24" Widescreen CRT (FW900) From Ebay arrived,Comments.

[Derupter]

IMO it's better to wait for issues to be confirmed/fixed before trying something like that. It's only an early disponibility so far, mainly in Germany.

I agree
Well at least the 7 wires with good solder on picture 7 are the most important.
They are the two main lanes,AUX channel and the hotplug(violet wire)
To the other side there is the main power(Red wire) and GND, i don't see which wire is connected on pin 5
If all on the PCB is good then it should be only the shit male connector that doesn't make good contact with pins on graphic card.
Meanwhile I contacted Delock to explain the problem

 

[etienne51]

Thank you again for the useful informations.
Looking the male connector of the adapter,how much space is between a pin row and the other?
In this photo is 0.80mm (this is the female connector): https://commons.wikimedia.org/wiki/File:DisplayPort_(connector).PNG
In picture 5 and 7 i see good wire soldering and the PCB seems good,but in picture 6 it's very bad,the red wire seems doesn't touch the pad and make contact only with solder.
About the Displayport cable,there is only one type and if done well it must be able to support all DP modes.
1.1-1.2 cables are only marketing bullshit,look here
https://www.displayport.org/cables/how-to-choose-a-displayport-cable-and-not-get-a-bad-one/

The soldering on pic 6 looks ugly indeed, but the contact is just as rock solid as the others. There is no issue at all here.

This is my monitor connector. It's Amphenol branded, so I imagine it would be hard to get a better quality one.

As I said, I broke the Delock connector, so this is not really a comparison...

Here is the interesting part.

The picture you sent me is the female connector on the graphics card. On the cable side, I have about 1.4mm of gap. Both my monitor and the Delock adapter seems to have the same spacing.

I think the issue might be with the pins, and with the outer shape of the connector. The pins on the Amphenol connector seems a lot beefier, and the spacing between the two rows is a bit shorter as well.

Allright for the 1.1 / 1.2 cables. I'll get a cable today, and see what I can do.

 

[etienne51]

I agree
Well at least the 7 wires with good solder on picture 7 are the most important.
They are the two main lanes,AUX channel and the hotplug(violet wire)
To the other side there is the main power(Red wire) and GND, i don't see which wire is connected on pin 5

On the other side, the red wire is indeed power, the rest is GND. On pin 5 you have the braid of the cable. It's a bit hard to see on the photo I took.

 

[etienne51]

Ok, work to do later today.

This is a DisplayPort cable a gave to a friend months ago, but he switched to DVI Dual Link, so I'll be using this cable.

Fortunately, the shell of the connectors are just clipped in place. It was very easy to open the connectors and check the colors of each wire.

I'll just remove the glue on the solders, so I can check for continuity before soldering each wire on the Delock PCB, just to make sure I'm not using the wrong wire. It won't do a thing to the connector.

This cable has been working just fine for a 144Hz display, the cable is rated for DP1.2, the connectors have a locking mechanish. So it should work fine here, as long as the circuit of the adapter is indeed good.

Again, for comparison purposes, this is a closer look at the pins.

 

[Blutrache]

Hopefully they make some kind of revision, because atm it looks quite messy.

 

[etienne51]

Oh my god...

After hours of struggling...


(Click for full resolution)

IT'S ALIVEEEE !!!! WOW. It's been years I've been waiting for an adapter finally cappable of doing that!

(If you're wondering, yes, above the screen it's local time... I will definately get some sleep now.)

Oh, and it definately does 1920x1200@96 as it was already confirmed before, which is probably the resolution I'll be mostly using.

Tomorrow I'll write more about the rather awful soldering job that I did. It is a mess, but it is a working mess. And as you can see, now it works as intended...

A lot more details coming tomorrow....

Oh and by the way, my monitor lost focus once today... Hell... I thought it was okay now, since it has been for a while. It seems to happen once in a while. I'm going to need a new flyback (or entire board) in the months to come I guess.

 

[Strat_84]

Nice one !

(and about the out of focus problem, you'll probably fix that with a D board because the problem comes from something somewhere on it, but I'm pretty sure blaming the flyback for it is a blatant mistake. I'm still working on the electronics of my GDM-5410, maybe more answers to come in the next months. )

 

[Derupter]

Well done!
Same result as mine sample 2304x1440 71 Hz 336.14 MHz with CVT
Did you tried 72 Hz? It's 341.10 MHz and my sample is unstable.
Did you find the limit? My sample limit is from 340.3 to 340.5,after that it starts noise on windows borders,near 341 it starts to lost sync,after 341.5 black screen,at 339-340 it's perfect.
Reducing Front and Back porch i tested 74 Hz,but it can do more.
2048x1536 75Hz and 1920x1440 85Hz need CVT timings,with GTF they are unstable because pixel clock go over 340 MHz (GTF is used by default with these resolutions)
I have some questions for you,i have to do a report to Delock and i try to collect informations.
In the picture 10 that mess is normal? I mean all the DP plug are like that? How were the solders on the new cable?
Did you measure the continuity between the pins of the old connector and the wires?
If you still have the old connector,can you make if possible a photo like the third of your post 15194,but with the connectors in parallel side by side,i want to see crearly the differences between the rows of pins.
In the link i gave you about DP cables (displayport.org) there is an exception,RBR cables,economic and long cables usually used for projectors,i hope they did not use it,is there anything written on the old cable?

 

[etienne51]

As promised, a more in depth writing about the fix.

Did you measure the continuity between the pins of the old connector and the wires?
If you still have the old connector,can you make if possible a photo like the third of your post 15194,but with the connectors in parallel side by side,i want to see crearly the differences between the rows of pins.
In the link i gave you about DP cables (displayport.org) there is an exception,RBR cables,economic and long cables usually used for projectors,i hope they did not use it,is there anything written on the old cable?

Unfortunately, I managed to break the old connector in half. I didn't apply too much pressure on it, trying to remove some glue I guess, and all the wires just snapped all at once. These wires are so thin, it's really easy to snap them...
But I can tell you, all the solders were strong. It definately made perfect contact from the pin inside the DP connector, all the way to the electronics. I have no doubts about that.

Now about the fix. So first, the pads arrangment on the PCB, and how it's wired to the DisplayPort connector.

The digits are the pin numbers of the DisplayPort connector. On the right, we can see which wire goes to which pad on the board. Below is the description of each pin (graphics card side).



This is the new, reliable connector I'm using now. Good thing it was not molded, I was able to open it up very easily.

The glue was covering the solder points, so I had to cut it just enough to have access to each wire, and check for continuity. So I could figure out which wire leads to which pin.

Bad news, there was no wire connected to Pin 20 (3.3V power line)... I had to desolder one of the unused wires, one that was thick enough (it's a power line, not a signal line), and connect it to Pin 20 so I could get power for the Analogix converter IC. I chose the unused Pin 12.

Now the cable itself, this is what it looks like on the inside.

This is a really good DisplayPort cable, with an outer shielding braid, connected to ground. Just after that, aluminum foil, followed by all the wires and the 5 shielded twisted pairs.
Now for the balanced data lines, as I said, these are also shielded, this time only with aluminum foil, wrapped in clear plastic. One bare ground wire is also contained inside each balanced data lines assembly.

For comparison, below you can see the original cable from Delock.

Honestly, I think this cable is okay. It also has the shielded twisted pairs, and double outer shielding. One difference though, the wires are thinner, but it's probably just fine here. This would be an issue on really long runs of DP cable, but here we are talking only a few inches (about 8in / 20cm).

So I think it's really the connector's fault. All the issues we have on the Delock adapter, are probably due to the really bad DP connector.

And btw, this is the ferrite core on the Delock cable.

Nothing to say about that, really. I don't have enough knowledge yet on the subject. But as someone said before, I've never seen EMI filtering on DisplayPort cables, but since this is a digital to analog adapter, it might help a bit in lowering the noise floor on the analog signal out, though.

Anyway, back to the subject. Here you can see the complete wiring. I connected a few more wires to GND, compared to the stock wiring. It probably doesn't change a thing, but it definately won't degrade anything either.

It looks a bit messy, it was not easy to solder these wires, without melting too much of the insulation, since I had to hold them with tweezers. The wires were too thick to fit in the grooves of the little plastic guide, and it broke in half anyway...

But then I double checked everything for shorts, and just plugged it to my GTX 1080 and FW900! Voilà, it worked first time, and 100% stable!

Now I need to use a glue gun or something to cover all that and make sure nothing moves near the PCB when bending the cable.

I also ran the adapter with no shielding, it works fine. But I will try to make a small metal box for it, and 3D print something to hold everything in place correctly on the inside.
Or I could go cheap with something similar as what the chinese did with the stock Delock adapter, but I don't have copper foil and I can't solder the shielding braid to aluminum foil. So I'll see what to do about that soon.

 

[Derupter]

Thank you for the pins diagram,i agree with numbers,did you connect all GND?
The wire of pin 20 in not present because is a standard VESA,bad cables have that wire connected and it has caused a lot of problems with graphic cards and monitors.
If you haven't already covered the chip,can you measure the temperature at 270 MHz and at 340 MHz?
What do you think about the quality of the image?
I was wondering if something can be done to improve the contact of the original connector,don't know,something like bend the pins towards the center,what do you think,it's a bullshit?

 

[etienne51]

No, I only connected the GND wires that were associated with the 3 balanced data lines in use, plus the ones required by the original Delock design, and of course the outer shielding braid.

Oh, I didn't know that about pin 20. Makes sense.

Well, I've been using the FW900 for the past few hours, I see absolutely no ghosting, which is a really good thing. I'm running 2304x1440@71Hz, this is the highest I can go without fiddling with any settings. I will definately try to optimise that! The image is really sharp, I have nothing to complain about so far.

I have a few older GPUs, from the CRT era, and a couple old school gaming computers for a few older games. I really don't touch them often, in fact it's been probably a year I haven't powered up any of these rigs.
But do you think I could technically get the best image out of the CRT on one of these cards? Just for quick comparison, I'm not sure I would see any difference anyway, but I could try!

- Matrox Millenium G200 16MB + SLI of 3dfx Voodoo 2 12MB (1998)
- 3dfx Voodoo 5 5500 64MB AGP (2000)
- Nvidia Geforce 2 Ultra 64MB (2000)
- ATI X850XT Platinum Edition AGP (2004)

These were iconic cards back then, especially the 3dfx!

About bending the pins, I really don't recommand that. These are small probably fragile and not really easy to reach. The tip of each pin slides in a small hole in the plastic mass of the connector. If you try to bend one pin, it might slide out of this hole, and then you would have a hard time putting it back in without any irreversible damage. It also does not seem to bend easy.

 

[etienne51]

Oh, and I forgot to add! As I said I've been running 2304x1440@71Hz, 337.392MHz (GTF default settings from Nvidia Control Panel, see attachment) for a few hours. I just used a cheap infrared thermometer to check the chip, it reads 51°C / 124°F.

I'm not sure it's the right temperature, but what I can say is I can keep my finger on the chip without burning myself. It's definately hot, but not burning hot.

 

[Derupter]

Well,Matrox is famous for its quality DAC,ATI is good even,but Matrox is the best.
Thanks for temperature measurement,well it is low,i expected higher values,that chip is amazing.

 

[etienne51]

Yes, me too. I was already planning for a heatsink, but it definately won't be useful at all.

 

[jbltecnicspro]

etienne51 - you da man! Or woman! This is definitely promising.

 

[Chief Blur Buster]

Oh, and I forgot to add! As I said I've been running 2304x1440@71Hz, 337.392MHz (GTF default settings from Nvidia Control Panel, see attachment) for a few hours. I just used a cheap infrared thermometer to check the chip, it reads 51°C / 124°F.

I'm not sure it's the right temperature, but what I can say is I can keep my finger on the chip without burning myself. It's definately hot, but not burning hot.

#1: Try 72 Hz instead of 71 Hz with a minor tweak.

That screenshot shows headroom to go up to 72 Hz while keeping horizontal scanrate. Try reducing Total from about 1500 to approximately 1478 in order to keep your "106.50 KHz" horizontal scanrate number while getting enough headroom to bump to 72 Hz.

If you're dotclock limited rather than horizontal scanrate limited, then you can instead keep 1500 vertical total and instead decrease horizontal total to get the dotclock headroom for reaching 72Hz instead of 71Hz.

So you have two choices of achieving 72 Hz without glitching-out:
A. Decrease Vertical Total from 1500 down to a number between 1475-1480 to be able to increase Vertical Hz from 71->72Hz without raising scanrate (106KHz) nor dotclock (337MHz).
B. If you're actually dotclock limited instead of scanrate limited (your monitor can do more scanrate, but is simply being limited by dotclock), keep your blanking interval the way it is (Vertical Total 1500) and instead decrease your Horizontal Total instead to get the dotclock headroom you need to increase refresh rate from 71->72Hz without needing a higher dotclock.

72 Hz is better because 24fps 4K/HD movies will play smoother.
---

#2: Try 4K 60Hz interlaced! The modified adaptor looks like it can handle it!

4K60i is only ~280 MHz dotclock (+/-10%) and only ~67KHz horizontal scanrate (+/- 10%) -- both lower numbers than you're successfully cramming through now!

Also, you might have the dotclock room to attempt 4K interlaced!!!!!!!! You're pushing enough dotclock now (337 MHz) and doing 4Ki60 would only require 67 KHz horizontal scan rate (less than 108 KHz you are doing now) and a dotclock of ~280 Mhz which is LESS than your 337 Khz. Sure, the CRT won't be able to fully resolve 4K, but you play 4K videos with absolutely no digital scaling (potentially better quality) .... and 4K60 interlaced is technically within the specifications of the FW900.

4K interlaced would be interesting to try!

 

[etienne51]

#1: Try 72 Hz instead of 71 Hz with a minor tweak.

Ok I ran some tests for 72Hz, this is what I came up with. I can do 72Hz both ways.

A. I get a display by decreasing Vertical Total to 1482, stable. Black screen for 1483.
B. If I decrease Horizontal Total down to 3131, I have a display, but it's unstable. At 3029 it seems to be perfectly stable.

#2: Try 4K 60Hz interlaced! The modified adaptor looks like it can handle it!

What software would be best on Windows 10 x64 to try that? The Nvidia control panel just won't accept the resolution. It says my monitor does not support it.

I remember Toastyx CRU, I'm going to give it a try.

etienne51 - you da man! Or woman! This is definitely promising.

LOL. Man, that would be, haha. And yeah, despite the issue with this adapter, the Analogix IC seems really interesting. Considering the price, it's a nice option to have while waiting for HDFury, if they finally release their new 400MHz adapter one day.

 

[pr0ton]

As promised, a more in depth writing about the fix.

....

Or I could go cheap with something similar as what the chinese did with the stock Delock adapter, but I don't have copper foil and I can't solder the shielding braid to aluminum foil. So I'll see what to do about that soon.

"In comparison, the DP plug of the Delock adapter is loose, has no lock, probably the worst possible DP connector one can come up with."

So that says it all. I can return mine and get it replaced for probably the same experience, or find a quality DP1.2 cable and do your fix . What is the new length of your DP cable now? I've been thinking about keeping the analog VGA cable as short as possible, maybe only a VGA -> BNC adapter.

In my opinion the difference in DACs isn't visible, Nvidia or AMD (FirePro). Somebody else in this thread noted that the bottleneck in image sharpness/quality lies in the monitor electronics itself and not the DAC, makes sense to me.

Great research btw Shame on Delock, but at least there is a solution after all these years.

 

[Derupter]

#2: Try 4K 60Hz interlaced! The modified adaptor looks like it can handle it!

4K60i is only ~280 MHz dotclock (+/-10%) and only ~67KHz horizontal scanrate (+/- 10%) -- both lower numbers than you're successfully cramming through now!

I think you have calculated it with LCD timings,here with CRU the pixel clock is 346.37 MHz

Ok I ran some tests for 72Hz, this is what I came up with. I can do 72Hz both ways.

A. I get a display by decreasing Vertical Total to 1482, stable. Black screen for 1483.
B. If I decrease Horizontal Total down to 3131, I have a display, but it's unstable. At 3029 it seems to be perfectly stable.

What software would be best on Windows 10 x64 to try that? The Nvidia control panel just won't accept the resolution. It says my monitor does not support it.

I remember Toastyx CRU, I'm going to give it a try

First you need to find the limit of the adapter using standard timings,look for noise on windows borders (small windows like right click on black desktop background),use CRU or nvidia panel and when you are near the limit increase the pixel clock no more than 0.1 MHz at a time.
When you know the safe limit, i can give you some settings to test.
My sample limit is around 340.50 MHz,so to be safe i set max 340.
What do you think about change only the connector?
Something like this: http://eu.mouser.com/ProductDetail/Molex/47271-0011/?qs=sGAEpiMZZMtAYTMy7wxAr0NBVnPXu1wDmhPM7eDMO60=
In this page there are all the components: http://eu.mouser.com/Search/Refine.aspx?Keyword=538-47271

"In comparison, the DP plug of the Delock adapter is loose, has no lock, probably the worst possible DP connector one can come up with."

So that says it all. I can return mine and get it replaced for probably the same experience, or find a quality DP1.2 cable and do your fix . What is the new length of your DP cable now? I've been thinking about keeping the analog VGA cable as short as possible, maybe only a VGA -> BNC adapter.

In my opinion the difference in DACs isn't visible, Nvidia or AMD (FirePro). Somebody else in this thread noted that the bottleneck in image sharpness/quality lies in the monitor electronics itself and not the DAC, makes sense to me.

Great research btw Shame on Delock, but at least there is a solution after all these years.

I suggest you to return the adapter,if it does not cost too much,i made a report to Delock but if no one return these things,they don't move.
For now i have only two feedback for Delock.
We'll see what they have to say,i don't think it's a Delock fault's,they chose a model and the chinese manufacturer did that,maybe Delock should have told them to use better components than those used for the usual shit low performance adapters.
If you want to do the job and you don't have the DP cable,it is better to buy a DP cable extender or an adapter like DP Male to Female because these cables have the 20 pin wire for power,so you don't need to do the extra work.
About the quality of the DAC,it most depends from what there's between the DAC and the VGA connector.
On graphic cards the dac is integrated in the GPU,from here to the connector the distance is long,on
some cards there are also power circuits which pass around the DAC lines (to the left of the GPU).
I have seen cards with a good image only on 2D,switching to 3D the screen was full of interferences.
This is an advantage of the external DAC,short track and it is alone.

 

[Strat_84]

First you need to find the limit of the adapter using standard timings,look for noise on windows borders (small windows like right click on black desktop background),use CRU or nvidia panel and when you are near the limit increase the pixel clock no more than 0.1 MHz at a time.
When you know the safe limit, i can give you some settings to test.
My sample limit is around 340.50 MHz,so to be safe i set max 340.
What do you think about change only the connector?
Something like this: http://eu.mouser.com/ProductDetail/Molex/47271-0011/?qs=sGAEpiMZZMtAYTMy7wxAr0NBVnPXu1wDmhPM7eDMO60=
In this page there are all the components: http://eu.mouser.com/Search/Refine.aspx?Keyword=538-47271.

Beware, with Mouser you need to order at least 50€ or pay 20€ of shipment fees, that makes a connector pretty expensive if you don't need anything else.

 

[lm92]

I have also bought a DeLock 62967, same problem here running an AMD R9 290X, with both a FW900 and F520.

Cable felt very loose and everything above a certain resolution either displayed black or intermittently, with the monitor constantly clicking.

At working resolutions, AMD's control panel indicates 2 x 2.7Ghz links, non working modes indicate 2 x 5.4Ghz links.

DeLock's support hasn't been very helpful, and in the end they suggested returning the adapter. I did try using pliers to tighten up the connector (it is very maleable) and sure enough it is very secure on the port now, but the problem remains. Could it be that they used a standard cable pinout and pin 20 is not wired on the connector?

 

[Derupter]

Beware, with Mouser you need to order at least 50€ or pay 20€ of shipment fees, that makes a connector pretty expensive if you don't need anything else.

I put that link only to show the model,i did't check the economic conditions but good to know,do you know any store that sell these things with good prices?
Anyway i think it is very difficult to solder on those things,but it has the advantage that you don't have to open the adapter.

I have also bought a DeLock 62967, same problem here running an AMD R9 290X, with both a FW900 and F520.

Cable felt very loose and everything above a certain resolution either displayed black or intermittently, with the monitor constantly clicking.

At working resolutions, AMD's control panel indicates 2 x 2.7Ghz links, non working modes indicate 2 x 5.4Ghz links.

DeLock's support hasn't been very helpful, and in the end they suggested returning the adapter. I did try using pliers to tighten up the connector (it is very maleable) and sure enough it is very secure on the port now, but the problem remains. Could it be that they used a standard cable pinout and pin 20 is not wired on the connector?

Another user with the same problem,where did you buy it?
The working mode is HBR,the non working HBR2,i think they have used a shit connector that is enough for low bandwidth transfers but insufficient for the capatibility of the chip.
On the box there is written 2560x1600 60 Hz (270 MHz) so if it can't handle that,they have to return all that things to China.
Without pin 20 the adapter simply doesn't work (no power)
Don't worry i contacted Delock with a very good report with pictures,will see.

 

[lm92]

I put that link only to show the model,i did't check the economic conditions but good to know,do you know any store that sell these things with good prices?
Anyway i think it is very difficult to solder on those things,but it has the advantage that you don't have to open the adapter.



Another user with the same problem,where did you buy it?
The working mode is HBR,the non working HBR2,i think they have used a shit connector that is enough for low bandwidth transfers but insufficient for the capatibility of the chip.
On the box there is written 2560x1600 60 Hz (270 MHz) so if it can't handle that,they have to return all that things to China.
Without pin 20 the adapter simply doesn't work (no power)
Don't worry i contacted Delock with a very good report with pictures,will see.

I bought it from direkt.jacob.de. Honestly returning the adapter would be a pain so I'm not sure I'll bother. jacob.de is all in german, which I don't understand. I've managed to buy it by translating the page, but contacting their support would be definitely more annoying. Also, I'm from Portugal and probably shipping costs back to germany would be half the price of a new adapter anyway..

Kinda sucks that DeLock advertises 2560x1600 60 Hz and they clearly didn't bother testing the final samples before putting it on sale. I have a very old Tripp-Lite adapter that's only rated for 1600x1200 60Hz and yet the cable is much thicker (it uses exactly the same connector that etienne51 used).

 

[lm92]

Derupter, my reply to you is waiting moderation

On a side note, anyone knows if it is possible under Windows 10, to render something on a card and display it on another?

I mostly use Linux, and I have a single slot HD4850 sandwiched between two 290X's, and the FW900 connects to that using it's native VGA output. Linux simply detects the FW900 as if it was connected normally and I can run 2304x1440@80Hz no problem. Any 3D content renders on the primary 290X and gets sent to the HD4850 via PCIe, so the limited performance of the 4850 is not an issue. As long as it is at least a 8x lane there is no lag or slowdowns, even while playing games that load the 290X.

Under Windows 10, no matter what I try it does not detect any display not connected to the 290X..

 

[Derupter]

With a proper driver it should be possible,but they will never do it.
Actually Delock tested it,but at standard resolution which i think isn't 2560x1600 60 Hz.
My first sample was the test,i was happy with that,so they started the production with the new aesthetic,no one expected such a thing.

 

[Strat_84]

I put that link only to show the model,i did't check the economic conditions but good to know,do you know any store that sell these things with good prices?
Anyway i think it is very difficult to solder on those things,but it has the advantage that you don't have to open the adapter.

Along with Mouser, I'd say Farnell, maybe Digikey (not recommended for people living in the EU since there are extra customs duties they do not advertise). RS Component doesn't seem to have Displayport connectors.
That kind of spare part can actually be found from distributors also selling electronic parts, the problem isn't the price of the part itself (1-2€) but the fact the distributors all charge quite massive shipment fees for small orders.

 

[lm92]

With a proper driver it should be possible,but they will never do a such thing.
Actually Delock tested it,but at standard resolution which i think isn't 2560x1600 60 Hz.
My first sample was the test,i was happy with that,so they started the production with the new aesthetic,no one expected such a thing.

Yeah, I'll be waiting to see what Delock tells you. I'm pretty sure retooling the manufacturing to a new connector will take a while, so if that's the case I might not return it and have a go at buying and soldering a new connector (not that I have the soldering skill to do it ).

Keep us posted!

 

[Mumble]

SO I dont have to play catch up to much, do we have a list of all the IC's I can email moome?

forwarding the information to moome from the CRT projector forums, the one who makes the HDMI boards for the Barco 909s and such.

He also the one I posted about the 12-bit HDMI to RGBHV for $250 that does 1920x1200 @ 75 hertz.

When I told him we found an adapter with an IC tha tcan do 2304x1440P at 71hertz, he seems pretty interested.

I know one is the Analogix ANX9847.

 

[spacediver]

That sounds like a good idea.

The ADV7123 came up earlier - 330 MHz that does 10 bit per channel. Sounds like a better option than the analogix since it's 10 bit and not 8 bit. If he could put together a DP-VGA converter using that, it could be just what we need.

 

[Strat_84]

Don't forget the cost issue ... The Delock adapter costs about 20€ and should be more than decent if it is fixed.

Of course it would be nice to have something performing even better (or performing properly without having to rebuild it, in case Delock doesn't react). But it also has to be at a reasonable price, an adapter in the 200$ range would be highway robbery for a tiny piece of PCB with a few SMD components and a couple of ICs,

 

[spacediver]

I'd pay extra if it meant I had 10 bit control over my LUT. That's hugely important to me. But I do recognize that it's not important to everyone, and perhaps even most people. So if we were only to have one adaptor, and it was a choice between a 20€ 8 bit one and a $200 10 bit one, clearly the 8 bit would be better. Course, I'd like the option of both

 

[Mumble]

12-bit colours, not 10, but his is handmade and custom case go back a few pages and you can see it. he is also shipping out some South east asia country so im assumign that also covers shipping, People at AVS and the CRT projector forums think his HDMI expansions cards are worht the price.

 

[Enhanced Interrogator]

That sounds like a good idea.

The ADV7123 came up earlier - 330 MHz that does 10 bit per channel. Sounds like a better option than the analogix since it's 10 bit and not 8 bit. If he could put together a DP-VGA converter using that, it could be just what we need.

I don't know nothin' about no electronics... but how possible is it use two of these chips in tandem, with a ~660mHz oscillator to rapidly alternate odd and even pixels between them, to get a combined output that gets near 660mHz, or at least greatly exceeds 400mHz?

I know I already asked this question in the past but nobody responded

 

[Chief Blur Buster]

I think you have calculated it with LCD timings,here with CRU the pixel clock is 346.37 MHz

Hmm -- you are right. That said, many CRTs can take a hybrid timings that's somewhere partially between the LCD timings and CRT timings. The FW900 generally can handle some reductions in porches.

My sample limit is around 340.50 MHz,so to be safe i set max 340.

That said, if it's 340MHz versus 346MHz, you can easily strategically shrink porches to get the necessary dotclock room. I would be shocked if the FW900 CRT couldn't handle a 2% variation.

 

[Strat_84]

12-bit colours, not 10, but his is handmade and custom case go back a few pages and you can see it. he is also shipping out some South east asia country so im assumign that also covers shipping, People at AVS and the CRT projector forums think his HDMI expansions cards are worht the price.

It may well be a nice piece of equipment, I can't judge that, I don't have much details about it. But what I'm warning about is people exploiting the fact buyers are not necessarily well informed of the value of electronic to charge ridiculous amounts of money for their products.

 

[Mumble]

"

It may well be a nice piece of equipment, I can't judge that, I don't have much details about it. But what I'm warning about is people exploiting the fact buyers are not necessarily well informed of the value of electronic to charge ridiculous amounts of money for their products.

oh not disagreeing, I like options too, plus I left images and email to contact, he seems nice and prett straight forward. Still waiting on his HDMI 2.0 ones, as I seens only a HDMI 1.3 in a barco 909 at a friends house. I still hope HDfury releases thier. but this adapter is nice once fixed, it seems, getting one soon, I plan to get the pci express USB C and USB verison as it is ALT-DP and Both the S8+ and Moto force Z2 support that. It be a little, I am currently moving again.

I did find this for a Ras pi or other SBC that support the pi hats. http://www.retrotink.com/, thought it was neat. does go into quite alot of detail.

also this thread, havnt read it all yet, https://ez.analog.com/thread/46833.
"
The ADV7619 can only do 170 MHz pixel clock on a single output bus. To achieve 228MH pixel clock you need to use both output buses where the even pixels are on the low bus and odd pixels are on the high bus. You effectively get a 48 bit pixel bus comprises of a 24 bit even pixel bus and 24 bit odd pixel bus. On the side note the 7619 will only output 8-bits per color on the double wide bus. You could interleave the odd/even buses in an fpga.Splitting the source into 2 streams just compounds the problem, then you have to resynchronize the frames"

So who ever though of doing that dual IC chip in a FPGA may be possible.

 

[Derupter]

I retested the adapter with my old AMD 5850,last time i made the test the image was totally unstable after 110 MHz,this time before the test i cleaned the female connector with a bit of isopropyl alcohol.
After this the adapter is stable up to 180 MHz which is the max with the DP 1.1 of the 5850.
So,before testing the adapter make sure that the connector of the card is clean.
Probably this week i will order the final version,i'll let you know how it will go.
I still waiting for Delock reply.

 

[Derupter]

Hmm -- you are right. That said, many CRTs can take a hybrid timings that's somewhere partially between the LCD timings and CRT timings. The FW900 generally can handle some reductions in porches.


That said, if it's 340MHz versus 346MHz, you can easily strategically shrink porches to get the necessary dotclock room. I would be shocked if the FW900 CRT couldn't handle a 2% variation.

Well,actually i tested resolutions up to 357 reducing both horizontal and vertical blanking so 346 is easy,but my card doesn't like that resolution,the max i can set interlaced is 2720x1530 (like the artificial limit AMD put on my 7950 for downsampling).
I tested it with native DAC,if the card accepts 4K interlaced it should works without problems even with this adapter reducing only the horizontal blanking.

 

[Stryker7314]

Want a delock for my fw900 but not in its current broken state.

Will this introduce any lag? I'm assuming not, but don't know for sure.

 

[Derupter]

No,there isn't input lag,i tested it with SMTT 2.0