• AVR Freaks

Optically isolated USB with external transceiver ?

Author
zabocks
New Member
  • Total Posts : 9
  • Reward points : 0
  • Joined: 2006/09/26 08:19:55
  • Location: 0
  • Status: offline
2006/09/26 08:33:04 (permalink)
0

Optically isolated USB with external transceiver ?

[8|] Anyone with experience in implementing an actual external transceiver with an optical insulation barrier with the 18F2455/4550, as suggested on figure 17.2 of the Microchip doc ?
 
What is the throughput and any recommended (tested) components ?
For a 12Mbps USB, I'm contemplating a solution based on USB1T1105A or MAX3454 USB transceiver.
Are 6N137 optocouplers fast enough for the job ?
How about any of the new IL717 or TI ISO150 or AD ADUM1100/2400 non-optical coupleurs ?
 
I cannot isolate on the other side of the 18F : would have too many i/Os to isolate, not to mention the analog signals coming to the A/D... 
Thanks in advance guys
#1

12 Replies Related Threads

    zabocks
    New Member
    • Total Posts : 9
    • Reward points : 0
    • Joined: 2006/09/26 08:19:55
    • Location: 0
    • Status: offline
    RE: Optically isolated USB with external transceiver ? 2006/09/28 10:50:33 (permalink)
    0
    So, no one ever tried to implement an external USB transceiver with a 18F2455/4550, with or without optical couplers ??? 
    Or no one of you PIC specialists tried to play with real components in order to implement the FIG. 17-2 of the doc ?
    If any one out there, please help me with this medical device project -hence the insulation barrier required. 
    #2
    Pacer
    Super Member
    • Total Posts : 1171
    • Reward points : 0
    • Joined: 2004/12/01 09:29:20
    • Status: offline
    RE: Optically isolated USB with external transceiver ? 2006/09/28 11:32:29 (permalink)
    0
    It might be worth checking out the www.usb.org forum. Similar questions do get asked there occasionally, and a search may give you some guidelines. Otherwise maybe you should just look at the data sheets (PIC and opto-isolators) and try to draw you own conclusion. 
    #3
    zabocks
    New Member
    • Total Posts : 9
    • Reward points : 0
    • Joined: 2006/09/26 08:19:55
    • Location: 0
    • Status: offline
    RE: Optically isolated USB with external transceiver ? 2006/09/29 10:49:56 (permalink)
    0
    Since we are on Microchip blog, I was expecting somebody in the community or at least at Microchip to help with that, since  they made it possible with their chip design...
     
    Thanks for the advice, I will go and check on usb.org, but I can expect general answers  on USB there, not  practical ly related to microchip's implementation...
     
    So, may be some Microchip guru will read my thread over the week-end... and give me a guru's advice ???
    Thanks in advance anyone. 
    #4
    aksispd
    New Member
    • Total Posts : 1
    • Reward points : 0
    • Joined: 2006/10/27 12:14:05
    • Location: 0
    • Status: offline
    RE: Optically isolated USB with external transceiver ? 2006/10/27 12:21:48 (permalink)
    0
    Hi all!
     
    I have the same problem, developing medical device and need the isolation. The most I found is this:
    http://www.edn.com/article/CA6347257.html but not PIC specific.
    Any professional support? Any basic suggestion (external transceiver or internal hardware)?
     
    Thanks in advance.
     
    PS: Where did  you find the schematic? In 18F4550 datasheet?

    aksi
    #5
    Skippy
    New Member
    • Total Posts : 1
    • Reward points : 0
    • Joined: 2006/11/29 15:41:58
    • Location: 0
    • Status: offline
    RE: Optically isolated USB with external transceiver ? 2006/11/29 15:58:54 (permalink)
    0
      I recently went through the same design issue with an industrial control project. I'm not an expert, but I'll pass along what I learned. Everything I've found concludes that opto-isolating the USB bus lines is a perilous and a costling exercise. Frankly I haven't seen any attempt (circuit design) at doing it. There are 3 drawbacks. (1) The cost of opto's that support bit rates > 1Mbit are expensive. (2) The bus is bidirectional so you need 'back to back' optos on each line plus you need a 'direction' bit that controls them. There is no direction bit. Presumably a dumb USB slave could generate such a signal via gpio since it knows to always listen and only talk when the host requests, but it's not a gimmee (sp?). (3) The worst is the unmatched capacitance one will add to the differential signal wires. This cancause unequal skew on the D+, D- lines. The bit times are 83nsec's so the receiver might get 'confused'.
      WHat designers typically do is to place the USB transciever on the 'outer' portion of the isolation barrier then opto-isolate the interface from this chip back to the host micro. This is usually RS232 or SPI. Unfortunately a nice micro with a USB onboard has this drawback. Hope this helped and Good Luck!
       
    #6
    SugarCreek
    Senior Member
    • Total Posts : 180
    • Reward points : 0
    • Joined: 2004/09/10 16:40:56
    • Location: Indianapolis USA
    • Status: offline
    Re:Optically isolated USB with external transceiver ? 2011/06/03 09:28:43 (permalink)
    0
    (coughing and sneezing as I blow 4+ years of dust off this topic...)
     
    I came across application-notes at http://www.analog.com/library/analogdialogue/archives/43-06/USB_isolator.html and http://www.maxim-ic.com/app-notes/index.mvp/id/3891, but these both promote their own solutions with different paradigms than that illustrated in figure 17.2 of the PIC18F4550 specification.
     
    One "drawback" mentioned in one of the application notes, as well as in this thread, was that there needed to be a "direction bit."  However, is that not what the UOE\ signal provides?
     
    I have a potential application for which the external transceiver could be an attractive option.  I personally like the partitioning illustrated in figure 17.2, and I have to believe that the Microchip designers had some particular circuitry (or even support chip) in mind when they came up with that particular complement of signals.  Furthermore, as was pointed out, there are many nuances that one has to get right (such as impedances, rise-times, etc.) when crafting such circuitry from scratch, so application examples would be extremely helpful.
     
    In the four years following the OP's query, surely there must have been some developments in this area.  Is there a single-chip, cost-effective solution to implementing an external transceiver using the signals (VPO, VMO, UOE\, RCV, VP, VM) as described in the PIC18F4550 datasheet?  Or, in what other directions has the industry moved to support isolated/external USB transceivers, and how do newer PICs support those?
     
    External transceivers aren't just a matter of isolation, but also of cable-length.  I've heard of "USB extenders" that allow USB communication over long distances, perhaps over twisted pair, but I've never heard evaluations of their actual performance nor what electrical approach is used.  Anyone familiar with these?
     
    Thanks and regards!
    #7
    chinzei
    Super Member
    • Total Posts : 2250
    • Reward points : 0
    • Joined: 2003/11/07 12:39:02
    • Location: Tokyo, Japan
    • Status: offline
    Re:Optically isolated USB with external transceiver ? 2011/06/04 01:02:53 (permalink)
    0

    USB PIC, which still exposes an external transceiver interface as FS/LS USB device, is a rare USB MCU family on current industry scene.

    These external transceivers fit to PIC USB well.
    TI        TUSB1106
    Fairchild USB1T11A
    http://www.microchip.com/forums/m364302.aspx

    For USB MCUs without any external transceiver interface, ADUM4160 is a handy option.

    As of board layout, ADUM4160 should be placed near to the D+/D- pins of a USB MCU, to make D+/D- traces as short as possible. Noise margin of these  traces is not so good. For the external transceiver (and opt-/magnetic-coupler) option on PIC, you may layout them as usual digital lines.

    External transceivers aren't just a matter of isolation, but also of cable-length.

    You may have interest on these products

    - Fiber optic USB extender
    http://www.metrotek.com/usb.htm

    - USB over Ethernet
    http://www.eltima.com/products/usb-over-ethernet/

    - AnywhwerUSB
    http://www.digi.com/products/usb/anywhereusb.jsp

    - Bluetooth - USB dongle
    http://www.microchip.com/forums/tm.aspx?m=489233

    - Wireless USB (WUSB)
    http://www.usb.org/developers/wusb/

    Tsuneo
    #8
    SugarCreek
    Senior Member
    • Total Posts : 180
    • Reward points : 0
    • Joined: 2004/09/10 16:40:56
    • Location: Indianapolis USA
    • Status: offline
    Re:Optically isolated USB with external transceiver ? 2011/06/04 07:30:55 (permalink)
    0
    Domo arigato!
     
    I had done a search of the forum for "external transceiver" and found only this thread.  The other message discusses an "external USB transceiver", so apparently I had my search-criteria too narrow... pink
     
    The Fairchild part appears to be of limited availability anymore, while the TI part appears to be still active.  I'll research them both, to see if I can use them (or at least learn from them).
     
    Regarding "USB extenders", I should have been more specific.  I'm particularly interested in approaches that isolate/amplify the physical layer of D+/D-, transcoding their symbols into a physically more robust transmission of the USB protocol, without actually transcoding protocols.  Furthermore, since so many buildings are pre-wired with CAT-5/6 or other twisted-pair copper (often "dark", yet unused by Ethernet), I would like to focus on that medium initially.  (In short, perhaps I'm looking for something that does for USB what RS-422/485 did for RS-232.)
     
    The TI and Fairchild parts, along with the AD and Maxim devices, give me a start.  As Pacer said, it may eventually come down to crafting something out of standard parts, but I'd like to re-invent as little of the proverbial wheel as possible.
     
    #9
    Antipodean
    Super Member
    • Total Posts : 1724
    • Reward points : 0
    • Joined: 2008/12/09 10:19:08
    • Location: Didcot, United Kingdom
    • Status: offline
    Re:Optically isolated USB with external transceiver ? 2011/06/05 10:08:50 (permalink)
    0
    SugarCreek 

    As Pacer said, it may eventually come down to crafting something out of standard parts, but I'd like to re-invent as little of the proverbial wheel as possible.
     


    You may find it easier to have a dedicated USB PIC running from the USB power, and then use a manchester coded signal (or some other pulse modulatable signal format) through transformers to do the isolation from your main PIC.
     
    That way you would not need to deal with high frequency signals across your isolation barrier, and can use an isolation mechanism that can be shown to meet your medical isolation requirements by using components already specified for this use.
     
     

    Do not use my alias in your message body when replying, your message will disappear ...

    Alan
    #10
    mad4764
    Starting Member
    • Total Posts : 44
    • Reward points : 0
    • Joined: 2010/01/11 06:52:03
    • Location: 0
    • Status: offline
    Re:Optically isolated USB with external transceiver ? 2011/06/07 04:26:49 (permalink)
    0
    Difficulty I see I see using optocouplers is the best rise and fall rate is in the order of microseconds for signal propgation. USB wants nanoseconds.

    Out of the USB 1.00 specs, bear in mind it is much faster now!

    The rise and fall times must be between 4ns and 20ns, and matched to within 10% to minimize RFI emissions and signal skew

    Don't think you can get optocouplers to perform to that level. Or should I say why I am slowly progressing to a USB link from one that already uses optocouplers. For switching a device on or off is one thing so I do not believe it would work, use wireless intead.

    That was possibly Microchips thought from the electrically isolated concept.
    #11
    DarioG
    Allmächtig.
    • Total Posts : 54081
    • Reward points : 0
    • Joined: 2006/02/25 08:58:22
    • Location: Oesterreich
    • Status: offline
    Re:Optically isolated USB with external transceiver ? 2011/06/08 11:12:25 (permalink)
    0
    Just wanna mention AD's ADUMxxxx devices for isolating USB connection Smile

    GENOVA :D :D ! GODO
    #12
    chinzei
    Super Member
    • Total Posts : 2250
    • Reward points : 0
    • Joined: 2003/11/07 12:39:02
    • Location: Tokyo, Japan
    • Status: offline
    Re:Optically isolated USB with external transceiver ? 2011/06/09 09:38:57 (permalink)
    0

    The rise and fall times must be between 4ns and 20ns, and matched to within 10% to minimize RFI emissions and signal skew

    Rise/fall times aren't so much matter. The wave is shaped by Schmitt buffers after opto-/magnetic coupling. Rather, propagation delay of the coupler is the matter.
    Split SE0 from differential signaling, and isolate these signaling, respectively. In this way, signal skew on differential signaling doesn't depends on the coupling.

    Tsuneo
    #13
    Jump to:
    © 2019 APG vNext Commercial Version 4.5