Neon/LDR Tremolo Repair

[Rhodesplyr]

A Gibson, but not a guitar amp: I'm working on a 1968 Gibson G-101 combo organ (made by Lowrey for Gibson; The Doors "Waiting For The Sun") that uses a neon/dual LDR circuit both for Percussion and "Repeat." This circuit is one of the most common things to need repair in these organs. I've read conflicting accounts of how best to fix it, so I wanted to check with people with greater expertise with neon lamps and triggering circuits.

On this G-101, the circuit works, but poorly. The Repeat/Tremolo rate is inconsistent, so I suspected the oscillator/triggering circuit, but I found via my scope that the triggering signal has consistent timing and a nice square waveform. It appears to be the neon bulb itself that, for reasons I don't understand, can no longer follow the triggering signal. (Others have observed the same thing in other G-101s.) It will fire consistently for, say, four times and then lag. I can watch the bulb's flash and the scope trace at the same time, and they're not in synch.

The neon/LDR assembly was obviously designed to be replaced as a unit, but since they're long gone, the usual method of fixing it these days is to replace the neon lamp. The catch is that nowhere in the Gibson/Lowrey documentation are the lamp specs given. It's a T2 style lamp with flying leads. Mouser lists NE-2 type lamps with varying design currents, mostly with a 90V breakdown voltage and 120V rating, but, though it's not linked specifically to this organ model, one Lowrey parts site lists an NE-23 neon lamp, rated at 75V.

The reason I'm elaborating all this is that people report widely varying degrees of success with lamp replacement, and I'm wondering if it has to do with not choosing the correct lamp for the application. The sites that deal with this repair are pretty fuzzy on this point.

What I find via my scope, going directly to the lamp leads, is a ~95-100V peak to peak square wave. So, given that triggering voltage and the circuit (attached below), what should I look for in a neon replacement lamp?

Also, one of the complaints about this tremolo circuit is that it tends to bog down at faster speeds. I know that a big part of that in any tremolo circuit is the LDR response rate, but given that I'm leery of getting into the brier patch of LDR replacement, is there anything I can do without radically redesigning the circuit just to make the lamp's performance as snappy as possible? For example, the 100V DC supply is running just slightly high, around 105-110V. Should I stiffen the supply with more capacitance and/or zener limit it to 100V?

On the other hand, if there's a pre-made neon/dual LDR assembly available for another piece of gear that might work in this one, I'm open to experimenting with an upgrade.

Sorry for the long description. I admit a large amount of ignorance around what parameters to pay close attention to in order to make a neon/LDR optocoupler work best, and I'm trying to use this repair as a way to fill in that gap.

Here are the relevant schematics. Q5, the driver transistor, crosses to a standard 2N4410.

Percussion:Repeat.pdf

G-101 Power Supply.pdf

You can see the lamp/LDR assembly on the linked document below, but none of the Combo Organ group discussions have really shed light on why some replacement lamps work better than others.

http://www.combo-organ.com/G101BulbReplacement.pdf

[Enzo]

This figures into serveral of your items. The NE2 bulb has been around forever, but it is an indicator lamp, not a precision component. Various people make them. SO there are differences between manufacturers, plus batch to batch variations and even variation within a batch. SO one bag of them might fire at a higher or lower voltage than another.

As they age, they blacken inside their glass. And over time, the glass seal is not perfect, and air molecules can get inside the glass. If you just want a light to tell you the power is on, none of that matters, but if you want some consistent circuit action, then it matters.

You can go to CHicago Miniature, Shogyo, and others to see the range of neon bulbs available.

Once conduction starts on a neon bulb - the firing voltage is reached - you can reduce the voltage and still have conduction. There is probably a term of it - quenching voltage maybe?

SO your circuit no doubt need the lamp to fire and extinguish repeatedly. You can set up a simple bench test circuit, and characterize a few bulbs, then compare the circuit they run in. Does the voltage range in it reliably cross the thresholds?


None of this circuitry is precision, just my opinion, but having voltage 5-10% high ought not matter. AFter all, the Q5 circuit is just a blinker.

So you scope the voltage across the bulb, and it peaks nicely but the bulb doesn;t always fire? You say it is not in sync, DO you mean the scope trace has no relation to the bulb action, or do you just mean the trace may pulse, but the bulb sometimes ignores it? If that is the case, can we at least say every time the bulb flashes it is in sync with a pulse?

Look at it this way, if you cannot get the bulb to fire reliably, then nothing that follows will work well either. If we get the bulb blinking the right way, then either the photocell will do its job - sorry they will do their jobs - or it won't, and that will be the next step.


Oh, and is that 100v clean? Does it sag hard with each bulb firing? Does it hold up? Those filter caps are 40 years old after all.


Look at the circuit a moment. Q5 drives the lamp. What exactly appears at the collctor Q5? That is your square wave? The top of the square wave ought to be 100v or whatever your supply is. What is the bottom of the square? In other words how close to ground will Q5 pull it? I('d bet pretty close. You have the 10k resistor to limit current. I'd be betting you could just ground Q5-C and turn the bulb on steady. WHy? You could then measure the resistance action in the two photocells. Pretty hard to do that when the circuit is pulsing. Shorting Q5 C to E will not harm it.

As to why inconsistent results? WHo knows what they did? Inside the NE2 are two anodes. Put AC across it like on a pilot light, and both glow. But on DC only one glows - no I don;t remember which polarity glows, easy enough to find out on your bench. This matters because it is an optical system. And there are sensors on two opposite sides. You have to be careful when you install a bulb that the dark anode doesn't obscure a sensor from the glowing one. If someone just glued a bulb in there, they may cause a shadow and get poor results. So line up the bulb so each sensor can see both anodes.

The lamps are not precision, so as I mentioned, the firing voltage one guy gets may not be the same down the road in another shop.

[Rhodesplyr]

Enzo,

Thanks for the suggestions.

The bulb will fire in sync with the scope trace a few times in a row, but then, there's a delay in firing. It's not just missing a beat. With the two anodes, I suppose it behaves like a capacitor until it fires. It's almost as if the bulb is charging, but there's a lag before it actually fires. It seems odd to me, but, as I said, I've heard other techs describe the same phenomenon in the same organ.

I used a dual trace scope for my test and put the two leads directly across the output to the lamp, using the A-B function. That's where I observed the 100V p-p pulse.

After posting, I tried alligator-clipping a new 100uF capacitor between 100V and ground, and it made the effect's attack a good deal more crisp--definite improvement. However, it didn't change the lamp's inconsistent firing.

Since the lamps are not expensive, I'm leaning towards just buying a batch and testing to see if one works better than another.

David