P2-ES psu woes

cheezuscheezus Posts: 178
edited 2019-03-19 - 07:47:02 in Propeller 2
I've finally come to the point where I need to use an external brick to power my project. This isn't a surprise given the large 5v current (500ma) as well as the 3v3 @150ma. I've got all pin jumpers set to VIO so to use the switching reg. I'm using pins 0-24 so these groups will probably need to stay @ VIO. I haven't moved these jumpers yet since I haven't found a need, haven't had a chance to mess with VGA or anything..

I've been testing things with ~40ma on the 5v rail, not sure the 3v3 draw. But after connecting the larger 5v draw, the P2 no longer responds. I haven't done much troubleshooting yet since I've been surprised things worked so far on the desktop's USB current alone. Like I said I have a 5v brick (phone charger) I plan on using as "supplemental" supply but I'm wondering if there's an edge-case that needs to be prevented?

I'm thinking the power up sequence should be
PC-usb -> P2-usb

and power down is
P2-usb -> PC-usb

but I'm not really sure? Sorry if this has been answered before. Thanks in advance!

Comments

  • The sequence you describe is the safest.

    Here's sharing some background info...

    It shouldn't matter, but there 'might' be an edge case that occurs when the P2-ES has a heavy load attached during the time that the USB port is negotiating. (when plugging in the PC-USB cable and waiting for the PC to signal that 500mA can be consumed). If the negotiation takes too long, there is a body diode in one of the USB switches that might be passing too much current for too long.

    It's speculation, and we've designed it out just to be safe on the next Eval board. But it wouldn't hurt to follow your sequence just in case.

    Another possible reason for failure of those switches (as was experienced by a couple of forumistas) would be an unexpected short between 5V and VIO on one of the accessory headers. Those pins have also been protected for the next Eval board. Ensuring those two pins don't get shorted on the revA Eval is certainly something to take care of.

  • VonSzarvas wrote: »
    The sequence you describe is the safest.

    Here's sharing some background info...

    It shouldn't matter, but there 'might' be an edge case that occurs when the P2-ES has a heavy load attached during the time that the USB port is negotiating. (when plugging in the PC-USB cable and waiting for the PC to signal that 500mA can be consumed). If the negotiation takes too long, there is a body diode in one of the USB switches that might be passing too much current for too long.

    It's speculation, and we've designed it out just to be safe on the next Eval board. But it wouldn't hurt to follow your sequence just in case.

    Another possible reason for failure of those switches (as was experienced by a couple of forumistas) would be an unexpected short between 5V and VIO on one of the accessory headers. Those pins have also been protected for the next Eval board. Ensuring those two pins don't get shorted on the revA Eval is certainly something to take care of.

    Thanks for confirming that. In hindsight I should have designed some solder jumpers that would allow me to isolate different sections. As it is, I have one big power net with multiple possible sources. Great learning experience for the next few board layouts though :smiley:
  • Best the other way round. Get power on the board before bringing up the connection. Otherwise the PC's USB management will automatically attempt to power everything from there. And likewise in reverse, pull the USB comms before down-powering the board.

    I've gone with a mains powered USB hub which has handy per port on-off switches. Eliminates all the replugging.
    "There's no huge amount of massive material
    hidden in the rings that we can't see,
    the rings are almost pure ice."
  • evanh wrote: »
    Best the other way round. Get power on the board before bringing up the connection. Otherwise the PC's USB management will automatically attempt to power everything from there. And likewise in reverse, pull the USB comms before down-powering the board.

    I've gone with a mains powered USB hub which has handy per port on-off switches. Eliminates all the replugging.

    Evanh is right.

    That is what I thought I read in your first post, but somehow must have been standing on my head this morning.

    Connect the aux power first (P2-USB). Then hook up the PC-USB connection.
    Disconnect in reverse.
  • Okay, makes sense now. I saw some people having issues with the usb switches and knowing my luck...

    I'm still trying to figure out my "tangle" of power. I really wish I had a way to isolate the 5v (and 3v3) from the P2es to my pcb but that would require using lots of jumper wires. Hindsight is 20/20 though and I still don't have a GOOD reason to do another board run.



    glad I didn't hit post before I tried plugging things in



    I'm not sure when, or how but I think I damaged the AUX usb switch U501. Not real sure what happened although I'm pretty sure it's not the high current draw or 5v / vio short. Symptom is when powered from AUX, usb AUX active led lit but no power to 5v common. I'll have to get my meter out later and double check.. So I guess this post is kinda pointless now! My bad...

    I have to figure out what the BEST solution is for now, and for permanent. I'm considering powering from 5v common on 2x6 header, from tbd psu.. But to do this I'd need to either not use the PC usb port, reprogram ftdi to ?float? cbus3 or pull U502? I have a couple prop plugs floating around so that seems like an easy solution to programming without using the onboard ftdi. Main issue now I think is finding the right 5v supply.
  • evanhevanh Posts: 6,882
    edited 2019-03-18 - 12:26:45
    Damn. I'd say U501 is a gonna. I wrote mine off just as easily. No idea how it happened.

    I've desoldered U501 on my board and bridged across pads from AUX to 5 Volt Common. Has worked well. Be sure to double check soldering is correct as it's a small working area with plenty of nearby surface-mount contacts.

    The down side is it permanently disables the other USB switch.

    An alternative approach is bridge across F401 near PC-USB connector. This will boost the strength of PC-USB supply a lot if the cable is rated for it. Certainly a lot easier to do than removing U501.

    PS: F401 is called a PTC resettable fuse but it's really a thermistor like device. The operating resistance is terrible, so bridging it out makes the supply from PC-USB much stronger.
    "There's no huge amount of massive material
    hidden in the rings that we can't see,
    the rings are almost pure ice."
  • It may be that the power-up surge with a low impedance supply is simply too much for those switches. When bridging out F401 and using a strong cable, there is the possibility of taking out U502 the same way as U501. I'll be interested to know the answer if you choose this path.
    "There's no huge amount of massive material
    hidden in the rings that we can't see,
    the rings are almost pure ice."
  • cheezuscheezus Posts: 178
    edited 2019-03-18 - 14:02:27
    I'm looking at the datasheet and I'm wondering if this has something to do with it:

    " Discharge Function The discharge function of the device is active when enable is disabled or de-asserted. The discharge function with the N-MOS power switch implementation is activated and offers a resistive discharge path for the external storage capacitor. This is designed for discharging any residue of the output voltage when either no external output resistance or load resistance is present at the output. "

    fb_usb_sw.png

    https://www.diodes.com/assets/Datasheets/AP22802.pdf

    To me this presents a direct path from 5v_common to vss through the switch. I'm guessing the discharge function is only activated on a falling edge condition. Seems the edge case here would be "falling edge on 5v_usb_aux, while powered from 5v usb" ?

    *added block diagram
    1299 x 624 - 28K
  • Okay, if that's the case then it'll be a short to Vss on the 5 Volt Common now I presume. Attempting to use PC-USB should result in U502 current/thermal limiting. Except F401 will droop too much and cause an under voltage oscillation instead.

    "There's no huge amount of massive material
    hidden in the rings that we can't see,
    the rings are almost pure ice."
  • evanh wrote: »
    Okay, if that's the case then it'll be a short to Vss on the 5 Volt Common now I presume.

    It's what I would expect but doesn't seem to be the case. Still need to do some probing to confirm. PC usb switch seems to be okay, and P2 usb switch seems to have failed open.

    Right now I'm thinking about bypassing 5v_aux to 5v_common as you have, except I'm going to try a jumper from the unsoldered 2 pin connector to a 5v on 2x6 headers. Looking at the store, the powerpal looks nice right now. Give me an excuse to finally order that WX :smile:
  • cheezus wrote: »
    ... I'm going to try a jumper from the unsoldered 2 pin connector to a 5v ...
    That's what I was using when I destroyed U501 ... and still am using. It won't give you power without the bypass. However, an accessory header could be used without the bypass. I hadn't thought of this back then.

    Note: When soldering the Vss pin of that 2-pin AUX header it takes a lot of heat before it'll wick right through.

    "There's no huge amount of massive material
    hidden in the rings that we can't see,
    the rings are almost pure ice."
  • cheezuscheezus Posts: 178
    edited 2019-03-18 - 14:40:25
    evanh wrote: »
    cheezus wrote: »
    ... I'm going to try a jumper from the unsoldered 2 pin connector to a 5v ...
    That's what I was using when I destroyed U501 ... and still am using. It won't give you power without the bypass. However, an accessory header could be used without the bypass. I hadn't thought of this back then.

    Note: When soldering the Vss pin of that 2-pin AUX header it takes a lot of heat before it'll wick right through.

    I really wish there were test points for some of the nets. Getting to 5v_USB and USB_PWREN# doesn't look very easy. I guess the bypass is going to be best for now, time to change the tip in my iron. Can't do much of anything until I find a 5v supply that I actually trust and although I don't THINK it had anything to do with the dumb phone brick... Kinda stuck deciding how overkill my 5v supply really needs to be.

    #whyengineersdrink

    *edit

    I'm trying to think this through fully. You COULD power through an accessory header, but if you don't have the bypass (either method should work) then won't there be contention when powered from 5v_usb? The jumper avoids the "permanent" disabling of 5v_usb, but still turns U502 off.

    I'm trying to figure out how to SAFELY test this. Willing to sacrifice U502 "for the good of the cause" if I can figure out exactly HOW to isolate and replicate this error. I THINK..

    1- plug usb-pc, allow enumerate
    2- apply power to AUX, turns off U502
    3- U502 turning off enters discharge mode, low resistance path 5v_common to ground
    4-???
    5- U502 should be blown!
  • Oh, good point. I may have tossed the idea out, in another form, early on too. Can't remember now.
    "There's no huge amount of massive material
    hidden in the rings that we can't see,
    the rings are almost pure ice."
  • I know I had it powered from PC-USB after removing U501 without the bypass.
    "There's no huge amount of massive material
    hidden in the rings that we can't see,
    the rings are almost pure ice."
  • I've done some testing and confirmed U501 faulty. Enable is only @ 2v8 but is above the 1v5 high logic level. I've soldered the AUX header in place, boy that took a ton of heat! Bypassing 5v_aux to 5v_common works and my display is no longer complaining about 4v8 it WAS getting. I do find it interesting that if I remove the bypass jumper when only powering from PC something seems to latch into the disabled state. VAUX LED is also dimly lit with jumper and after jumper removed. If usb is enumerated without the jumper everything is fine...

    I'm considering removing U501 since it's known faulty to see if latch up condition continues. Either U501 is "leaking" OUT -> IN or U503 is "leaking" somewhere.

    Just thought I'd update since I have some actual answers. Now to find a nice 5v regulated supply :D
  • I'm fairly confident I've solved this issue for the most part. The PowerPal an is almost PERFECT fix, save for the part of the added breadboard. It would really be nice if the next rev of the ES could accommodate this, considering I *expletive deleted* hate using usb cords as power cords.

    again, my (probably less than) 2c. you know, inflation and all...
Sign In or Register to comment.