Question on P2 breadboard work
photonverse
Posts: 9
in Propeller 2
Hi,
I am a hobbyist and trying to build a minimal P2 breadboard setup. As shown in the attached picture, all the TQFP-100 pins have been soldered, and I drilled a hole on the bottom plane so that I can solder a wire to the ground pad later. Now to complete the rest with minimum wiring effort, I was wondering if I could:
(1) Use only the internal RC clock source,
(2) Supply VDD 1.8V to physical pin 2 (P0~P3) and pin 94 (P60~P63), and 3.3V to VIO_0_3 and VIO_60_63 only without bypass capacitors
(3) Connect TX/RX (P62~P63) to Prop Plug
If I program (without SPI flash) that uses only in range P0 to P3, would that work?
Thanks in advance
I am a hobbyist and trying to build a minimal P2 breadboard setup. As shown in the attached picture, all the TQFP-100 pins have been soldered, and I drilled a hole on the bottom plane so that I can solder a wire to the ground pad later. Now to complete the rest with minimum wiring effort, I was wondering if I could:
(1) Use only the internal RC clock source,
(2) Supply VDD 1.8V to physical pin 2 (P0~P3) and pin 94 (P60~P63), and 3.3V to VIO_0_3 and VIO_60_63 only without bypass capacitors
(3) Connect TX/RX (P62~P63) to Prop Plug
If I program (without SPI flash) that uses only in range P0 to P3, would that work?
Thanks in advance
Comments
Sorry, it's a bit harsh. But there is just too much wrong with this to comment further.
Okay. I must have got it all wrong. Thanks for your comment though, otherwise I wouldn't even know about that.
You really are best to get one of the pre-assembled P2 boards. eg p2 Edge.
You need to also connect TEST and RESN pins, and the VDD pins are commoned, but the 3v3 VIO pins are not. So you do need to also connect the PLL/OSC (IIRC V2831)
The VIO and VDD pins should be decoupled, with shortest lead lengths.
On your breakout, you might be ok with as many VDD pins as you can physically get access to on the rear side, all decoupled with SMD caps.
A local 1V8 regulator would be safer and a local 3v3 regulator would also be recommended.
Furthermore, I would not risk connecting only a single VDD pin up even if they are "commoned" internally since this is not the same as a very low resistance common that comes from 1oz copper.
If you want I am sure I can mail you a blank P2D2 pcb which you can mount your P2 onto and it brings out all 64 I/O etc.
A few months ago, I constructed a minimal Propeller 1 on breadboard, following "Propeller Education Kit Labs" manual, and I felt I learned so much more by making one by myself (I do have Activity/Activity WX, QuickStart board). In the same spirit, I wanted to build a P2 breadboard circuit. After seeing a P2 Eval Rev C board picture, and I noticed there are so many small SMD components placed around the P2 chip (which I felt somewhat intimidating). It was my very first time to solder a 0.5mm pitched 100 pin SMD successfully in my life (checked continuity for all 100 pins)! Yes, I am aware of P2 Edge and Jonny Mac board which I plan to purchase when they become available.
I thought if I limit the clock frequency with the default internal RC oscillator 20 MHz with PLLx1 (i.e. final frequency to 20 MHz or lower) and try to toggling just one pin for LED blinking, then I thought I could get away with dealing the bypass capacitors and the thermal issues.
Thanks again to Cluso99, jmg, Peter for your response, and I will take your advice and save my other P2 chip for PCB designed by others (I purchased two P2 chips and now only one is left) .
Solder the center pin (preferably via a heavyish wire (you could use a resistor lead) to a tiny piece of breakout board at the center of you main pcb. Use this pcb to solder your GND power connection(s) and your bypass and bulk capacitors (later) between here and the 1V8 and 3V3 0.1" breadboard pins. Connect all your 1V8 pins together and all your 3V3 pins together using the 0.1" headers.
Then underneath connect a few 100nF bypass capacitors (thru hole) and between the center pin pcb and the 0.1" 1V8 and the 3V3 pads on the main breakout pcb.
Where you connect the 3V3 and the 1V8 to the main pcb, you should connect a bulk capacitor to that pin and the center ground board. A value of 4u7F to 22uF should be ideal.
Then you can connect LED(s) - each pin to the LED and the other side via a 1K to 10K (I use 3K3 often) resistor to the center ground board.
Now you have a "ratsnest" board and you should be able to download some code to it. You'll need to connect a PropPlug to your board to do this. I'm guessing you have one since you build a P1 board.
Post some pics before you power your setup so we can check your not going to destroy anything.
The code is easy to flash a few leds. If you have trouble, just ask on the forum.
for what its worth, if running of RCSlow or RCFast (~25 MHz) I think you'll be fine. There's a 1v8v VDD ring around the inside and its designed for the heavier currents. There's more informatioin in this thread,
https://forums.parallax.com/discussion/167444/power-routing-troubles
In case any of you aren't aware of ChipQuik, I'll mention it now. ChipQuik makes removing smt parts much much easier than any other technique I've tried. Since the stuff melts at such a low temperature, you have a very good chance of getting a part off a PCB without out damaging the PCB or the component.
Forum member Leon recommended this years ago and I've successfully used it many times myself.
Components with soldered center pads will still be difficult remove even if using ChipQuik. It your P2 doesn't work on the present PCB, you may still have a chance of using it on a different board.
It'd be really cool if you could make that work!
Every voltage input pin on the P2 does need to be connected to a cap, probably.
But, one cap can serve several pins.
Look like you have some good advice above.
Maybe with something like a PropPlug you could make it work...
@Cluso99
Your guess is correct. Since I already soldered the P2 chip on the adapter, I thought the only option is to push forward and see if I can make it working. Taking your advice, I connected all VDD pins and VIO pins and *a few 100nF bypass capacitors* as you mentioned (I interpreted that if I plan to use only a few pins, then I may use only some capacitors to the most relevant pins. I will only use RX/TX and P0 for LED test). Unfortunately, I already drilled a small 3/32" (2.38mm) diameter hole and soldered a wire through there, so I didn't use any extra breakout board. For the initial testing, I will use my bench power supply. For the RESET pin, can I directly plug to PropPlug's reset connector? If I can, then I plan to add 10K resistor to 3.3V with a jumper option so that I will remove the jumper when I program, once done the programming place header to connect the pullup resistor.
@Tubular
Thanks for sharing the link. It answered many questions that I had in mind. The discussion from the link convinced me that I should connect all the VDD and VIO pins although VDD pins are all connected in a ring as part of the custom layout.
@"Duane Degn"
I thought my adapter project would be either "sort of working" or failure, but you gave me a hope that I may be able to save the P2 chip. Yes, indeed the soldered GND would be the problem, but if my project fail, I will try ChipQuik as you suggested.
Also, thanks @Rayman and @DigitalBob for encouraging remarks.
I have attached the adapter pictures that I have worked on. I guess it is a horrible sight to some, but I am sharing it anyway here. While connecting all VDD and VIO header pins, I felt I am constructing monorail or working on dental braces. And stacking the capacitors in such a mess made me to realize that indeed using a TQFP100 adapter is not a feasible way. Now I know how much effort will be saved by using someone else's P2 PCB design. I will truly appreciate the designer when I use one.
Once the banana plugs I ordered from Amazon arrives tomorrow for my bench power supply, I will test if it works (I don't think I will be able to add more capacitors to all the rest of VDD and VIO pins).
Probably you can connect the prop plug reset direct to the reset pin, but you will need the 10k pull-up too. I say probably because I’ve seen some reset timing problems with the different loaders. The P2 specs a series 100nF whereas the P1 uses 10nF. This affects the timing.
You didn’t mention any bulk capacitors - you will need a couple as the power supply is some distance away.
I placed a 10K resistor between RESET and VIO rail and placed a 5V LED with built-in resistor between P0 and TEST (GND). When I issued ">" + SPACE + ESC, sequence in the PuTTY console, then P0 pin showed up with status "d". I don't know what it meant, but I thought it was something since only P0 was connected a LED. With issuing command "0 BLINK", it started blinking! I am so pleased that all the effort didn't go wasted.
Again, thank you so much everyone who helped me. I have attached the photo taken with LED blinking. I will start leaning TAQOZ tonight!
Typical cogexec is PGM0, and typical hubexec is PGM5. So on RCFAST you should be at about 33 mA on the 1.8 V rail (VDD).
You are exactly right. My bench-top supply reading is 0.03A. Your power rating table is very handy. I had no idea about the max current consumption, so I had arbitrarily set the limit to 400mA for VDD.
I don't know which of your program (PGM0 ~ PGM5) corresponds to TAQOZ "0 BLINK" command, but I am guessing that although BLINK is using only smartpin operation (according to TAQOZ documentation), TAQOZ itself could be running on a cog/hubexec, so 30mA makes sense to me.
I got hooked on Parallax and at that time the P1 with the same P1 Breadboard Build. The P2 is even more fascinating but documentation for most tools is still in progress, as are the tools them self.
@"Peter Jakacki" is the guy responsible for the ROM based TAQOZ, he has a quite newer version to put in flash or on a sd card.
But the ROM based TAQOZ is very useful to validate your hardware. And yes the "d" you mentioned lsio shows you states that the pin 0 is down (low) and not floating.
have fun,
Mike