Modding the JVC TM-A13SU Monitor

[Bryan]

My favorite replacement for the old school (1702-like) monitors is the JVC TM-A13SU. This monitor is a descendant of the TM-13U which is the monitor the 1702 is based on. Like the monitors of old, this one features a standard resolution P22 tube.

 

The TM-A13SU is lighter and easier to carry than most 13-14" CRT monitors and has automatic PAL/NTSC switching (it appears to do SECAM as well!). It also has an OSD, front input switching, and both composite and s-video inputs. If you're looking for an authentic experience from a newer monitor, this is the one to get...with one caveat:

 

Unfortunately, the TM-A13SU employs line blending in both NTSC and PAL mode. This causes weird interactions on horizontal color boundaries and although it is a common feature of PAL sets, it's usually not seen here in NTSC-land. Also, this unit uses comb filtering to remove the chroma carrier from the picture which can cause dot artifacts when used with computer video where every line can have unique colors (as opposed to camera video, where color changes don't tend to happen abruptly on line boundaries).

 

Looking through the user and service manuals I don't see any way to override this feature, but it appears to be a setting (N-COMB) sent from the onboard microprocessor to the TB1226EN video processor. My goal is to install a small microcontroller which can override settings as they are sent. I plan to use the smallest Atmel AVR I can.

 

Anyway, here's the service manual for the TM-A13SU.

TM-A13SU.pdf

[Bryan]

By the way, these can frequently be had for next to nothing as they don't have the general appeal of older computer monitors or the higher-end professional models. I just won a backup unit for $5.50.

 

eBay Auction -- Item Number: 141862843850

[Bryan]

I opened the monitor today and installed wires to monitor the SDA and SCK signals (I2C). I also put wires on to pull 5V from the microprocessor's regulator. While I was in there I made a dump of the settings EEPROM and installed an 8-pin socket since you're royally screwed when your device gets amnesia. There's also a picture here of the guts of the luma comb filter module for NTSC mode (MD201). The white thing with black dots in the back is a delay element. It literally works mechanically by sending waves into the wafer and picking them off at another point.

 

It was nice to see all name brand caps in there (mostly Rubycon). Here's the contents of the EEPROM:

00000000  ED 00 00 23 00 DC F2 00 00 05 03 00 FF FF FF A5     #            
00000010  00 00 00 00 00 00 00 00 00 00 00 FF FF FF FF FF                  
00000020  93 57 85 87 82 46 43 4D 32 77 45 FF FF FF FF FF   W   FCM2wE      (SIGNAL BLK/WB BLK)
00000030  08 32 1F 7F 0D 04 FF 00 00 FC FF FF FF FF FF FF   2               (DEFLECTION BLOCK)
00000040  06 3F 3F 3F 3F 00 1E 32 32 0C 05 00 01 01 04 01   ????  22        (CONTROL BLOCK)
00000050  02 02 02 02 96 0F 01 FF 00 00 01 00 00 01 00 00                  
00000060  00 00 00 04 06 01 00 01 01 02 02 01 00 00 00 00                  
00000070  00 00 00 00 00 07 02 00 02 FF FF FF FF 4A 56 43               JVC
00000080  00 00 00 00 00 01 00 00 00 00 FF FF FF 4A 56 43               JVC
00000090  FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF                  
000000A0  FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF                  
000000B0  FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF                  
000000C0  FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF                  
000000D0  FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF                  
000000E0  FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF                  
000000F0  FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF                  
00000100  FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF                  
00000110  FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF                  
00000120  FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF                  
00000130  FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF                  
00000140  FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF                  
00000150  FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF                  
00000160  FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF                  
00000170  FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF                  
00000180  FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF                  
00000190  FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF                  
000001A0  FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF                  
000001B0  FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF                  
000001C0  FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF                  
000001D0  FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF                  
000001E0  FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF                  
000001F0  FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF                  

[Bryan]

One last update for today. Here's the culprit.

 

Register $1E (decimal 30) is written with $80 (128) as soon as a valid signal comes in. Bit 7 is the comb filter bit. The microprocessor constantly polls the TB1226EN for status changes then sends a giant configuration block once a new signal is detected. The task will be to snoop on the line to detect NTSC signals and then force bit 7 low when register 1E is sent. Once I do this, the dot patterns along the top and bottom edges of colored objects should vanish.

 

EDIT: I've added the TB1226EN datasheet.

TB1226EN.pdf

[Bryan]

Been working on the program...

 

The ATtiny's USI has the ability to sniff an I2C bus (so you don't have to bit-bang it...much). The program will recognize the video status being read from the TB1226EN, and register 1E being written. I've also decided to make the mode toggle by holding down the PHASE (tint) button on the front for 2-3 seconds. The current mode will be stored in the AVR's EEPROM.

 

The plan is to make it an 8-pin chip you glue to the underside of the board and run 6 wires. I'll ship 'em cheap to anyone who wants one.

[+Stephen]

This is an insanely cool project.

[Bryan]

The i2c sniffer/injector program is working and I can now compare the results. The attached picture shows the how the top line of a color change is affected. In the top 3 strips, the color is either very pale from being mixed with a black line, or is mixed with the color above it. The 2nd set is greatly improved in this regard. I noticed there were still visible dot patterns in the top lines, however, which means that the filter module is causing it. Because its output goes into a video switch, I can probably control it on the fly as well.... stay tuned.

 

(this is all that needs to be done for perfect edges in S-video. The dots only appear in NTSC composite mode.)

 

EDIT: I was looking through the I2C traffic and in NTSC mode, the TB1226EN's chroma trap is disabled, meaning it relies solely on the external comb filter (and sure enough, bypassing it causes lines to appear in the screen). The trick will be to bypass the external one, and enable the internal one for comparison. Oh, SECAM mode is also disabled. Don't know if it would work if turned on or not.

[_The Doctor__]

damn fine modification,

How is the monitor for 80 columns... is the dot pitch fine enough to avoid the crappy moire patterns the commodore monitors produce, I hope this has the better dot pitched tube.

[_The Doctor__]

did you wire up a y/c cable for it...

[Bryan]

damn fine modification,

How is the monitor for 80 columns... is the dot pitch fine enough to avoid the crappy moire patterns the commodore monitors produce, I hope this has the better dot pitched tube.

 

Thanks! IMO this tube is a little bit nicer than the ones in most '80s monitors, but it's still consumer-level and not high-resolution by any means.

 

did you wire up a y/c cable for it...

 

Yes. The only problem in s-vid was the line blending that made the top line of everything look stupid. When I get back to the UAV upgrade thing, I'll get better pictures of all modes. Right now I'm using an XEGS for testing.

[Bryan]

I got the comb filter disabled and the color trap enabled so we've finally got old-school monitor mode. This is the result. There are no more dots or weird edges. It's amazing how many options are actually present in a modern(ish) monitor even if you're not given access to them. I'm going to make these options available to the user so you can try them. Time to go to bed!!

[RevEng]

Excellent hack, Bryan. Enjoying every bit of this!

[Bryan]

Excellent hack, Bryan. Enjoying every bit of this!

 

Thanks. I didn't really know how many people would be interested in something like this, but I figured the JVC was the perfect, easy to find replacement for the aging Commodore/Amdek/Sakata/Teknika... monitors with multi-system compatibility to boot. IIRC, these were made into the mid 2000's and I still see NOS ones on ebay from time to time. However, these little color issues have been driving me nuts for a while and testing the UAV board with it made it even more annoying. This thread was my way of getting off my butt and fixing it.

 

I prefer not to use professional SMPTE-C monitors (which includes most of the metal-box ones) for regular use with home computers because although they have a very clean picture, they also have a different look from consumer CRTs (just as LCDs do). The reason they exist is that the chemistry of the phosphor favors a better color spectrum (gamut) over intensity and they never really achieve that CRT glow. You can crank them a bit brighter, but it's hard on the tube.

 

Anyway, my idea at the moment is that there will be a wire going to the PHASE button (far left) and one to the Power LED. When you hold PHASE for 2 seconds the mode will change and the LED will blink 3 times to show the selected mode:

 

1. NTSC Filter Mode, 2. NTSC Blending Mode, 3. PAL Blending Mode

 

A long blink means the default selection is in effect:

 

1. NTSC Comb Filter, 2. NTSC is Blended, 3. PAL is Blended

 

A short blink means the alternate selection is in effect:

 

1. NTSC Color Trap, 2. NTSC is Not Blended, 3. PAL is Not Blended

 

Each press of the button advances the mode for the video type being viewed and the results are saved in EEPROM (I may put NTSC filter and blend on different buttons rather than sequence through all 4 modes).

 

-Bry

[+Stephen]

I got the comb filter disabled and the color trap enabled so we've finally got old-school monitor mode. This is the result. There are no more dots or weird edges. It's amazing how many options are actually present in a modern(ish) monitor even if you're not given access to them. I'm going to make these options available to the user so you can try them. Time to go to bed!!

Wow - that looks awesome.

[Bryan]

I've gotten a little more work done on this. I've rewritten the I2C sniffer program in AVR assembly, and it will have a few more adjustment options like the ability to adjust the parameters of the color trap filter. At this point I mostly need to finish the user control (menu) part of the code. When it's done it will be a 8-pin chip with 7 wires to the board (2 of them with resistors).

[Almost Rice]

I got this monitor back in 2010. I got it because it could do PAL.

[Bryan]

Well, things just got a lot simpler now that I'm finally going through all the I2C traffic. The datasheet for the video chip is really cryptic but PAL line blending can't be disabled, so I'm dialing it back to the original mission of fixing the NTSC modes.

[Joey Z]

wait, does the line blending in NTSC mode mean that the APAC modes would work?

[Bryan]

wait, does the line blending in NTSC mode mean that the APAC modes would work?

The mixing definitely improves the look of APAC modes. I'm working on having it where you can turn it on or off separately for composite and S-video modes.

[Mathy]

Hello guys

 

It's a pity SCART wasn't popular in the US. While not every SCART connector has I2C, there are pins reserved for it. That would mean that you wouldn't even have to open the monitor to get to the I2C bus.

 

Sincerely

 

Mathy

[Bryan]

Hello guys

 

It's a pity SCART wasn't popular in the US. While not every SCART connector has I2C, there are pins reserved for it. That would mean that you wouldn't even have to open the monitor to get to the I2C bus.

 

Sincerely

 

Mathy

Interesting. Would it give you the internal chip-to-chip bus of the monitor (I can't see why anyone would expose that), or some kind of external control interface?

[Mathy]

Hello Bryan

 

I guess it would. Otherwise there would be two I2C busses inside the device, which in my opinion doesn't make sense.

 

Before the video cassette recorder went out of production, many had the ability to read, in which order the TV stations are stored in the TVs memory and vice versa. (although, of course, many manufacturers used their own protocol) If that's possible, it should be possible to set some other stuff.

 

Sincerely

 

Mathy

[Bryan]

In this monitor, there wouldn't be much use for it because the bus master is a MPU that just crams the registers in the video controller every few milliseconds so all you could do is watch it.

[Bryan]

My design is finished! My final control scheme is:

 

PHASE - Toggles Line mixing mode. The LED will flash once for ON and twice for OFF

CHROMA - Toggles the comb filter. The LED will flash once for ON and twice for OFF

CONTRAST - Saves the current settings. The LED will flash 4 times quickly

 

Each button must be held down 3 seconds

 

While using Composite video, you can press PHASE, CHROMA, or CONTRAST

While using S-Video, you can press PHASE or CONTRAST

 

The mixing setting is saved separately for Composite and S-Video.*

 

I will post the source code shortly (after final testing and comment clean-up). If anyone wants a pre-programmed AVR, I've got some spares to send out for $3 + shipping.

 

-Bry

 

 

* Right now I detect S-Video mode using the I2C messages. I could instead allow separate settings between input A/B but it would require another wire to the board and I'd have to disable the RESET pin on the AVR, meaning it could not be easily reprogrammed after that.

[Bryan]

Here's the project with source and hex file. I'll post installation pictures soon.

 

CPU flags are set to Brown-out at 2.7V and Internal RC @8MHz 6CK/14CK + 64ms.

jvc_mod.zip