| Author |
Message |
redfox001
Joined: 16 Mar 2009
Posts: 2257
Location: The Netherlands
|
| Posted: Wed Feb 24, 2016 6:05 pm Post subject: |
|
|
Yes they where caled foolproof diodes but fools can buy new parts except for those hybrids 
_________________
701s->runco933->8500ultra->cinemax->9500mp->919 splitpack + cinemax
|
|
| Back to top |
|
 |
redfox001
Joined: 16 Mar 2009
Posts: 2257
Location: The Netherlands
|
| Posted: Wed Feb 24, 2016 6:07 pm Post subject: |
|
|
| the dude wrote: | | Remember that 4 pf is not too bad for this. This is a low impedance circuit with low voltage. The neckboard video amp output is high voltage, so the same capacity will require much more energy to be charged to that voltage. |
Yes but I see three that makes it 12 pf and that is the same as my probe and I do see the effect of the probe!
_________________
701s->runco933->8500ultra->cinemax->9500mp->919 splitpack + cinemax
|
|
| Back to top |
|
|
redfox001
Joined: 16 Mar 2009
Posts: 2257
Location: The Netherlands
|
| Posted: Wed Feb 24, 2016 6:56 pm Post subject: |
|
|
Perhaps you did not measure a drop on the driver at 1080p@48 might be just below the limits? I measured at 1080p@60
_________________
701s->runco933->8500ultra->cinemax->9500mp->919 splitpack + cinemax
|
|
| Back to top |
|
|
mp20748
Joined: 12 Sep 2006
Posts: 5689
Location: Maryland
TV/Projector: 9500LC Ultra / Super 02 and 03 VIM
|
| Posted: Wed Feb 24, 2016 7:18 pm Post subject: |
|
|
| the dude wrote: | This time Mike is absolutly right  ... |
Why that is so nice of you to acknowledge that. You really should have said that about everything else, but I understand.
Anyways, you still never answered my question. I'd still like for you to explain to me since you say I'm wrong about it... How is it that the entire industry is using the word bandwidth, and using it where it does not relate to -3db at all?
The bandwidth calculators do not use -3db, neither do any of the pattern generators. Nor any of the very complex and sophisticated Blending and other Video Processors. I really want you to answer why when it comes to anything OUTSIDE of the video chain itself, It appears that -3db is not the RULE (standard). I'm sure I missed something so help me out...
Maybe -3db is just a little too elementary for me, considering how long I I've been using it over the years to include that I've rarely used it recently. And not only do I have a FET Probe, I have about 12 scope probes. Three scopes (one digital) had four. Also have two pattern generators (had three). And along with all of this, I have various and all kinds of RIGs to connect and impedance match it all. Not sure what I may be missing, but that's just a small part of what I've been using to MASTER the video chains I've IMPROVED already..
So can you please answer my question on why that word "bandwidth" is so widely used when you say it should only be used for -3db
Btw, I'll throw you a bone here:
The Barco 909 Hybrid chip does not have a part number on it, because it was a proprietary design. Barco had it designed for the 909. It was believed that the original design goal had the Sanyo 180mhz module in mind, but the 180mhz modules available may have been an issue.
Me, I would have stuck with the original 1209 neck board for modification, but that's just me..
|
|
| Back to top |
|
|
redfox001
Joined: 16 Mar 2009
Posts: 2257
Location: The Netherlands
|
| Posted: Wed Feb 24, 2016 7:54 pm Post subject: |
|
|
Sorry multilanguage. Pictures here for 1209 mod very simple
http://www.curtpalme.com/forum/viewtopic.php?p=390438
_________________
701s->runco933->8500ultra->cinemax->9500mp->919 splitpack + cinemax
Last edited by redfox001 on Wed Feb 24, 2016 8:00 pm; edited 1 time in total
|
|
| Back to top |
|
|
El Duderino
Joined: 23 Jan 2011
Posts: 4653
Location: Portland, OR
|
| Posted: Wed Feb 24, 2016 7:57 pm Post subject: |
|
|
| mp20748 wrote: | | If my video chain is -3db 200mhz, what is the DISPLAYED bandwidth? |
A chain is only as strong as its weakest link, eh? Note that -3db isn't a brick wall or cliff face either. You may still see some 1:1 grill bars on something like a SMPTE pattern, even without having full BW. The grill bars should ideally have precise edges and transition fully from full black to full white. If you're short on some BW, the edges won't be as sharp or the excursion will be from light gray to dark gray, but you may still be able to resolve them.
Consider the photos of the o-scope here. It's a 300Mhz rated scope, but it shows a Vellemen probe. That probe likely isn't a 300Mhz probe, so neither is the scope when that probe is attached to it. The pig tail leads are also much too long for a good HF measurement, with way too much loop area. The passive probe must also be compensated to the scope front end, to even have a hope of achieving accurate HF measurements.
There is also the capacitance of the probe. If you guys are concerned about a few pF of a clamp diode, when you put a passive scope probe on the circuit you're likely introducing much more capacitance than the clamp diode is.
Making good, accurate HF measurements isn't just about using capable equipment, it's also needs good technique and isn't trivial. When looking for things like peaking or rolloff, if you're not using good practices and technique, then you'll be hard pressed to know if what you think you're measuring is from the circuit under test or skewed measurements.
Last edited by El Duderino on Wed Feb 24, 2016 8:02 pm; edited 1 time in total
|
|
| Back to top |
|
|
redfox001
Joined: 16 Mar 2009
Posts: 2257
Location: The Netherlands
|
| Posted: Wed Feb 24, 2016 8:01 pm Post subject: |
|
|
Probe is 250 mhz
Probe is compensated
Leads are short
Better would be active probe.
But do not expect rectangular signals.
1:1 should be a sine with a perfect scope 1 sample
2:2 should be a wave with two bumbs 2 samples one harmonic
3:3 should be a little more bumbs 3 samplesor two harmonics
With 2:2 there are two samples no way the dac can make it a rectangle. A rectangle needs lots of harmonics so lot higher sample rate.
Considering this the measurements are not so bad.
_________________
701s->runco933->8500ultra->cinemax->9500mp->919 splitpack + cinemax
|
|
| Back to top |
|
|
El Duderino
Joined: 23 Jan 2011
Posts: 4653
Location: Portland, OR
|
| Posted: Wed Feb 24, 2016 8:09 pm Post subject: |
|
|
| redfox001 wrote: | Probe is 250 mhz
Probe is compensated
Leads are short |
'Short' is relative. IMO, those 1-2" pig tails are way too long, with way too much loop area for a good and accurate HF measurement.
|
|
| Back to top |
|
|
the dude
Joined: 11 Sep 2013
Posts: 179
|
| Posted: Wed Feb 24, 2016 8:14 pm Post subject: |
|
|
| Quote: | | The bandwidth calculators do not use -3db, neither do any of the pattern generators. Nor any of the very complex and sophisticated Blending and other Video Processors. I really want you to answer why when it comes to anything OUTSIDE of the video chain itself, It appears that -3db is not the RULE (standard). I'm sure I missed something so help me out... |
A pattern generator just generates a pattern. If you use DVI out there is no degeneration of amplitude at the generator side because that are just binary numbers encoded on the TMS data lanes. And for your analog out the same is true as long as the output amp is sufficient to handle the bandwidth (for a instrument like that it is for sure). Your generator can generate a picture with 300 Mhz pixel clock for instance. On the analog out a 1:1 pattern would be the same amplitude as a 2:2 pattern on the generator side. Your generator is build for +0 db at max. clock. Maybe thats what be manual trys to say (it could handle a 360 Mhz pixel clock at the analog out at -3 db, but 300 Mhz is the max. clock, so output will be for sure +0db).
So you can use the generator to find the frequency where the amplitude is just -3 at the end of the video chain. Thats up to you. If you find out that at 260 Mhz pixel clock for instance the amplitude of a 1:1 pattern at the end of your video chain is -3 db, then you can say your video chain has a analog bandwidth of 130 Mhz.
And of sure a video processor has no -3db point. In the digital domain there is just the sample rate.
|
|
| Back to top |
|
|
redfox001
Joined: 16 Mar 2009
Posts: 2257
Location: The Netherlands
|
| Posted: Wed Feb 24, 2016 8:15 pm Post subject: |
|
|
Ok thanks. Shorter leads makes sense. I do pick up a lot of noise esspcially on the lower side of the driver. But the 12pF mainly leads to a roll of that is why I only look how two measurements with the same 12pF look relational to eachother.
But those probe leads that is a very good point.
_________________
701s->runco933->8500ultra->cinemax->9500mp->919 splitpack + cinemax
|
|
| Back to top |
|
|
redfox001
Joined: 16 Mar 2009
Posts: 2257
Location: The Netherlands
|
| Posted: Wed Feb 24, 2016 8:18 pm Post subject: |
|
|
Good point about the -3db. If one uses the term bandwidth it should be able to have a -3db point. A pixel clock is not bandwidth.
_________________
701s->runco933->8500ultra->cinemax->9500mp->919 splitpack + cinemax
|
|
| Back to top |
|
|
mp20748
Joined: 12 Sep 2006
Posts: 5689
Location: Maryland
TV/Projector: 9500LC Ultra / Super 02 and 03 VIM
|
| Posted: Wed Feb 24, 2016 8:26 pm Post subject: |
|
|
| El Duderino wrote: | | mp20748 wrote: | | If my video chain is -3db 200mhz, what is the DISPLAYED bandwidth? |
A chain is only as strong as its weakest link, eh? Note that -3db isn't a brick wall or cliff face either. You may still see some 1:1 grill bars on something like a SMPTE pattern, even without having full BW. The grill bars should ideally have precise edges and transition fully from full black to full white. If you're short on some BW, the edges won't be as sharp or the excursion will be from light gray to dark gray, but you may still be able to resolve them.
Consider the photos of the o-scope here. It's a 300Mhz rated scope, but it shows a Vellemen probe. That probe likely isn't a 300Mhz probe, so neither is the scope when that probe is attached to it. The pig tail leads are also much too long for a good HF measurement, with way too much loop area. The passive probe must also be compensated to the scope front end, to even have a hope of achieving accurate HF measurements.
There is also the capacitance of the probe. If you guys are concerned about a few pF of a clamp diode, when you put a passive scope probe on the circuit you're likely introducing much more capacitance than the clamp diode is.
Making good, accurate HF measurements isn't just about using capable equipment, it's also needs good technique and isn't trivial. When looking for things like peaking or rolloff, if you're not using good practices and technique, then you'll be hard pressed to know if what you think you're measuring is from the circuit under test or skewed measurements. |
This is the kind of response I've been looking for. It explains some of the problems with measuring -3db and its accuracy. It also mentioned the second and more common standard and the importance of properly understanding it.
.
Last edited by mp20748 on Wed Feb 24, 2016 8:31 pm; edited 1 time in total
|
|
| Back to top |
|
|
redfox001
Joined: 16 Mar 2009
Posts: 2257
Location: The Netherlands
|
| Posted: Wed Feb 24, 2016 8:29 pm Post subject: |
|
|
It is also not a good idea to judge bandwidth at 1080p@42 You could use 720p@42 and conclude there is not problem in the whole chain
_________________
701s->runco933->8500ultra->cinemax->9500mp->919 splitpack + cinemax
|
|
| Back to top |
|
|
the dude
Joined: 11 Sep 2013
Posts: 179
|
| Posted: Wed Feb 24, 2016 8:38 pm Post subject: |
|
|
I have this clips for my probes. If i really need to make good measurements i can use them to connect the probe directly to the pcb. For many things a very short ground wire is not sooo bad as long as you be aware of the ringing artefacts and take it into account. On the probe tip there should be no wire, thats not nesserary with the right tip clamps.
@Redfox: get some proper Tek probes. They are good quality and there are also many nice accessories.
| Description: |
|
| Filesize: |
55.16 KB |
| Viewed: |
4742 Time(s) |
|
|
|
| Back to top |
|
|
El Duderino
Joined: 23 Jan 2011
Posts: 4653
Location: Portland, OR
|
| Posted: Wed Feb 24, 2016 8:40 pm Post subject: |
|
|
| redfox001 wrote: | Ok thanks. Shorter leads makes sense. I do pick up a lot of noise esspcially on the lower side of the driver. But the 12pF mainly leads to a roll of that is why I only look how two measurements with the same 12pF look relational to eachother.
But those probe leads that is a very good point. |
It's not just the length, but the loop area. Consider the area in the 'loop' shown on the good vs bad technique. Of course, shorter will also mean smaller loop area, but you could also likely minimize some loop area by twisting your pig tails together. This may also increase the capacitive loading however.
You many also consider using two matched compensated probes, and make a differential measurement, assuming your scope has an add and invert vertical functions. Short of an expensive J-FET probe, this is likely about the best you'll be able to achieve.
|
|
| Back to top |
|
|
redfox001
Joined: 16 Mar 2009
Posts: 2257
Location: The Netherlands
|
| Posted: Wed Feb 24, 2016 8:43 pm Post subject: |
|
|
Very good point! Relative measurement with two probes.
@dude I was looking for those clips.
And remember do not argue with a fool. He will take you to his level and there he has much more experience
_________________
701s->runco933->8500ultra->cinemax->9500mp->919 splitpack + cinemax
|
|
| Back to top |
|
|
redfox001
Joined: 16 Mar 2009
Posts: 2257
Location: The Netherlands
|
|
| Back to top |
|
|
the dude
Joined: 11 Sep 2013
Posts: 179
|
| Posted: Wed Feb 24, 2016 9:08 pm Post subject: |
|
|
For exaple.
I have P6133, they are rated 150 Mhz but for normal mearurements thats Ok. If you are really want so see actual rise time of a pixel you have to get a active one anyway
|
|
| Back to top |
|
|
El Duderino
Joined: 23 Jan 2011
Posts: 4653
Location: Portland, OR
|
| Posted: Wed Feb 24, 2016 9:18 pm Post subject: |
|
|
|
|
| Back to top |
|
|
mp20748
Joined: 12 Sep 2006
Posts: 5689
Location: Maryland
TV/Projector: 9500LC Ultra / Super 02 and 03 VIM
|
| Posted: Wed Feb 24, 2016 9:30 pm Post subject: |
|
|
| El Duderino wrote: | |
Exactly, and before you start. Clean all connections (to scope and probe contacts with cleaner). And when probe is connected, wiggle the probes cable, if the image changes that's not good.
Scope measuring for -3db is THE only way to properly look at the video chain. The multiburst pattern in comparison is a joke.
|
|
| Back to top |
|
|
|
|
