|
|
Post by radiohead314 on Feb 15, 2016 12:02:30 GMT -5
Hi I am back with a new name. Most of you know me as Redfox314 but I got tired of that and deleted my account. When I heared that another member was gone too I made a new account so her we go. I am first going to post a few pictiures of multiburst testpatterns and how they show the bandwidth of the Barco video chain. I am using a 909 but there are other Barco models having only different neckboards. And someon has even modified those neckboards on his 1209. My switcher is modified with a 5 GHz transistor on port 3 and an opamp (the 5166 in the place of the hfa1100). My RGB driver is modified with an 5166 in the place of the hfa1100. No further modifications. This is the multiburst patern www.w6rz.net/phase0.pngIt has 4 vertical blocks that go from 4 pixel on 4 pixel off to 1 pixel on 1 pixel off. The last part generally is losing amplitude in the chain. Measuring the output of the Moome the 1 pixel part is compensated for later loss or peaked.   Entering the RGB driver. Not much changed.  Untitled by Radio Head, on Flickr  [/url] Untitled by Radio Head, on Flickr Entering the RGB output amplifier   |
|
|
|
Post by radiohead314 on Feb 15, 2016 12:02:58 GMT -5
So what I see is that the RGB driver kills a lot of bandwidth. I also made pictures of the different peaking setting on the driver. Peaking low frequencies watch the two bars to the left  Peaking mid frequencies watch the third bar  Peaking high frequencies perhaps a little overshoot of the third bar?  |
|
|
|
Post by radiohead314 on Feb 15, 2016 12:03:10 GMT -5
I think peaking off or peaking high frequencies is the best but in fact peaking high should peak higher frequencies. Prhaps a mod is useful here? This is how it looks in the tube with peaking off. Not so bad? It sort of confirms the burst picture before the neckboards. So are those neckboards really that bad?  |
|
|
|
Post by radiohead314 on Feb 15, 2016 12:04:18 GMT -5
The probe does have an effect on the bandwidth. So the burst are probable better with a better probe. The probe is a parasitic capacitance of 15pF. Probe away  Probe there  |
|
|
|
Post by radiohead314 on Feb 15, 2016 12:19:10 GMT -5
Now next thing will be to measure just before the tube but those voltages are much higher.
|
|
|
|
Post by radiohead314 on Feb 15, 2016 14:56:21 GMT -5
First I am going to make some calculations. Barco uses a LC series combination to peak on the RGB driver.
The lowest peaking has C=12p L=3u3 that give resonant fres=25.3 MHz
The middle has C=8p2 L=1u5 that gives fres=45.4 MHz
The high has C=2p2 L=0u47 that gives fres=157 MHz
Now my 1080p@60 has a pixel clock of 148.5 MHz so that give a frequency of 74 MHz
Now it seems to me the high frequency peaking is much to high
|
|
|
|
Post by radiohead314 on Feb 16, 2016 5:32:10 GMT -5
Ok I would like to measure the burst signal on M4. It comes from the hybrid amplifier that is fed with 190V. Black is close to 190V I think. So if I set low contrast I would have a signal going between 190V and 100V or so. If I use a voltage devider network with 1 MOhm resistors getting it down to 20 V or so than would that be possible using a 15pF probe? M9 seems to be a grounding point.  |
|
|
|
Post by radiohead314 on Feb 16, 2016 7:23:48 GMT -5
I remeasured m204 before the hybrid and something went very wrong. The peaking I saw earlier is gone. I do not know what happened but now I get the same response I also got from the driver and that makes more sense. Not much changed in the picture so it was wrong measurement or something. Also the peaking on the driver does not work like I figured have to experiment a little more.  |
|
|
|
Post by radiohead314 on Feb 19, 2016 8:11:48 GMT -5
Conclusions so far are 1-my scope is not up to task and a new Tektronix old CRT scope is on its way to me  2-I measure a drop of bandwidth on the RGB driver I want to confirm that 3-It is difficult to measure after the hybrid but from my measurements nothing confirms that the drop is only there Someone else measured with a better scope and concludes that the drop is on the hybrid alone. Hopefully Gjaky gets to some measurements too so we know for sure what is happening. |
|
|
|
Post by Casethecorvetteman on Feb 19, 2016 14:19:23 GMT -5
Greg Eisemann did all this 15 years ago. Its known where the bottle necks are.
|
|
|
|
Post by radiohead314 on Feb 20, 2016 5:48:29 GMT -5
I have seen the Eissemann modifications and he does nothing on the neckboards. In fact I measured and they improve almost nothing because they do nothing about the bottlenecks that are probable the ad835 and the hybrid on the neckboard. Both of them are still there on Gregs modifications. The opamps he changes do very little on objective bandwidth improvement. He might however improve the picture making noise levels lower I did not look into that yet but that makes sense. Also he tells stories about using better coax that make no sense at all from an objective viewpoint. No bandwidth drop is expected on a small coax wire compared to the ad835 or the hybrid. We do these measurements also because we want to verify the stories coming from many sources. The first result I achieved is to improve on the peaking on the rgb driver. First thing was to find out what it does and second thing if it can be improved upon. Yes a very little bandwidth is won there. But solving this problem is not easy and I do get why Eisseman did not solve it too. We might not be able to. The bigest improvement you can get on the Barco is to use the moome instead of the hdfury. It seems to peak but the end result is cleaner and sharper. You can easily verify looking at the smpte bandwidth testpicture (you know that one  ). Now you can see some real improvement |
|
|
|
Post by radiohead314 on Feb 20, 2016 6:23:16 GMT -5
Here you see a good scoped measurement of a hdfury and a moome using multiburst. Just for information. Coming from Craigr. Fury at 1080p@60  Moome at 1080p@72 !!!  |
|
|
|
Post by Casethecorvetteman on Feb 20, 2016 13:52:27 GMT -5
Depends which HD Fury youre talking about and which MOOME. If its the external MOOME or HD Fury 3, theyre both garbage. The HDF4 and internal Barco MOOME give about the same image, and hardly much different to RGB-HV
|
|
|
|
Post by radiohead314 on Feb 20, 2016 15:46:25 GMT -5
|
|
). Now you can see some real improvement 
