I've heard this from friends who repair pins in SF... That it "smells" of power or ground related issues. While no one has a specific suggestion, the ground could be suspect. When we restored the cabinet, we ran new ground braid, and followed the original path, with one exception: We eliminated all the unnecessary loops that the original game had put in the main cabinet. The backbox, though, is 100% identical.

If I were to run a jumper from the rectifier board to the lamp driver board, what exactly do you mean? You're talking about disconnecting the power and ground from the board connectors, and running it directly somehow to the rectifier?

Oh nevermind, I see what you're saying...

No, this didn't happen... I didn't change out the rectifier board yet. What I mean is, we haven't done a swap...But yes, the original one worked before the restoration.

Ok, first I'm going to try bypassing the ground path and doing a jumper from the LDB to the rectifier ground, just for giggles. If that doesn't work, the next step is a rectifier swap, and I may wait to do this until I have a confirmed working transformer/rectifier pair I can temporarily borrow just to see if it changes things.

I seems to me that what I would do is test the voltages of those three ground wires. Then I should disconnect them from the LDB J1 connector housing. Then run three jumpers to those points and connect them to the rectifier board ground. Then run the socket test. Does that sound right?

Well already there are some differences for sure. The original is a AS-2518-49. The repro is closer to a AS-2518-18, but they are supposed to be compatible. The one main difference is the feature lamp fuse on the -18 is a 10amp fuse, and on the -49 is a 20amp fuse. So I'm running the feature lamp bus on 10amps, but I can't see the difference that would make other than blowing the fuse if I push it too far. If you look at connection point E9, you can see that the original board had a place for two wires to terminate. For the repro, I twist those wires together and go into a single connection point.

Just for reference, here is a photo of my original board:

IMG_3277.JPG

IMG_3278.JPG

Here is the new board:

IMG_3259 2.JPG

IMG_3257 2.JPG

Do you recommend that I remove the current ground connections first?

That's helpful to know, but you're using an original Bally -18 rectifier or a repro?

I’m pretty sure I changed nothing. I wouldn’t even know how to change that honestly.

Sure... better than the ones I posted above?

Ahhhh I see sorry. Yes I’ll pull it out and take a look!

Yep, wired for 115v.

66ABFB7B-0EC6-4322-B48D-D1736B792058.jpeg

Ok done.

765DF8E6-ADF9-4FF6-9429-2BC75867DD13.jpeg

No difference that I can see.

I don't have handy a 10A AGC fuse, but I did have a 10A MDL, so I put that in there. I hope that's ok. No difference in behavior.

Zero.

Done. No impact.

It's hard to see if it's marginally worse. The rate of the flicker seems about the same. It's interesting that the flicker rate isn't super random, it almost seems to pulse by design (which we know it does not).

...which also leads me to believe that the pulse is related to something on a beat/waveform that is consistent.

I love doing this.

before_after_resto.jpg

The challenge is that the flicker happens with incandescent bulbs and with two separate lamp boards... and two separate MPU boards... so it has to be something else.

In my opinion, it’s something in my power wiring that I’ve overlooked. My next move is to swap the rectifier/transformer with a known working combo (not to solder myself). Let me absolutely get that ruled out. There is nothing left to try that I know of.

I'm not sure exactly if we've tried this. When you say a simple socket test, what do you mean? How would I do that? I've taken a socket, connected it directly to the SCR gate leg and to ground to make sure it stays on/latched, and that worked fine. I've also tried a socket, connected one lead to the single pin associated with that SCR and the other to the rectifier TP1, and that flickered in lamp test. My assumption here is that the SCR has the capability of latching based on that set of experiments. I've also swapped the SCR with a known working one, same results.

When we checked the waveforms hitting the SCR, there is no question a spike comes at the normal rate to make it latch, but it seems to not be strong enough for every two cycles, then successfully latches for two cycles. It's like something interferes with the voltage on a rhythm. It looks like it's not perfect (meaning, it's not precisely two on and two off), which leads me to believe whatever is doing this is on a slightly different cycle, but close.

waveform_info.jpg

I actually take that back, it's precisely two on, two off for the most part...sometimes one on, two off.

I have tried various increasing resistor values across even an incandescent (or LED) bulb on these flickering lamps, no change, if that's relevant.

One other clue that we uncovered in the other thread...

Most of the lamps that are address 1 or above on the decoder clear up after the game heats up for 5-10 minutes and stop flickering.

It's only the address 0 lamps (rollover F and SPA target) that keep flickering after warm up.

This means something heats up and changes the behavior. What changes when heated?

-Solder joints (but that would have to be the rectifier, since we swapped out the other boards)
-Resistors on the rectifier?
-Resistance increases?
-Something expands and moves a short circuit out of short?
-Metal pin or female connector trifurcon expands making better contact?

What would impact current flow? Pin connectors? Which power wire am I taking about here?

Well, based on the earlier Quench post, I attached a jumper from "the lower leg of either white ceramic resistor on the rectifier board" to the GND test pad on the LDB. It didn't seem to matter. I didn't bother re-terminating the three ground wires to the LDB, but I can do that next if you think it might have something to do with it.

When you say I hooked the 6VDC lamp bus right at the rect board...what do you mean exactly?

Oh I see...

I’d have to change the bulbs to LEDs to do that test since the bulbs that cleared are the same ones that don’t flicker with 44s. I’ll do that and hit it with the hair dryer test.

Still localizing... meanwhile, another glamour shot:

C00BDC88-B25C-4C89-9094-FBA0AD388097.jpeg

Heat gun tests have failed so far. After about 5 minutes, some incandescent lamps stop flickering when on. I have been presuming it was a warm up period in which something gets hot. I can't imagine another relationship. I have not replicated this with a socket directly connected to just one blinking LDB pin and TP! on the rectifier...I can make it blink, but not go solid, even with time.

I'm going to try this with LEDs and see if some of them clear up as well (they have resistors via the Alltek and the 6.3v rail).

I've now removed the connector housing from the rectifier J3 and replaced the whole thing with a molex style connector with obsessive, careful crimps, all identical. The connector has excellent contact on all pins. This includes the three ground returns from the LDB.

Just for giggles, I replaced all three ground returns to the rectifier on the LDB connector as well.

No change to any behavior.

barakandl just double checking: Am I correct in saying there are no differences between your -18 and the original -49 that could *possibly* have anything to do with this, right? I’ve literally got nothing left to try.

Exactly, that's what I meant in my last post. This is my next move. I don't think I can just re-install it though, it had been hacked up in a way I'd have to rebuilt it somewhat and bullet proof it first, then I could use it in the original way. For example, certain power/ground lines had been directly soldered to the back of the board, and the ground was to the test pin, etc. It was a mess. However, as you saw from the photo, it's in good enough shape it can be cleaned up.

I ordered a kit from Big Daddy to clean it up. I'm not sure if I'll be using the Kulpa method or just following along a thread here on Pinside.

I just worry I'll go through this whole process and achieve no different result. It's worth it though, because then I'll have a backup rectifier.

HA! I think given the full scope of this thread, we do a good job of "glass half full." And old buddy used to complain that I didn't give up enough.

This is good advice. I think the only challenge with this approach is that every time the transformer is re-soldered to a rectifier, you're cleaning up the ends and shortening the leads, even by a small amount. I think I can do this, but probably only once without jumpers. In my perfect world, I would be able to do this with only non-permanent connections, but these earlier rectifiers were ugly in that way.

I understand your point. The connections in question were routed from the back of the board directly to a new cable connector to avoid the burnt pins. To simulate that would require I remove those specific trifurcons from the connector housing and somehow bypass the pins/traces on the board back to those solder points. That is a LOT of archaeology and the original bypasses have all been removed from the old rectifier. I took a lot of photography but I'm not confident I know which is which, and where the bypasses landed, so that is why I have been leaning on the idea of bringing the original back up to spec instead. I'll be honest though, I'm still not 100% sure the old rectifier is the way to go here. I've spent the better part of my career troubleshooting hard engineering problems, and I still feel like we're not answering the "why" question properly.

Given the nature of the problem (which is super weird), I'd frankly be happier pulling a transformer/rectifier combo from a friend's Future Spa that works, put it in mine, and see if that works before we go down this archaeological rabbit hole. That would isolate the rectifier board or not. If it turned out to be rectifier board related, we should find out why, what the actual differences are between my new one and the other one.

I agree with that. I should have done the rectifier swap before the teardown. I've generally believed that when there are volumes of text for a decade about the design flaws and failures of a specific piece of equipment leading to all sorts of mysterious problems, cutting out those flaky devices generally has been a good strategy...Even after a restoration...This is the first time I've seen it backfire. I doubt very highly there is anything wrong with the new rectifier...more likely I wired something wrong.

...it's also worth noting we've had people post here saying they are using an original Bally -18 rectifier board in place of the specified -49 without problems. It's not a barakandl reproduction but it's pretty much the same thing (the reproduction has improvements, but operates the same). That's more evidence it's our wiring!

Agreed.

If the bridge had a bad diode internally, wouldn't that result in far more problems than we are seeing? Seems unlikely.

Given that this happens with two separate (original Bally and Alltek) MPU boards... How on earth could my clock be going faster now than before the restoration?

The Alltek is designed to be adjusted between the Bally clock and the (faster?) Stern M-200s, whereas the Bally doesn't adjust. It's a jumper, I'll look into it.

I'll check that out today. Note that I'm currently using the original Bally MPU. I used both successfully before the restoration with no problems.

I’d say you figured out how to precisely replicate my problem. That is super bizarre. I’m still trying to wrap my head around how both boards would be running too fast. Going to check CPU pin 37 speed next.

Ok barakandl I hate to be a complete idiot but I must be missing something in the procedure. DMM set to Hz, black lead to ground, red lead to p37 on the 6800P on the Alltek, right?

Alltek reads .553 MHz.

I'd really prefer to use the Alltek, I'll ask Dave over there what he thinks about this.

Here's the infuriating thing: We know that these CPUs (Bally MPU or Alltek MPU) work with Future Spa in other scenarios. Why on earth would this be a different clock speed, even faster?

What external factors would speed up your clock!? While I agree with the premise that this would create this problem, I can't imagine how this could possibly be happening.

Let me ask you a different question: If we presume that timing is the issue...and that a faster clock would create this problem...could the opposite also be true, meaning could something else be moving slower/taking longer than normal?

Here's another clue, by the way. I noticed that I can make a lamp flicker more if the gate assembly coil is firing, drawing power. What's interesting about that coil is that it holds indefinitely until either the switch fires or the ball ends, so it's a good test. When it's firing, all the lower address lamps we've discussed fall into the flicker.

We've been able to replicate this flicker with a single socket, connected to a single lamp pin and the TP1 of the rectifier, with ALL other lamps and switches removed, so it's fair to say that the draw there or playfield wiring to lamps wasn't related to the cause. However, it's interesting to me that when the coil fires, power changes and the threshold gets crossed.

It is hot to the touch in attract mode.

I also have a nice 43vdc on TP2 on the MPU.

Hypothetically, what might interfere with when the ZC detector triggers? Let’s assume 43vdc is stable when the test begins. Would some kind of noise or change on that 43vdc source impact the accuracy of the detector?

We know the MPU detector is working normally because we’ve swapped it and it’s not a board issue. Therefore, that suggests to me that under certain conditions something is impacting how it functions. Something on the 43vdc?

My scope is dual input. I’ll give it a shot.

Here is my best shot of recreating that measurement:

Screen Shot 2018-10-23 at 5.11.41 PM (resized).png

Screen Shot 2018-10-23 at 5.16.16 PM (resized).png

Screen Shot 2018-10-23 at 5.30.39 PM (resized).png

Screen Shot 2018-10-23 at 5.32.03 PM (resized).png

How does this compare? It's worrisome to me that the waveforms are so messy.

Ok, I hope I'm doing this correctly. I have the - leg hooked into the ground braid in the backbox. I have the + leg hooked into the bottom leg of the large ceramic resistor on the rectifier. Here is the result:

Screen Shot 2018-10-23 at 9.52.20 PM (resized).png

- leg on the ground braid, + leg on the top leg of the ceramic resistor:

Screen Shot 2018-10-23 at 9.54.42 PM (resized).png

- leg on same ground braid, + leg on another segment of braid elsewhere in the backbox:

Screen Shot 2018-10-23 at 9.56.38 PM (resized).png

I keep setting the offset, and it seems to just make it longer and longer. Not sure if I'm doing offset right.

Screen Shot 2018-10-23 at 10.15.29 PM (resized).png

I presume you also want power on during that snapshot (it was).

Done as requested:

Screen Shot 2018-10-23 at 10.29.43 PM (resized).png

000.0.

000.0

Done:

Screen Shot 2018-10-23 at 10.50.27 PM (resized).png

What on earth!?!?

Here is TP1:

Screen Shot 2018-10-23 at 11.24.33 PM (resized).png

Here is TP3:

Screen Shot 2018-10-23 at 11.26.21 PM (resized).png

Here is the TP3 with 5V/div so you can see the full waveform over a couple cycles:

Screen Shot 2018-10-24 at 9.15.11 AM (resized).png

Here is a close-up of when it gets closest to zero volts:

Screen Shot 2018-10-24 at 9.16.08 AM (resized).png

Well, first off, I get a very different looking waveform. Here is what it looks like:

Screen Shot 2018-10-24 at 9.21.55 AM (resized).png

But yes, if you zoom in, you can see the same noise here, though it does go to zero volts:

Screen Shot 2018-10-24 at 9.23.20 AM (resized).png

I've made a cable, but I'm having some difficulty getting the TP3 to register any voltage with it. Strange. I hear the transformer power up, but no voltage... I have black on 6, white on 7, ground on 10, right? I'll keep looking for what I'm doing wrong.

Unbeknownst to him he is also teaching me electronics better than any college course I could take.

I realize now I messed this up, and was recording TP3 in this configuration. This is the top of the large ceramic resistor with F1, F2, F3 and F5 removed:

Screen Shot 2018-10-24 at 1.44.49 PM (resized).png

Here is my attempt at getting more waveform on a screen:

Screen Shot 2018-10-24 at 1.46.34 PM (resized).png

A couple cycles:

Screen Shot 2018-10-24 at 2.01.53 PM (resized).png

Ok, more progress. Here is the waveform across the large ceramic resistor with a clean power cord going directly from our house power into the rectifier. Same results:

Screen Shot 2018-10-24 at 2.05.08 PM (resized).png

It makes a barely perceptible hum, more pronounced just as it turns on. I recall it being the same before the restoration as well. Screws are tight.

HTML:

https://reference.digilentinc.com/reference/instrumentation/analog-discovery-2/reference-manual

PDF:

https://reference.digilentinc.com/_media/reference/instrumentation/analog-discovery-2/ad2_rm.pdf

I can’t find a 10x setting or switch.

No work was done on the MPU, also I’ve swapped MPUs with no change in results.

One thing that absolutely could be important is that I did personally do the work of soldering the transformer to the rectifier board. I’m wondering if anything in that solder work could result in this noise.

I'm afraid not, I have no BNC adapter or other clips of any kind. It basically came with nothing but female jumpers coming off the scope itself. I've been hacking together my own probes using male-to-male jumpers.

Is this safe to do since my scope can only handle +/- 25V? It has a 50V max peak-peak rating, whatever that means. Assuming the answer is yes, I'll rig this up tomorrow morning (my time zone).

When you say bad grounds on the 12v regulator, you're saying on the rectifier board itself? We're seeing this noise with everything disconnected (including all those returns) from the board. I ask because I like potential directions where it could be caused by my bad soldering skills.

I do have a big resistor assortment pack on hand. What's the ideal level you want? I can just put them in-line on my hacked probe wires.

E11 and E12:

Screen Shot 2018-10-25 at 8.13.53 AM (resized).png

2ms/div:

Screen Shot 2018-10-25 at 8.56.00 AM (resized).png

200us/div:

Screen Shot 2018-10-25 at 8.56.35 AM (resized).png

Here is zoomed out. Got the sine wave this time.

Screen Shot 2018-10-25 at 8.55.29 AM (resized).png

By the way, the way I did that was to set it to 2ms/div time base, 5v/div on the channel, and then do a recording. Once I've done a recording, I can change the time base. Should I be re-recording?

1v/div:

Screen Shot 2018-10-25 at 9.16.38 AM (resized).png

It plugged it back into the cabinet wiring. I can do it again with the direct connection...I had assumed since we saw the noise both times it wasn't related. Should I re-record?

Here is the same recording bypassing the cabinet wiring and direct to the outlet:

Screen Shot 2018-10-25 at 9.26.45 AM (resized).png

Same location. Or put another way, the machine moved around my house as we completed the work, but right now it's plugged into the same location it was when it worked prior to teardown.

If you mean could the transformer have been damaged, moved, jostled, etc., anything is possible. For a game like this, the transformer is mounted to a back plane along with the rectifier. It's then moved as a single unit in/out of the machine. I assembled it on the bench and tested it on the bench. However, I can't stress enough, I found the whole through-hole soldering of wires to the rectifier awkward at best. Whiskers, shorts, solder blobs... I didn't find any, but anything is possible.

Meanwhile I'll try to find a period-correct transformer/rectifier combo to swap with, and also get my other rectifier board ready for some kind of swap as well.

I have a TOM, a TOTAN, a BoP, and a BM66. Probably the BoP is the oldest but still 10 years newer than the FS. I'm not sure if that's helpful.

Unfortunately, I don't think my scope can handle 120V (unless I misunderstand my scope specs, which say 50v peak to peak). If I put some kind of resistor on each lead, I could just plug my scope right into the AC power coming off the circuit and check the waveform for noise, right? Any advice on what level resistor?

Also, I'm more inclined to think it has something to do with the way I attached my transformer to my rectifier...

HOLY #$%^!

I need to sit down for a second.

I have an elaborate solar panel setup in my house. I'm a beta tester for some newer inverter technology. Unfortunately, as part of the beta test, some equipment failed back in December of 2017. We had a few months of downtime as we re-worked some of the system.

We acquired the Future Spa in January, and began working on the project in February. This is all during the time the solar panels and inverters were offline.

The machine then was torn down and the restoration began.

I've had various ups and downs of the inverter system, which is fairly unique. It's a great system, but I've been helping the company work out the bugs. We brought it back online and fully stable in August and early September.

We then completed the restoration, all after the panels have been fully operational, as well as the batteries and inverter system.

So, bottom line, before the teardown, no inverter. After the restoration, inverter.

I do have that filter, but I suspect we would get the same results. I think we have our smoking gun at this point, don't you? I wonder how I can do a better job filtering the inverter noise from my power system.

THAT I can do. I have a UPS for certain systems in the house. Long story. I'll make that happen.

I will test with the UPS first. If the problem goes away, I'd rather focus on cleaning my power...Unless you think a better EMI filter would help this game?

I do have various AC power packs probably lying around, and DC units. Can you explain a bit more about how I would use this?

I can power down the inverter, but as I'm part of a beta program, I need to contact the company before I do that (long story, we're monitoring and building data for this machine learning system for power shifting). Meanwhile, I plugged an extension cord into a UPS but I'm still getting noise. I need to figure out if I'm hooked into the UPS correctly, stay tuned.

This seems like such an amazing solution...however somehow I'm still getting noise even through my UPS. This doesn't say much, the UPS is old, and who knows how good a job it's doing of line conditioning when not on battery. I could pull power and have it run on battery... Hmm...

Anyway, this is plugged into the UPS, fuse F1, F2, F3 and F5 removed, line power bypassing the cabinet wiring/EMI filter, and captured across the large ceramic resistor:

Screen Shot 2018-10-25 at 8.47.44 PM (resized).png

Same results with cabinet wiring... I think I'll just pull the plug on my source power. (Yikes)

Ok... well this was a weird experiment.

When grid power was removed from the UPS, and the UPS was running 100% on battery, I then turned the Future Spa on (both directly into the power or also through the cabinet wiring and line filter).

The transformer made a LOUD hum this time. There was something goofy going on.

When I measured the same waveform across the large ceramic resistor, I got these weird waveforms:

Screen Shot 2018-10-25 at 9.14.23 PM (resized).png

Screen Shot 2018-10-25 at 9.15.50 PM (resized).png

Returning to grid power brought me back to that "V" with noise.

Anyone care to explain this to me? Note that I do have other devices also plugged into the same UPS...and I can't turn them off.

Yeah, that was some weird output when disconnnected from the grid!

Tomorrow I’ll shut down the inverter. I now have the procedure that won’t break anything. If it’s proves to be noise from that, it’s progress!

The specs on the UPS say that on battery it should output a “simulated sine waveform.” That doesn’t look like a simulated sine waveform.

To be clear, the super weird waveform is from my UPS system’s simulated sine wave. My machine and the power of the house never sees that sine wave because it’s power backup only for my telecom gear (and has been tested fine in that gear for many years). I only tried it to see if I could get a clean sine wave without any inverter noise.

My inverter noise (if that is the culprit) is a separate thing, and I’ll try shutting that down today and see if the noise goes away.

None of those options are possible. The neighbors are just too far. Shutting down the inverter for a few minutes is easy.

Oh, believe me, this is hysterical to me as well. If this is the cause, essentially 1979 Bally failing to precisely predict the future when people have solar panels and prototype inverters, I’m going to be thrilled.

What’s bugging me is that my UPS also has an AVR (automatic voltage regulator). There shouldn’t be noise passing through that source. The power is converted from AC to DC then back again.

I have transformer packs, I have to see if I have ones that just reduce the voltage. I must somewhere.

Here are a couple shots I took trying to get signal out of a 120W AC to 32W DC transformer pack.

Screen Shot 2018-10-26 at 8.21.11 AM (resized).png

Screen Shot 2018-10-26 at 8.20.52 AM (resized).png

That transformer was wonky so I've switched to a different one. This one is 120V AC in and 24V AC out. Here is the wave. It's weird.

Screen Shot 2018-10-26 at 8.37.24 AM (resized).png

Here is zoomed in:

Screen Shot 2018-10-26 at 8.39.49 AM (resized).png

Here is the label:

IMG_3430 (resized).JPG

I'm going to shut down my inverter next, because...why not. If it shows a clean signal, we'll keep moving on from there. This is fun.

Here is my results with no inverter. This is measured across the large ceramic resistor on the rectifier. F1, F2, F3 and F5 fuses removed.

This is my regular wall outlet power:

Screen Shot 2018-10-26 at 9.19.10 AM (resized).png

Zoom in:

Screen Shot 2018-10-26 at 9.19.40 AM (resized).png

Here is plugged into my UPS line conditioner (not on battery):

Screen Shot 2018-10-26 at 9.12.29 AM (resized).png

What does this tell us?

Yep. Here we go. I have to rebuild that connector but shouldn't take too long.

No flicker.

We solved it.

I'm speechless. I need to take this all in.

Before we go any farther with this...

We need to thank some folks here. Tons of you helped out and all deserve thanks for supporting this process. However, I need to single out @quench, who really took an insane amount of time helping me understand all this. Secondly, @barakandl, who not only took the time to really dive deep on some troubleshooting, also CLEARLY makes a rock-solid rectifier board, which has now been proven by countless waveform analysis recordings. Folks like @zitt, @bjm-maxx, @vid1900 and so many others.

I think what this proves (beyond a shadow of a doubt) is that for any service professionals responding to early Bally lamp latching, before spending a ton of time fixing an MPU board or adding resistors, ask them before you truck roll if they have solar panels and an inverter.

What a journey. Wow.

Now I have to figure out how to clean my AC power in my house. Any suggestions?

I know right? Can we just have Quench get the MVP award for most patient electronics expert of all time?

The company that makes my equipment (sort of a more sophisticated Tesla Power Wall) is shipping, but while we're adding features and taking notes I'll keep it private. They make a fantastic product. I'll work with them for some possible solutions that can be integrated into their next revision.

However, for my current setup, I'll need something inserted into the architecture that doesn't create additional load or other issues. The solar company that helped install the panels, while awesome, are out of the loop in this case. We went outside their normal equipment to do this.

For a home, though, there has to be a solution that is more residential-scale. I'm surprised this isn't more of a common situation that there wouldn't be some easy solutions.

Thanks, totally agree on the learning. It doesn't hurt that this Future Spa is pretty fun to look at while you're doing it.

I'm thinking that since it seems to only effect old Bally machines, perhaps I should get an isolation transformer for the machine for now, and just get the inverter company to fix this.

Well, you can't really use isolation transformers for a whole house. You can do massive UPS systems or line conditioner systems...but your mileage may vary.

Your best bet is usually eliminating the noise. That's why I'm trying to get the inverter company to dig in to this one.