CurtPalme.com Home Theater Forum

[hansilili]

I have to correct myself from th eabove. When 1920 digital pixels are converted to an analog signal, then this causes aliasing which in fact adds information. This is why in theory it is even better to upscale the 1920 horizontal pixels and shift the aliases to frequencies higher than the bandwidth of the grafic card converter.

Also I admitt to scramble up bandwidth, rise time and max scan rate. Still have to do some learning here.

[gjaky]

[hansilili]

gjaki, thanks for clarifying (again, because you mentioned before but my ignorance won ). With that in mind I reread the first two pages of this thread and understand much better now.
By now I found other sources saying the 10PG as the 1292 have significantly better beam spot size than what the electronic parts in the video signal chain can actually reproduce. The limiting factor in my case is the video amp. I followed your links to Mark_A_W but did not suceed to find the source thread. Is he not around here anymore, used to be a very helpfull guy for this community.

[kabuby77]

The best way to avoid bandwidth problems is to install an internal moome card(if you can). It replace the original imput ic with limited band (150/200... MHz )

[gjaky]

[CasetheCorvetteman]

Gabor, im not sure if you missed this earlier, but his 10PG is actually now a 6PG PLUS with 9inch tubes. Im not sure if that changes anything from your post about 4 posts up where youve got 65MHz in the chain.

[gjaky]

[macgyver655]

Your leaving a valuable part out of your equations. The advertised spec of the speeds of any chip are NOT a chips true maximum capability. The advertised spec is a percentage of the actual manufactured and tested speed. This is to be sure that the device when sold and used in various harsh environments will still produce the advertized speed. The true speed of a device could be 25, 50 or even a 100% or more then what it's spec'ed at. So your 65mhz speed chip could actually have a usable speed of 100mhz or even much, much more.

[kabuby77]

No, that's no correct. Specification on datasheet are the result of test over many chips. They are with a tolerance associated. The manufacturer undertakes to maintain the characteristics of devices within these specifications. I'm sorry but it is impossible to have a complex device like video amp with a tolerance of 100% over speed. In reality what happens is that the bandwidth is specified for an excursion to 100% of the signal. The magic word is GBP Gain Bandwidth Product. This number is a constant that characterizes the device. For example if your VPA chip can handle 100 Volt pp at 65MHz , in theory at 50V pp it can reach 130 MHz: 100 x 65 = 50 x 130 . This is one of the reasons why it is best not to overdo the brightness.

p.s. I simplified to the maximum with the intention to understand, without claiming that it is exactly how it works

[macgyver655]

[kabuby77]

Yes, but the main thing is the second. The speed value is given for 100% brightness, if you keep the projector darker the speed will increase. As said before there are some bottle necks into the video chain, like video input chips , and they can cancel the advantage obtained keeping the brightness low.

[macgyver655]

I disagree, but I'll leave it at that. I have no desire to debate this.

[macgyver655]

Here, rather then debate it I figured I just give you the facts. It took me a couple datasheets to find the description. This is Sanyos 120mhz chip (as advertized). Now note under Features where it says "typical bandwidth". Now go to page 2 where it says Maximun frequency. Note the MAX as 160mhz. I believe that's @ 35%. Nuff said...

[AnalogRocks]

[AnalogRocks]

[macgyver655]

[kabuby77]

[gjaky]

In my opinion, most of these chips are based on the same circuit, or even the same die. They did a preselection in the factory: the better performers got higher rates, the worse ones speced at lower rates, like in CPU manufacturing. Since there were VPH08 chips availabe over the VPH06s, I doubt VPH06s could do over 80MHz, in this case I also think the engineers took advantage of the fact that VPH06s are standing relative high suplly voltages so they got plenty headroom with under driving the circuit.

[gjaky]

Turning back to the original goal of this thread...
While the above mentioned diode switching circuit is working as I expected it causes further problems. The diodes causing non linear distortions. Tried with different diodes in parctice my selected super fast germanium diodes did the job the best, but on my test circuit still caused distortions even with 480i PAL signal observed on an oscilloscope.
Then I looked at the XG's service manual and found that they cut off the blanking outout of the CXA1779P at the last last buffer stage with paralelling the output transistors drived by a 'blanking slice' signal, unfortunately the output transistor are also very hard to find and high quality parts (2SC3613, 2-3,5GHz transition freq, high current, high power)
And again comes an interesting part, I took a look at the older PG schematics and found that the PG plain's GAIN CTL (PWC-3715) has those parallelized transistors at the output, instead of PG plus's and xtra's.
The wise guys are saying the PG plain GAIN CTL is not compatible with the PG xtra's, however I compared closely the two schematics and found that the only difference is in this extra brightness control which was removed from PG xtra, everything else including pin assignment, used transistors are all the same on both board, that is why I want to buy a GAIN CTL board for experimenting, ironically the PG plain's GAIN CTL board could be the key ingredient for tuning up a PG xtra.

[macgyver655]

[macgyver655]