@dgately
Thanks for trying with an FT232.
Seems we have found a limit with the CP2102
The FT232 is an overpriced chip, with the raw chip costing around 3x the cost of the complete CP2102 board posted. Guess you get what you pay for
And you have to be careful of the FT232 fakes out there too.
I have a Click board ("Uart 3") that has a CP2102N on board. Let me know if you want me to courier it so you can test with it
The CP2102N supply is a bit of a joke, Mouser show nearly 200,000 on order across the variants. I'm not sure whether Silabs are doing something right or wrong, its hard to tell.
@ozpropdev said:
FYI
I was able to get "debug_plot_hubram.spin2" to work with CP2102 by dropping the baud to 115200..
Anything higher fails.
DEBUG_BAUD = 115200
]
Brian,
Just been looking for the instructions to add this into the DEBUG_demo.spin2 to try. Do I add it into the CON section like this (doesn't do anything)
CON
_clkfreq = 100_000_000
DEBUG_BAUD = 115200
PUB go() | i
waitms(1000)
repeat i from 0 to 9
debug(udec(i))
@ozpropdev said:
FYI
I was able to get "debug_plot_hubram.spin2" to work with CP2102 by dropping the baud to 115200..
Anything higher fails.
DEBUG_BAUD = 115200
]
Brian,
Just been looking for the instructions to add this into the DEBUG_demo.spin2 to try. Do I add it into the CON section like this (doesn't do anything)
CON
_clkfreq = 100_000_000
DEBUG_BAUD = 115200
PUB go() | i
waitms(1000)
repeat i from 0 to 9
debug(udec(i))
@ozpropdev said:
FYI
I was able to get "debug_plot_hubram.spin2" to work with CP2102 by dropping the baud to 115200..
Anything higher fails.
DEBUG_BAUD = 115200
]
Brian,
Just been looking for the instructions to add this into the DEBUG_demo.spin2 to try. Do I add it into the CON section like this (doesn't do anything)
CON
_clkfreq = 100_000_000
DEBUG_BAUD = 115200
PUB go() | i
waitms(1000)
repeat i from 0 to 9
debug(udec(i))
That's correct Ray.
Thanks Brian.
Don't know why it didn't work the first time I tried it ;(
Curiously, sometimes I just get a blank debug screen ie no output.
Not sure if it's a W10 problem???
No errors are reported and everything looks normal, excepting the blank debug window.
@Cluso99 said:
@dgately
Thanks for trying with an FT232.
Seems we have found a limit with the CP2102
The FT232 is an overpriced chip, with the raw chip costing around 3x the cost of the complete CP2102 board posted. Guess you get what you pay for
And you have to be careful of the FT232 fakes out there too.
Chip has fixed the debug_baud setting problem with pnut35j. It will take a bit for it to get into PropTool - Jeff is aware of the problem.
Now, I've found with testing that currently the max reliable operation for the cheapie CP2102 boards is 921_600 baud. At 1_000_000 it drops characters. The CP2102 specs say 1Mbaud and the CP2102N specs say 3Mbaud - the CP2102 boards I've seen/have seem to be CP2102 not CP2102N chips.
Note I can download with loadp2 at 2_000_000 baud without apparent problems, and the initial setup with the DEBUG still occurs at 2Mbaud. So it may be a buffer overrun issue that locks the CP2102 ???
The datasheet claims 1M on the front page but shows 921600 as the max later on.
This explains the reliable operation from 921600 down, however the 2M download success is a mystery.
It seems to be able to Tx above spec but not Rx?
When using a CP2102, use pnut35k or later with the following setup in the CON section. The PUB section is an example only. Note the restriction of 921_600 baud for the CP2102.
Also restrict loadp2 to this value to ensure reliable operation even tho I've not seen problems with 2_000_000 baud.
CON
_clkfreq = 200_000_000 or whatever
DOWNLOAD_BAUD = 921_600
DEBUG_DELAY = 100
DEBUG_BAUD = 921_600
PUB go() | i
repeat i from 0 to 9
debug(udec(i))
Tested mine with TAQOZ. Seems fine using SD card. Comms via Prop plug with flying wires to Retro board. Whole plug and wire ends in epoxy (my version of conformal coating and strain relief)
Works under Fedora33, putty 921600...
Flexprop is able to compile the code, but I am not able to see any output on the serial terminal.
Pnut can compile but it doesn't show any terminal output.
Don't know what could be the issue. Not sure if this is related to CP2102 (which I am using) or if Chip's code is using any feature only available on Parallax demo board that is not available on Retroblade2.
Flexprop is able to compile the code, but I am not able to see any output on the serial terminal.
Pnut can compile but it doesn't show any terminal output.
Don't know what could be the issue. Not sure if this is related to CP2102 (which I am using) or if Chip's code is using any feature only available on Parallax demo board that is not available on Retroblade2.
Presume you’re running Windoze 10? I’ll try and check it out tomorrow. If I don’t get a chance, sorry it will be a week before I get another chance as I’ll be away for a funeral.
Change baud to 921600 for CP2102. (1.2M and 1.5M work too)
If nothing is connected to msr_pin there will be no terminal output.
I have gone back to 2000000 for loading P2 with loadP2. I am using 115200 for the serial. These work with loadp2 and the W10 command prompt and CP2102 dongle with my boards. I've been loading 70K programs.
I made a stupid mistake with the pin on DP83848 board (and there was no signal present). The code was actually working, but as it was not sending any string (as ozpropdev said, Thank you!) I though it was not running.
I am using W10, and It works perfectly (I tested at both 115_200 and 921_600). Yes, flexprop will need LF (otherwise it will overwrite the same line again and again).
The program shows between 49,996,660.00 MHz - 49,996,740.00 for the DP83848 (50MHz oscillator).
One happy customer here!
Has anybody got a neat idea of how to make use of pins 16-26 when the vga header is in place? I was thinking of a mini pcb that would bring out the pins in line with pin 0-15 for use with a breadboard.
Any suggestions?
@Cluso99 said:
@aaaaaaaargh
Nice find. Are those vga pins connected on the add on board? Perhaps you could have placed the vga connector on the add on board instead?
Nah, too late now as the picture shows . He's got the VGA soldered in already and the extension board as well so probably too much trouble undoing it. But getting another Retroblade2 could solve this inconvenience nicely.
Comments
I think I am using loadp2 to download at 2MBd. I will check in the morning if I get a chance.
I dumped the CP2102 and went to a FTDI-based adapter and now, DEBUG works well with PNut & Propeller Tool...
I run WIN 10 in Parallels on macOS...
On the macOS side, I got the checksum errors with the CP2102 and solved that with the FTDI adapter, as well!
dgately
Do you have a CP2102N you can try, and maybe a P2D2 with the UB3 ?
@dgately
Thanks for trying with an FT232.
Seems we have found a limit with the CP2102
The FT232 is an overpriced chip, with the raw chip costing around 3x the cost of the complete CP2102 board posted. Guess you get what you pay for
And you have to be careful of the FT232 fakes out there too.
I'm looking through my collection of stuff right now for other CP2102 parts.
I have lots of P2 boards but not a P2D2.
I have a Click board ("Uart 3") that has a CP2102N on board. Let me know if you want me to courier it so you can test with it
The CP2102N supply is a bit of a joke, Mouser show nearly 200,000 on order across the variants. I'm not sure whether Silabs are doing something right or wrong, its hard to tell.
FYI
I was able to get "debug_plot_hubram.spin2" to work with CP2102 by dropping the baud to 115200..
Anything higher fails.
DEBUG_BAUD = 115200
]
Brian,
Just been looking for the instructions to add this into the DEBUG_demo.spin2 to try. Do I add it into the CON section like this (doesn't do anything)
CON _clkfreq = 100_000_000 DEBUG_BAUD = 115200 PUB go() | i waitms(1000) repeat i from 0 to 9 debug(udec(i))That's correct Ray.
Thanks Brian.
Don't know why it didn't work the first time I tried it ;(
Curiously, sometimes I just get a blank debug screen ie no output.
Not sure if it's a W10 problem???
No errors are reported and everything looks normal, excepting the blank debug window.
I've posted a bug report for pnut35i here
https://forums.parallax.com/discussion/comment/1519876/#Comment_1519876
DEBUG_BAUD = xxxxx doesn't work reliably at any speed.
FWIW I routinely download using loadp2 at 2MBaud using my CP2102.
On a whim, I had purchased the FTDI adapters (possibly fake) at: https://banggood.com/3Pcs-FT232RL-FTDI-USB-To-TTL-Serial-Converter-Adapter-Module-p-959210.html?rmmds=myorder&cur_warehouse=USA
about $4 US for each adapter...
Chip has fixed the debug_baud setting problem with pnut35j. It will take a bit for it to get into PropTool - Jeff is aware of the problem.
Now, I've found with testing that currently the max reliable operation for the cheapie CP2102 boards is 921_600 baud. At 1_000_000 it drops characters. The CP2102 specs say 1Mbaud and the CP2102N specs say 3Mbaud - the CP2102 boards I've seen/have seem to be CP2102 not CP2102N chips.
Note I can download with loadp2 at 2_000_000 baud without apparent problems, and the initial setup with the DEBUG still occurs at 2Mbaud. So it may be a buffer overrun issue that locks the CP2102 ???
The datasheet claims 1M on the front page but shows 921600 as the max later on.
This explains the reliable operation from 921600 down, however the 2M download success is a mystery.
It seems to be able to Tx above spec but not Rx?
When using a CP2102, use pnut35k or later with the following setup in the CON section. The PUB section is an example only. Note the restriction of 921_600 baud for the CP2102.
Also restrict loadp2 to this value to ensure reliable operation even tho I've not seen problems with 2_000_000 baud.
CON _clkfreq = 200_000_000 or whatever DOWNLOAD_BAUD = 921_600 DEBUG_DELAY = 100 DEBUG_BAUD = 921_600 PUB go() | i repeat i from 0 to 9 debug(udec(i))Tested mine with TAQOZ. Seems fine using SD card. Comms via Prop plug with flying wires to Retro board. Whole plug and wire ends in epoxy (my version of conformal coating and strain relief)
Works under Fedora33, putty 921600...
Hi,
I am trying to compile the frequency counter demo ( https://forums.parallax.com/discussion/170882/reciprocal-counter-demo/ ).
Flexprop is able to compile the code, but I am not able to see any output on the serial terminal.
Pnut can compile but it doesn't show any terminal output.
Don't know what could be the issue. Not sure if this is related to CP2102 (which I am using) or if Chip's code is using any feature only available on Parallax demo board that is not available on Retroblade2.
Presume you’re running Windoze 10? I’ll try and check it out tomorrow. If I don’t get a chance, sorry it will be a week before I get another chance as I’ll be away for a funeral.
@Ramon
Change baud to 921600 for CP2102. (1.2M and 1.5M work too)
If nothing is connected to msr_pin there will be no terminal output.
@Ramon
Where is the code for testing?
I have gone back to 2000000 for loading P2 with loadP2. I am using 115200 for the serial. These work with loadp2 and the W10 command prompt and CP2102 dongle with my boards. I've been loading 70K programs.
@Ramon
You will also have to add a "LF" character to chips code for it to display properrly on flexprop terminal.
This was verified using flexprop and retrobales2+cp2102. (Windows 10)
I am sorry to hear about your loss.
I made a stupid mistake with the pin on DP83848 board (and there was no signal present). The code was actually working, but as it was not sending any string (as ozpropdev said, Thank you!) I though it was not running.
I am using W10, and It works perfectly (I tested at both 115_200 and 921_600). Yes, flexprop will need LF (otherwise it will overwrite the same line again and again).
The program shows between 49,996,660.00 MHz - 49,996,740.00 for the DP83848 (50MHz oscillator).
Have TAQOZ up on my board using SD. Easiest. But does run a bit to the hot side.
Sorry, I wrote the wrong URL this is the correct link (code in first post) -> https://forums.parallax.com/discussion/170882/reciprocal-counter-demo/p1
For flexprop: (otherwise, output will rewrite the current line)
For PNUT (to avoid "DEBUG requires at least 10MHz of Crystal" and be able to use DEBUG, CTRL+F10):
One happy customer here!
Has anybody got a neat idea of how to make use of pins 16-26 when the vga header is in place? I was thinking of a mini pcb that would bring out the pins in line with pin 0-15 for use with a breadboard.
Any suggestions?
Found a solution, nice to have stuff in dusty drawers... this might well be the most breadboard friendly P2 yet...
Retroblade2 + VGA breakout board = Retroblade2X
When not using the retroblades onboard vga connector P27-P31 can be connected as well.
@aaaaaaaargh
Nice find. Are those vga pins connected on the add on board? Perhaps you could have placed the vga connector on the add on board instead?
Nah, too late now as the picture shows
. He's got the VGA soldered in already and the extension board as well so probably too much trouble undoing it. But getting another Retroblade2 could solve this inconvenience nicely.