Jvc gy-hd*** camera family re-attaching blue sensor.

[Bob Hart]

If one was to design and build an alternative attachment system for the many detached blue-channel sensors out there, would there be anyone interested I wonder in having theirs done, given the cameras are quite dated nowadays and their comparatively low value on eBay.

The camera will operate with the right side panel off and just by hand manipulating the detached sensor PCB after first attaching it roughly by the threaded jig end to thin cardboard, one can get pretty close to a good image. One has to be very careful to avoid the conductors on the PCB shorting across the the green channel PCB. This time, I dodged a bullet before I thought to put an insulating strip in there but permanent damage could be caused.

It might be possible to reattach the sensor with adhesive to the outside onto the black-painted portion of the prism but I suspect the new mount system may have to be by the jig strip as larger glue bonds may shrink and pull the sensor off alignment.

A large monitor screen would be needed for final trim to get rid of blue edges.

[Bob Hart]

Footnote:

There was a little white indicator arrow popped up on the LCD screen which I had not ever seen previously and suggested which direction the sensor was not in vertical alignment. One time I got it very close, the image seemed to come in by itself and the arrow extinguished. I cannot find anything in the service manual about it.

I wonder if these things automatically track the images across the sensor to align the colour channels so long as the sensors are aligned within a certain tolerance. It would be clever stuff for the times the camera was made. Apparently there was also an automatic function evolved for adjusting the split screen effect which aggravated the early adopters of the cameras.

If there are any retired JVC camera techs out there, I would be interested in your responses.

[Bob Hart]

For the curious among readers, from appearances, it sees that the sensors are attached by temporary mounting strips and wedge ends to an assembly jig. There are eight holes altogether some threaded, two elongated as if to allow the mounting strips to slide for horizontal alignment and skew slightly for roll adjustment.

These appear to be there for assembly of the PCBs to the actual sensors as the holes are inaccessable due to the PCBs being in the way. I am beginning to suspect that the sensors are as accurately aligned to the PCBs as possible in fine jigs so that as little adjustment as possible is needed for the final assembly to the prism.

The wedge shapes in each of the temporary mounting strips are possibly there for an accurate clamping scheme with external adjustments to align the sensors then glue spots being added then the sensors offered up to the prism and clamped there until the glue sets.

A scheme to re-attach a dislodged sensor could be to use three of the screw holes at ends of the mounting strips to support one end of the loose sensor to the attached sensors and the wedge shapes at the opposite ends of them used for support into a home made matching support. The build accuracy would have to be better than five microns - big challenge.

[Bob Hart]

After messing around with a small plate dummied up out of thin cardboard mastered to a good but electronically faulty sensor block and manually adjusting the opposite end of the blue sensor PCB which remained floating, it seems that it may be possible to remount the blue sensor mechanically without adhesive although with adhesive would be a better fix.

Micro-adjustments may be possible by using screws with cone ends or countersunk heads as wedges which is a similar arrangement to the very old Sony EIAJ video head alignment system. The screws would attach through the mechanical mounts to bear against the original jig strip which remains attached to the sensor. I think those do double duty as heat sinks. I have yet to work out a method.

Any ideas from any JVC techs would be welcome.

[Bob Hart]

For what it is worth, here is a rather shoddy phone pic of the LCD screen of the image yield with the blue sensor relocated. I used roughcut bridges made from cereal packet cardboard with a razor blade, mastered to another intact sensor block. It was the labour of about 30 minutes with the sensor relocated by fingers alone and no adjustment screw system.

Misalignment remains on the left side of the image with a blue margin on the lower edge and yellow margin on the upper. Obviously much finer adjustment is needed for the defects not to be seen on a large display screen. You will observe in the bottom left of the screen an upwards arrow which indicates the direction the blue sensor needs to move for alignment.

One wonders how much the issue may have been little more than a lack of the will to make it happen on the part of JVC. Common sense suggests that replacement of the defective assembly is the preferred option because there then remains little chance of comeback due to an ineffectively executed repair, an expensive business to do properly when costs of labour and clean rooms are concerned.

A re-attachment regimen would also need special jigs and techs being trained worldwide. Replacement of the entire sensor block probably makes more sense.

The JVC GY-HD*** camera family had its teething problems but the fine miniature execution of everything built inside is sweet to observe. there was a lot of hard design work done. One imagines a bunch of little inscrutable techies with thick-rimmed glasses working slavishly for the corporation. One hopes they derived a certain amount of pleasure and pride in what they built at the time.

[Bob Hart]

For those who have the endurance to remain curious, any bridge pieces which are made have to be vented to allow the small fan to push air through. The little processors on the PCBs the sensor chips are mounted to can get hot enough almost immediately to be on the pain threshold.

The bridge pieces are best made of formica or laminex sheet which is a thin thermosetting resin similar to bakelite. I have established with these it is possible to get a good alignment but perfection will only come from screw adjustments, not finger pressure and barely tightened screws.

Once I have established a position for the loose sensor I will make up different mounts attached to four conveniently located screws on the rear face of the mount to confirm to the restored position of the sens and use that instead of bridging all the sensors together.

[Chris Hurd]

Thanks for the updates, Bob -- somebody will need this info someday.

Not me.

But somebody, somewhere will find it all highly useful.

[Bob Hart]

A small update.

Here is an illustration of a temporary support for the detached sensor which enabled me to establish just where in space it should be.

The next step is to build a different support attaching to two pairs of screws and is microadjustable and yet enables enough clearance that the small dobs of adhesive can be spotted upon the four corners of the sensor to prism junction.

The material needs to be viscous so that it does not creep or wick in between the sensor and the prism through capillary action.

The adhesive appears to be a non-brittle material similar to white water cleanup bathroom sealer.

The bridgepiece is cut from formica or laminex which is a material similar to thermosetting resin or bakelite. It is insulative to avoid any chance of electromagnetic fields being introduced into the original steel jig pieces which function also as heatsinks.

EMF seems to have been an issue because magnetic strips have been taped to the ribbon cables.

The bridgepieces block airflow so cannot be a permanent solution.

The far bridgepiece has been made to be a precise fit to another intact sensor block with three screws. There are elbows in the far jig pieces which have threaded holes which are conveniently are a match to the camera body screws.

This cannot be an accurate exemplar because each sensor itself is very slightly different and must be aligned before the adhesive is applied in factory. It does however get very close. The bridgepieces are cut to be a very slight interference fit to the jig pieces.

[Bob Hart]

Obviously much finer alignment is needed. The blue margin is a result of the sensor not being quite flat on the prism surface in the lower right corner.

[Bob Hart]

Here is a small piece of motion vision of the detached blue sensor being slipped into near-alignment. You'll need to be quick to spot it.

This clip is just a quick test of the temporary jig arrangement for re-positioning the detached blue sensor after using an intact sensor block as a master for making the bridgepieces. The next step is to make a three-axis micro-adjustment adjustment which moves in a fine controlled manner yet allows access to add viscous adhesive to the four corners of the sensor at the original adhesive locations.

YouTube

Uh-oh. Looks like I somehow get the clip link thing wrong. Here is the link with characters each end so that it appears in text not as a blank youtube screen. You will have to highlight copy and paste the link without copying the # marks

#####https://www.youtube.com/watch?v=lP1VN1WvIIc######

[Chris Hurd]

Trying a different linking method...

YouTube (Short URL)

*edit* -- Blast, that didn't work either. Looking into it.

[John Nantz]

Bob - This has been some really impressive work.

Here is another idea, granted its a long shot, just thinking out loud. Since you’re really into cam stuff, there is one area that might be able to use your expertise. The JVC LS300 is a really great cam and it has one weak point that many have brought up and that’s the LCD screen. If the screen could be upgraded as a retrofit it would make this cam really shine.

Given the price of new TVs has really plummeted one would think that the small LCD screens would have taken a drop too, It would be interesting to know if there is a manufacturer out there that makes a really good screen that could replace the existing one.

If there isn’t some sort of “standard” size then another thought would be if the flip-out part could be re-designed to accommodate a slightly larger and better screen, a la a small Atomos 1500 nit, that would be super.

And maybe at the same time, deal with the “loose hinge screw” issue at the same time, either with a different screw, screw and nut with lock washer, or using “Loctite” or similar product.

After I took one small handicam LCD screen apart (it was a Sony with a broken ribbon cable) and seeing how it goes together (didn't look complicated), the rest might be just a matter of getting parts to connect up, both from a mechanical and technical standpoint. Curious how “standard”, or not, the screen data is. TV connections are standard but LCD screens probably not so much due to different data. And then there is the potential legal aspects of doing this. It would be a plus if JVC would team up to help.

[Bob Hart]

John Nantz. I don't know that camera model. My sense is that the chipset in the camera itself has been programmed to work only with that particular LCD screen as built in to the camera.

I am not a camera tech, just somebody who as us aussies say, "has a go". The only reason I am wandering around inside these cameras now is that they are of little actual worth and the risk of loss is not so much of a heartbreak.

Their primary reason for being is to recover an old tape archive and they can do that without the camera head being able to "see".

There would likely have to be firmware devised for the LS300 camera to drive another small monitor.

If it is the LCD display panel itself which fails. it is likely that there may be a part out there to be found. Time has gone by though and spares inventories are not kept as long as once they were.

The controller PCBs in the monitors for my SI2K lay down and died. The monitors were actually a third party product intended for backs of tour bus seats and as standalone monitors for a number of applications.

I found a product number on the little PCB and as luck would have it, with a web search I found replacements from yet another third party supplier.

Initially, the monitor's USB controller did not work. Then I found the latest driver and installed that, which fortunately on the SI2K one could do because it was an embedded Windows XPe system. and things went fine for a while. Subsequently the monitors went west again, possibly a weakness in their design.

[Chris Hurd]

Originally Posted by Bob Hart

Subsequently the monitors went west again, possibly a weakness in their design.

Things go west in Australia? Up here they go south. As in "subsequently the monitors went south again." It makes me wonder how and why certain regional colloquialisms evolve the way they do.

By the way, how does one go west from Perth??

[John Nantz]

Chris - you beat me to it! And said it better than I would have. “West”? That is a new one on me, too. They probably use the term “West” because its a lot further to Cape Horn than it is to Antarctica.

Bob - You’ve gone far further with more perseverance and knowledge than I could have so that’s the reason I sprung this idea and question on you.

With regard to JVC, I really liked the one I had, the GZ-HD7 and really hated to sell it. Model was released in 2006 but I was given it from our son-in-law in 2010 and used it up until a couple years ago, that’s a long time in camcorder life spans. It just plain worked really good. I really liked it (if you can’t tell).

The GY-LS300 is a really cool cam and probably only something that JVC would bring to market. After reading the thread about cinematic and ENG cams, the LS300 is kinda like both. It is small enough to be used hand-held yet has interchangeable lens capability. Why buy a new cam and pay for a new lens every time? It doesn’t make good economic sense. With the LS300 and a very inexpensive adapter I could use my Nikon lenses or any number of other lenses out there with perhaps a different adapter. The frosting on the cake is it can use prime lenses and I really like that. However …

The LCD screen is a concern. I do a lot of outdoor video so being able to see what you’re shooting and make setting adjustments is important, and I’m not alone on this. The part about the chipset and compatibility with the LCD screen is something I wondered about but it seems like it would take someone really knowledgeable about this aspect of the cam design. It would likely have to be someone involved in the actual cam design and that person probably wouldn’t be allowed to divulge anything. Bummer.

Well, it appears this idea went South, or is that West?

Edit: Just checked the Ninja V and it only has 1000 nits while the larger Flame has 1,500 nits.

Edit: My wife's letter scale had a problem where the batteries "leaked" and corroded it's innards. I used a battery terminal cleaner to help remove the corrosion but it leaked down into the inside of the scale. Took the scale apart, cleaned it, and it still didn't work and I suspected it could be something with the battery contacts. Disconnected the really small wires and with a couple AAAs on the table finally got it to "light up". Next step is to check the battery terminals with a VOhm meter then, if OK, solder the wires back onto the battery terminals. Wife complained that I spent too much time on this because you could buy a new 0.1 gram scale for $13. Well ... what's the fun in that??? At least this way I got to see how it works.