New blog at dancharblog.wordpress.com


DisplayPort->HDMI dongles/adapters – active vs passive re-visited


Moved to https://dancharblog.wordpress.com/2014/10/07/displayport-hdmi-dongles-active-vs-passive-re-visited/

Comments (13)
  1. Danchar4 says:

    @ost Thanks – clarified now. Well I clarified as much as I could given how misleading the "active" and "passive" marketing words already are.

    Type-2 isn't dead, its just that if you're going to get the higher rated chips to run at 300MHz, you may as well just make it an active adapter that can run on ANY output port on your video card, not just the ones that support DualMode/Displayport++.  So basically you'll see adapters that are both type-2 compatible AND "active". I think all the new Accell brand adapters work this way. I should break one open to see which chipset they are using…

    I expect many "Active" adapters are using this chipset which are $3 in bulk:

    http://www.paradetech.com/products/displayport-format-converters/ps171-dp-to-hdmidvi/

    While the type-2 non-active chipsets are about $1.50 in bulk:

    http://www.eetimes.com/document.asp?doc_id=1315835

     

    So at retail you can expect a price difference of about $5-6 between a type-2 and an active adapter of same quality assuming similar overall part count on the board.

     

  2. Tim says:

    Thanks for clarifying between these two types of DP++ adapters.

    Some manufactures even didn't describe there product is Passive or Active at all which is pretty confusing.

  3. Jeroen says:

    Thanks for putting up this article. It's a good one!

    regards,

    Jeroen

  4. Juan Navarro says:

    Thank you very much for this detailed article. This has been of a great value, and together with this video from AMD [1], I got to better understand what is going on between the DP and HDMI/DVI devices.

    However I have a big problem (in the sense of completely undertanding this topic): my experience does not match the theory. I am working with an AMD FirePro W9000, and connecting FIVE screens with a passive adapter DP to DVI, and everything works without problems at 1920×1080@60Hz.

    This is the passive adapter cable that I'm using: StarTech DP2DVIMM6 [2].

    I have read tons of documents from AMD where they state that only two outputs at the same time can work as DualMode/Displayport++. If this was actually the case, then I should have 3 non-working screens, but that is not the case and I don understand why.

    Any thoughts or knowledge about this?

    Best regards

    [1]: http://www.youtube.com/watch

    [2]: http://www.startech.com/…/DisplayPort-to-DVI-Cable-Adapter-Converter~DP2DVIMM6

  5. Danchar4 says:

    Hi Juan, thanks for the comments. I don't know why your setup works. My guess is that the Startech adapter is actually "active" after all. Perhaps they sell two different models, one advertised as active for high cost and one advertised as passive for lower cost. But maybe they figured they only needed to design one adapter to fit both scenarios since the internal parts cost (BOM) isn't very different.

    Its also possible that Startech went with a subcontractor for these adapters and that subcontractor consolidated the specifications across several adapter brands to "active" at some point and Startech simply didn't update their documentation to match the change.

    Just be glad it works and don't mess with anything 🙂

    Cheers,

    Dan

  6. Sam says:

    Could you guys suggest any Windows or 3rd party tool allowing to check the adapter model and chipset version of any attached to a PC video card DP-to-HDMI or similar adapter?

  7. Danchar4 says:

    Sam that's a great question. I'm not aware of any tools that will tell you what chipset is inside an adapter – I don't think there is a protocol/specification for this sort of thing. You could reach out to the folks at vesa.org or displayport.org and see what they have to say though.

    There may be ways to deduce the chipset based on certain observed behaviors by setting different screen modes one after the other, but you'd have to test several different adapters to find the patterns.

    The only tool I know of that would come close is dispdiag.exe built into windows. This tool spits out a bunch of display topology information including all the enumerated monitor modes available per EDID/DisplayID protocol. You need a separate log parser tool available from the MS NDA developer program to read the binary log file. I highly doubt it provides any insight into the adapter though – my bet is that it will tell you whether an adapter is present or not and that's about it.

  8. Sam says:

    Hi Danchar4,

    Do you happen to know the exact name of that parser tool? Google is very quiet on this. Can you share a copy of it? 🙂

  9. Danchar4 says:

    Sam, you could try posting a question in the MSDN HW development forum:

    social.msdn.microsoft.com/…/home

    or HW testing/certification forum:

    social.msdn.microsoft.com/…/home

  10. Grant says:

    Hey, thanks for a really practical article. Do you think it would be possible to hardware test an active adapter to determine if it is still functioning correctly? – I'm thinking say a bus pirate or an arduino looking at voltage and clock level outputs – maybe not even plugged into a card?

    I've built several 5×1 eyefinity rigs and the active adapters seem to be a real weak point, my guess is they don't handle well the hot-plugging needed to get the screens to register and configure properly and eventually they just get fail. So each time we move our portable gear I make sure I've got spares on hand. And now I've got about a dozen name brand, certified adapters that are outside of warranty and possibly unreliable, unfortunately with all the auto-switching on the video cards, the drivers, catalyst and windows, its damn near impossible to determine by switching out on a video card – even with a known good adapter theres just too much automatic reconfiguring to isolate the effect of a single adapter. I'm mainly using HD7970 with 2xDVI and 4x full size DP sockets. Any wisdom to offer on this ? thanks in advance !

  11. Danchar4 says:

    Grant, I'm not aware of a device specifically designed to test these dongles or a simple way to build one. I know VESA.org / DisplayPort.org maintains a list of service providers that have expensive machines to help certify cables, video cards etc for signal integrity. VESA has a pretty in-depth list of specifications for the signal requirements.

    I might investigate this as a power supply issue first. The dongles run off the 3.3v coming out of the video card and if that's not clean power you can get a lot of issues on hotplug. Do you have a quality P/S in the host machine? Have you been able to reproduce the issue on alternate video cards or an alternate host machine, or different monitors?

    May I ask what size/type/resolution monitors you have? If low-res, you might try running a few monitors off a powered DP MST hub that has built-in HDMI/DVI conversion and see if that makes a difference. If perf/throughput isn't a big concern, you could try running a few monitors off a Displaylink-based solution as well.

  12. ost says:

    I have some more adapter experience to share..

    You can identify passive (at least type2 adapters) by reading "extended edid" at i2c address 0x80. Not sure what tools you can use for this. Maybe there is a api where you can use to that.. You can find "leaks" of the specification of these databits in datasheets, like here: http://www.ti.com/…/detailed_description (under 8.5.1 DP-HDMI Adaptor ID Buffer)

    Tbh, I doubt todays graphic drives cares to read this.

    So far I prefer passive adapters, cause the active has a difficult task of recovering hdmi clock from the displayport lane rate. Since the division factors can be anything, even dynamic, this becomes risky.

    But even the passives need reliable low jitter dual mode source. I suppose there is a risk that the graphic cards drivers forget to turn of the spread spectrum clock..

    Also worth mentioning that FPGA sinks are specially sensitive to hdmi jitter, as their gigabit tranceivers does not handle hdmi/tmds signaling well, particularly not when special patterns in the image cause long 11 or even 12 bit long static (0 or 1) data. Typical problem is when the first hor pixel is white (255) on std cvt-rb blanking (11 bit) or black if hsync is inverted (causing 12 iirc).Its a pattern to be aware of if you test against fpga sinks.

Comments are closed.

Skip to main content

New blog at dancharblog.wordpress.com