1103 replies to this topic

[mrarcade]

Everything was ok until I fired the solenoid with the config tool. The power supply still works. It has 24v and I can fire solenoids with the 24v which has a higher amp rating.

[Suikazz]

I finally started testing dof. Not having any luck. I am afraid the pwm ics from china are bad. nothing but input buttons and nudge will work. Strobe will light when - is hooked to pwm -, not with strobe pin. same with small led, they will light when I hook direct to -, but not through any jp8 pin. flashers are the same. I am testing with a 3 led 5v pushbutton lamp. Oddly it was much brighter when hooked to small led than when hooked to flasher. hook direct to - and I get dim led. nothing through the flasher pins.

 

When I tested the knocker (with diode, + to banded end) it worked once and the psu I hoped would work went into shutdown, It only has 1.8a available for 48v. I turned the knocker off and reset the psu. The knocker seemed stuck on, when I plugged the psu back in the knocker would fire even though it was set to OFF in the config tool.

 

Is the strobe controlled by a pwm ic? If not, I will have to inspect my soldering etc again.

 

Up too late, time to sleep.

let me guess:

You got ICs with marking 77A357K4E on them? I have around 50 of them and none work.

All short immediately on Pin1 or have a constant 2.xxxV running through

try to get a refund immediately with aliexpress (i assume you bougth them from there)

[fhjui]

These are Fake-ICs and a fraud from the seller.

I have repaired a pinscape-board for a friend of mine. He had bought the same ICs. All are not working (of cause)..

When you wipe a qtip with aceton over the top surface, the qtip is black and removes the marking. He tries to get a refund from Aliexp.

[mrarcade]

I bought 8 on eBay from China. 7 are marked 58ZY097, 1 is marked 73M89AT then smaller superscrpit underlined G4.

 

I have put the unique one into main board IC1. Strobe works and I tested one of the flasher outputs and it works. Looks like I have 7 junk and one good TLC5940NT.

[mrarcade]

I replaced Q1 and now the knocker works with 24v.

 

Now I just need 3 more TLC5940NT (actually would like 7 so I have backups)

[mjr]

I replaced Q1 and now the knocker works with 24v.

Now I just need 3 more TLC5940NT (actually would like 7 so I have backups)

 

That's all good progress - the quality ratio with the TLC5940NT's is awfully frustrating, though.  How does the knocker sound at 24V?

[cerimorgan]

Just giving a second opinion on knockers and 24V, I much prefer it to 48V. I bought the B-10686-1 Williams/Bally replay knocker and at 48V, it scares my wife a floor away from the pinball cabinet. 24V sounds plenty loud to me

[mrarcade]

I currently have the knocker mounted to the back wall of the body. It is a self contained knocker where the top of the bracket is over the coil and that is what the plunger hits. It doesn't have the loud sound that I remember. Older games had the knocker mounted in the head and the plunger would actually strike a metal plate mounted to the inside top of the head. I think that setup might make it louder at 24v.  I just tested the knocker straight from the power supply. 48v doesn't sound much louder than 24v at all. I will try mounting the knocker in the head with the top plate flush against the top of the head and see if it sounds more like I remember.

 

I can work on getting the knocker sounding its best, mounting the strobes, and make a thin flasher bar. I can wire all that up and it should all work with just IC1. I bought some proto boards from Amazon so I can make flipper/magnasave rgb led boards. I bought 5mm square piranha rgb leds. I will be using the small led ports once I have more TLC5940NT's. I know resistors aren't required since the small led is limited to 30ma, but should resistors still be added to make the different colors "even"?

 

Forward Current Per Channel: 30ma

Forward Voltage (typical) Red: 2.1v

Forward Voltage (typical) Green: 3.8v

Forward Voltage (typical) Blue: 2.7v

 

[mjr]

I think a lot of the knocker sound does come from percussion against the wood, so hopefully it'll sound better when you get it mounted there.  You really should be able to get the same effect by having it strike against any cabinet wall - there's nothing magical about the backbox roof.  Most of the 80s/90s machines mount it there, but TAF and a few others mount it in the main cabinet.

 

I don't think you'll need to tweak the LED brightness with resistors, but I guess you can experiment to see.  RGB LEDs do tend to have pretty different brightness curves per color segment.

[mrarcade]

You really should be able to get the same effect by having it strike against any cabinet wall - there's nothing magical about the backbox roof.
 

The knocker has to be mounted vertically since it has a gravity return, if I want it to "strike" the surface then the backbox roof is the only surface I can use.

[mjr]

The knocker has to be mounted vertically since it has a gravity return, if I want it to "strike" the surface then the backbox roof is the only surface I can use.

 

True - the cabinet-mounted TAF knocker has a spring.

[mrarcade]

 

 

I slapped together an EAGLE plan for an adapter board, in case you run out of possible ebay sellers.

 

http://mjrnet.org/pi...DIP-adapter.zip

It's untested, so use at your own risk, and you might want to give it some scrutiny to double-check my work.  It requires the HTSSOP28 version of the TLC5940, which Mouser has in stock, and a pair of SMD 14-pin 2.54mm pin headers. I designed around Preci-Dip 350-80-114-01-899101; other similar products might fit, but it would probably be better to tweak the board to match the substitute footprint rather than try to use it directly.

 

 

I moved this quote into this topic from my Stellar Wars build.

 

I want to try this. I tried the websites you suggested, but they require gerber files to get quote/order. One of them has base shipping of $26. As always, I would like to do this is inexpensive as possible. Is it possible to rotate the surface mount chip 90 degrees and use through hole 14 pin dip headers? Or use the smaller VQFN TLC5940RHBR? Then I would feel comfortable soldering the headers on first, then doing the surface mount ic. If rotating the ic or using the TLC5940RHBR won't work, can you provide gerber files so I can get quotes for assembled units?

 

I am going to download Eagle and test if I can use it.

Edited by mrarcade, 30 January 2022 - 07:30 PM.

[mjr]

can you provide gerber files so I can get quotes for assembled units?

 

The Gerber files are vendor-specific, so it's too much work for me to upload them for all the vendors.  But you can easily generate them with the free EAGLE - all of the vendor sites have their own CAM files that you download and feed into a dialog in EAGLE - the vendor site should have step-by-step instructions for their particular setup.

 

 

 Is it possible to rotate the surface mount chip 90 degrees and use through hole 14 pin dip headers? Or use the smaller VQFN TLC5940RHBR?

 

You could try, but this is the best geometry I could come with.  You could also try the VQFN, but that might be too hard to hand-solder.  The HTSSOP chips are already a little tricky, but I've found them doable.

 

(The reason that I had to use the SMD pin headers and wacky chip rotation is that the through-hole pin headers don't leave enough space for routing the traces.  The only way I could get it to work in the available space was to get the pin headers entirely out of the way on the top layer, which obviously isn't possible with through-hole headers.  If these boards end up being viable, the thing to do would be to get one of the fabs that offers PCB assembly to make a fully assembled batch.  The cost-effective way would be for somebody to order a hundred or so as a prospective group order, like I've been doing with the blank expansion boards.  This would be a bit more of an up-front investment, though, and I shudder to think of the Quality Control burden that whoever did this would be shouldering.  The China fabs are a bit hit-or-miss.)

Edited by mjr, 30 January 2022 - 10:54 PM.

[mjr]

One of them has base shipping of $26. 

 

That is typical with the China fabs, I'm afraid.  With China sourcing, you can either pay about $30 for DHL 3-7-day shipping, or pay $15 for China Post 30-45-day shipping.  For every board I've ever had made, the cost has been 40% for the nominal production cost and 60% for the shipping charges.

 

For a small board like this, you should have them made by oshpark.com.  They're domestic, so they charge 10X what the China fabs charge for the nominal production cost, but OSH Park includes free shipping.  They charge by the square inch, so they're often a better deal for small boards like this. This board is 0.6525 square inches, so it'll cost you $5 x .6525 = $3.26 at OSH to have 3 of them made, or about $1/board.  It's outrageously expensive compared to the pennies per board if you had 1000 made by a China fab, but if you're only having three of them made, you save 95% on the shipping charges.

Oh, and another nice thing about OSH: they accept .brd files directly.  I still generate Gerbers for them because their default CAM rules for some reason don't include the part name/value text on the silkscreens, but if you can live with that, it only takes about 60 seconds to order a board through them.

[mrarcade]

I ordered 3 yesterday, I double checked the 7 tlc9540 and they are definitely bad.

Edited by mrarcade, 31 January 2022 - 01:17 PM.

[Suikazz]

I ordered 3 yesterday, I double checked the 7 tlc9540 and they are definitely bad.

Same here, ordered 20 new, old 50pcs were all bad.

[mrarcade]

I have a question about the small led forward voltage resistor. If I have it set up for 30ma, with 30ma leds in the 5 buttons on the front of the cabinet, and the flipper rgb leds are 20ma, will the rgb leds be too bright and fail prematurely? Or if I set it up for 20ma will the button lamps on the front of the cabinet be dim?

Edited by mrarcade, 09 February 2022 - 01:02 AM.

[mjr]

I have a question about the small led forward voltage resistor. If I have it set up for 30ma, with 30ma leds in the 5 buttons on the front of the cabinet, and the flipper rgb leds are 20ma, will the rgb leds be too bright and fail prematurely? Or if I set it up for 20ma will the button lamps on the front of the cabinet be dim?

 

Probably yes and yes.  I think the most rigorous approach for a mix of LEDs like this is to set the common resistor on the main board (R5) to match the LED with the highest current draw, and then add individual current-limiting resistors in series with the lower-rated LEDs.  The R5 setting sets the maximum current per circuit, not the minimum, so you can get a lower current on any individual output using a resistor.  Just wire it as though you were using a simple voltage supply with no current limitation.  That approach forfeits the nice "no resistor needed" feature of the small-LED outputs, but it at least lets you use them for a heterogeneous mix of LEDs like you're doing.

 

Another alternative might be to set R5 for 25mA, and just sort of split the difference between your two LED types.  I wouldn't suggest this if they were 10mA and 100mA respectively, but 20/30 are close enough that I think you can get away with it.  Going 5mA over the nominal rating will probably somewhat reduce the life of the smaller LEDs as you suggested, but probably not by much - I bet if you look at the data sheet for those LEDs, 20mA is only the nominal working current, and it'll list the Absolute Maximum as somewhat higher, probably more like 50mA.  So maybe the LED will last for 28 years rather than 30.  And as for the bigger LEDs, running them at 85% of the nominal current shouldn't make much difference to the perceptual brightness.  LED brightness isn't linear in power the way that incandescent bulbs are - in the operating zone, the brightness-vs-current curve is usually almost flat, such that increasing the current only barely increases the brightness, with most of the excess power converting to heat.  So by the same token, reducing the current will mostly just make it run cooler without much reduction in brightness.  And perceptual brightness has the very opposite non-linearity - you have to cut the physical light power by about a factor of 2 to see a substantial step change in brightness level.  So between those two factors, running a nominal 30mA at 25mA shouldn't make a lot of difference in the way it looks.  But you can always build a little test circuit with resistors to compare two of them side by side, one at 25mA and one at 30mA, to make sure you find the difference acceptable.

Edited by mjr, 09 February 2022 - 03:04 AM.

[mrarcade]

Wish me luck. The 3 boards arrived from OSHPark and the 3 TLC5940PWPR & 3 PreciDIp 350-80-114-01-899101 arrived. I had my first experience with surface mount. I used a toaster oven because I had to solder parts to both sides of the board and the boards are so small. I was afraid if I used hot air, did the legs first, then flipped and did the ic that the solder holding the legs would melt and the legs would fall.

 

I masked the 2 solder pad edges on the outside of the ic footprint to keep the solder where I needed it and also giving me what I hoped was a thickness to shoot for...the thickness of the blue masking tape. I used a toothpick to apply the solder paste to the pads. Then I set the ic on making sure I orientated it with the dot on the ic lining up with the do on the pcb. Then I flipped the board over and applied the paste to the pads for the 2 x 14 pin headers. I set the headers on the paste and then sat the whole assembly ic side down onto the front edge of the toaster oven pan. I had the toaster oven with quartz elements heated to 150-200 degrees. I slid the pan into the lower slot so I could easily see the assembly and closed the door. I let is sit about 30 seconds then turned the temp knob up until I heard the "click" meaning the heating element was turned back on. Then I waited until I heard the click meaning it was turning off. As soon as I heard the click I bumped the temp up again. I repeated this until I saw the paste start to turn silver. then I left the temp alone and watched until it looked like the solder was just starting the bubble or boil, then I turned the temp down to 150-200. I waited about 10 seconds then opened the door. I waited about 30 seconds then using a hot pad glove I removed the pan and set it on the bench. I waited about a minute then I moved the assembly to the bench. After another minute I could handle the pcb.

 

First try I got some pretty big solder globs. two on one side and one on the other. I repeated the process, but this time I applied the paste looking through a lighted magnifier. The second one had two slight solder bridges. The third had one slight solder bridge, I cut one of the bridges, you can see how I cut way too far. Luckily I didn't cut the trace. It took some work getting the solder bridges taken care of. I used a fine tip soldering iron to pull the solder to the ends of the fingers and kept cleaning the tip. Eventually I cleared all the solder bridges.

 

I tested for continuity between adjacent pins and had none. I just hope I didn't get the ic's too hot.

 

I will be testing them tonight or tomorrow.

 

 

Edited by mrarcade, 13 February 2022 - 05:30 AM.

[mjr]

Looks good on the adapter chips - fingers crossed they testing goes well.  I've had that experience with solder bridges on these fine-pitch parts myself.  They just don't make these things for human hands any more.  Fortunately you have genuine TI chips this time, and I've always found their products to be pretty darned rugged physically, so I'm optimistic that they survived the repeated heating cycles.