Jump to content
Due to a large amount of spamers, accounts will now have to be approved by the Admins so please be patient. ×
  • 0
IGNORED

RGB hacking a consumer TV - Samsung Hi-Tron TV-488


buttersoft

Question

For a guide on how to do this right, see here - http://www.aussiearcade.com/showthread.php/86693-RGB-hacking-consumer-CRT-TV-s

 

EDIT: This hack is not a good thing to do. I have since learned that using the RGB outputs of this chip might throw the auto-beam-current limiting out of whack, which means the HV can drift. So, yeah, more X-rays and possible damage. Thank you to Tim Worthington for pointing that out. There's a better thread on shmups about RGB hacking. Tim was also kind enough to point out a way I might be able to reprogram the EEPROM that controls the set-variables, and I'll update the thread if it works for me. It's not going well so far.

 

UPDATE: Nope, the sets EEPROM was a Samsung special, with a one-time-programmable lower buffer. Did not expect that. The EEPROM theory is sound though, and it'd be interesting to play around with another set like this.

 

 

To the admin: I'm kinda asking for help improving this effort. I hope it's ok to post it. If not, pls (re)move it :)

 

 

WARNING WARNING WARNING WARNING WARNING

Even when unplugged, a CRT tube contains enough energy at high tension to kill you! And the power supply section of a TV or monitor might too. DO NOT attempt to de-case or modify or even touch a CRT tube unless you know how to do so safely. Emulating actions described in this write-up may lead to death, injury or damage. Neither I nor the AussieArcade site take any responsibility for this whatsoever.

WARNING WARNING WARNING WARNING WARNING

 

Part 1

To Begin

This is an account of hacking RGB into one particular Australian consumer-grade PAL CRT TV, that had only composite and RF inputs accessible from the outside – the Samsun Hi-Tron TV-488. Fair warning: I’m long-winded. If you want, you can skip to Part 2 and jump right into the action.

 

I can’t say how similar this set is to any other models or brands, as I’ve done this once. There are whole boatloads for me still to learn. I don’t expect every claim I make to be correct, and I don’t claim the method I use for the hack is perfect, or that there aren’t more obvious, easier ways.

 

I’ll try to explain why I did what I did in an FAQ, but if you can see something I could have done better or easier, let me know! The community as a whole can benefit from more information about this whole idea.

 

(I’m going to assume you can solder, that you know a little about electronics, video signals in particular, and that you already have a 15kHz source putting out R,G, B and Composite Sync of some form. Composite Vid and Luma might also work for that, maybe through an LM1881 chip if needed.)

 

I intend to write 3 parts. Doing this should help provide more information than if I simply recorded where I ended up. The method is actually pretty simple, but by itself may not explain much. I realise it’s going to be teaching some people to suck eggs, and for that I apologise. Other write-ups of RGB hacking seem to be heavy on pictures but light on details. No criticism is intended there, they're a lot better than there being no information at all!

 

The parts are:

1 (you’re reading it). The Preamble above, the premise, and some general info as I understand it.

2. What I actually did

3. FAQ explaining why I ended up taking that route, drawbacks, etc.

 

 

The premise is that every CRT TV has at its core an integrated circuit, or chip, nicknamed the “Jungle” IC. Whether this is true for all sets, or perhaps only for ones from about the mid 80’s onwards, I don’t know.

 

In short, the Jungle IC is responsible for taking video signals and de-matrixing them back into R, G & B to send further into the workings. It normally handles the sync, and a lot of other things too. If you feed it RGB to begin with, it adjusts this and passes it along, internally or externally as needed. And you can feed it RGB, somehow. At least, that’s what gets bandied about. Some sets actually have the Euro-SCART cutout piece ready to go, with composite inputs in the outline just sitting there mocking you. Either way, the Jungle IC will probably take RGB. For one, most sets need RGB internally. For two, it was usually cheaper to just use the same chip and PCB worldwide, and alter the casing by region. It’s also likely that the extra components needed to filter the RGB signal between the SCART input and the Jungle IC are missing to save extra cents. Those missing parts are probably breaks in the circuit, so finding the RGB feeds near the back of the set, where the other AV inputs are, isn't likely to do you much good.

 

There’s not one type of Jungle Chip or anything; it’s a generic term for the video chip inside a CRT TV, and lots of companies made them. Most IC’s in consumer electronics are helpfully marked with their maker, model no., and orientation. Googling the datasheet will help you identify what you’re looking at.

 

Finally, CRT TV’s are from an era when surface mount components weren’t really used (Yes? No?). All the ones I’ve opened up use through-hole resistors, capacitors, et al. This makes the boards roomier. Components are easy to change and/or solder to/around, and it’s easier to follow traces on the PCB with your eyes. It also leaves a lot of room for markings and labels, the reading of which can help you nut things out.

 

That’s enough background for now. The next post will describe what I actually did.

 

- - - Updated - - -

 

WARNING WARNING WARNING WARNING WARNING

Even when unplugged, a CRT tube contains enough energy at high tension to kill you! And the power supply section of a TV or monitor might too. DO NOT attempt to de-case or modify or even touch a CRT tube unless you know how to do so safely. Emulating actions described in this write-up may lead to death, injury or damage. Neither I nor the AussieArcade site take any responsibility for this whatsoever.

WARNING WARNING WARNING WARNING WARNING

 

 

Part 2

The Hack

1. The Set – Samsung Hi-Tron TV-488

(Also called TV4886X/XSA.)

 

I rescued this 20" set from a local E-Waste bin. Powering it on, it showed the worst composite video image I’d ever seen, with the red streaking into everything around it and shimmering badly, for one. But it worked, and it was nice and bright with a great black-level.

 

(Apologies for terrible internet pic, mine no longer has its shell)

m05Dgxs.jpg

 

 

2. The Chassis

Decasing the set shows the two main IC’s

0IEwMXF.jpg

 

The Zilogic IC I will talk about in the FAQ in Part 3. There’s also an audio chip under that white plastic strut at the bottom of the picture, identifiable by its one-sided form factor that I forget the name for.

 

 

3. The Jungle IC

The TDA8841 was clearly marked, though I had to google the datasheet to determine it was the jungle chip, of course. That datasheet game me the chip’s pinout.

rdMVqvf.jpg

 

The datasheet:

https://my.mixtape.moe/hiflqk.pdf

zHSxr6C.jpg

 

The datasheet also noted that the R, G and B outputs of the chip (pins 19, 20 & 21) were at 2Vp-p (or 2 volts DC, peak-to-peak, rather than the expected 0.7Vp-p for RGB). Taking the chance that this 2Vp-p output was just amplified RGB, as the datasheet implied, I used two THS7314 SD video amplifier chips DC-coupled back to back - the first to take the signals on my R, G, & B lines from 0.7V to about 1.5V, the second to go from 1.5V to about 2.5. The outputs of the first stage are plugged directly into the inputs of the second. I then added 1K linear potentiometers to the output lines of the second stage, one for each line, in series, so I could control the voltage. A wire from each potentiometer’s output went to the R, G and B output pins respectively. The pic above shows those wires soldered in place.

 

After that, all I had to do was feed C-Sync into the frontside AV input of the set, and switch to the AV channel so the set found the feed.

 

And voila, RGB picture :)

 

(Pic below is during testing, red line only connected, along with sync. OSD shown for no particular reason)

23i6oJa.jpg

 

- - - Updated - - -

 

WARNING WARNING WARNING WARNING WARNING

Even when unplugged, a CRT tube contains enough energy at high tension to kill you! And the power supply section of a TV or monitor might too. DO NOT attempt to de-case or modify or even touch a CRT tube unless you know how to do so safely. Emulating actions described in this write-up may lead to death, injury or damage. Neither I nor the AussieArcade site take any responsibility for this whatsoever.

WARNING WARNING WARNING WARNING WARNING

 

 

 

Part 3

FAQ

This is where I’m going to answer specific questions with specific answers, with the hope of shedding light on RGB hacking in general.

 

 

Why go to the trouble?

I needed a tube for a GroovyMAME cab, and I’m broke ATM. Besides, I wanted to see if it could be done, and to learn how.

 

 

Is the picture any good?

Yep. Not quite perfect, though, I’ll admit. The much older tube I have in another cab is just slightly crisper, though nowhere near as vibrant. It was probably owned by a little old lady who only drove it to church on Sundays. I’ve also changed out some of the neckboard cables because the ribbon shit Samsung used had to be crosstalking, and that cleaned things up a little too. Any further suggestions on noise reduction are welcome.

 

Doing this on a decent TV to begin with would have been a much better idea. I guess I didn’t expect it to work ;) (and for a better set it may not…)

 

 

Why are you hacking into the RGB outputs of the TDA-8841 jungle chip, why not feed the jungle chip through its RGB inputs?

That would be a better way round, yes. I tried, both with and without the blanking input on pin 26 being fed from 0 to +12V. The blanking input works fine, but the RGB still doesn’t show. (With the method I did go with, the blanking input is not necessary as there’s no video signal going in that needs blanking, whether composite or anything else.)

 

The set seems to be controlled by the Zilogic chip pictured in post 2. I suspect it interfaces with the TDA-8841 via I2C to set the RGB permissions, does the service menu, changes channels, and a whole host of other things I can only guess at. I tried feeding the inputs stencilled as AV and SCART with 5V, and then with 8V, but nothing happened. I’ll keep tracing the PCB though, and see what I can find out. But the Zilogic is probably programmed, I don't know because I couldn't find the datasheet for it, and that’s where I top out. This is also why I didn’t bother trying insertion over the RGB mask for the menus.

 

Samsung might have commissioned a really cheap batch of these low-end Jungle chips with the RGB inputs turned off like the YPbPr inputs already are, though I doubt that because the menu insertion is there. Then again I might be an idiot. It might be straightforward and I just failed to figure it out. Hmmm, maybe trying a combination of the above insertion inputs at once is the key...?

 

 

What are the disadvantages to doing things this way?

The OSD menus show up, but do nothing. I’m not feeding anything into the TDA-8841’s inputs for them to influence. I can adjust gain and contrast via the potentiometers in the R, G & B lines, and brightness via the screen pot on the flyback. For bias control I could probably level-shift the video signals somewhere, but I’d have to figure out how to actually do that :)

 

The V-size and similar controls would still work if I had the remote and knew how to get to the service menu. They’re probably accessible via the Zilogic chip too, of course.

 

Also, feeding Sync into the AV input, and then the RGB into the outputs of the Jungle chip means they’re out of step. The TDA-8841’s datasheet claims a process delay of 20ns, I think it was. Either way my picture is shifted about 2-3mm left (earlier) on a 20” set. This will be easy to correct or move with ArcadeOSD from the crt_emudriver package, which happens to be what I’m using.

 

 

Why did you use the satellite AV feed for sync, why not the main AV feed on the rear of the set?

I don’t have the remote; I couldn’t switch to it. The auxiliary AV input on the front was at the end of the channel range, after chan 13 and before chan 1 in the loop, and I can change channels with the buttons on the chassis. I did try feeding voltage into the Zilogic at the pin marked AV, but nothing happened. Or nothing I could determine.

 

 

Does this mean you’re stuck with having to push a button every time you turn the set on?

The set remembers the last channel it was on before the AV channel. I’m going to use two timed relays in series, soldered across one of the buttons. One to close 1 second after power-on, the second to open 1.1 seconds after. That should give me a .1 second impulse, and change channels for me. I’ll post back with results after the relays arrive.

 

http://www.ebay.com.au/itm/191620243964?_trksid=p2057872.m2749.l2649&ssPageName=STRK%3AMEBIDX%3AIT

 

 

Why not feed C-Sync directly into the Jungle IC like SCARTHunter did?

My Jungle IC doesn’t seem to have that function, though I might have glossed over it. Or it’s through the Zilogic again. I did trace the composite video line from where it enters the set, and it split in two – one side to the TDA-8841’s composite input, one to a bank of transistors I haven’t finished tracing. I’m fairly certain the TDA-8841 is outputting the horizontal and vertical pulses to keep the video in sync, but damned if I’m certain how it takes the input for those just yet.

 

 

What if there is a video signal going in/what if I feed the set composite video for sync?

Try pumping 5V into the blanking input on pin 26 of the TDA-8841 Jungle IC. Or using an LM1881 to get clean sync, or any other type of sync-stripper. But if you’re attempting this you shouldn’t need to pay top dollar when a cheap IC will do.

 

 

Why did you DC-couple the THS7314 video amplifiers?

With AC-coupling I didn’t get the output voltage needed, and I didn’t want to add a third stage and introduce more noise. Pullup resistors did not work, and the THS7314’s I had on hand top out at a +5.5V supply. DC-coupling worked perfectly. TBH there are no components at all in the circuit before the potentiometers after the final outputs. Just RGB into one chip, out into the second chip, then out to the pots and from there to the output pins of the jungle chip. (Along with +5V and ground into each amp, of course). Added to which, the set has no earth. Thus the ground can be common between the set, the amps and the laptop producing the VGA signal, with no fear of ground loops. I want to explore the CMRR more though.

 

EDIT: In addition, the PC, the amp, and the TV set all share a common ground, and the only path to Earth is through the PC's PSU. With that reference level providing black, i'm not actually sure AC coupling would achieve anything beneficial.

 

 

No 75R resistors to ground? (No terminations?)

Nope. I'm not using standard 75R impedance lines after the first THS7314; or not that I'm aware of. Anyway, the picture isn’t getting any ghosting. I tried resistors in various places and there was no discernible difference in picture quality, where trying previously to build a PC CRT amp I was basically seeing double without them. Also, see the DC-coupling question above for brightness reasons.

 

 

Can I use the same method with my TV?

If it’s the same TV, sure. You’ll need to know your video and sync levels, of course, and how to adjust them. And how to be safe around a CRT, which I cannot stress highly enough.

 

If it’s a different set, I honestly have no idea. Go slowly. Be careful.

 

If it’s a different set using one of the TDA-884X/TDA-885X IC family for its jungle chip, then… perhaps. These chips basically output RGB at elevated levels. The TV-488 set runs the RGB lines through a single resistor each, and then to the neckboard. Simple as.

 

I believe the electron guns in a CRT TV set need to be driven at about 100V each (Yes? No?) but this level shift seems to happen on the neckboard itself for the set I used. For a different set this may happen on the PCB, though I think that would be unusual. Or the neckboard might be taking some slight variant of RGB. It may be dependent on the number of pins on the tube. I’d love more information myself.

 

 

Is there any risk in doing all this?

Other than the big red electrocution warnings all over this post, you mean? Yes, there is. I wasn’t able to measure the current coming out of the TDA-8841 under normal conditions. The THS7314 amps are limited to a 90mA output though, and the video signals coming out of a PC are sometimes double that. Also, I might be irradiating myself more highly by delivering too powerful a signal to the neckboard. I really, really don’t think so, knowing even the small amount I do about CRT design, but I’ll admit the risk is there.

 

In short, I can’t really advise you. All I can say is that you should read all relevant datasheets very carefully, trace the paths on the PCB over and over to make sure you know what you’re expecting, etc, after first acknowledging the following disclaimer… :)

 

WARNING WARNING WARNING WARNING WARNING

Even when unplugged, a CRT tube contains enough energy at high tension to kill you! And the power supply section of a TV or monitor might too. DO NOT attempt to de-case or modify or even touch a CRT tube unless you know how to do so safely. Emulating actions described in this write-up may lead to death, injury or damage. Neither I nor the AussieArcade site take any responsibility for this whatsoever.

WARNING WARNING WARNING WARNING WARNING

Edited by buttersoft
Link to comment
Share on other sites

13 answers to this question

Recommended Posts

  • 0
I may have implied you didn't provide enough info on your blog, but it was your posts that mainly led me to believe it was possible, and from there to give it a go :)

 

I definitely didn't provide enough info!!! [emoji14]

 

I have had a step-by-step guide in draft form for a year now... documentation always takes ages and I've been too busy building cabs to actually finish it. Maybe one day...

 

Anyway, it's great to see others experiencing some success with this. [emoji3]

 

 

 

Sent from my SM-A300Y using Tapatalk

Link to comment
Share on other sites

  • 0
Great write up, I saw a 34CM model of this TV a couple of weeks ago wish I grabbed it now. On one of my rescued TVs (Akai) to keep it in AV I put a piece of Blue Tac on the AV button on the TV and now it starts up in AV mode each time.
Link to comment
Share on other sites

  • 0
Great write up, I saw a 34CM model of this TV a couple of weeks ago wish I grabbed it now. On one of my rescued TVs (Akai) to keep it in AV I put a piece of Blue Tac on the AV button on the TV and now it starts up in AV mode each time.

Yeah admittedly i haven't tried that yet. No AV button on my set, so I plan to solder some wires to one of the channel buttons and try crossing them at startup. I figure it's either going to do nothing, or keep rolling channels. If it works though, i'll have wasted $4 on the timed relays :)

Link to comment
Share on other sites

  • 0

Thanks for the write up.

 

I'm guessing on a 100Hz TV, that the Jungle IC will be before the 100Hz processing? Just thinking out aloud on how to utilise a Loewe 100Hz set that I have (which has a nice dark tube, but is otherwise no good).

Link to comment
Share on other sites

  • 0
Thanks for the write up.

 

I'm guessing on a 100Hz TV, that the Jungle IC will be before the 100Hz processing?

 

Hmmm, interesting question. I'd guess, and it is just a guess, that the Junlge IC is last. I reckon a consumer set (not a true multiformat set) will take any input and turn it into 31kHz (or whatever higher sync rate it uses) by linedoubling it (and/or upscaling it), and probably feed identically-timed signals to the Jungle chip. Whatever you feed in, the circuitry turns into 31kHz+.

 

I suspect that there's no way the circuitry provided in the set is going to deliver 15kHz to the tube, and if you managed to bypass the doubler or whatever it is and do so, the set is not going to be very happy. That said, I don't know what a set does with Sync internally, so I could be entirely wrong. Obviously it scans the beam back and forth across the tube using the deflection yoke as an electromagnet, timed to the sync pulses, but what pulls the beam down each h-line, what controls the retrace? Not the timing, but the geometry of it, how far and at what angle. What controls that voltage?

Link to comment
Share on other sites

  • 0
Thanks for the write up.

 

I'm guessing on a 100Hz TV, that the Jungle IC will be before the 100Hz processing? Just thinking out aloud on how to utilise a Loewe 100Hz set that I have (which has a nice dark tube, but is otherwise no good).

Yep, and the deflection circuitry in that set will be designed around a 31.5kHz scan rate, itll do 50Hz and 60Hz, but will line double the 15.7kHz image.

 

- - - Updated - - -

 

Hmmm, interesting question. I'd guess, and it is just a guess, that the Junlge IC is last. I reckon a consumer set (not a true multiformat set) will take any input and turn it into 31kHz (or whatever higher sync rate it uses) by linedoubling it (and/or upscaling it), and probably feed identically-timed signals to the Jungle chip. Whatever you feed in, the circuitry turns into 31kHz+.

 

I suspect that there's no way the circuitry provided in the set is going to deliver 15kHz to the tube, and if you managed to bypass the doubler or whatever it is and do so, the set is not going to be very happy. That said, I don't know what a set does with Sync internally, so I could be entirely wrong. Obviously it scans the beam back and forth across the tube using the deflection yoke as an electromagnet, timed to the sync pulses, but what pulls the beam down each h-line, what controls the retrace? Not the timing, but the geometry of it, how far and at what angle. What controls that voltage?

It wont really make any difference if the image processing is done before or after, almost none of those sets will scan to anything other than 31.5kHz or very close to it, some will do 33ish kHz for 1080i, but either way, youre stuck with the absolute s*** excuse for a line doubler/field multiplier/deinterlacer the manufacturer saw fit to throw in there to reduce visible scanlines and flicker. Just makes a terrible blured image and adds a heap of lag. Never seen one that look good in a TV, even the Barco LIMO PRO in my Cine 9 is pretty average, but far better than any others ive seen.

 

A good image processor is still a couple grand, a fair one could be had for around $500. The ones installed in consumer grade TVs is an absolute joke, and they shouldve been dual sync instead.

Link to comment
Share on other sites

  • 0
The ones installed in consumer grade TVs is an absolute joke, and they shouldve been dual sync instead.

 

Would've been more expensive I guess?

 

Oddly enough I was persuing another forum today in which a particular shitposter claimed that to make a set multiformat/multisync it needs multiple windings on the flyback, into which concept i'd lump multiple taps on the one winding or set thereof.

 

That would no doubt make things more expensive, but it also feels like complete bs to me. Anyone know - does a tri-sync chassis/set change voltage greatly when syncing to it's different rates? Surely not, surely it's just the circuitry...

Link to comment
Share on other sites

  • 0

The flyback has f***all to do with it, it feeds a given voltage to the tube, and thats all it does.

 

Majority of my experience and knowledge is with front and rear projection CRTs and the calibration of them, and im yet to see one that changes HV level when scanrates change. There is not a huge amount of difference in the workings of direct view sets and projection sets as far as that side of it goes. Ive calibrated HV level on a few of them, and it makes no difference what the scanrate is.

 

Some CRT projectors do have different windings on the deflection yokes that are switched with different scan frequencies, but its more complex than that, and most of them are multisync, they will generally lock to anything between their given range.

 

Increasing HV level will reduce raster size, as the increased beam speed toward the tubeface requires higher deflection force. It will also affect beam focus, which on a direct view set is almost always electro-static, and adjusted with a trimpot.

Link to comment
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
  • Recently Browsing   0 members

    • No registered users viewing this page.
×
×
  • Create New...