HDMI to VGA works on one Pi only


9 posts
by kg001 » Tue Dec 18, 2012 2:49 am
Guys this one has me stumped...

I bought a HDMI to VGA converter for my Grandson to go with his PI as a chritmas present. & decided to test it before giving it to him....

I initially tried it on my old 256Meg PI & it worked perfectly running to a newish ASUS LCD Monitor. Although it came up in very high resolution.. Was easily fixed by setting the HDMI Group to 1 & mode to 4 in config.txt.

I then changed over to a 512Meg Rpi (which normally performs incredibly well) & was greeted by a black screen. after a day spent trolling the forums & changing configurations I am no further advanced that when I started. Both Pi's are set up identically with a hard 5.2v direct wired PSU
Both are running the latest firmware & Raspian updates both work perfectly HDMI to HDMI without any issues. Running: tvservice -s reports
state: HPD high|DVI mode|HDCP off|composite off (0x120016), 1280x720 @ 60Hz, progressive
On both Pi's

I need to get my head arroud this before I give a lemon as a Xmas present.

Has any one any Ideason where to go from here!

Cheers & thanks in advance Nev B
Posts: 14
Joined: Wed Jul 25, 2012 1:06 pm
Location: Pt Augusta South Australia
by The Bishop » Tue Dec 18, 2012 8:27 am
I have two 512 meg RPi's and I can't get the HDMI/VGA to work on either of them.
Posts: 1
Joined: Tue Dec 18, 2012 8:16 am
by obcd » Tue Dec 18, 2012 9:44 am
How are those hdmi2vga converters powered? The older Pi's had a diode that provided power to the hdmi port. It could only handle 80 - 100mA I think it was removed on the newer Pi's. If the converter needs that voltage to work, it won't on the newer pi's. Self powered hdmi2vga converters are recommended.
Posts: 890
Joined: Sun Jul 29, 2012 9:06 pm
by mahjongg » Tue Dec 18, 2012 10:22 am
No the newer PI's have the exact same diode still, its rated at 200mA, and drawing more through it will just overheat it.
There are adapters so that you can add more power to a HDMI cable.
User avatar
Forum Moderator
Forum Moderator
Posts: 5752
Joined: Sun Mar 11, 2012 12:19 am
by kg001 » Tue Dec 18, 2012 11:49 am
Hi guys.

Thanks for the info I will check the voltage at D1 and see if there is a difference between the two pi's. They should be the same as both had 5.2v between p1 & p2 when they were set up & moved to their permanent position a few days ago. I will post back with what I find.
Incidently it won't be the poly fuse dropping volts as they are both powered with the polyfuses by-passed & a 1.3A trip cct in line instead.

Cheers Nev B
Posts: 14
Joined: Wed Jul 25, 2012 1:06 pm
Location: Pt Augusta South Australia
by -rst- » Tue Dec 18, 2012 12:52 pm
And your monitor only has the VGA connection (yes, I know there are some such)?

If you would happen to have DVI, HDMI-to-DVI would probably work better and would be just a 'dumb', relatively cheap cable...
http://raspberrycompote.blogspot.com/ - Low-level graphics and 'Coding Gold Dust'
Posts: 900
Joined: Thu Nov 01, 2012 12:12 pm
Location: Dublin, Ireland
by kg001 » Tue Dec 18, 2012 3:02 pm
I have just finished checking voltages arround D1 on both the 256m & 512m PI's the results have left me dumb-founded! With no HDMI load both units measure 5.19 & 5.17 v respectively on the input to D1.
and show a fairly normal forward voltage drop. of arround .4v. Under load it got quite interesting.
Remember that the setup in both cases was identical only the model of the Pi changed! The input voltages remained the same within a few milivolts But the output side of D1 was a very different story. The 256m Pi measured 4.29v on the output pin if D1 (about normal for this device fully loaded) but the 512m PI measured 1.24V. The next move was to grab a 1n914 diode out of my spares kit and bypass D1 by holding the Cathode of the 1n914 on pin3 ( D1 cathode 5v out to the hdmi socket) fed with 5v via a clip lead. then booted the PI and everything worked as advertised. My conclusion from this is that
1. The adaptor is drawing the full 200ma that the diode is rated for or a little more.
2. Anyone using one of these adaptors has probably just blown the warantee. (if they admit to it!) In My case the point is mute because the heatsink mods alone have killed that.
3. The BAT54 in the 512M Pi is almost certainly either from a bad batch or a different manufacturer. & has totally different characteristics. Note: My 512m Pi is out of China not the Sony factory in UK. Perhaps JamesH or one of the hardware guys should be involved at this point.
4. Interestingly the diodes do not seem to be blown as I would have expected as the HDMI socket still works normally with a HDMI monitor & they test normally with an Ohm Meter.

The answer in my case the answer will be to source a higher rated diode & replace or Bypass D1.
if my shaky old hands will let me.
Thanks Everyone for your input.
2 Questions:
The Bishop... Do your 2 512 PI's have a made in China sticker on them?
Majongg... Can you bing this to the appropriate person's attention please?

RST Yes hdmi to dvi would be better but my grandson who the adaptor was desined for has only VGA available.and in my case the plan was to run headless but the 512 PI performs so brilliantly that I see little point in firing up the 6 core phenomII power hungry beast to use as a dumb terminal while I am programming & the only easily swapped monitor is VGA only.

Cheers & Thanks all
Posts: 14
Joined: Wed Jul 25, 2012 1:06 pm
Location: Pt Augusta South Australia
by johnbeetem » Tue Dec 18, 2012 7:31 pm
kg001 wrote:... The 256m Pi measured 4.29v on the output pin if D1 (about normal for this device fully loaded) but the 512m PI measured 1.24V. The next move was to grab a 1n914 diode out of my spares kit and bypass D1 by holding the Cathode of the 1n914 on pin3 ( D1 cathode 5v out to the hdmi socket) fed with 5v via a clip lead. then booted the PI and everything worked as advertised. My conclusion from this is that
1. The adaptor is drawing the full 200ma that the diode is rated for or a little more.

It may be that your HDMI to VGA adapter draws significantly more than 200 mA and your 256 MB RasPi just happens to have a BAT54 that exceeds the spec, while the 512 MB RasPi is behaving normally. One way to make sure is to remove D1 and use an ammeter (set to measure at least 1A) to see how much current the adapter really draws. If it's more than 100 mA (highest recommended current) then replace the BAT54 with an NXP (or equivalent) PMEG2010AET as mahjongg suggested in a different thread. See the RasPi Wiki for more detail and links: http://elinux.org/RPi_Verified_Peripher ... rter_boxes
User avatar
Posts: 942
Joined: Mon Oct 17, 2011 11:18 pm
Location: The Coast
by mahjongg » Wed Dec 19, 2012 11:53 am
Be sure that the relevant persons at the RPF are aware of this issue, and I have even advised them to switch to a better diode, like the PMEG2010AET. If it will happen I don't know, there is also the "little problem" of an 1/3 increase of current consumption when you use such an adapter! In fact the makers of such adapters are consciously making something that shouldn't be allowed to be connected to the HDMI bus, which specs the maximum current to be drawn from this supply voltage at 50mA! In fact this "supply line" is only intended for the small EEPROM inside the monitor that contains the monitors characteristics, and those typically need only 10mA.

A much better solution would be a powered adapter, and only the cheapest ones don't have it! If you use one of these the solution is to buy a small and cheap ($3) adapter that combine a male and a female HDMI connector with a (barrel) 5V power input.

Why would the RPF redesign their hardware to support stuff that breaks the HDMI's commissions technical guidelines (rules)?

Lets start recommending the small adapter, (plus separate 5V power supply) as the preferred solution for people having bought an unpowered VGA converter already.

You can buy such an adapter here: http://dx.com/p/hdmi-male-to-hdmi-female-adapter-w-power-input-port-black-155361
User avatar
Forum Moderator
Forum Moderator
Posts: 5752
Joined: Sun Mar 11, 2012 12:19 am