CurtPalme.com Home Theater Forum

[RogerH]

Moome,

I have a problem that has taken me several days to troubleshoot.

I am unable to get a 1080i output on my new EXT-FULLHD v2, received a little over a week ago. It has taken a few days of testing with multiple display units and ultimately an oscilloscope to narrow the problem down to the EXT-FULLHD v2.

More specifically, the output from the Moome is not properly interlaced when I give it a 1080i HDMI input, and the output looks and behaves like a 540p signal.

On a scope you can see that the horizontal sync pulses are in the same location after every vertical sync pulse, instead of being shifted by half the horizontal sync period on alternate vertical sweeps to give the proper interlace. At least that's what I think it should look like, unless I am missing something.

What should be a 1080i output looks almost exactly like a 480p output (except for the slightly different horizontal sync frequency), so I assume it is 540p. There is no interlace when displayed, just rock-steady lines with black lines between them. The image is also slightly blurred (above and beyond the coarse scan lines), possibly because the correct image information for the second field is being scanned on top of the first field instead of where it should be.

Over the last couple days I have suspected everything, but everything has been cleared except the EXT-FULLHD v2. The display units tested include a projector, a CRT multisync monitor, and finally a borrowed LCD TV with a VGA input to report what it thinks the Moome is outputting. 480p, 720p, and 1080p all report correctly on the LCD TV. But when the HDMI source (a Panasonic BMP-BD85K Blu-Ray player) is set to 1080i and run through the EXT-FULLHD, the TV falters for a moment and then displays a small image which it reports as 640 x 480, probably the closest thing it can do with a 540p input.

As a final check to rule out the Blu-Ray player, I confirmed that it is indeed putting out all the correct resolutions on the HDMI cable, using the HD TV to verify and report each resolution. The EXT-FULLHD v2 is definitely putting out something like 540p with a 1080i input.

I can successfully run both 720p and 1080p (with 1080p cropped about 4% horizontally and a bit smeary on green), but I would like to give 1080i a fair evaluation since some have recommended it as the best choice for this projector.

Since I observed that the picture on other resolutions does degrade (darken and lose red) on the EXT-FULLHD v2 near the limit of extreme right picture shift, I did reset the horizontal shift to zero just to check whether the interlace came back. It never did, so it is nothing related to hroizontal shift that I can see.

I am hoping that you have seen this before, and that there is a simple fix or maybe I have something set wrong, so I don't have to send it back.

Your help would be appreciated. I will email you directly as well so I can give you additional contact information.

Thanks.




,

[RogerH]

I wanted to acknowledge that Moome has contacted me a couple days ago via email and is currently looking at my oscilloscope plots compared with his own hardware.

I will update this post when the source of the problem or a fix for the missing interlace on 1080i is found. For now, 720p and 1080p are working fine (1080p is stunning, actually).

[AnalogRocks]

Thanks for the update.

[AnalogRocks]

Thanks for the update.

[RogerH]

I thought I would add another update regarding my missing interlace on 1080i, just in case anyone is following this before buying one.

Moome has contacted me again after reviewing my information and looking at his own units. He did not give any details, but he did indicate that he has found a problem and that he is waiting for a response from a chip vendor.

I still do not know if this problem is unique to my particular EXT-FullHD or if more units have it and it had not been caught yet, but Moome is on it.

[Jackal29a]

I don't know if the problem I had is related but in my case Moome had to change firmware to allow for a much bigger horizontal correction/shift. I told him that with the first unit whenever I tried to output 1080i50 I got something similar to what I got outputting a PAL DVD in a 1080i60 timing with my VP50. Maybe there is something wrong and affects also 1080i50, even more so as there is no official 540p50 resolution.
I'm very interested in what he finds out.
Regards.
Javier

[RogerH]

How recent is your converter? Mine was shipped in late July, so I was assuming it was up-to-date with the latest firmware at the time, but I just don't know...

[Jackal29a]

I received mine on the 17th of July and indeed it had the new firmware plus an RC.

[RogerH]

It was suggested that I document here in the forum the problem that I am having at 1080i with the Moome EXT-FullHD that I received in late July (2010). Anyone else that has one of these converters from this time period may want to check out its 1080i operation if it has never been looked at. I still have no idea if this is an isolated problem or something common to the current production run.

I would be interested in hearing from anyone that can confirm 1080i operation on other units from this period.

RogerH

The following is my detailed explanation of the problem sent to Moome on 8/10/2010:



Hi Moome,

Thank you for your reply. I will try to tell you what I know without making this too long, but I don't want to leave anything out. If you have any ideas or things that I could look at to help solve the problem, please ask (I'm an engineer and I have some limited test equipment here at home). As you may know I did post my original message to you in the Curt Palme forum too, but if we can resolve this through email I will post a resolution summary to the forum when we are done.

There should be 7 photos attached here. They may not show up in the correct order, but they are captioned with explanations.

I have taken 3 pictures of the EXT-FullHD driving my monitor screen at 480p, 1080p, and what should be 1080i. I also took 4 photos of oscilloscope traces of the 1080i H and V sync pulses coming out of the EXT-FullHD and my Faroudja NRS for comparison. I only have analog scopes at home, but the photos do show what I am trying to explain.

The equipment I have been testing includes the following:

1. NEC E21 CRT RGB monitor (good to over 100kHz)

2. NEC XG-852 projector

3. Panasonic Blu-Ray player DMP-BD85 for HDMI source

4. Faroudja NRS Scaler used in passthru mode

5. Sony 1080p LCD TV to confirm my HDMI source

6. Moome EXT-FullHD

Everything except the EXT-FullHD has been independently tested and found to be working correctly- all sources are correct and all display units show the same problems with the EXT-FullHD output on 1080i. Other output resolutions seem to work perfectly.

Before I get started with my information, I wanted to answer your question about H-shift. I have tried all positions, but I never get any interlace on 1080i through the EXT-FullHD. And the problem of brightness/color loss near the far right shift is not really an issue, although I do need to operate very close to that setting to minimize left image cutoff on my projector.

The output mode is still set up for RGBHV H(-) v(-) as shipped.

So here is a summary of what I have see on my unit:

EXT-FullHD- 1080i Interlace Problem

I have included 3 screen shots of the EXT-FullHD driving an NEC multisync monitor. The results are the same with my NEC XG-852 projector, so I did not bother to take pictures of that. All photo exposures were 1/6 of a second, so interlace fields (if there were any) would be captured.








Note that the 1080i image should look as smooth as the 1080p with the long camera exposure, but it looks more like the 480p screen shot. The 1080i image usually looks a little fuzzy (although you don't see it in the photo) because I believe that the interlace field-which contains slightly different video information from the first field-is being scanned directly over the first field. So when I say it looks like a 540p image, that is not entirely accurate, because it has a slight 30 Hz jitter that a true 540p signal would not have.

If you look at the scope photos (Fig 1 and Fig 2), you can see why the interlace field lines do not fill the gaps between the lines of the first field on the EXT-FullHD output. The scope in all cases is triggered on vertical sync.

Fig 1 shows the trailing edge of an arbitrary vertical sync pulse from the EXT-FullHD (which I called the first V sync) and the relationship of the horizontal sync pulses to it.



Fig 2 looks at the next vertical sync pulse after Fig 1. Compare Fig 1 to Fig 2, and you will see that the H sync pulses have the same time relationship to the V sync trailing edge, which I think is the source of the problem. It seems that the H sync pulses following every other vertical pulse should be displaced in time by 1/2 the horizontal sync period. This would displace the second video field up or down by the proper vertical distance to interlace properly with the first field.



For comparison I have included Fig 3 and Fig 4, which are traces of sync pulses coming from my Faroudja NRS scaler set up to transcode the same Blu-Ray signal source into 1080i. The EXT-FullHD is not part of this test. You can see from the photos that the H sync pulses of alternate fields are correctly offset by 1/2 of the horizontal sync period.





I have repeated the tests many times and many different ways, and they all confirm that the EXT-FullHD is not outputting a correctly interlaced field when given a known-good 1080i source.

If you have any additional tests you would like to have done, please ask. I want to get the 1080i working properly, but it does work at 720p and 1080p and I would have a hard time giving the unit up for any length of time. It really does work amazingly well at those settings.

[Ron W]

To me, I find it quite bizarre that the unit would function properly in the 720p and 1080p resolutions yet not with 1080i. I have a 540p/1080i Toshiba RPCRT with a Video EQ Pro in the chain and as far as picture quality is concerned, my Moome converter (purchased in June/2010) works beautifully and that is also comparing it to the latest rendition of the Fury. One would think if one feature of this sort wasn't working properly(1080i) either would anything else(1080p/720p). On the surface, I would have thought something in the video chain would be causing this issue but since you have done several troubleshooting tests to deal with that it would only seem logical to assume the problem is with the EXTV2-HD converter itself.

In order to solve this mystery, it looks like Moome has got quite the challenge ahead of him.

[Jackal29a]

Roger,

I imagine when you say 1080i you mean 1080i@60Hz, could you please test 1080i@50Hz? There are many free PAL DVDs available on the web, like the HCFR pattern diskn (http://www.homecinema-fr.com/colorimetre/release/DVD_V2.0.0.ISO) .
Cheers,
Javier

[kal]

Interlaced (i) signals behave differently than progressive (p) so it's possible that one may work correctly but no the other.
I have nothing else to add here other than the fact that I was the one that recommend that Roger document his findings here as it's often easier to get to the bottom of something if more than one set of eyes can see the results, and the issue is clearly defined. Disjointed emails are often not the best way to solve issues.

That being said, I have to say I *LOVE* when another Engineer documents something! Kudos to you Roger! This is documenting a problem and then some!

I'll send an email to Moome so that he can chime in here with an update.

Kal

[kal]

Interlaced (i) signals behave differently than progressive (p) so it's possible that one may work correctly but no the other.
I have nothing else to add here other than the fact that I was the one that recommend that Roger document his findings here as it's often easier to get to the bottom of something if more than one set of eyes can see the results, and the issue is clearly defined. Disjointed emails are often not the best way to solve issues.

That being said, I have to say I *LOVE* when another Engineer documents something! Kudos to you Roger! This is documenting a problem and then some!

I'll send an email to Moome so that he can chime in here with an update.

Kal

[RogerH]

Yesterday evening Moome came through with a modification that involves removing a chip resistor and adding a jumper wire. I made the modification this morning (it took about 15 minutes) and have been testing it today. It's an easy fix, but you do need a needle tip on your soldering iron.

It appears that I now have interlace on 1080. There is some slightly imperfect nesting of the interlace fields that I originally thought was my monitor, but it does the same on the XG-852. I don't see any flaw in the timing of the H-sync pulses with respect to the V-sync on my scope. The H-sync pulse train correctly starts with a 1/2h delay after the V-sync for every other field. The imperfect nesting could still be something I am doing, and I only have one HDMI source to test with (a blu-ray player).

Moome did seem to suggest that the cost of fixing the 1080i would be the inability to change sync polarity. My unit came to me set up with H(-) and V(-), but after the mod it is now H(-) and V(+) on all resolutions, so every resolution change on my projector came up as a new input. I'm not clear if this will be fixed later.

So the good news is that progress is being made, but I need to do more testing. I have a list of questions for Moome, but I'm just going to give him a short update tonight. I took photos of the modification, and one of us will document the fix in case anyone else has the same issue.

RogerH

[Jackal29a]

Roger,

Have you tested YUV or only RGB?

Javier

[RogerH]

***NOTE: There is a revised modification in a later post that retains configurable sync polarity***

As I mentioned previously, Moome supplied modification instructions for my EXT-FullHD v2 to enable it to provide a proper interlaced output signal with a 1080i input.

In case this was not an isolated problem with my particular unit, I thought I would post a couple pictures of the modification.

The repair took me about 15 minutes, but you are dealing with small components and feature sizes, so you will need a very pointed soldering tip and possibly a magnifier or eye loupe. Use an ESD mat or work on a metal surface to prevent static buildup or discharge. Remove any cat fur or balloons from the work area...

And yes, you might want to get Moome's blessing before you do this to avoid any warranty misunderstandings. He of course offered to do the repair himself, but shipping is always a last resort for me.

1) Open the converter case by removing the VGA connector nuts followed by the four philips screws on the bottom of the case. Remove the cover.

2) Locate chip resistor R4 (between a cylindrical capacitor and a pin header). Remove the jumper from the pin header to give your soldering iron a little extra room.



3) Avoid overheating the board in the next step. Remove R4 quickly using a blob of solder on the tip of your iron to envelope both pads at the same time. If you are fast and the resistor survives you can save it by tacking it to an unused solder pad on the right side of the board.

4) Prepare a length of single conductor wire (I used 30 gauge wire-wrap wire). You could probably use a piece of magnet wire if you don't mind scraping insulation off the ends.

5) Solder the wire from the left end of R86 to the left pad from which R4 was removed. Check with a magnifier to make sure you are not bridging something with an oversized solder ball.



6) Put the jumper back on the pin header and reassemble the case. You are done.

So far the fix seems to work pretty well as long as you don't need negative vertical sync polarity. I have another minor interlace issue that I suspect might be my own, so until I find out otherwise I think this is a pretty easy solution.

Hopefully Moome will chime in here if he has something to add (or subtract).

RogerH

[RogerH]

I thought I should update this post with what I consider to be the final solution that Moome supplied to me a couple weeks ago for the missing 1080i interlace. I made the modification to my EXT-FullHD v2 and have verified that all resolutions (including 1080i) now work and the sync polarities are again configurable as they should be. The first modification detailed in previous posts had forced vertical and horizontal sync pulses to be positive.

Do this at your own risk, and check with Moome first if you are concerned about warranty issues.

The basic steps are:

1) R4 is left in place (I had removed mine for the first mod so I had to put it back)

2) Lift pin 62 of U73. This is tough without some significant magnification, and the pad is probably going to detach from the board in the process. Just make sure there are no shorts to adjacent pins when you are done.

3) Add a jumper wire from the left pad of R86 to the top pad of R31.



This modification appears to fix my initial problems, and I assume that Moome will be incorporating the changes on subsequent units (but you may want to ask if considering purchasing one).

Any other minor issues that I thought I might still have (including imperfect nesting of the interlace fields and jitter at 1080p) have been confirmed to also be present using an HDFury 3, so the EXT-FullHD is cleared.

[Jackal29a]

BTW, Moome told me this problem only affects RGB output, with jumper 2 up (YUV) everything should be OK.

[Rolls-Royce]

I have one of the first batch of EXT-FULLHD V2s. Should this mod be done to it, or is it only applicable to later production units?

[RogerH]

I bought mine in July of 2010. I don't see any reason why this problem would have been isolated to only my unit, but you may want to ask Moome if yours is affected, or test the 1080i resolution yourself. I don't think he is monitoring this thread.

Now somewhere along the line my unit has lost the black boost and H-shift reset functions. Moome claims that this should be unrelated to the modifications needed to fix the 1080i problem, but I definitely wouldn't do the mod unless absolutely required. Personally, I'm getting a little tired of spending any more time troubleshooting this thing and I have no confidence that if I send it back that I won't just get another one with problems.

[Rolls-Royce]

Thanks for your response, Roger. If I may ask, what test pattern were you using to generate those scope displays, and which line (R, G, B, H, or V) did you have the scope on? I have a scope and would like to do the same check on my V2 that you did on yours.

[RogerH]

There is no need to look at R, G, or B, so no particular test pattern is required for using a scope. Any source material will do, but make sure the source is set to output 1080i into the EXT-FullHD.

I was using an old dual-trace analog scope, so you may have a better way to do this if you have something more suitable. My top trace was taken directly from the V output from the EXT-FullHD and fed to channel 1 of the scope, and the H output of the EXT-FullHD went to channel 2. Display both H and V traces simultaneously. I had to set the scope to trigger on the vertical trailing edge (channel 1). I then had to set the horizontal sweep time to display exactly two vertical sync pulses, and then I used the B-delay sweep to zoom in on one area of the traces at a time: first the trailing (right) edge of the first v-sync pulse, and then the second v-sync pulse. In each case you will be observing the timing of the H-sync pulses in relation to the v-sync trailing edge. Specifically, you are looking for the h-sync pulses to be offset by 1/2 of the horizontal period on alternate vertical sweeps.

You need to display two v-sync pulses and zoom in with B-delay to make sure that you are comparing two directly adjacent vertical pulses.

Lack of interlace should be very obvious on your analog display, however. If there is no interlace, a 1080i input to the EXT-FullHD will give an output that looks almost like 480p (since it was actually 540p). If interlace is working, the line spacing and fill (if you ignore field flicker) should look nearly the same as 1080p.