| Author |
Message |
redfox001
Joined: 16 Mar 2009
Posts: 2257
Location: The Netherlands
|
| Posted: Sun Feb 14, 2016 8:57 pm Post subject: |
|
|
The input resistance of my probe is 10M Ohm. Also I think that with 150MHz pixel clock 75MHz bandwidth scope should be enough.
Oh wait you did not use the probe?
My probe is 100MHz should be enough to. But some rolloff is expected.
_________________
701s->runco933->8500ultra->cinemax->9500mp->919 splitpack + cinemax
Last edited by redfox001 on Sun Feb 14, 2016 9:10 pm; edited 1 time in total
|
|
| Back to top |
|
 |
the dude
Joined: 11 Sep 2013
Posts: 179
|
| Posted: Sun Feb 14, 2016 9:10 pm Post subject: |
|
|
|
Exactly. I just plugged the 5 BNC cable into the scope. The scope can be terminated internaly for RF measurements, then the input impedance is 50 Ohm.
|
|
| Back to top |
|
|
redfox001
Joined: 16 Mar 2009
Posts: 2257
Location: The Netherlands
|
| Posted: Sun Feb 14, 2016 9:11 pm Post subject: |
|
|
| the dude wrote: | | Exactly. I just plugged the 5 BNC cable into the scope. The scope can be terminated internaly for RF measurements, then the input impedance is 50 Ohm. |
Ok that does minimise the capacitance? Is it better than a 100MHz probe?
_________________
701s->runco933->8500ultra->cinemax->9500mp->919 splitpack + cinemax
|
|
| Back to top |
|
|
redfox001
Joined: 16 Mar 2009
Posts: 2257
Location: The Netherlands
|
| Posted: Sun Feb 14, 2016 9:17 pm Post subject: |
|
|
How long are your bnc cables? Longer than 3m? 100MHz is about 3m.
_________________
701s->runco933->8500ultra->cinemax->9500mp->919 splitpack + cinemax
|
|
| Back to top |
|
|
the dude
Joined: 11 Sep 2013
Posts: 179
|
| Posted: Sun Feb 14, 2016 9:19 pm Post subject: |
|
|
|
It removes reflections. It is the most direct and precise way to view a signal, so yes- its better than a probe. But for video we would need 75 Ohm. This type of connection is not suitable for measurements in a existing circuit because it add a high load.
|
|
| Back to top |
|
|
the dude
Joined: 11 Sep 2013
Posts: 179
|
| Posted: Sun Feb 14, 2016 9:20 pm Post subject: |
|
|
|
A few cm. This is the Fury BNC adapter, ist just a few cm of cable between the VGA connector and the BNC.
|
|
| Back to top |
|
|
redfox001
Joined: 16 Mar 2009
Posts: 2257
Location: The Netherlands
|
| Posted: Sun Feb 14, 2016 9:27 pm Post subject: |
|
|
Ok but setting 20MOhm and 15pF parallel to the video with probe compensation is not totally useless I think. How does the multiburst look like with the probe?
_________________
701s->runco933->8500ultra->cinemax->9500mp->919 splitpack + cinemax
|
|
| Back to top |
|
|
the dude
Joined: 11 Sep 2013
Posts: 179
|
| Posted: Sun Feb 14, 2016 9:43 pm Post subject: |
|
|
|
It would look like the unterminated signal. The termination is not required because of reflections like in a real video cable. It is required because of load conditions. The driver is designed to drive into a 75 Ohm impedance cable. The interesting fact is that the frequency response without load is different.
|
|
| Back to top |
|
|
redfox001
Joined: 16 Mar 2009
Posts: 2257
Location: The Netherlands
|
| Posted: Sun Feb 14, 2016 9:50 pm Post subject: |
|
|
| the dude wrote: | | It would look like the unterminated signal. The termination is not required because of reflections like in a real video cable. It is required because of load conditions. The driver is designed to drive into a 75 Ohm impedance cable. The interesting fact is that the frequency response without load is different. |
Yes ok but if you connect the fury to the bnc of the barco than the termination is 75 Ohm. The probe parallel should give a reasonable measurement or not?
_________________
701s->runco933->8500ultra->cinemax->9500mp->919 splitpack + cinemax
|
|
| Back to top |
|
|
the dude
Joined: 11 Sep 2013
Posts: 179
|
| Posted: Sun Feb 14, 2016 10:16 pm Post subject: |
|
|
|
Yes, in a already terminated circuit the probe is Ok as long as you don't want to make the actual waveform visible (remember the picture from the switcher output, the pixel signal looks terrible). I will try to make some measurements at the switcher and driver output the next days.
|
|
| Back to top |
|
|
redfox001
Joined: 16 Mar 2009
Posts: 2257
Location: The Netherlands
|
| Posted: Sun Feb 14, 2016 10:44 pm Post subject: |
|
|
| the dude wrote: | | Yes, in a already terminated circuit the probe is Ok as long as you don't want to make the actual waveform visible (remember the picture from the switcher output, the pixel signal looks terrible). I will try to make some measurements at the switcher and driver output the next days. |
Yes I am going to measure all boards I have to see how linear they are. We really discoverd a new measuring method here to better understand what we see.
_________________
701s->runco933->8500ultra->cinemax->9500mp->919 splitpack + cinemax
|
|
| Back to top |
|
|
redfox001
Joined: 16 Mar 2009
Posts: 2257
Location: The Netherlands
|
| Posted: Mon Feb 15, 2016 8:46 am Post subject: |
|
|
| the dude wrote: | Some interesting tests on the Fury 2...
The fury was connected directly to the scope input via the BNC adapter, the scope was terminated internally with a 50 Ohm load. The hsync cable was connected to chanel 2 and used as the trigger.
At 48p the 1:1 pattern is almost the same amplitude as the first pattern. I watch movies at 48p when using the HTPC, so this is fine!
At 60p the amplitude drops a litte.
Now the fun beginns... 50 Ohm is not video impedance. The scope has only a internal 50 Ohm terminator. When this termination is turned off something interesting happens: The amplitude of the highest frequency increases. This indicates some kind of compensation maybe. It may be possible that at 75 Ohm the results look a little bit better at 60p. Something to be investigated.
Two things of the CIR test are unknown: What type of scope was used? What bandwidth is it? Maybe less than 300 Mhz? And what termination was used? |
Reading this again and reading what Gjaky wrote I am thinking the folowing. Gjaky wrote to do a direct measurement with coax and use proper termination 75 Ohm in most places. But my internal impedance is 20 MOhm so I think I should get a T-connector and use a 75 Ohm closing connector or a resistor? Does your scope have a switch for that? You could go external too?
_________________
701s->runco933->8500ultra->cinemax->9500mp->919 splitpack + cinemax
|
|
| Back to top |
|
|
redfox001
Joined: 16 Mar 2009
Posts: 2257
Location: The Netherlands
|
| Posted: Mon Feb 15, 2016 8:57 am Post subject: |
|
|
In your case using the not terminated input is probable most right. The termination might introduce a paracitic capacitance? I do not think reflections are a problem at this short cables.
_________________
701s->runco933->8500ultra->cinemax->9500mp->919 splitpack + cinemax
|
|
| Back to top |
|
|
redfox001
Joined: 16 Mar 2009
Posts: 2257
Location: The Netherlands
|
| Posted: Mon Feb 15, 2016 2:35 pm Post subject: |
|
|
Moome
Entering the RGB driver
Entering the RGB output amplifier
_________________
701s->runco933->8500ultra->cinemax->9500mp->919 splitpack + cinemax
Last edited by redfox001 on Fri Feb 19, 2016 1:05 pm; edited 1 time in total
|
|
| Back to top |
|
|
redfox001
Joined: 16 Mar 2009
Posts: 2257
Location: The Netherlands
|
| Posted: Mon Feb 15, 2016 2:37 pm Post subject: |
|
|
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
Peaking mid frequencies
Peaking high frequencies
_________________
701s->runco933->8500ultra->cinemax->9500mp->919 splitpack + cinemax
|
|
| Back to top |
|
|
redfox001
Joined: 16 Mar 2009
Posts: 2257
Location: The Netherlands
|
| Posted: Mon Feb 15, 2016 2:39 pm Post subject: |
|
|
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?
_________________
701s->runco933->8500ultra->cinemax->9500mp->919 splitpack + cinemax
|
|
| Back to top |
|
|
redfox001
Joined: 16 Mar 2009
Posts: 2257
Location: The Netherlands
|
| Posted: Mon Feb 15, 2016 4:09 pm Post subject: |
|
|
The probe does have an effect on the bandwidth. So the burst are probable better with a better probe.
Probe away
Probe there
_________________
701s->runco933->8500ultra->cinemax->9500mp->919 splitpack + cinemax
|
|
| Back to top |
|
|
the dude
Joined: 11 Sep 2013
Posts: 179
|
| Posted: Mon Feb 15, 2016 7:46 pm Post subject: |
|
|
Nice test!
The 1:1 field is quite dark in my setup. It must be the neckboards, at the rgb driver output the 1:1 amplitude of the multiburst pattern is full range at 60p (Yes, really!).
The probe was proper compensated of course. And i trust my scope, the amplitude stays the same with the probe connected to the internal calibrator when changing timebase.
Even peaking does not affect the intensity of the 1:1 field on the screen. Since the peaking circuit affects the output stage of the driver the weak component must be on the neckboard. Guess its the hybrid. Seems to be much better than the 1209s video pack, but far away from beeing a true 180 Mhz driver. Maybe we should play with the peaking network 
Next step would be to measure the input of the hybrid.
|
|
| Back to top |
|
|
redfox001
Joined: 16 Mar 2009
Posts: 2257
Location: The Netherlands
|
| Posted: Mon Feb 15, 2016 8:02 pm Post subject: |
|
|
I made 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
_________________
701s->runco933->8500ultra->cinemax->9500mp->919 splitpack + cinemax
|
|
| Back to top |
|
|
redfox001
Joined: 16 Mar 2009
Posts: 2257
Location: The Netherlands
|
| Posted: Mon Feb 15, 2016 8:28 pm Post subject: |
|
|
If we change the middle to 3pF the frequency is 75MHz. I will try that if I did not make an error.
The problem is of cause that it is an optimisation for 1 resolution but I think that is ok.
_________________
701s->runco933->8500ultra->cinemax->9500mp->919 splitpack + cinemax
Last edited by redfox001 on Mon Feb 15, 2016 8:33 pm; edited 1 time in total
|
|
| Back to top |
|
|
|
|
