I have designed a circuit board to read 2 pressure sensors and send the data serially over USB. I'm using an FTDI FT232RL chip for USB to UART communications. I have built up 4 boards so far and all have worked as intended at first. But only at first. After several minutes they no longer show up in device manager as a COM port. Unplugging and re-plugging works briefly on a few of them, one refuses to be detected anymore. I have tried updating my FTDI drivers, restarting windows, etc. No luck.


The circuit is essentially an Arduino Pro Mini and a Sparkfun FTDI Breakout with two pressure sensors. I used those two breakout boards to verify the circuit and software worked before ordering PCBs. I have uploaded pictures of my schematics and board layout. I am not using the oscillator shown, instead I'm using the ATmega328PB's internal oscillator. Also, the MAX31865 and its accompanying resistor and capacitor have not been added to my boards yet.

Since I had all these boards working, at least briefly I know the circuit works, but I have no idea what would cause a COM port device to vanish. Since the com port comes back briefly when I re-plug a device I know the FTDI chips aren't broken (at least not completely).
Another weird thing is that when the device vanishes, the power LED turns off. I have verified with a multi-meter that the board has power when it is no longer detected and that there is 5V going to the non-LED side of R2.
please help. I am flummoxed.

 

*ugh* device manager/com port problems.

It's only a shot in the dark, but: During recognition (when you first plug it in), is the FTDI sending anything? I have personal experience seeing that situation causing problems. The COM port did not "vanish", but instead, Device Manager dropped the idea of the FTDI board and decided that a ballpoint pen mouse (or some such thing) was now there. Is anything showing up anywhere in device manager concurrently with the disappearance?

In any event, the easiest way to test it is to disconnect TXD on the FTDI during plug-in and until device manager has decided what it is. Then connect TXD. If the sensor proceeds to work, you know that is the problem. Again, it is only a shot in the dark.

 

Sounds like they are on the brink of the power requirement for that particular USB port.

If you meander over to "Device manager" and select the USB device in question, you can see what it is consuming.. If it's over 300mA the port may barf... Usually they just go off and a simple disable port / enable port brings it back to life..

Some PC ports have a real miserable power output and a hub is needed..

Your circuit certainly doesn't look like it consumes anything like that, but a short?? Well it's an easy test..

 

hi,
I have had the same problem as Raymond, it sees the port as a ball point device.
I have deleted the ball point driver in Devices.
E

 

Thanks for the replies! I tried all your suggestions, unfortunately to no avail. Perhaps its back to the drawing board on this one...

 

Maybe you don't have high enough voltage on Vcc. I have been burned on this in the past.

From the data sheet:
1. The minimum operating voltage VCC must be +4.0V (could use VBUS=+5V) when using the internal clock generator. Operation at +3.3V is possible using an external crystal oscillator.

 

To self power full speed.
You don't show reset hooked up it should have a 4.7k top and 10k bottom resistor divider. Hooked to pin 1 of usb and the 10k to ground.
the center to the reset pin.
And where you have the fuse bad Ideal that should be a ferrite bead.

Your more then likely having the FT show up as a non readable usb device.

You have to use one or the other you don't have anything hooked up to turn it on.

Or the PWREN if your letting the uC turn it on.
The Pwern 10 k to VCC or the reset pin with them floating the chip will just come on and turn off best way to say whats happenning here.

 

Check out one of these. A USB experimenter must have.

https://www.ebay.com/itm/USB-MultiF...916632&hash=item1a3dce9fe4:g:Za4AAOSwmudaOyUP

 

I have four of these there all made like this the PWREN is hooked to VCC of the USB with a 10 K

 

Next could help:
Make an used power overview.
Disconnect USB from PC
Remove the USB fuse from your board.
Connect an external 5V to the VCC and observe the current during a few minutes.
Then connect PC via USB and observe external current and PC USB behaviour.
This will give you information on power usage.
the next step will depend from results.
Remove the all power from board except to the ftdi and test usb could also exclude the trouble maker(s).
Let us know what the results are.
Good luck
Picbuster

 

Does the board have a ground plane? The PCB layout you show looks to be a 2 layer board with the GND removed, it could be that the boards susceptible to noise causing some latch up. Also an issue could be the location of the decoupling capacitors C4 & C6 which are doing very little in terms of why they are in the circuit, You want these located as close to the power legs of the FTDI device to minimise the current loop (this assumes you have a gnd plane.

I also notice Pin 3 of you micro is GND but not connected and Pin 6 is Vcc and also not connected. Is this mean't to be the case?

 

He has to turn it on In his sch he clearly shows it's not on one of the power on pins has to be used the
PWREN or the RESET with a voltage divider

Most use the PWREN I've made these on strip board and they work fine.
The RESET Makes the chip come on it high speed the chip talks in USB 2.0 with it.
Or use the PWREN pin as I posted there in the data sheet

The TS clearly shows in in this he has not done it as in the data sheet.

 

He has to turn it on In his sch he clearly shows it's not on one of the power on pins has to be used the
PWREN or the RESET with a voltage divider

Most use the PWREN I've made these on strip board and they work fine.
The RESET Makes the chip come on it high speed the chip talks in USB 2.0 with it.
Or use the PWREN pin as I posted there in the data sheet

The TS clearly shows in in this he has not done it as in the data sheet.
View attachment 152369

Got ya, good spot.

 

Yell the ts needs to take a good look at this