YesAffinity Blog

[BillyGaines]

WTG Chris!  It was well overdue!  Awesome score!  lit

[YesAffinity]

Thank you, Billy, good sir! 

[YesAffinity]

Well, after two days of non-stop 'cading and standing on feet, I am glad to be home and off my feet.  The event was fun, and it is good to see DK getting some hype.  It was a bit disappointing that yesterday's head-to-head between Wes and Robbie didn't get much fanfare, not even so much as an announcement on the loud speaker.  At one point in the afternoon, as my buddy and I were making rounds, we realized that they were both sitting down playing on the "head-to-head" cabs set up at the main attraction area on the arcade floor. Once they both were getting to upper teen levels it became evident that they were going head to head, and as they were both reaching level 20, a few people were watching in anticipation of both of them passing the kill screen point.  They did, and both players finished at level 24 something.  Very impressive!

For me, I was honored to meet Billy Mitchell and Walter Day.  My runs were mediocre to decent throughout the weekend, hovering in the 500-600k region.  Then, I finally had a good run as a last ditch effort before calling it quits last night (Saturday night).  Almost got a second kill screen under my belt.  L=21-5, at 882,100 points.  Nerves, hunger, all over achiness from being on my feet for 2 days, and having to go to the bathroom all played a part in losing focus and not seeing the game through to KS.  Looking back at the video my friend took, the barrel groupings that overtook me for loss of last life, although they seemed unmanageable under the pressure, could have been very easily overcome.

Time to get back to practicing, and incorporating 1M strategies.

[YesAffinity]

Inaugurating the next generation.  My 6 year old clears L1-1.  This is the second or third time she's done it, but first time she's done it on demand.

http://www.twitch.tv/yesaffinity/v/19181609

Next stop, L1-2!

[YesAffinity]

I finally got around to swapping in the replacement monitor that I picked up from HSD (on this forum) a couple months back.  The replacement had a piece of painters tape on it with "Play Choice" hand-written on it.  The chassis is freshly capped.

The replacement was surprisingly easy:

Ground the monitor, discharging any stored voltage.  Disconnect the 4 affected connectors, including the 110V plug.  Remove 4 bolts securing the monitor in the cab.  Pull the monitor out through back of the cabinet.  2 more bolts to swap the upper retention bar from the old monitor to the new monitor.  Drop the replacement monitor in through the front of the cab, re-secure the 4 bolts (the upper retention bar might needs some slight adjustment to get the upper bolts that secure the monitor to line up with their respective threaded holes).  Reconnect the 4 connectors including the 110V plug.

:edit: corrected the part about removing the monitor through the back of the cab.  I tried removing it through the front at first, but it wouldn't go.  Might be possible, but pulling it out the back is much easier.

[YesAffinity]

I power cycled 3 or 4 times, prior to pushing the cab back into its corner of the man cave (which is not as simple a task as it sounds, but I won't go into all the details).  Once I had it back in its rightful place, and had it running for a couple minutes, the screen went wonky.  Appeared to be a sync issue.  Fearing that one of the caps was defective, or some worser issue that would end up requiring me to swap the original back in and then spend hours/days/weeks of troubleshooting, I pulled the cab back off the wall, grounded the monitor and went about adjusting all the manual adjustments.  Finally, the last possible adjustment, which I believe is a pot on the main chassis board, and must be sync adjustment, cleared it right up.  15 minutes more of screwing around with all the adjustments (since I now had a mirror set up so I could see the monitor from behind the cab), and this bad boy is gleaming.  No more weird vertical columns of waviness under moving objects like barrels and Kong.  No more skewed ladder rungs.  Colors and crispness like the original monitor hasn't seen in probably years.

And, as an added bonus, lo and behold, under the monitor chassis is the Nintendo nameplate.  The one that is supposed to be above the rear door was missing when I bought the cab.  Based on feedback from good folks here, I was 90% confident that my cab was an original DK.  That confidence is now 100%. 

Next improvement will be capping the original chassis and swapping it back in.

[YesAffinity]

Been a while since I updated the ol' blog here. Upping my PB during the DKOO breathed some new life in to me, as far as enthusiasm for DK goes.  I've been committing the time to continuing the pursuit of ever greater accomplishments in the game, but the enthusiasm just hasn't been there most times.  Well, it's back...at least for now.

I took the week off work, cuz I had a bunch of vacation days accruing.  There's never a good time to take off, but it's been so ridiculously busy since the beginning of the year, I needed a breather.  And, of course, what better way to spend the time, while the kids are at school, than hobbying.

I decided to finally bullet-proof all the cabling in my direct feed setup, and do some cable management.  I originally patched all the cabling together with quick disconnects, as a means for troubleshooting while developing the concept.  But, now that the solution is tried and true, I grew ever more leary of the sleeping beast giant that is the potential for failure at each quick disconnect (both the crimps to the cables, and the QD connection itself).  In fact, I lost blue on the direct feed, a couple months back, which was remedied by simply shaking and pulling on cables.  This only served to reinforce concerns I already had in the back of my mind.

So, with wire strippers, heat shrink, some spare wire, and my solder iron, I set about to remove all the QD's, strip back the wire where the QD's had previously been, then solder the splices and seal 'em up nice with some heat shrink.

After a couple hours, I re-connected all the cabling, and started wire management.  The wire management was coming out nice, and I kept telling myself "don't get too far without firing up the cab and testing".  But, the wire management was going so cleanly, you know how sometimes the wires just seems to manage themselves, all you have to do is grab zip ties fast enough so that the way things are rolling out doesn't fall out of your hand?  Well, it was one of those jobs.  Everything looked money.

And then I turned everything on to test....no green.    Jiggled wires, tugged on wires, no quick remedies to be had.  So, back out the cable bundle comes, after clipping about 10 perfectly placed zip ties and watching the neatness dissolve back into a jumble of unmanaged wires.

A bit of continuity testing, and lo and behold, a connection at a butt connector was failed.  I didn't bother replacing the butt connectors with solder splices, cuz those are already solid enough in my mind.  Not this one, though.  A bit more stripping, soldering and heat shrinking, and back into the cabinet I go.

Mind you, I only pull my cab maybe 4' off the wall to work in it, having to climb over a desktop computer and behind the cabinet every time I need to get in there.  A) The A/V cables for the direct feed out to my capture setup don't reach very far once they get to the cab.  B) I don't like moving my cabinet any more than I have to.

So, I re-make all the connections.  And, of course, this time the cable management doesn't go as smooth and I have the sense the end result isn't as good as the first attempt looked.  But, it's a marked improvement over what it had been for the last however many months since I put it all in to the point of being satisfied with the capture results and not having the desire to touch it again until recently.  While I was in there, I also took the time to do a little management of the wiring that connects to the PCB, as well as spending the time fully dialing in my monitor.

[YesAffinity]

Dialing the monitor in is no small task.  I've been annoyed at the "dead space" or borders around the screen, which I assumed could be filled by centering the image and increasing horizontal size and vertical size.  A somewhat true assumption.

I got the plastic tool kit from Bob Roberts, so started with checking B+ voltage. (measuring at pin BJ).  It was at about 112V.  Supposed to be at 108V.  Stuck the tool in to VR601, and turned it counter clockwise, and voltage came down to about 110V, but then wouldn't go any further no matter how much I turned.  Turning clockwise, voltage went up to about 111V, but then no further no matter how much I turned.  I'm sill not sure what did the trick, but apply pressure in different directions (up, down, left, right) while turning, I eventually got it to shoot down to around 100V, and then shoot back up to 113V.  I then slowly dialed in either direction, until getting it to 108.1V.  Good enough, I suppose!

I then jumpered V CENT JUMPER, and messed with VR403.  After a few minutes, I had the image about as perfectly centered as any human is capable of.  I then spent a few minutes VR354, to get the image expanded to the left/right extents of the screen.  Keep in mind that DK is technically a rotated game, so V adjustments actually adjust horizontally, and H adjustments actually adjust vertically.

The last thing I wanted to tackle is filling up dead space at top and bottom of the image.  This has always been an annoyment, since swapping in the replacement monitor that is currently running in the cab (see previous blog posts for that nice little local pickup off craigslist).  The image has always had more dead space at the bottom of the screen that the top.

So, I screwed around with L453.  This actually does very little, almost impercetible, for increasing the image size.  This is corroborated by posts I read on KLOV.  L453 is no maxed out on my monitor, and it increased image size by may 1/4" at the most.

I then tweaked on VR352.  This actually doesn't seem to center the image, comically enough.  It does more for increasing H SIZE than does L453.  At least on my monitor, adjusting VR352 causes the top of the image to stay relatively static, while the image expands downward.  So, I maxed out VR352, as well.  I then messed with VR451 (H HOLD), and think that helped center it downward ever so slightly, but it's still not perfect.

And with that, the image is expanded perfectly to the left/right extents of the screen, and expanded and centered as much as possible in the top/bottom directions, and there is still about 1/4" of dead space at the top of the screen, and 1/2" at the bottom of the screen (picture attached).  It's still an annoyment, but guess I'm just going to have to live with it.

Lastly, I messed with the colors to get them a bit sharper.  I won't go into all the details.  But, I did also try the service switch (SW301) on the chassis and turns the screen black (no service screen available on DK, apparently).  I also tweaked the sub-brightness via VR301.  Not sure if this was an indirect result of the color adjustments I did, but the "between board" screens and the DK title screen have a slightly off-black, almost a slight hint of yellow, to them.  Board screens are money though.  Perfect coloration, and on off-coloration to the black background.  A bit of a peculiarity, that.

So, if anyone knows how to remedy the dead space at top and bottom, I sure would be grateful for some help.

Lastly, also attached is a good diagram of all the image adjustment features on a Sanyo 20EZ chassis.  I'm sure most of you have seen this or know where to find it, but I thought I would post it as an added resource.

[YesAffinity]

And now as an unintended consequence, the direct feed video is shifting ever so slightly back and forth, from left to right.  This has historically been an issue that goes away after a couple minutes of the cabinet being on.  It seems that there is a correlation between cold weather and it taking longer for the condition to go away.  However, the cab's been running for about 30 minutes now, and its about 70 degrees here in San Diego.  So, it seems I get to spend some more time figuring out which adjustment triggered this now being a persistent problem. 

[YesAffinity]

I've spent quite a bit of time troubleshooting since discovering the "twitchy" video issue.  Here is an example of the issue occurring (DK attract mode):  http://www.twitch.tv/yesaffinity/v/56284949.  The issue gets worse, the longer the cabinet is powered on.

It seems like the issue is coming down to the Avermedia HDMI capture card I'm using, but I'm not 100% certain.  I can take the HDMI output from the DVDO Iscan VP20 video processor that is ahead of the capture card in the circuit, and plug the HDMI cable into an HDTV, and there is no twitch.  So, I've got a Black Magic HDMI capture card on the way.  Fortunately, now that UHD is starting to gain traction, the 1080p capture cards are coming down in price.

I'm cautiously optimistic that this will resolve the issue.  I'm concerned that it really comes down to a sync issue.  Talking with JROK when I was putting this whole setup together, he acknowledged that the output from his encoder is not a specific standard (such as NTSC), and many capture devices don't sync well with the signal.  I found this to be true in some of my early testing.  To be fair, the JROK encoder works great for displaying on a replacement monitor in your cab, or in theory, could be used to create a supergun or other portable arcade player.

At the same time, I'm perplexed as to why this issue has arisen as a result of adjusting the 20EZ.  The output for the direct feed is in theory simply an output parallel to that which feeds the monitor, unless there's some kind of backfeed effect through the game board and the edge connector, which is affecting the direct feed video.

I'm also concerned that the VP20, while apparently displaying an output normally on a HDTV, has started failing, and the capture card is sensitive to whatever is failing.  I did buy it used on ebay "as is", so who knows how much life it has left in it.  The VP20 is otherwise performing flawlessly, for processing for all of the various sources I feed through it (Dreamcast, Xbox 360, Sega Saturn, Playstation, Wolfmame WinXP PC, to name a few).  But again, it could come down to the abnormal signal from the JROK encoder that the VP20 is no longer handling well.

Well, more to come on that in a couple days, when the Black Magic card arrives.

Simultaneously, I was excited to discover that OBS now offers a more advance version, OBS Studio, and that version offers the ability to rotate source video.  So, no longer do I have to use VLC to rotate the direct feed video via the Avermedia card, and then do a window capture of VLC, with OBS.  OBS Studio also seems to be handling the timing of sources better than OBS vanilla did, as well.  On OBS vanilla, I spent a significant amount of time tweaking the delay of audio and video from various sources that I am displaying in the single stream window (webcam audio and video, direct feed video via Avermedia card, and direct feed audio via "line in" on my PC sound card).  OBS Studio also gives that ability to re-size sources by dragging within the stream window.  This is hugely useful!  OBS Studio is now the one-stop shop.

[YesAffinity]

Well, with the ol' direct feed acting up, and while I wait for a replacement capture card to arrive, I decided to take it back to the old school.  Being OCD, the twitchy video was bugging the crap out of me...just seeing it in my peripheral vision and between boards was throwing my game off.  Plus, I don't want people to have to watch that.  Maybe it's not truly a big deal, but to me it is.

So, back to the trusty ol' webcam feed we go.  A couple DeWalt 'C' Clamps from Home Depot (which I already owned), and a couple zip ties, and it's about as good as it's going to get, for a temporary solution.

And the cool thing is, I can still have the direct feed audio in the stream.

[YesAffinity]

Well fellow DK’ers, I’m pleased to announce that the problem appears to be solved.  If you remember, I noted that when jumping V CENT JUMPER, the image centers more precisely on the monitor, and when removing the jumper, the image shifts slightly to the right, and I found it impossible to get it perfectly centered without V CENT JUMPER shorted.  Well, I resolved to jumper it permanently, so that the image would be centered on the monitor, and had a hunch that this also might have had something to do with the twitchy direct feed video.  Turns out it did.

I jumpered V CENT JUMPER, and lo and behold the twitch went away.  If you will notice (in the attached), doing so actually pins it to the far right of the available image space, and while this is by no means a thorough or scientific explanation, in layman’s terms, the image has nowhere to shift to.

All very peculiar, I must admit.  Why adjusting the monitor chassis affects the edge connector output is indeed a perplexity.  But, at the end of the day, all is working better than it was prior to start of the whole effort to dial in my monitor and tidy up the direct feed cabling.  In addition to the image being as large as the chassis will allow, I now also have a simplified stream solution via OBS Studio, which is significantly more functional than OBS vanilla.

Now, I’m noticing a growing issue with the sound output in the cab.  Sounds like hammer smashes and Kong’s stomping have a consistently tinny sound to them, with the sound becoming muffled and static-ey at times.  It has been a very slow, intermittent development of the issue, but it is starting to occur with increasing frequency.  Time to study up on the troubleshooting flow chart, and I’m guessing, to do some minor cap replacement.

[YesAffinity]

Here is the screenshot from OBS Studio, illustrating what I believe to be part of the reason that the issue is solved.  Interestingly, image being pinned to the right side of the viewable area is not fully against the border, whereas the day before it was.  The image is a pixel or two off the edge.  Either way, it seems to be resolved now, so I guess all's well that ends well.

[YesAffinity]

Sent my original Sanyo 20EZ chassis off to Arcade Buffet a couple weeks ago, to get a cap kit, reflow, and anything else he deemed needed to be done.  It arrived back to me yesterday, and of course, I had to get it swapped in today.  I had a bit of trouble adjusting brightness, particularly VR301 sub-brightness.  I stick one of the plastic tools from Bob's plastic monitor repair tool kit (similar to this one https://www.amazon.com/Jonard-TK-AT5-Piece-Alignment-Tool/dp/B006C4BXLS?ie=UTF8&camp=1789&creative=390957&creativeASIN=B006C4BXLS&linkCode=as2&redirect=true&ref_=as_li_ss_il&tag=arcaderepair-20) and turn and turn and turn.  Nothing will happen for most turns, but then randomly brightness will shoot up or shoot down, and then I can't get it to adjust again for a bunch more turns.  If anybody knows a trick to adjust VR301, please share.

That aside, I've got the picture pretty well dialed, but do want to spend a bit more time getting it just perfect.  I also discovered a H POS and V POS adjustment, on the the DK game video board.  Not sure how I never noticed those before, but they were just the right adjustments needed to perfectly center the image vertically and horizontally, and fill the screen from side-to-side.

Lastly, the image "twitch" I described above has continued to persist, even though I thought at one point I had fixed it.  I got a Sync Separator, and put it in line, between the output of the DK direct feed setup and the video processor (DVDO VP30).  It cleaned up the twitch to some degree, but there continued to be some correlation between higher ambient temperatures and the presence of the twitch condition.  It was all very inexplicable, but the good news is that with the refurbed original chassis back in the cab, this problem in the direct feed setup is gone.  I have deleted the sync separator, and have seen no mysterious twitch all day, even in the midst of 95 degree ambient temperatures here in sunny Escondido, CA.

[YesAffinity]

Finally got around to doing a cap kit in the original PP-7B for my blue cab.  I had pulled it out a while ago and replaced with a switching PSU kit, in the effort to eliminate shake in the monitor video.  This seems to be a problem for many Sanyo 20EZ's (see various threads on KLOV about this very issue).  Not only did the switching power supply not fix the issue, but it turns out this is also an issue caused by running Nintendo cabs on switching power supplies.  Would have been good information to have YESTERDAY. 

Anyway, I had swapped the original PSU back in temporarily, to take the switcher to a friends house for some troubleshooting on one of his cabs.  It seem the original had developed some greater issues while sitting in storage, as it was now producing a full-blown left-to-right wave through the image.  That was enough for me to get a bee in my bonnet to do the cap kit.  Plus I finally dropped the money on a legit desoldering gun (Hakko FR300), so am no longer intimidated by bigger cap and rework jobs.

About 2.5 hours later (an hour of that just for figuring out how to disassemble the power supply and then re-assembling it when done), and the problem is solved.  Not only that, but other little imperfections and anomalies that were persisting in the video image have cleared up.  In fact, it seems the game is running a bit faster, too.  That could just be a perception thing cuz it was late and this was all being done after working a 10-hour day, but on first run after letting the machine get good and warm, it was like "whoa, it's playing faster and more fluid".

Anyway, if any of you experience monitor issues and a recap of the monitor chassis hasn't cleared them up, I highly recommend ensuring an original Nintendo power supply is powering the cab, and doing a cap kit on it.  Ian Kellogg and https://www.arcadepartsandrepair.com/ are both good sources for reliable and high quality cap kits.

Now to continue the never-ending pursuit of 1M....