-interest-please/p1?new=1
As an alternative to the VGA connector version, and the other version of 2x10 pin 0.1” pitch header for P16-23, what about if I fitted 2x RJ45 sockets?
Each RJ45 would connect 8 P2 pins so we would get the whole P16-23 set. Any ground would need to be provided by the P2 sacrificing a pin.
The RJ45 can be used with a cheap CAT5 cable (you probably have some of these anyway) and an RJ45-VGA Adapter (fleabay ~$2.50 for a pair). This solves the VGA connection - it uses the twisted pairs in the cat5 cable for the RGB pins with the pair being GND which would be supplied by the P2 pins.
https://google.com.au/search?q=rj45+vga+pinout&client=safari&hl=en-au&sxsrf=ALeKk01FhBRFtLWAC9A9zIF_74-zmdgwTg:1602359293043&tbm=isch&source=iu&ictx=1&fir=ADbz54C7hIb27M%252C4lsfIvpzBecD7M%252C_&vet=1&usg=AI4_-kROUW35KF2qsDby1h-slu1J2O-eLA&sa=X&ved=2ahUKEwjyitvD5arsAhVPzzgGHSiGBIIQ9QF6BAgKEEU&biw=1261&bih=819&dpr=2#imgrc=ADbz54C7hIb27M
https://ebay.com/itm/2pcs-lot-RJ45-Network-Cable-Adapter-VGA-Female-To-RJ45-Female-Adapter-LAN-CAT5-C/222100439562?hash=item33b637c20a:g:fiQAAOSwnFZXWjfh
Now we can have two VGA’s if required. We also have the full P2 pins on the connector/cable for some other uses (remember, you may need to provide GND on one of those pins, or else you will need to provide an extra GND wire from one of the headers).
What do you think?
Using the shield would require both shielded RJ45 socket ($$) and shielded CAT5 cables (not readily on hand, and $$).
My main objective is to get a cheap interface to VGA. Anything else is a bonus. The CAT5 cable and VGA adapter are readily available and cheap and the CAT5 is more flexible than the VGA cables and doesn't put the big VGA connector directly on the pcb.
@Peter,
Don't think I will have the space for 0.1" headers and RJ45. I want to keep the pcb to 2" in length and I am not in the mood to redo the layout even if I could squeeze it in - you know the problems with this - I remember someone (who shall remain nameless) asking you to squeeze the 0.050" pitch to 1.0"
The dual RJ45 requires more pcb area (depth) than the VGA so the pcb must grow in length by 0.1" to 2.1"x1.5" providing I can squeeze the tracks in in this space
I had to move the existing headers out 0.050" to get them to match the 0.1" grid of the new 1x16 across the bottom edge.