CurtPalme.com Home Theater Forum

[the dude]

Hi Folks,

last weekend i managed to install Barco 909 neckboards in my 1209s! Since the ABL and blanking signal and supply voltages are exactly the same this was easier than expected. Only G2 voltage supply was an issue. The picture with the 909 boards ist really stunning I also replaced the HFA1100 Op Amps on the switcher and RGB driver board with THS3201 (1.8 Ghz) a while ago, so now my set should have more bandwith than a stock 909. Unfortunally i don't have the proper equipment to prove that.

If anyone is interessted in this stuff, i could try to document how this can be done. In fact the only board that must be modified slightly is the G2/Diagnostic board.

Some screenshoots:

[Nashou66]

Nice, what did you have to do to the surrounding circuits to get the THS 3201 to behave properly. In a marquee it is a pain in the ass.

Nashou

[km987654]

Documenting these mods would be great. Look forward to it.

Screen shots don't always convey the improvement. What would you describe as the differences?

[dummyload]

Not to spoil the party but the hfa1100 is not the problem (bandwidth wise) ,on the switcher and driverboard.

[km987654]

[mp20748]

[dummyload]

what I mean is that the bandwidth of a circuit is determines by the component with the lowest value.
In this case it is the ad835 that has a -3db at 250mhz and there are two in serie so even that number will be lower, so there will be little difference replacing a 850mhz opamp in the chain with a 1,3ghz one. The sony g90 also uses the ad835 and also lacks bandwidth ,the marquee doesn't use the ad835 but uses the ad834 which has -3db at 500mhz and that's why the marquee can be moded and tuned to the level it is at now. The only way to get a real difference in bandwith is to replace ad835 with ad834 but there lies the problem they are not compatible components ,so the circuit has to be changed significantly and there are 7 of them in total in the chain (switcher and driver) and then all opamps have to be tuned to get max bandwidth . And then you still have to deal with the neckboard wich is a total different animal.

In dude case the neckboard of the 909 made the difference I think.
MP you already know this.

[hansilili]

Hey Dude,

a guy in another forum I'm active in (aka rumpeli) is thinking (phantasizing) about upgrading his 1209s to a 909. In Germany that still means massive invest. How would you rate the upgrade.

Offtopic but much more intreisting to me - give us some more details about your apparently curved screen. Can you post some pics of your built in the "Home Theater Design & Construction" board?

Thanks

[mp20748]

[dummyload]

@mp
I agree rf is a game apart , theory and practice can vary big time ,everything affects the outcome (quantity of soldering ,for example) and reliable measuring is also tricky. By the way , very nice job on the marquee.

[mp20748]

[the dude]

Let me try to give some answers...

THS3201 vs. HFA1100

I know that this isn't the bottleneck. The real bottleneck was the neckboards for sure, and then the gain control multiplier AD835. But the op amps are much easier to replace since the THS3201 is pin compatible and very similar except for better bandwidth and much better rise time. Thats the reason for replacing it first. The "only" problem was to get rid of the oscillations, this was much harder than expected.

Gain Control AD835

I plan to replace this, but this is more complicated. I have ordered ADL5391 also, but was scared away just by the size of the package I plan to make a very tiny pcb with this chip on it some day to replace the AD835, but that would not be easy. The soldering really scares me. I think i need a microscope or there is no hope for it.

Gamma corrector AD835

I believe the second AD835 (I203/I403) is not an issue. In fact it is not in series with the signal, it just applies a litte of the video signal multiplied with itself to the inverting input of the second op amp (I204/I404). So even if this signal is missing completly, gamma will be slightly off but no information in the image is missing. You don't really need gamma correction on fine structures, if this stage forms a lowpass you loose almost no quality in my opinion. Different frequency responses may introduce some ringing here, but i believe that is barely visible. I would keep this stage as it is. Especially with the stray capacitance in mind that a replacement circuit that is mechnical just a litte bit larger would cause on the inverting input of the current op amp (stabilty issues, this is *VERY* critical). Also the inverting input of a CFA is current driven, so the output impedance of this multiplier may also be critical. Maybe not. Just keep it, it only equalizes the gamma response of the different phosphor types for the individual colors.

THS3201 and oscillations

I had issues with that on the rgb driver board and it took days to get rid of it. A combination of this did the trick:

- playing with the feedback resistor network (this sets gain AND bandwidth)

- adding a rc lowpass to the output (between the output and the load, of course the fb resistor need to be connected straigt). It was something like 25pf / 4,7 Ohm, i don't remember the actual values, i tried many things. But i have noticed them somewhere. It is important to place the series resistor as close as possible to the output lead, there should not even be a milimeter of pcb track between the output and the resistor.

- Removing parts of the groundplane (not sure if this is really helpfull, it was just a idea i tried)

- And finally most important: Removal of ANY peaking component. Its a bad idea to add a capacitor to the inverting input of a cfa like barco did. This is a NO GO.

Maybe some of this is bull**** - im not an expert for CFAs, but it worked for me.

The Screen

Yes, its curved. Its a Kappa Line board painted with aluminium paint.

The image quality with the 909 boards

Well, its hard to be objective without a direct A/B comparsion. In general i would say the image is sharper in general. Also it seems to look cleaner than before, i guess this is because of massive peaking of the original VPH08. Remember, this video pack was rated for 85 Mhz only and Barco added a lot of peaking! And there seem to be slightly more details visible in darker areas (this may be a subjective impression, but i think so). Maybe because because ABL is done via G2 and G2 is more precise now? With the old boards there was a quite visible difference between 1080p48 and 1080p60. Now 1080p60 is almost as sharp as 1080p48.


I will try to explain how to replace the neckboards soon, stay tuned!

[mp20748]

[redfox001]

There are ADL5391 breakout boards see google (LFCSP breakout). Perhaps with one of those you can make a miniboard mod?

[the dude]

[redfox001]

Small wires to input and output are no problem with the ad835. The soldering can be done with a small tip and flux. Or hot air station.

But you have to add some decoupling caps somewhere. Perhaps connect the breakout to a small experimentation print?

[gregstv]

I removed the Gamma chips a while ago. Added a jumper from pin 4 to pin 5. This helped reduce noise. Noticeable on the blue mainly.
I also added a 22pf cap from Pin 8 to ground (gain control) on the AD835. This reduced the noise level quite a lot on the 808s.
There are a few other changes I made that also helped. They are listed in the SEOS thread.

I have some mods on the Focus Astig board that SEOS have done by the looks of it. 3 SEOS boards and all have a cap in series with a resistor from the output pins on the STK chips to ground. I will post more on this if anyone is interested.

[km987654]

Hey Greg

So did SEOS modify the H-Shift and focus board?

[gregstv]

Hi Keith,
I have boards that are different revisions. I haven't seen anything that looks modified.
Greg.

[redfox001]

Checkout this one for some testpatterns to see how the bandwidth is now.

http://www.curtpalme.com/forum/viewtopic.php?p=333943