[KennyG]

I'm dealing/ learning with a new display system, the Ampro 3600...so far I can say it's convergence is rock solid.

It does however have a trait that has me puzzled, and this happens to everyone of the 3600s we've had up and running (five so far)

When you get into an H-width adjustment between 88 and 96 the picture starts to distort real bad, it's never the same setting from pj to pj, but always the same effect.

My guess it the H drive circut is clipping...really no big deal I suppose, just back it down until it stops, then back it down a couple more notches...However...


I'm wondering if shortening the Horizonal Front and Back porches will stop this.

I"m I correct in thinking that by shortening both porches retrace time is also shortened?

I realize the guns need a certain amount of porch to stabilize...the back porch would be important for stabilization prior to drawing the active image...and the front porch to prepare the gun to blank and sync.


My question is how much time, or how many pixels are needed for these functions?

I suppose it varies with projector types?...and is this stabilization a function of the tube's amplifier?


I would think the shortest porches would be the easiest on any pj???

After all, any drawing by the guns that are not active image would be a waste (other than stabilization pixels/time) Wouldn't it????


Yes, my brain is starting to run wild again.

 

[Chuck Williams]

Kenny,


What in the world are you doing five 3600's?!


I've never seen the problem you're describing with these units. You should be able to overscan the glass on the CRT's with no distortion.


Your analysis of the front and back porch timing seems correct, but how are you going to change this? Does your source allow you to do this? If you can shorten the retrace time, give it a try. How much can you go? The projector will let you know if it's not happy.


BTW, one way to cheat and get more picture size is to slide the deflection yokes back a little on the necks of the CRT's.


Chuck

 

[KennyG]

Hey Chuck,

I only have two of them Bob Wood has a couple and Andre Fraser also has a pair of them.

I'm using an HTPC, which if you go into the "Advanced timings & Resolution" menu allows for Porch adjustment.

Yeah, it's very strange, but not one of these will hold a stable image high enough into the H-width adjustment range to even get close to the tube face edge.

 

[RobertWood]

Iro, Boy and His Ampro and Jurgen are you out there? All of you have 3600's. Are you also experiencing what Kenny is talking about?

My 3600's hor size can be set no higher than 87 before I start to get distortion in the image.


My earlier Ampros have a hor width pot (or coil) that allows some adjustment. I'm not seeing that on the 3600. Does anyone else?


Bob Wood

 

[ThurmanWitte]

Kenny and Bob,

Are you seeing this in the upper left corner of the picture and does it look like the picture is "tearing" ?


If so, then it also occurs on both my 3300 and 4300 with an HTPC feed. It does not happen when I am feeding a HDTV signal from a HDD200 set top box with TWC cable. I experimented with the HTPC porch settings to widen the image and it helped the tearing in the upper left corner also.


I'll check my porch settings tonight.


Bob,

Where is the pot or coil you are talking about?

 

[RobertWood]

Wow, Thurman, you're right. It's only with my PC output. My Dish 6000 doesn't do it. I can go all the way to the max without the tearing.


Yes, it's a horizontal tearing and in my case it's at the bottom edge of the image (because I still have it table mounted I reckon).


p.s. the hor width adjustment pot was on the HOT module on my 4200 (the same model Ray is using).


Bob

 

[ThurmanWitte]

That is weird. I would expect it to be in the same place, because I thought the picture was drawn starting in the upper left corner.


The HOT board was changed for the late models I think. When Ray and I were discussing it a while back I could not find the pot on the 3300/4300 board.

 

[Cesiumdeth]

Not too much to add here... but with my Ampro 2000G hooked up to an HTPC every time I click on something or click anywhere on the screen what looks to be a piece of a test pattern or something shows up in the upper left hand corner for a brief moment, then disappears until I click again. Doesn't have any problem unless I click the mouse. I have the horizontal in the high 80's. Would like to hear more on this mystery.


DAN

 

[mp20748]

Guys,

I'm not familiar with the 3600, but have have had some fun with the other models.


- Can you confirm this problem in both bands (31Khz/70Khz).


- If the problem exist in the high band only, can you measure the power supply voltage on the deflection board while making the adjustment to see if the voltage sags.


This sounds like power supply (main)... There should also be one on the deflection module that regulates and controls the two different bands.

 

[ThurmanWitte]

Mike,


Don't know about the 3600, but the 3300/4300 schematic shows 4 ranges on the horizontal board: >20,>28,>48 and >68 kHz.


Both of my feeds fall within the 28-48 kHz range (33 for 1080i HDTV and 45 for the HTPC). Would the higher ranges still need to be examined?

 

[mp20748]

Yes, there are two bands, one is usually never the problem (low band), it is somewhere around 60 Khz and below. The other is the high band, it is usually more of a problem because it involves a MOD circuit (modulation), that circuit works very close with the deflection processing circuit on the processor board (it may have a different tag on the Ampro). This circuit adjust the amount of voltage to the deflection circuit based on the band and scan frequency, this is all done based on the input Horiz scan rate. The manual should show the deflection circuit B+ supply at different frequencies, that voltage should be checked and confirmed and it should also be checked for sags while varying the width.


This may or may not apply to the Ampro. If you have a manual (schematic)send me a copy and I'll try to help out further...

 

[ThurmanWitte]

Mike,

I emailed you the schematic.

On page 4 of the schematic there are three areas labeled "LF", "HF" and "FREE RUN". They are located in grid coordinates B8, B7 and C5 respectively. Is this what you are talking about?

 

[Ray Cendroski]

Mike,


You're giving me some hope that an explanation might be at hand. The scenerio you describe is exactly what I've been seeing with my Ampro 4200. I know the old/new models have some design differences, but hopefully the Ampro approach is the same for all.


The breakup in my case is a function of the scan frequency. Back when I was running 32khz with a line doubler, the width was no problem (exactly as you mentioned). I could overscan the full width of the crt face by using the remote, and the "Width" pot on the HOT board as Bob mentioned. That circuit is a summing junction that takes an analog signal from the registration board and combines it with the pot such that increasing one setting or the other can cause the picture to breakup.


When I went to a HTPC, I started bumping into the 45khz switch point "width problem" by moving up to a scaled 720p image. The higher the frequency goes from there, the worse the breakup becomes. I know it's not the HTPC, because I get the same results when I switch in the internal 64 khz test pattern with the HTPC dis-connected.


I've sort of gotten around the problem by connecting a jumper on the 45 khz detect circuit so that the high band circuit doesn't switch in. That at least lets me run solidly at 720p. The resulting image doesn't breakup at higher scan rates, but it's not a pretty picture either. It's obvious that high band circuit is there for a reason.


I've pretty much ruled out the HOT board as the source of the problem. Mine was repaired and tested to spec by the factory just before they went under. Recently I substituted a known ok HOT board and the result was exactly the same breakup above 45 khz.


At that point I sort of gave up, because I can't understand the circuit operation from the limited info on the Ampro schematics. You know what I mean - no voltages or wave forms marked at test points, and limited descriptions of operation.


I'll take a look to see if I can understand how the power supply fits in. Any other thoughts you can come up would be much appreciated. You've really been a big help on these kinds of problems over the last couple of years.


Thanks,

Ray

 

[mp20748]

Ill get back as soon as I finished studying the diagram... in the mean time get your meter ready...

 

[ThurmanWitte]

Here is the info on the settings I am currently using (this is on the Elsa Gladiac software) for 1704 x 960 @ 48 Hz:


Pixel Clock: 9.111 ns, 109.758 MHz


Horizontal (pixels/ns)

display: 1704/15525

frontporch: 64/583

sync width: 192/1749

backporch: 304/2770

total: 2264/20627

48.48 kHz


Vertical (pixels/µs)

display: 960/19802

frontporch: 1/21

sync width: 3/62

backporch: 46/949

total: 1010/20833

48 Hz

 

[KennyG]

Man, I wish I'd have waited to start this thread until I had my new pj together and up on the ceiling...it's hard to keep up with everything.


Here's what my lastest thoughts on these porch setting are.

Perhaps the most important perimeter we are concerned with is bandwidth...now, lets say we are using a resolution of 1280X720.

In this case we have a total of 921,600 active pixels...to keep things simple, let us say our H-front & back porches each have 100 pixels (200 hundred total per H scan line)...and our V-front & back porches have 5 pixels (10 scan lines total)


This means our total pixel count is now at 1,080,400...or 15% higher!

I would think bandwidth would actually be calculated on this total pixel count rather than the smaller active pixel count???


Anyway, by shortening all four porches, wouldn't the guns actually draw at a slower rate? After all, your active pixel count stays the same, and your refresh rate stays the same.

Less H-porch means less "H blank time", and less "H active time"

Less V-porch means less "V active time", and less "V blank time"


Since the actions involved here are actually time related as they relate to the projector, perhaps number of pixels isn't important.


The real questions for Horizonal porch are:

1) How long does it take for the pj's H sync pulse to shut off, and the guns to stabilize before drawing the active image (in ns)...this will tell us the minimum H back porch needed.

2) Once the gun reaches the end of the H active time, how long does it need to start the "H sync pulse" (again in ns) this would be the needed front porch.


Same with Vertical porches:

1) How much time does the pj need between the end of the V active time and the start of the V sync pulse? this will give us the minimum V front porch.

2) How much time is needed to start the V active time after the V sync pulse has ended? This will give us the minimum back porch.


You would think that Vertical porches could be very short, since it takes several nanoseconds to draw a Horizonal line, I'd think both of these porches could be as small as 1 or 2 pixels.


So...where do we find this kind of info?????

 

[Boy & His Ampro]

It's little annoyances like this that prevent me from going HTPC. I still really like the picture the ISCAN Pro gives. Everything is plug & play and I don't have all the extra colour adjustments and software conflicts. (and 1 remote). Kal had his HTPC over about a year ago and there were slight trade offs using either source albeit that was before Kal swaped out his video card for the latest Radeon and his new player software. I would be interested in seeing it again. I know Iro switched to HTPC and says it is a huge difference, but I really like keeping things simple. Now if I could hook up my new Imac...

Brad

 

[KennyG]

Well tonight I set aside a little time to play with these settings...other than move the picture around the screen it didn't visually do anything!

I lowered my H-back porch from 260 pixels to 200 (but also dropped it as low as 60 just to see what it did...nothing)

Lowered my H-front porch from 164 to 120...played with the sync width (which I don't understand)

The one thing I did do was remove a Horizonal line that was drawing off the bottom of the screen, this was done by increasing the V-front porch...I had the feeling the guns weren't getting enough time to shut off, so I added an extra pixel (line) to the V-front porch...the run away line disappeared.

Again I played with the V-sync width and couldn't figure out what it was doing.

I image is no more or less stable than before...but, the H scan freq did drop...so that's a good thing.


Just what is "Sync width", I realize it's the amount of time the sync pulse is active (both H & V)...but I'm not understanding what that means

If I have a V-front porch of 4 pixels...and a V sync pulse of 3 pixels...just what is happening in this time/pixel space???

 

[proudx]

adjust the horizontal geometry porch settings in powerstrip.


Keep front and back porch below 80, sync width should have little effect.


This will def widen the picture. I have seen this problem on a Barco 1209.

 

[KennyG]

The one thing that worries me is the H-back porch...this time/pixel count is used to stabilize the gun as it starts to draw it's next liine...is 80 pixels enough to allow for this stabilization before the active image starts drawing?

The Ampro will do it, I had it down that low last night, just to see what happened. (Nothing)

What I really wonder is how wide the Sync width has to be, since I don't really understand how this works.

If the H-sync width is actually the amount of time the H-sync pulse is active...and this activity takes place during H-blanking...then how do we determine the optimum Sync width? Should it be half the length of H-blanking...one quarter???


Perhaps I will set aside more time to lower this settings even more tonight.