LED Flicker

[2Moons]

Hello Pinball Folks:
I'm working with a Bally Lost World. The controlled lamps flicker when I replace them with an LED. Bally applies voltage to all controlled lamps. To get the lamp to light, they switch on a ground path through an SCR located on the AS-2513-23 lamp driver module. If you take a look at the attached schematic, Q14 supplies the ground for the 1k lamp. If I put an LED in the 1k socket it flickers when it's turned on.
If I clip lead the gate of Q14 high to TP3, the led glows steady & bright. When I look at the output on pin 11 of the U1 decoder, it appears that it is only going high 2/3Rd's of the time. The LEDs are quick enough to show the flicker. The incandescent lamps are so slow to react that they show no apparent problem & it's a great way to save electricity. It appears to me that the only way to solve this problem & get an LED to work in this vintage Bally machine is to redo the soft ware on the MPU board & get rid of this energy saving fetcher.
Have any of you run in to this problem before, or more important found a way to sole it? Please take a look at this & let me know what you recon.
Thanks 2Moons

[jay]

I would not want to fool with the software or goto to any great trouble for that matter.

CHF probably has some ideas on this that are far beyond mine......

If you want a steady light I would drive the LED chain off of a double pull double throw relay using a latching circuit.

+----- SCR source -----------------------------------------  + Relay coil {} - -------------- Gnd 
                                                                    |                                                                     
                                                                    +------------------
                                                                                       |
+ ------Non Scr Power --------------------------+---------/switch 1/-----+
                                                         |
                                                         +---------/switch 2/-----+--------------Led----Led -----Led-----etc-------------Gnd

The SCR turns on the relay but then the nonSCR power source keeps it on. If you are switching grnds vs hots I am sure you could reverse this.

[channelmaniac]

Wouldn't it be simpler to use an RC network to dampen the pulse and remove the flicker?

[brichter]

Wouldn't it be simpler to use an RC network to dampen the pulse and remove the flicker?

Yes, it would.

[poppo]

It appears to me that the only way to solve this problem & get an LED to work in this vintage Bally machine is to redo the soft ware on the MPU board & get rid of this energy saving fetcher.
Have any of you run in to this problem before, or more important found a way to sole it? Please take a look at this & let me know what you recon.
Thanks 2Moons

It's not really an energy saving feature as much as the CPU needs to 'strobe' the lamp driver board (J14 P13). Besides lighting up the switched lamps, there are other things going on. In order for the lamps not to be lighting at the wrong times when the address and data lines are active, a strobe signal is sent to the board. This turns on the driver chips only when the CPU is talking to the lamp driver board. With incandescent lamps, this happens fast enough that you don't notice. With LEDs you see the flicker.

As for the solution, you can take a look at this http://www.allteksystems.com/ultimate_lamp.html It is deigned to eliminate LED flicker. However, I replaced just a few lamps with LEDs and they still flicker. I have an e-mail to Alltek asking about it. I'll post when I get a reply.

Take a look at the on-line manual for it. This board has a jumper that you connect to the switched voltage line to somehow eliminate the flicker. I'm only assuming it somehow keeps the driver chips 'latched' between strobes, although I don't see how. I'm going to look at the board and see just what that line does. Seeing how the board is compatable with the the standard one, there can not be much different.

[brichter]

I still think the RC bridge would be the simplest, just need to know the voltage, current, and frequency of the strobe.

[poppo]

Ok, after looking at the Ultimate Lamp driver board, I see that the jumper for the 6.5vdc goes to a bunch of resistor banks that are tied to the OUtPUTS of the SCRs. I have indicated the way these are wired in red below. I do not know the value of the resistors. Thus each output is pulled up to 6.5vdc when the jumper is used for LEDs. I am not a SCR expert, so all I can think of is that somehow it helps keep the SCRs latched between strobes. Other than that, the board is electrically identical to the standard lamp driver.

[KirkLasVegas]

Why not simply add a 10-100Mfd capacitor across the LED? It would fill in the time constant,no more "Flicker" or strobing.
10 to 100 Mfd 15 volt electrolytic caps are cheap. Can be soldered across the lamp sockets......


Kirk

[brichter]

I'd use a resistor to keep the response time fast...

[KirkLasVegas]

I'd use a resistor to keep the response time fast...

Huh???

[brichter]

Wouldn't want the cap keeping the LED lit for a millisecond longer than it was supposed to be on for, would we? 

[KirkLasVegas]

Wouldn't want the cap keeping the LED lit for a millisecond longer than it was supposed to be on for, would we? 

Yes, once the LED is forward biased. But if the pulse is narrow, the cap will take time to charge and the LED WONT light.
The reason the LED flickers is it's fast response time.the filament type lamp takes time to reach temperature,you dont see the flicker.
Some series "R" can help, but too much and the LED wont be bright.Added a capacitor across the LED wont affect the ON
brightness and will allow it to stay on a few miliseconds longer.The cap will absorb the fast strobe pulses on the bus.

Kirk

[brichter]

I understand about the pulse width and the LED's fast response time, as well as the filament heating/cooling cycle time. Don't forget the latency caused by the cooling time here, it's not just about the heating time. You get it on both ends of the cycle, slow on, slow off..  

Whatever the waveform supplying the lights, it will be enough to charge the cap with a resistor of the correct value in the circuit and bias the LED, as the LED requires substantially less power than the filament bulb did. Just need a scope on the leads to determine the supply characteristics. 

[poppo]

Why not simply add a 10-100Mfd capacitor across the LED? It would fill in the time constant,no more "Flicker" or strobing.
10 to 100 Mfd 15 volt electrolytic caps are cheap. Can be soldered across the lamp sockets......

Kirk

Actually, I had already tried that first, and it just made the flicker more noticable depending on the value of the cap.

Yes, once the LED is forward biased. But if the pulse is narrow, the cap will take time to charge and the LED WONT light.

Or stay off too long as I noted above.

[poppo]

Well, after two e-mails to Altek about their Ultimate Lamp Driver board, I have yet to get a reply.  However, I found that it does indeed work. The lamps I had been checking were being driven by the auxilary lamp driver board and not the main one.  So, the pull-up resistor on the outputs of the SCRs does indeed fix the flicker problem.

[2Moons]

Hello NLG:
Let me do a recap of where we are at and where we are heading. Once upon a time there was a led that flickered when it was installed in an older Bally or Stern pinball machine. The LEDS flickered on any machine that used a Bally As2518-14, AS2518-23, or Stern LDB-100. These lamp driver modules were used from 1977 thru 1989. These boards use SCRs to switch # 44 Lamps on and off in pinball machines. This has worked well for many years, until someone tried to use a led instead of a #44 lamp. The LED flickered. I discovered that the reason the led flickered was that the LED drew no current below it’s forward bias point that it would not hold an scr latched. If a resistor is added in parallel with the led, it would draw enough current to keep the scr latched and there was no flicker. (Life is Good). Through testing I found that a 470 olhm resistor was more than enough to keep any of the scrs that I tested latched. The 470 olhm resistor can be added in the socket of the led, soldered across the lamp socket, or soldered on the lamp driver module. The electrons don’t know or care where the resistor lives, as long as it will cause a 1ma load to the anode of the scr. All 3 ways have been done and they all work well.   This is not rocket science; all that is required is to add 1 Jumper wire to the lamp driver module. I have attached a file to this email that will show where the control lamp buss is located on the power supply rectifier board. On Balley machines this unregulated DC is fed through the machine to the outside or barrel of every controlled lamp in the machine. The tip of the lamp sockets go back to the lamp driver module and to an scr that switches on to make a ground path. If you hook the jumper between the buss wire for the resistor matrix on the lamp driver module and the DC supply on the rectifier board, you are good to go. I attached a photo. 
This is the prototype i made for Alltec. It's not pretty but i sure works well.
If you need LED stuff I suguest you look here. http://www.cointaker.com/category.sc?categoryId=97
Ron Googe
(818) 848-6805

[poppo]

Let me do a recap of where we are at and where we are heading.

I'm a little confused because the current Alltec boards already incorporates those resistors as I indicated a few posts back. So what is the 'where we are heading'?