Mad as a hatter for old Whelen stuff

[JohnMarcson]

I got this box of airport related Whelen stuff and in it was a foam paddle on a swivel with wires coming out of it. The foam was brittle and flaked away revealing a Mercury switch. When the padle is pushed backwards it triggers the electronics in the beacon and illuminates it. Not sure if it was a wind measurement device or what, but it's got good old fashioned mercury inside which makes the wife happy I didn't just go breaking it open.

Here is the foam paddle before it disintegrated


And here is the Mercury switch inside

 

[Stampeed Valkyrie]

float? simple switch to trigger levels?

 

[JohnMarcson]

float? simple switch to trigger levels?

Not sure if it was a float, an air resistance measurement device, or what really. When enough force moved the paddle the light blinked. Maybe for a marine application like bilge tank levels? It was in with a bunch of airport stuff, but I'm having trouble thinking of how it would benefit an aviation application. Maybe it was for wind gusts? A flot makes more sense, but there is no sign of water damage. The foam looked more sun damaged than anything.

 

[NPS Ranger]

Maybe a contact sensor for one of those aircraft pushback tugs? When the tug moves into position to contact the wheel assembly, it triggers the light so the driver knows to stop and hook up.

 

[JohnMarcson]

Maybe a contact sensor for one of those aircraft pushback tugs? When the tug moves into position to contact the wheel assembly, it triggers the light so the driver knows to stop and hook up.

Also possible. The "float" is so light it almost seems like air would trigger it. I even searched my oldest Whelen "non vehicle" stuff.

 

[kitn1mcc]

Almost looks home made or a one off

 

[JohnMarcson]

Almost looks home made or a one off

It was a whole system, no way it was home made. There our electronics inside the flashing beacon that are specific to the mercury switch. Now whether it was a one-off or a special order, that is certainly a possibility. The bigger possibility is I'm missing a large portion of the system.

 

[toon80]

I've seen float switches used to monitor dry bilges level on barges, so someone would notice it's technically sinking and pump it out, even tough it was relayed with the automatic bilge pump.
Many applications possible.

 

[JohnMarcson]

I've seen float switches used to monitor dry bilges level on barges, so someone would notice it's technically sinking and pump it out, even tough it was relayed with the automatic bilge pump.
Many applications possible.

That is how I am leaning. A water level warning for a place that should not be too wet or wet above that level. Whelen made freeze alarms with and without pilot lights as well as portable fire alarm and strobe boxes for events. Their "non-vehicle" lines were pretty varied.

Another thread mentioned that mercury switches were used as a park gear sensor sometimes. This could be an add on "not in park" sensor or something too. The list goes on, Whelen made too many one offs.

 

[dmathieu]

Just because mentiond, here are the various models and vintages of Freez-Alarms

 

[toon80]

Dan Mathieu,

Whenever I see those freeze alarms of your, I always think of the first Ghotbusters movie when the inspector shuts doen the grid power, and everything goes to hell.

 

[JohnMarcson]

So I did some more playing around with the device....

It seems each time mercury switch is pushed down and completes the circuit the device supplies voltage to the lamp socket for a short period of time. The lamp does not stay illuminated, and will not illuminate again until another full off/on cycle occurs. It is acting like each "trigger" gets you on "capacitor full" of energy, not a full constant supply voltage. I took the mercury switch out of the equation and just shorted the leads of the switch it was opening and closing and got the same behavior.

I shot a video, it will post a link when edited.

@RS485 ideas?

 

[JohnMarcson]

 

[RS485]

Can you possibly get part numbers of 1,2,and 3?


Also: does the circled part of the PCB appear to have been mechanically severed after-the-fact?

I'm thinking this might just be a defective flasher circuit.
Also, I can't make out what exactly the base of the socket is connected to. I see that tab coming off it but it obscures where its actually connected in the circuit.

If this really is a device meant to indicate on just the rising edge of something, the only application I can imagine is, if one is waiting for the toilet bowl to fill up to get a nice fresh flush, that brief confirming flash would be welcomed

 

[JohnMarcson]

Here is how the socket works. Straight voltage from input to the "housing" and black lead after traveling through the board and components to the center contact.


F2N5134 943



F2N5142 931



MMPS U01


Removal of something or just two solders that were too close ground off?


Here is a breakdown of the different wires

 

[JohnMarcson]

Data on the components

 

[RS485]

Thanks for all those details!
The reason I'm interested in the separation that appeared ground off was that it looked like that might be where the mercury switch was "inserted". Mercury switch "connects" at or near zero ohms.
You might then ask, "So why didn't they just insert the mercury switch between the power supply and the device?" -- I don't know.
_maybe_, once triggered, it was supposed to say _on_ - or maybe stay on and flashing?
(Still trying to piece together the circuit...more to follow if/when things make more sense).

Wondering how much current your LED head draws...the MPSU01 NPN transistor looks to carry the highest rated current of 2A. I assume it is controlling the lamp. The other transistors (NPN and PNP) look to be only 0.5A and look to do some kind of switching.

 

[JohnMarcson]

Thanks for all those details!
The reason I'm interested in the separation that appeared ground off was that it looked like that might be where the mercury switch was "inserted". Mercury switch "connects" at or near zero ohms.
You might then ask, "So why didn't they just insert the mercury switch between the power supply and the device?" -- I don't know.
_maybe_, once triggered, it was supposed to say _on_ - or maybe stay on and flashing?
(Still trying to piece together the circuit...more to follow if/when things make more sense).

Wondering how much current your LED head draws...the MPSU01 NPN transistor looks to carry the highest rated current of 2A. I assume it is controlling the lamp. The other transistors (NPN and PNP) look to be only 0.5A and look to do some kind of switching.

The LED is what I use for low amp testing because it will start to illuminate at .2 amps and draws .68 MAX. It will also glow dim at low voltage unlike some LEDs, so it's good for testing. It is not reaching full brightness, I measure .4 amps peak when at full supplied voltage.

I wish I had the original bulb, it was shattered. The base was nothing special, single contact on the bottom, 795 style nubs as the socket would indicate. Very narrow diameter socket. The interesting thing is that when the red lead is positive and the black negative the bulb housing is positive and the bottom inner single point contact is ground, which I guess really doesn't matter with some bulbs. I can't really think of a bulb that would be worth much for signaling at half an amp, the incandescent bulbs i have connected haven't illuminated at all.

Can we be sure this is 12vdc? Could it be higher voltage DC like 24, 28, or 36? Any evidence of it being AC? (no signs that I saw except the 115 stamped on the board).

To me it acts like it is given "a capacitor load" each time it is triggered, and then it then dissipates with use . You can get current back fast, there isn't much "load time". It takes a few seconds to drain, then it's ready to go immediately. Or if you break the contact it is ready again as soon as you reconnect the leads.

The black wire "in" and the black wire to the switch are connected. The red wire "in" and the red wire connected to the switch enter different spots on the board. I assumed the red "input" went to a charging circuit and the red from the switch went to what triggered it. The switch is full input voltage, the measurements at the switch and in the bulb socket mirror each other.

It looks like input ground and switch ground are simply tied to each other. Then Input power goes right to the housing of the bulb. It is The red lead from the switch that seem to actually be the patchway. There are 3 transistors, one diode and one capacitor; the rest are resistors from what i see. Let me know what I missed on this tracing.



Here are some see throughs with and without mirroring





Mirror 1




Mirror 2

 

[RS485]

This is _great_ detailed info! but I'll need some time to process it.
FWIW, I don't think its AC -- the 2N5134 and 2N5142 don't want to see much voltage diff between collector and emitter (base being the "switch") and there seems to be no protection against reverse polarity -- i.e DC is expected.
The NPN controlling the lamp however, will tolerate 30v between collector and emitter.
After mapping out the entire circuit, we might find a voltage divider that ensures the "control" side of the circuit only sees a voltage less than ~10V...but the bulb and MPSU01 are allowed to see up to ~30V (based on the data sheet).
Still, the way I see it is either:
a) the observed behavior is of a defective circuit, meaning we need to "fix" it to understand its context or,
b) the observed behavior is correct, meaning we need to figure out what circumstances this device found itself in so as to be useful.

At this point, I need to use the great info you provided and map out the circuit. That's the only way I'll be able to understand it.

Good news! -- looks like the three "active" parts (the transistors) are still out there and available. So maybe I can make up a circuit to tinker with.

If you're inclined to tinker, I'd be curious of what effect happens when piggy-backing another cap on the one cap that's there....i.e. does it stay on longer? does it take longer to turn on? Very curious...

 

[JohnMarcson]

This is _great_ detailed info! but I'll need some time to process it.
FWIW, I don't think its AC -- the 2N5134 and 2N5142 don't want to see much voltage diff between collector and emitter (base being the "switch") and there seems to be no protection against reverse polarity -- i.e DC is expected.
The NPN controlling the lamp however, will tolerate 30v between collector and emitter.
After mapping out the entire circuit, we might find a voltage divider that ensures the "control" side of the circuit only sees a voltage less than ~10V...but the bulb and MPSU01 are allowed to see up to ~30V (based on the data sheet).
Still, the way I see it is either:
a) the observed behavior is of a defective circuit, meaning we need to "fix" it to understand its context or,
b) the observed behavior is correct, meaning we need to figure out what circumstances this device found itself in so as to be useful.

At this point, I need to use the great info you provided and map out the circuit. That's the only way I'll be able to understand it.

Good news! -- looks like the three "active" parts (the transistors) are still out there and available. So maybe I can make up a circuit to tinker with.

If you're inclined to tinker, I'd be curious of what effect happens when piggy-backing another cap on the one cap that's there....i.e. does it stay on longer? does it take longer to turn on? Very curious...

I'm always up to tinker. An extra capacitor would be an interesting test. What more capacitance does to the behavior would be a good metric.

I also can't get the "damaged" area next to the ground out of my head. It looks like there could have been an additional connection, but the system seemed very complete as far as space management and holes for cables etc. I'm trying to think of a simple component that would go there (like a diode between the two wires) that would alter the behavior. The way the grounds are just tied together via the board seems odd, but I don't know what I'm basing that on not knowing the purpose of the device. But yes, let's tinker.

 

[RS485]

Wondering if this might be a 6V circuit. The path to the "Base" pin of the MPSU01 is pretty simple (follow the green path though a voltage divider).

Assuming that PNP (2N5142) is switched ON, then the 12V supply would produce 8+ volts at the base of MPSU01 -- more than spec. But at 6V input, it would produce 4.3V at that base.

I go here:

Voltage Divider Calculator

Try our easy to use Voltage Divider Calculator. Enter any three known values and press Calculate to solve for the other.

ohmslawcalculator.com

and plug in 377 (47 + 330) for 1st resistor, and 1000 for 2nd.

 

[JohnMarcson]

Wondering if this might be a 6V circuit. The path to the "Base" pin of the MPSU01 is pretty simple (follow the green path though a voltage divider).
View attachment 244111
Assuming that PNP (2N5142) is switched ON, then the 12V supply would produce 8+ volts at the base of MPSU01 -- more than spec. But at 6V input, it would produce 4.3V at that base.

I go here:

Voltage Divider Calculator

Try our easy to use Voltage Divider Calculator. Enter any three known values and press Calculate to solve for the other.

ohmslawcalculator.com

and plug in 377 (47 + 330) for 1st resistor, and 1000 for 2nd.

You are correct At 6vdc it supplies constant current when activated.

My tester with direct 6vdc



My tester when hooked to the socket of the device at 6vdc and "mercury switch" activated


Volts at socket


Volts over the closed switch leads


With the device powered by 6V the socket gets the full 6v when the low voltage mercury switch is closed. Now I am left wondering what bulb it used and why bother with this circuitry at all? Could you not just use the switch and a 6 volt bulb? What is the circuitry gaining?

The voltage going through the mercury switch is approx. 2vdc. That means the circuit board is essentially a relay switching a 6 volt load with 2 volts? So we have a float with a low voltage mercury switch inside that illuminates a 6vdc lamp when lifted into the closed position. The circuit is really a slightly more complex low voltage switching relay using transistors, a capacitor, and resistors to trigger a 6v lamp with a low voltage switch. This was cool to figure out, but the kid in me wanted it to blink, although the idea behind the float probably make more sense not blinking.

Also, I need to remember to use this resistor color code tool too.

The lamp holder looks like BA9 based on the measurements, the bulb shards and base were consistent with that lamp type, and the base looked a lot like

Lamp Type: T2-1/2 Miniature ​

Volts: 6V, Watts: 0.84W​

which would be a .14 amp draw, well within what this is putting out.

​

 

[RS485]

Here's my estimation of the schematic:

I'm really interested in the full part number for "D1" and the voltage and cap rating of C1.
_if_ the diode is a 1N4896 (as per this datasheet ), its a very precise 12.8V zener diode...and I *think* it would explain why it turns on, then turns off, then stays that way...the zener would bypass the ability for Q3 to drive Q2 (which, in driving Q1 causes the lamp to turn on).

I hate to say it but if it really is a 6V circuit, there could have been damage. Also, I still am not sure the LED array is acting just like the intended lamp. I wish we knew exactly what lamp was there.

Taking a step back, I think one possible use-case to consider is filling a tank. Maybe, the float switch was meant to just indicate to the the person filling the talk, "it's full".

I'm going to build me one of these Its so simple yet so elusive.

 

[RS485]

You are correct At 6vdc it supplies constant current when activated.

My tester with direct 6vdc
View attachment 244118

My tester when hooked to the socket of the device at 6vdc and "mercury switch" activated
View attachment 244117

With the device powered by 6V the socket gets the full 6v when the low voltage mercury switch is closed. Now I am left wondering what bulb it used and why bother with this circuitry at all? Could you not just use the switch and a 6 volt bulb? What is the circuitry gaining?

The voltage going through the mercury switch is approx. 1-1.5 vdc. The circuit board is essentially a relay switching a 6 volt load with low voltage. So we have a float with a low voltage mercury switch inside that illuminates a 6vdc lamp when lifted into the closed position. The circuit is really a slightly more complex low voltage switching relay using transistors, a capacitor, and resistors to trigger a 6v lamp with a low voltage switch. This was cool to figure out, but the kid in me wanted it to blink, although the idea behind the float probably make more sense not blinking.

Also, I need to remember to use this resistor color code tool too.

The lamp holder looks like BA9 based on the measurements, the bulb shards and base were consistent with that lamp type, and the base looked a lot like

Lamp Type: T2-1/2 Miniature ​

Volts: 6V, Watts: 0.84W​

which would be a .14 amp draw, well within what this is putting out.

​

My point about the LED array vs. the lamp is this: At an electrical level, the lamp appears as a resistor and at 6V and .14 amps, it present a resistance of ~42 ohms (though I believe the resistance of an incandescent lamp is known to differ depending on being lit or not).
Could be that the LED array is presenting exactly what the design expects...but if it has some form of active power managment circuitry within it (like Whelen's seem to have), then it may not be a good stand-in for an actual lamp.

 

[JohnMarcson]

My point about the LED array vs. the lamp is this: At an electrical level, the lamp appears as a resistor and at 6V and .14 amps, it present a resistance of ~42 ohms (though I believe the resistance of an incandescent lamp is known to differ depending on being lit or not).
Could be that the LED array is presenting exactly what the design expects...but if it has some form of active power managment circuitry within it (like Whelen's seem to have), then it may not be a good stand-in for an actual lamp.

That's a good point, an LED is a diode that makes light (and as you pointed out usually uses other circuitry in critical applications) whereas a lightbulb is a resistor that emits light and heat. So like you said, not really the same thing. I would like to see an actual bulb in it. Maybe that resistance is required for it to do something else? Shut off at a point, reset at a point? Fully drain and turn off? I don't have anything 6 volt that would mimic an incandescent bulb, so I can add a 6vdc bulb to my next Amazon order.

My math or theory may be wrong...please correct--I guess I could find a bulb with a near equivalent resistance by means of changing the watts to match. To keep 42 ohms you would need the bulb to draw .28 amps at 12 volts (12v/.28a=42.8). Volts/amps=ohms. If we are shooting to keep the same resistance (42 ohms) , but with 12 volts (amps x volts=watts) 12v x .28a =3.36 w ... so a 3.4 watt 12 volt bulb?

 

[RS485]

Could be wrong but here's my reasoning:
Initially when mercury switch makes contact to provide ground to the circuit, Q1, Q2 and Q3 are off.
Current reaches the base of Q3 (after passing through the internal resistance of the lamp that is at this moment, off).
As well, C1 begins charging.
Q3 begins to turn ON thus forms a connection between its collector to ground via its emitter. This presents a negative potential the PNP Q1.
PNP transistors only turn on when their base is ~.7 volts _lower_ than their emitter (their high side).
As Q1 turns on, it passes current to Q2 which connects its collector to emitter so as to connect the lamp to ground...viola! we see light now and this path no longer "feeds" Q3. However, C1 is discharging to Q3's base...so for a while, C3 will continue to feed Q1 which will continue to feed Q2 and the lamp will remain on until C1 is sufficiently discharged.
I know this is not a perfect analysis...but if its in the ballpark, I think this _is_ a flasher circuit.
I can't see any reason to have C1 in the picture otherwise.

And _if_ it really is a 6V circuit, then some damage may have occurred to the transistors (I already ordered some , I'll send you a care-pack when I get 'em).

 

[JohnMarcson]

I appreciate that, that reminds me I still have a power supply package for you (that I don't expect back) but I think you will enjoy.

I found a 12vdc cruise lightbulb that on 6vdc on my power supply glows dimly but lights. It's pulling .2a at 6v. So that is ballpark-ish at 30 ohms. It does nothing when hooked up. Then I found an instrument cluster bulb that pulls .145 amps at 6 vdc, so even closer on resistance to what I think was the original bulb. Nothing.

I agree that you have described a flasher circuit, and I am left with the conclusion that it is in fact damaged. I ordered what I am pretty sure is the original bulb which will arrive tomorrow so we can see the results there too, and maybe when repaired. At least this is more interesting than being a method to switch 6v with 2v.

 

[JohnMarcson]

After reviewing the diagram and explanation, reviewing how PNP and NPN transistors work, and ensuring the bulb I have is close to what was used; I am fairly confident that (Q2) NPN MPSU01 is bad. More later, but following the schematic the device seems to be functioning theoretically until the point of connecting the socket to ground. I think conditions are met for Q2 to be triggered (current at the base and collector), and jumping "past it" illuminates the bulb. I don't think Q1 or Q3 are bad, but I could be overlooking something. Regardless that's the logical part to replace first IMHO.

Update- I desoldered, removed and tested this component and it failed all three multimeter tests that i was able to perform. So this gets replaced first.

Update 2- I had another NPN of a similar size and spec, although not the same exactly. The device now illuminates but doesn't quite flash. There is a voltage drop to Q3 when Q2 activates the bulb, but Q3 never loses its power enough to become only activated by capacitor, it stays activated via Q2 or some other problem pathway(keeping the bulb illuminated) so something isn't right. So it might just be I need the right bulb and/or the right NPN (Q2) and it will work. OR maybe I need to swap out Q3 or Q1 too. C1 works though as does the idea that connecting Q2 collector to emitter lights the lamp.

 

[RS485]

So if Q2 doesn't sufficiently connect, (i.e. if it isn't fully switching on), Q3 would still be fed. Which NPN part are you using for Q2's replacement?
You _could_ try to put a resistor across the C and E of Q2 to try to draw a bit more current though the lamp to try to starve Q3 - that shouldn't hurt anything.
(also, any idea on the part #for D1 or the C1's values? )

Also, I'm gettin' a few boards made:

 

[JohnMarcson]

It works..... standby for video. I replaced Q2 with ZVN 420 6AV, which was the only component that I had in my bin of electronics that both fit and had the same behavior a Q2. I still want to put the right part there.

 

[JohnMarcson]

As far as testing I removed Q2 and replaced it with a wire and my hand. I would power the device before connecting the wire from socket to ground (simulating Q2). I metered all sides of Q1 and Q3. Then I would connect the wire standing in for Q2 and meter Q1 and Q3. They seemed to function as if Q1 had tripped Q2 except that the light did not go out because Q2 wasn't powering it, so the loss of path to Q3 and subsequent Q1 turning back off still occurred but didn't have any effect on the lamp. From what I could tell I had essentially substituted myself holding the wire for Q2, except I was not controlled by the interactions of Q3 and Q1 (which occurred once per time I grounded the socket). I was still able to note that their "chain reaction collapse" occured when the path connected the lamp to ground, it just didn't do anything because nothing depended on it.

 

[JohnMarcson]

Video:

Note the pile of components on my bench that didn't work or fit near the end of the video.

 

[RS485]

NEET! Congrats!

 

[JohnMarcson]

NEET! Congrats!

I feel like I should still order the correct part before soldering it in. Your explanation of how this worked saved it from being discarded as "incomplete not working" and caused me to put the time into fixing it; your electronics knowledge is much appreciated.

 

[JohnMarcson]

Thanks to @RS485 the correct NPN transistor is in place at the Q2 spot. Everything is working like it should, Including the mercury switch. I wish the foam hadn't degraded, but at least we have an idea of what it should do. Thanks again to @RS485 for plotting out the design and keeping me from writing this off. A video will be up soon.

 

[RS485]

Your image flipping and part-level cut-and-pasting to illustrate both sides of the PCB were truly inspirational! Thanks for putting so much effort into capturing what was needed to understand the circuit design.

 

[JohnMarcson]

Thanks again to the community. This was another fun project that played out online.