View previous topic :: View next topic |
Author |
Message |
ronaldus
Joined: 25 Dec 2010 Posts: 183 Location: france
|
Link Posted: Wed Apr 03, 2019 6:50 pm Post subject: |
|
|
Wow Nidi,
You have a much better bandwidth then mine. What kind of HDMI converter do you use?
Regards,
Ron.
|
|
Back to top |
|
|
nidi
Joined: 17 Aug 2008 Posts: 303 Location: Switzerland
|
Link Posted: Wed Apr 03, 2019 7:30 pm Post subject: Moome Card |
|
|
its the latest marquee moome card running straight out of the DAC bypassing filters contrast chip and output buffer, and the image is 1080P 72hz 195Mhz.
|
|
Back to top |
|
|
thewolfman
Joined: 28 Mar 2011 Posts: 1311 Location: Sweden
|
Link Posted: Thu Apr 04, 2019 6:15 pm Post subject: |
|
|
Post some more screenies Nidi.. looked great the single one you posted over at the competition site.
|
|
Back to top |
|
|
ronaldus
Joined: 25 Dec 2010 Posts: 183 Location: france
|
Link Posted: Wed Apr 10, 2019 7:33 pm Post subject: |
|
|
Hi All,
I recapped the LVPS and but I didn't see any further improvements. Anyway the LVPS will outlive the whole projector now I hope.
What I forgot to mention was that after Gjaki's upgrade I didn't see any blooming anymore. Now maybe I have to re-adjust the G2 and drive settings after his mods but I'll do that at the end.
I'm now working on the last part of the chain: the moome box.
I did a quick measurement of the rise time. See attached pictures.
Unfortunately I had only a 50 ohm termination so I'll redo the measurements tomorrow with the correct termination.
The measurements are done on a 500Mhz bandwidth Tektronics scope that was standing in one of our labs.
I took a VGA to 5 BNC breakout cable with the red channel connected directly to the scope with a 50 ohm termination.
In the pictures you have the test pattern I used, the scope settings and the rise time with 1Mohm termination and 50 ohm Termination.
With 50 Ohm the rise time is about 4.9ns.
Regards,
Ron.
Description: |
|
Filesize: |
153.88 KB |
Viewed: |
8721 Time(s) |
|
Description: |
|
Download |
Filename: |
red channel 50ohm.jpg |
Filesize: |
486.85 KB |
Downloaded: |
300 Time(s) |
Description: |
|
Download |
Filename: |
red schannel 1Mohm.jpg |
Filesize: |
478.48 KB |
Downloaded: |
290 Time(s) |
Description: |
|
Download |
Filename: |
scope settings.jpg |
Filesize: |
458.35 KB |
Downloaded: |
287 Time(s) |
|
|
Back to top |
|
|
gjaky
Joined: 05 Jun 2010 Posts: 2790 Location: Budapest, Hungary
|
Link Posted: Wed Apr 10, 2019 8:24 pm Post subject: |
|
|
I think you did the measurement right. Don't care about the 50 Ohm load, properly terminating at 75 Ohm wouldn't change the rise time.
4.9ns rise time is pretty bad by the way, that translates to ~70MHz bandwidth. also note that the rise times are propagates: so if you have a source (signal) with rise time tr1 and a system with self rise time tr2 then at the output the accumulated rise time will be tr3=sqrt(tr1^2+tr2^2).
One can calculate the signal/system bandwidth by knowing the rise time: BW[MHz]=0.35/tr[ns]
Note: All these applies to your your signal measurements, there tr3 was 4.9ns, and the 500MHz scope should have a self rise time of 0.75ns, doing the math backwards gives a 4.83ns theoretical signal rise time, not much difference, since the 500MHz bandwidth of the scope is much higher than the signal of interest, but generally if the measured signal's rise time gets close to the measuring system's own rise time this effect is more considerable. This is why it is advised to have your measuring system's own bandwidth 3-5 times more than the signal of interest.
Anyway, a well behaving stock VNB (like yours) usualy have self rise time around 3ns, my daughterboard has a self rise time about 1.5ns, that results in approx. 3.4ns system rise time. To properly exploit this you'd need a source with <2ns self rise time.
_________________ projectors in the past : NEC 6-9PG xtra, Electrohome Marquee 6-7500, NEC XG 1351 LC ( with super modified Electrohome VNB neckboard !!!)
current: VDC Marquee 9500LC
The MOD: VNB-DB, VIM-DB
|
|
Back to top |
|
|
ronaldus
Joined: 25 Dec 2010 Posts: 183 Location: france
|
Link Posted: Thu Apr 11, 2019 7:17 am Post subject: |
|
|
Hi everybody,
As promised the rise time with 75ohm just for completeness.
rise time of the system 4.5ns
with your formulas we get around 80Mhz bandwidth.
So now it's time to find the outputs of the DAC on the moome box and see if we can connect that into the projector.
Regards,
Ron.
Description: |
|
Filesize: |
278.5 KB |
Viewed: |
8678 Time(s) |
|
|
|
Back to top |
|
|
mp20748
Joined: 12 Sep 2006 Posts: 5681 Location: Maryland
TV/Projector: 9500LC Ultra / Super 02 and 03 VIM
|
Link Posted: Sat Apr 13, 2019 2:25 pm Post subject: |
|
|
ronaldus wrote: |
So now it's time to find the outputs of the DAC on the moome box and see if we can connect that into the projector.
Regards,
Ron. |
How did that turn out?
|
|
Back to top |
|
|
ronaldus
Joined: 25 Dec 2010 Posts: 183 Location: france
|
Link Posted: Sat Apr 13, 2019 3:53 pm Post subject: |
|
|
Hi Mike,
I'm working on it. i first want to draw the schematic from the PCB in order to see where to bypass since there is also the gamma and contrast circuit involved.
Regards,
Ron.
|
|
Back to top |
|
|
mp20748
Joined: 12 Sep 2006 Posts: 5681 Location: Maryland
TV/Projector: 9500LC Ultra / Super 02 and 03 VIM
|
Link Posted: Sun Apr 14, 2019 12:41 pm Post subject: |
|
|
ronaldus wrote: | Hi Mike,
I'm working on it. i first want to draw the schematic from the PCB in order to see where to bypass since there is also the gamma and contrast circuit involved.
Regards,
Ron. |
Good idea to pluck the circuit, where you'll be able to draw out a diagram first. Because the idea of bypassing the filters is not a simple one with some DAC's. The filtering varies from a simple low pass feritte bead to a more complex Video Reconstruction Filter networks that can involve multiples of feritte beads and capacitors. And with the more complex, there is no simple means to bypass. And when bypassing, you have to remove all of the caps and beads, because the "network" itself is tuned. And in most cases with this kind out network, there will be an over emphasis of the high frequencies. meaning the DAC also boost the upper range of its frequency window. When simple bypassing this using a jumper, you induce that boosted high frequency into the video chain. Which also means you're adding a bit of high frequency distortion, that some seem to like, because it tends to appear as a slightly sharper image, when its really a distortion. That's why it's important to know what type of output you're dealing with.
You may be able to see this on your scope.
On the other hand, you may create the problem of inducing very high frequencies in the video chain that are not commonly seen as sharpness, but still plays out as a noise (hash). In this case, filtering is very important, because the higher frequency noise though not easily seen does create a problem.
On the Moome Marquee cards, it requires a simple low-pass bead. This beads purpose is to remove those higher frequencies about the required video frequency band. It's really needed on the Marquee Moome cards, with the exception being when you isolate the power rails of the DAC, that is that a good DAC will always have a separate Analog +/- and Digital +/-. The isolation require that you either decouple the rails using beads, inductors, caps. Or provide an independent regulator for both Digital/Analog supply. And in this case, the ground is also isolated. This form of isolating is done to lower or prevent the high frequency processing from getting into the analog section of the DAC via sharing the same coupled power rails and ground.
I'm only familiar with the DAC's Moome uses in the Sony and Marquee, but have no idea what's going on the external you're using. But would suggest to you to see what type of filtering is being used, whereas if a simple circuit or more complex. If more complex you'll have to understand the added sharpness is most likely distortion because of high frequency boosting. If a simple solution, the same would apply, but it's not as bad with the more complex if you bypass, but it's still will induce noise into the chain.
Those filters serve a real purpose, though most designs could use a more defined approach to suppressing the noise.
|
|
Back to top |
|
|
ronaldus
Joined: 25 Dec 2010 Posts: 183 Location: france
|
Link Posted: Mon Apr 15, 2019 7:45 am Post subject: |
|
|
Hi mike,
I have puzzled a little bit during the weekend and this is the result:
Description: |
|
Download |
Filename: |
Screenshot 2019-04-15 at 09.38.50.png |
Filesize: |
351.97 KB |
Downloaded: |
289 Time(s) |
|
|
Back to top |
|
|
mp20748
Joined: 12 Sep 2006 Posts: 5681 Location: Maryland
TV/Projector: 9500LC Ultra / Super 02 and 03 VIM
|
Link Posted: Mon Apr 15, 2019 8:40 am Post subject: |
|
|
That's clearly a single feritte bead being used as a low pass filter. That is the least complicated filtering method and by simple jumping the three beads, you should get good results. But I would still scope it at some point to see if there is any overshoot on the video.
Post back your before and after SMPTE pattern images.
.
|
|
Back to top |
|
|
gjaky
Joined: 05 Jun 2010 Posts: 2790 Location: Budapest, Hungary
|
Link Posted: Mon Apr 15, 2019 10:02 am Post subject: |
|
|
mp20748 wrote: | That's clearly a single feritte bead being used as a low pass filter. That is the least complicated filtering method and by simple jumping the three beads, you should get good results. But I would still scope it at some point to see if there is any overshoot on the video.
Post back your before and after SMPTE pattern images.
|
But if it is really using the THS8200 DAC then, no high hopes, since that DAC has 3.5ns output rise time specs.
_________________ projectors in the past : NEC 6-9PG xtra, Electrohome Marquee 6-7500, NEC XG 1351 LC ( with super modified Electrohome VNB neckboard !!!)
current: VDC Marquee 9500LC
The MOD: VNB-DB, VIM-DB
|
|
Back to top |
|
|
mp20748
Joined: 12 Sep 2006 Posts: 5681 Location: Maryland
TV/Projector: 9500LC Ultra / Super 02 and 03 VIM
|
Link Posted: Mon Apr 15, 2019 11:09 am Post subject: |
|
|
gjaky wrote: | mp20748 wrote: | That's clearly a single feritte bead being used as a low pass filter. That is the least complicated filtering method and by simple jumping the three beads, you should get good results. But I would still scope it at some point to see if there is any overshoot on the video.
Post back your before and after SMPTE pattern images.
|
But if it is really using the THS8200 DAC then, no high hopes, since that DAC has 3.5ns output rise time specs. |
Good point. But his question was dealing with jumping the filters, which was my focus.
Are you sure of it using the THS8200, that could be a 1280X1024 at best DAC?
.
|
|
Back to top |
|
|
ronaldus
Joined: 25 Dec 2010 Posts: 183 Location: france
|
Link Posted: Mon Apr 15, 2019 7:43 pm Post subject: |
|
|
Hi all,
No I'm not sure about that, but its an 11 bit dac with 80 pins housing and I don't find so many dac's with so many pins. On top of that the output pins at least correspond with the PCB.
I tried to bypass the bead of the green channel with a short wire (4mm) across the bead but that didn't change anything on the screen at least. I couldn't check it on the scope today.
Now I feel ripped off paying 240$ for a so so dac.
regards,
ron.
|
|
Back to top |
|
|
ronaldus
Joined: 25 Dec 2010 Posts: 183 Location: france
|
Link Posted: Mon Apr 15, 2019 8:16 pm Post subject: |
|
|
I tried also bypassing the complete amplifier section. I took out the bread FB8 and R2 and connected from R65 straight to pin 2 of the VGA connector. No change in image. absolutly nothing not in brightness not in sharpness.
I guess i have to find another dac.
regards,
Ron.
|
|
Back to top |
|
|
ronaldus
Joined: 25 Dec 2010 Posts: 183 Location: france
|
Link Posted: Mon May 06, 2019 3:46 pm Post subject: |
|
|
Hi All,
I just placed an order yesterday for the latest Moome internal card. let's hope that that one works better.
In the mean time I put my source to 1080i and I get fully resolved horizontal resolution
Regards,
Ron.
|
|
Back to top |
|
|
nidi
Joined: 17 Aug 2008 Posts: 303 Location: Switzerland
|
Link Posted: Mon May 06, 2019 5:41 pm Post subject: |
|
|
you will be surprised how good those new Moome cards perform.
I had to buy a new one after MP unmodded my old one.
|
|
Back to top |
|
|
mp20748
Joined: 12 Sep 2006 Posts: 5681 Location: Maryland
TV/Projector: 9500LC Ultra / Super 02 and 03 VIM
|
Link Posted: Thu May 09, 2019 5:11 am Post subject: |
|
|
nidi wrote: | you will be surprised how good those new Moome cards perform.
I had to buy a new one after MP unmodded my old one. |
Any original Moome Marquee card you would have had would most likely had been already modified. Only the very early was the exception. And I'm sure today that would be the case with any cards he still sells. The improvement was the implementation of low pass filtering, that removed a greater parts of the frequencies above the usable video frequencies. This of course was well tested and verified based on using the right filters.
Now why would I undo my own work, unless you'll talking about me later separating or better isolating the Digital and Analog supply rails, that is also known as "decoupling" and was also what the manufacturer of the DAC had recommended. Now if being able to look at this using the right test procedures, the manufacturer intent was the right one.
Aside from that and from I what I'm told, any later cards would still have the Low Pass filters implemented. Oh yeah and we did get the Low Pass design verified by STE of VDC and another design engineer to confirm that a greater part of the non usable high frequencies generated inside the DAC was being removed after the filtering. The later decoupling version did not require the filters at all.
http://www.curtpalme.com/VIM-FULLHD.shtm
Description: |
|
Filesize: |
61.03 KB |
Viewed: |
8166 Time(s) |
|
|
|
Back to top |
|
|
HaydnG90
Joined: 22 May 2006 Posts: 1335
|
Link Posted: Thu May 09, 2019 12:00 pm Post subject: |
|
|
mp20748 wrote: |
This of course was well tested and verified based on using the right filters.
Now why would I undo my own work, unless you'll talking about me later separating or better isolating the Digital and Analog supply rails, that is also known as "decoupling" and was also what the manufacturer of the DAC had recommended. Now if being able to look at this using the right test procedures, the manufacturer intent was the right one.
Aside from that and from I what I'm told, any later cards would still have the Low Pass filters implemented. Oh yeah and we did get the Low Pass design verified by STE of VDC and another design engineer to confirm that a greater part of the non usable high frequencies generated inside the DAC was being removed after the filtering. The later decoupling version did not require the filters at all.
|
Mike are you referring to the 120ohm@100MHz filters? Were these your mods which Moome implemented starting with the v2 cards? Which versions have your 'decoupled' design? v3 only?
|
|
Back to top |
|
|
mp20748
Joined: 12 Sep 2006 Posts: 5681 Location: Maryland
TV/Projector: 9500LC Ultra / Super 02 and 03 VIM
|
Link Posted: Thu May 09, 2019 1:07 pm Post subject: |
|
|
HaydnG90 wrote: |
Mike are you referring to the 120ohm@100MHz filters? Were these your mods which Moome implemented starting with the v2 cards? |
Yes. and it worked well for both 1080P 60 and 72 hz.
Quote: | Which versions have your 'decoupled' design? v3 only? |
Well, this change is not a Moome version but somethi8ng I had done myself on some V2's and use now. It follows what the chip manufacturer recommended and what is also a design practice on most integrated and Mixed Signal (analog/digital) chip designs, where they use special decoupling to separate the Digital from the Analog supply rails for the chip.
When done right it greatly reduces the high frequency switching noise (pulses) from getting into the final analog stages of the chip. Where a greater part of the noise that is stripped using the filters really come from.
Not really needed or you may not notice any difference on most CRT setups, so you shouldn't really worry about looking into it. The cards themselves are very well designed and are amazing performers right out the box and as-is.
|
|
Back to top |
|
|
|
|