What's new
On my Egret 2 I used emphatics guide to calibrate. Got decent results. Got better results with the arcadeotakus guide on my Nanao MS9.

What brad808 said is true. Every chassis+crt combo is a little different. A lot of these guides are general guidelines not exact formulas for maximizing.

Recapping is also a good idea like ShootTheCore states as capacitors can have a pretty significant impact on how well your tube CAN function. Another thing to keep in mind is a lot of these tubes+chassis probably have a lot of hours due to their commercial use + being over 20 years old.
 
Last edited:
yea the chassis are recapped. @ChuChu Flamingo what exactly made the difference for you in the arcadeotaku guide? i mean all of them are kind of ok and with cps2 the picture is decent most of the time. things that bug me most are if the picture is just too dark or the black is too grey. but again the question is the approach. high or low screen pot for example seems the be quite controversy
 
If both don't show good results you might be doing it wrong, your cps2 A-board might need recapping (I have one board here that has dimer video than my others), chassis need recapping, or your tube just has a lot of hours. The later might just need a tube cleaning on the guns but this is the last thing you want to do before tying all the other options I stated.

A quick glance at the differences between the two guides

1.)Emphatics guide basically states to put everything rrgb gain/cutoff/brightness/contrast at a neutral state/12 o clock then adjust flyback volts then everything RGB gain/cutoff/neutral/brghtness/contrast.


2.)The arcade otaku wiki states turn RGB gain/cutoff/brightness/contrast all the way down to 0 then adjust flyback then adjust rgb gains/cutoff/brightness/contast.

The difference on my E2 was that using emphatics guide the cutoffs were at 50%. flyback was lower. Arcade otakus wiki the RGB cutoffs were at around 20-25% and flyback higher.
 
Yea I know. Basically the visible result should be the same. But the question is why do they approach total different. What is the benefit over the other. One guy states something like higher screen would result in let’s say less spread mir on point beam.
 
I've had a few people ask me for more details on the procedure for using a colorimeter to calibrate a cab CRT. Using a colorimeter is perfect for OCD people like me that want their cabs calibrated perfectly to professional standards with no guesswork involved.

This is a summarized procedure. The complete procedure with tons of detailed information and screenshots can be read here:
http://www.curtpalme.com/forum/viewtopic.php?t=10457

I use a X-Rite I1 Display Pro colorimeter - it's around $100 - $150 used. There are less expensive alternatives like the Spyder 4 that you can pick up used for around $30 - $50 - it just doesn't send measurment updates to the laptop as quickly.

I use the SNES 240P Test Suite running on a PC, Raspberry Pi or MiSTer connected to the cab with a JPAC for the test patterns. You have to use a Windows laptop to run the colorimeter and calibration software separately from whatever you're using for the test patterns.

One-time only software setup:
1) Download and install the free ColorHCFR software from here: https://sourceforge.net/projects/hcfr/files/
2) Launch ColorHCFR.
3) Click on the Advanced menu and choose Preferences.
4) Click on the References tab. Set the Standard drop-down to "HDTV-Rec. 709", check the "Change White" checkbox and set the drop-down to "D65". Click OK.

Calibrating a cab

IMPORTANT NOTE ABOUT TERMINOLOGY

CRT manufacturers often confuse the labeling for Contrast, Brightness, Drive and Cutoff. On the MS8 and MS9 chassis, Contrast adjusts the proportion of the signals while Brightness adjusts the black level. On Wells-Gardner chassis, Contrast adjusts the proportion of the signals while Black Level adjusts the black level.
"Drive" and "Gain" are two interchangeable names for the upper-end Red, Green and Blue color gun adjustments. I use "Drive" in this guide, but yours might be labeled Gain.
"Cutoff" and "Bias" are two interchangeable names for the low-end Red, Green and Blue color gun adjustments. I use "Cutoff" in this guide, but Sony PVMs label it Bias.

After calibration is complete, the only dial you should ever adjust is the Contrast dial to account for differences between PCBs.

CRT and colorimeter setup
1) Connect your test pattern generator to the cabinet. These instructions assume you're using a MiSTer or laptop running a SNES emulator with the 240P Test Suite for your pattern generator.
2) Power on the cab. In the 240P Test Suite, go to "Test Patterns" -> "White Screen".
3) Wipe down the monitor glass with glass cleaner. Place the colorimeter directly on the glass in the center of the screen. Place a book or something heavy on top of the cable to stop it from sliding or falling off the screen.
4) Wait one hour with the screen on all-white to let the CRT and chassis warm up.
5) Connect the colorimeter to your laptop running ColorHCRF with the USB cable. Launch ColorHCRF.
6) Click on the "File" menu and choose "New". On the Generator Selction screen that appears, choose "DVD manual" and click Next. Select your model of Colorimeter from the sensor list, select "Do not use a meter correction file" and click Finish.
7) Your meter might now ask what kind of display you're calibrating - choose "CRT" or “Refresh Display”. Also, Reading Type should be set to “Display”. Click the Calibrate Meter button. Some meters will then ask you to display a white image of 80% IRE or higher. You're already displaying an all-white screen, so just click OK to finish the calibration.
8 ) In the main ColorHCFR window that appears, click on the Green Triangle button in the menu bar to start taking readings.

Adjustments - Ideally, you'll want to do this with the room as dark as possible.

Flyback Screen Dial

1) On the cab adjustments, turn Brightnesss and SubBrightness all the way down. Set Contrast and SubContrast to center. Set all three Drive adjustments to center. Turn all three Cutoff adjustments all the way down.
- Some chassis don't have Contrast, SubContrast and/or SubBrightness adjustments.
2) In the 240P Test Suite, back out and go to "Test Patterns" ->"White Screen". Tap R until the screen is all-black.
3) Turn the Screen dial on the flyback down (counter-clockwise) until the display is completely dark and pitch black.
4) Slowly nudge the Screen dial on the flyback up (clockwise) until the display becomes slightly visible from complete black.

Contrast and Brightness
1) On the cab adjustments, set your Drive, Cutoff, Brightness, Contrast, SubContrast and SubBrightness pots all to the center.
- Some chassis don't have Contrast, SubContrast and/or SubBrightness adjustments.
2) In the 240P Test Suite, back out and go to "Test Patterns" -> "100 IRE".
3) Look at the readouts in the lower-left corner of the Color HCRF window. The "Y ftL" measurement tells you how bright your screen is. You'll want it between 30 - 40 ftL. 30 is better for a dark room, while 40 is better for a brightly lit room. I usually target 35 Y ftL for my cabs. Adjust the SubContrast dial (if you have one; Contrast otherwise) until the Y ftL hits your target.
If your chassis doesn't have any Contrast adjustment dials, try adjusting the Red, Green and Blue Drive pots evenly - the Drive adjustments are essentially Contrast dials for each color.
4) Write down what your "Y" (not "Y ftL") measurement is in ColorHCRF.
5) In the 240P Test Suite, the L and R buttons on the controller adjust the IRE levels up or down in 10% increments. Tap L until you're at IRE 10.
6) Your target now is to set the SubBrightness pot (if you have one; Brightness otherwise) - not Contrast - such that the Y reading is 0.65% of what you wrote down on Step 3. So if you wrote down "47.387", 47.387 times 0.0065=0.308, so you adjust the SubBrightness pot until Y is as close to 0.308 as you can get.
If you can't go low enough, you can either nudge the "Screen" dial on the flyback down a hair (Screen is a coarse Brightness adjustment) or you can try adjusting the Red, Green and Blue Cutoff pots evenly.

Note that aged chassis components and tubes may have trouble hitting the high and/or low targets. In that case, just get as close as you can.

Drive and Cutoffs
1) Use the L and R buttons on the controller to set the IRE level to 80.
2) Look at the gauges in the lower-left corner of ColorHCFR. The Red, Green and Blue gauges show your levels relative to the target, while the yellow gauge shows your Delta E (deviation) away from the target. As the RGB gauges get close to 100%, the Delta E gauge will drop. Your goal is to get Delta E as close to zero as possible.
Adjust the Blue and Red Drive pots until the gauges are as close to 100% as possible. Adjust Blue first before Red. Adjusting Green will scew the Red and Blue levels, so you shouldn't touch the Green pot unless you can't get Red or Blue to adjust far enough to reach your targets. You'll notice as you adjust the Red and Blue levels closer to 100%, the Green will be pulled there as well.
3) Use the L and R buttons on the controller to set the IRE level to 30.
4) Adjust the Blue and Red Cutoff pots until the gauges are as close to 100% as possible. Adjust the Blue first before Red. Adjusting Green will scew the Red and Blue levels, so you shouldn't touch the Green pot unless you can't get Red or Blue to adjust far enough to reach your targets.
5) When you change the Drive pots, the Cutoff shifts and vice-versa, so repeat Steps 1 - 4, going back and forth between 30 and 80 IRE and adjusting Cutoff and Drive respectively until both of them are as close to 100% RGB and 0.0 Delta as possible on the gauges.

Now, the cab is pretty well dialed in at this point, but if you're extra picky, repeat the Contrast and Brightness steps one more time, because adjusting the Drive and Cutoffs will have shifted those values a little bit. If you do end up readjusting Contrast and Brightness, you'll also have to readjust the Drive and Cutoffs again afterward as well, but none of them will need very much readjustment.

Verifying the Calibration
If you'd like to check your cab CRT now and see how the calibration holds up compared to a professional CRT, here's how:
1) In ColorHCFR, click the Green Triangle to stop taking constant readings.
2) Click the Measures menu and choose "Gray scale", and click Yes at the prompt.
3) The software will prompt you to "set 0% grey level". Use the L and R buttons on the controller to set IRE to 0, and click OK.
4) After a moment, the software will now prompt to "set 10% grey level". Use the L and R buttons to set IRE to 10, and click OK. Repeat all the way up through the 10 incrementing IRE levels.
5) When the measurement is complete, look at the DeltaE row. Recall that DeltaE measures your variance from the perfect target. If all ten cells are green (Delta E < 2.0) then you're on par with a professional CRT. If some or all of the cells are yellow (Delta E between 2.0 and 3.0) then you're on par with a consumer CRT. If some or all of the cells are red (Delta E > 3.0) then you're off target - either the calibration is wrong, your chassis PCB needs servicing, or the color guns in your tube are worn out.

Note that calibration drifts as the tube and caps age - you may want to "tune up" your calibration every couple of years.

Here are screenshots of the calibration results on a Sony PVM I calibrated (note the row of green) and my MS9-29 before I recapped it (note the red deviation at the high end).
IMG_6870.jpgIMG_6880.jpg


UPDATE 5/13/2021

I've learned that the SNES hardware has a limitation where using it as a calibration target will cause the Gamma levels to skew a bit lower than they should above 50% IRE. The MiSTer SNES core is precise enough to the original hardware that it has the same limitation. Honestly, you'd probably never notice it for arcade games, but if you really want your cab to be calibrated to the very best it possibly could be, use a different test pattern generator. There is a free Raspberry Pi test pattern generator called PGenerator that you can use if you have a way to connect an R-Pi to your cab. There's a video on setting it up here:
View: https://youtu.be/D6z0wS5oRoE

Here are two calibration result pictures from the same MS9 chassis and tube - the first when calibrated with MiSTer SNES 240P Test Suite and the second with the R-Pi PGenerator. Notice that the yellow Gamma measurement line in the right-hand graph stays more consistently straight with the PGenerator.
9D018C4F-5BEA-4BFB-9C62-D411A0E60CAA.jpegE1PvB7rVUAMbSYX.jpeg

UPDATE 11/4/2021

Modified contrast adjustment instructions to accommodate chassis without a contrast adjustment pot.

UPDATE 2/2/2022

Added adjustment procedure for the Flyback Screen dial.
 
Last edited:
@ShootTheCore, would you recommend running the test patterns in 240p test suite in 480i during calibration?

I remember StickFreaks saying that in his CRT calibration videos but wasn’t sure if you’ve tested and found much of a difference. The thinking goes you’d get a brighter test pattern to read in 480i since it’s not “skipping” lines drawn on screen like a 240p signal.
 
H
@ShootTheCore, would you recommend running the test patterns in 240p test suite in 480i during calibration?

I remember StickFreaks saying that in his CRT calibration videos but wasn’t sure if you’ve tested and found much of a difference. The thinking goes you’d get a brighter test pattern to read in 480i since it’s not “skipping” lines drawn on screen like a 240p signal.
I’d recommend against it since arcade PCBs are going to be rendering in 240p and you’re calibrating the CRT with that use-case in mind. Generally speaking, your calibration resolution should match whatever material you plan on using the most with the display.
 
I used my MiSTercade, colorimeter and procedure tonight to dial in an MS9-29 tube and chassis. Final result is less than 1.0 Delta E across the board, which is PVM/BVM levels of color accuracy.

It’s tricky to take clean pictures of a CRT with a cell phone, but trust me - this thing is gorgeous! :thumbup:

E392115E-6179-47CD-8BB0-26DE7437EEDD.jpeg9D018C4F-5BEA-4BFB-9C62-D411A0E60CAA.jpeg6BC9E576-63BB-4A0B-972D-D5D2F61FE52D.jpeg38C7B6BC-E04D-4AD9-B82C-E470D68768E0.jpeg44923235-D75F-47D3-A0EE-B717277090F8.jpegAF0485C8-1CD8-4751-897B-89B6FEDED0A8.jpeg
 
I used my MiSTercade, colorimeter and procedure tonight to dial in an MS9-29 tube and chassis. Final result is less than 1.0 Delta E across the board, which is PVM/BVM levels of color accuracy.

It’s tricky to take clean pictures of a CRT with a cell phone, but trust me - this thing is gorgeous! :thumbup:

E392115E-6179-47CD-8BB0-26DE7437EEDD.jpeg9D018C4F-5BEA-4BFB-9C62-D411A0E60CAA.jpeg6BC9E576-63BB-4A0B-972D-D5D2F61FE52D.jpeg38C7B6BC-E04D-4AD9-B82C-E470D68768E0.jpeg44923235-D75F-47D3-A0EE-B717277090F8.jpegAF0485C8-1CD8-4751-897B-89B6FEDED0A8.jpeg
Awesome! I’m waiting for a good deal on a colorimeter so that I can get on this. I’m really bad at calibrating colors so your guide has me excited.
 
@ShootTheCore how did you connect your Pi to the cab?
There are some solutions out there to install over Pi to catch player input signals and output 15Khz video and audio throught GPIO port to a jamma edge, such as: RGB-PI or PI2Jamma. Check them out
 
I picked up an RGB-Pi, and while the cable is well made, it's designed to use their specific driver and card image that boots up into a RetroPi front end. I tried to swap the RetroPi software out for the PGenerator calibration software but haven't had any luck yet - it just hangs. I need to to tinker with it some more.
 
There is a "Studio" level, which is below the Pro. I don't know the difference besides about $100...
I’ll have to see if this over works with refresh displays but looks promising.
 

From what I'm reading, the Studio model has the same main sensor, but it is capped in firmware to be 5X slower than the Pro, and it also doesn't have the secondary sensor to accommodate for ambient light in the room. The Pro sends an updated reading to the laptop about every second, and the Studio about every 5 seconds.

IMHO, I suggest spending the extra money for the Pro up front or else you'll probably end up buying twice. Getting an updated reading only once every five seconds would be really annoying and would really slow down your calibration process.
 
Back
Top