Prop's programming pins should have been made active low!

Rayman

I'm using the Prop with a FTDI chip, using their circuit to control power to the Prop (uses a p-channel fet).

But, I noticed that there was still 1.6-2.0 V on Vdd, even when computer was "sleeping" and USB devices are commanded to go to low power mode.

What I just found is that this voltage isn't through the p-channel fet, but through the Prop's programming pins, connected to RX,TX on the FTDI chip.

I believe people have also seen this kind of problem with the DEMO board...

But, the FTDI chip can be programmed to be active low.

I believe it would have been better if the Prop programming pins were active low.

This problem isn't limited to the FTDI chip, but any serial connection to the Prop would give this bleed through power...

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
My Prop Info&Apps: ·http://www.rayslogic.com/propeller/propeller.htm

Phil Pilgrim (PhiPi)

Rayman,

A couple things:

1. FTDI's "soft start" circuit is virtually worthless. First, there's a start up glitch on the PWREN pin that activates the pMOSFET before it's supposed to. Second, the pMOSFET circuit itself is poorly designed and doesn't really provide a good soft start. Here is a thread that discusses this issue. The schematic in this document (PDF) shows my solution, arrived at after many iterations.

2. The problem with current leaking in or out via the I/O lines can be solved easily with an open-drain 74LVC buffer. The same schematic cited above includes such a buffer. (You can ignore the transistors after the buffer. They were put there so the Stamp programmer could turn off the echo on the programming port.)

-Phil

Rayman

Hey Phil, I remember that discussion...

But, I find that the FTDI circuit works well enough for my application.· I don't really need "soft start", but I do want to keep the power off until the PC approves the >100mA current.· I'm using about 150-200 mA, and things seem fine.

I may use your buffer solution.· Right now, it's more of an annoyance than a real problem because I don't think Vdd is high enough to actually make anything work.· I currently have 1k series resistors on the RX and TX lines, so I was really surprised to find this voltage.· I think Vdd just bleeds up until the circuit limits it.· I think it's below the brown-out threshold.

Another solution might be to use the USB power enable signal to gate TX line too...

But, don't you think reversing the polarity of the RX, TX lines inside the Prop would have fixed this issue?


▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
My Prop Info&Apps: ·http://www.rayslogic.com/propeller/propeller.htm

Phil Pilgrim (PhiPi)

Rayman said...
But, don't you think reversing the polarity of the RX, TX lines inside the Prop would have fixed this issue?

Theoretically, it could work. But as a practical matter, it would introduce other issues. For example, those wanting to use a MAX3232 to shift to RS232 levels would have to use an extra inverter. Also, each FT232R chip would have to be reconfigured via MPROG for the reversed polarity — an extra manufacturing step.

-Phil

Rayman

Yes, I imagine ease with the MAX3232 was the primary consideration. But, I think they would have to use MPROG anyway to do things like fix the serial number. But, I'm sure in their volume that FTDI would do this for them...

I did just play around and find that a single PMOS on the TX line solves the problem somewhat. But, I still have a problem that if I'm using USB for programming and not as primary power, that this voltage shows up when primary power is removed...

So, I'll just have to tell people to remove the USB cable when not programming in this circumstance...

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
My Prop Info&Apps: ·http://www.rayslogic.com/propeller/propeller.htm