Serial communication woes...

Wuerfel_21

Hello,

after a long wait, I've finally been able to assemble those PCBs I ordered (thread link).
Initially, reset-on-RTS did not work, but i could upload programs by pressing reset in the right moment. Since my soldering is pretty bad, i suspected a bad joint. This suspicion was true: the reset transistor's emitter wasn't actually connected to ground. After fixing this, it does reset when RTS is released (the program currently in EEPROM has video output). However, now it won't upload anything, even when I disable the automatic reset and manually press reset, as I did previously!

So far i've tried:
- Inverting RTS in FTPROG
- Changing max bus power from 90mA to 500mA in FTPROG
- Wiggling the cable
- Looking at the RX line on a scope - inconclusive (only have access to analog scopes)
- Checking conductivity from connector to propeller
- Checked for shorts to ground, 3v3 and 5v
- Messing with the delay setting in prop tool
- BST's upload/identify
- Connecting RX/TX together on the serial cable - loops back as expected

Any ideas?

Wuerfel_21

UPDATE: I got out a bunch of jumper wires and found out that serial apparently only works when the RTS output of the converter is left floating. I have no idea why.

UPDATE 2: I uploaded "Parallax Serial Terminal Demo v1.0" and got interesting results: When RTS is connected, i get an ð (0xF0) every second on the serial terminal. So the issue seems to be with the PC receiving data from the prop?

UPDATE 3: The actual status of RTS makes no difference: inverted, not inverted, on, off - same results

UPDATE 4: Using the jumper to disconnect reset-on-RTS from the reset line works, too. EDIT: As in serial rx/tx works. Uploading obvs. needs manually timed reset button press.

Phil Pilgrim (PhiPi)

Can you post your schematic?

Phil

Wuerfel_21

It was in the thread I linked, where I discussed the schematic / board layout, but I'll just post it here again for convenience.

DigitalBob

I'd have to see how the FT232 is actually wired in. Are you using a prop plug etc. What's with the 10K resistors on Rx,Tx. I use 150R when building my own plug that could be an issue. How about power 3.3v any volt regulators etc.

Wuerfel_21

I'm using the FTDI-232R-3V3 cable, which is described on page 11 of this document. All I/O on it is 3.3V, but it also provides 5V power to the board.

Phil Pilgrim (PhiPi)

That cable can provide 5V at 75 mA. The Propeller, under certain circumstances (i.e. depending upon how many cogs are running, output loading, etc.), can draw a lot more than that.

-Phil

DigitalBob

I'm a little confused with their cable what's with the blue and the black on 6.1 vs. 7.1. Maybe pin it in without the cable. The Prop plug uses pins 2 and 30 (pins on FT232RQ) for the Rx/Tx with 150R in series. Pin 32 RTS is not used instead pin 31 DTR is used for reset. Checkout the Prop plug schematic.

DigitalBob

Ok now I see page 10 4.1 (I skipped down from page 11)

Wuerfel_21

Hmm yes, i didn't check the power requirements there. However, I don't think that that is the root of my problem here, as the propeller itself runs fine with all cogs enabled at 80 Mhz. Only when the RTS_ENABLE jumper is set, serial communications become garbled. (But reset does work and the prop boots from eeprom as usual)

Phil Pilgrim (PhiPi)

Try adding a stiff pull-up (e.g. 1K) from the /RES pin to 3.3V. With /BOE grounded (check that solder connection), /RES is somewhat weakly pulled up internally, but it may still be subject to noise, especially if you're on the fringes of your power consumption vs. sourcing. And measure your 5V and 3.3V busses to make sure they're what you think they are.

-Phil

Wuerfel_21

I will try that next time i get some time to tinker on this.

Wuerfel_21

So I finally bothered to resolve this issue: A 2k resistor instead of the jumper seems to fix the problem. Now I've got to figure out how to turn that into a jumper-like thing that can be nicely installed (unlike the mess of wires i've got for testing)

Wuerfel_21

You know what they say about dire times? Keep that in mind as I introduce...

THE RESISTI-JUMPER
Only genuine with cold solder joints....



(no, it's not just a hacked up stacking header soldered to a random resistor. Where would you get that idea?)

Wuerfel_21

Also, as a bonus of sorts, here is a picture of the entire setup on my local hackerspace's workbench: