Panel lights fuse

[FlyBoyM20J]

The fuse that is inline with the Panel Lights rheostat keeps blowing, but only while the engine is running. 

Sitting in the hangar with engine off, a new fuse, there are no problems.  The dimmer works as expected and the post lights and instruments get bright or dim, etc.  But if I start up and start taxiing, the fuse will blow within minutes.

I've studied the schematic for my 1982 M20J and I see that the primary load on that fuse is the Radio Lite bus.  All the post lights draw power from that, as do the interior lights in the HSI, AP, annunciator, and many other panel items.

This is a rather long list of locations that might have an intermittent short.  Everything seems to work, when it's working, too, so I'm a little bit stuck.

Has anyone else ever seen this type of failure?

[N201MKTurbo]

I think you will find that vibration is causing a short somewhere.

[ArtVandelay]

You’ll need to start gently shaking wires till fuse blows, then you’ll have it narrowed down, careful, sometimes shaking 1 bundle of wires, causes another to shake.


Tom

[FlyBoyM20J]

Thanks folks...I realize this is a tricky one since the loads are all essentially temperature-dependent shorts themselves (DC lightbulbs).  Shaking is the way...I have a bunch of fuses so I could do this a few more times but I think I have an easier way.

I'm going to bypass the fuse and power this circuit at that point with a bench power supply set at current limit at, say, 1A.  Then I'll carefully shake this and jiggle that (in all innocence, gents), and listen/watch for the PS to hit CC limit.  Easier/less bother than blowing fuses, I think.

[RangerM20]

Where in the circuit is this fuse located?  I also had a similar one, but the glareshield light wires touched, and shorted out the lights.

[FlyBoyM20J]

Fuses are located on the bottom of the enclosure that houses the glare shield and panel light rheostats.  In my 1982 M20J, the fuses are both 5A and are easy to access from their holders there at the bottom.  The sch snippet I just pasted shows the location of the fuse on the output of the transistor (which, incidentally, appears to be drawn in backwards).

[Yetti]

Could also be a screw dropped on the transistor board that is rolling around

[jaylw314]

I can't explain it, but I've had 3-4 fuses blown on the panel lights in the first year I had the plane when I turned them on and turned up he rheostat quickly.  Since then, I've turned them on, and then slowly turned up the rheostat and I haven't blown a fuse in a couple years now.  I have no idea why that makes a difference, though--there shouldn't be a high capacitance circuit anywhere in a bunch of lights?

[FlyBoyM20J]

5 minutes ago, jaylw314 said:

I can't explain it, but I've had 3-4 fuses blown on the panel lights in the first year I had the plane when I turned them on and turned up he rheostat quickly.  Since then, I've turned them on, and then slowly turned up the rheostat and I haven't blown a fuse in a couple years now.  I have no idea why that makes a difference, though--there shouldn't be a high capacitance circuit anywhere in a bunch of lights?

I suspect that the reason a slow-blow fuse was spec'd for this circuit is that the initial resistance of DC light bulbs is low...like 2 ohms or something, if I recall.  It will more than double as they heat up but the initial current draw is high and might blow a fast-blow fuse.

In my case, I'm using fast-blow fuses and I do turn it on slowly, so that's not what blows it.  Something is loose somewhere, causing a short.  I suspect that I could run things OK if I don't turn the rheostat all the way up, thus allowing the transistor to limit the current through the fuse, so all that would happen when the short occurs is that the lights would dim out.  But I really want to fix this, so I'm going to keep at it until I find something I can move around that causes the short.

 

 

[FlyBoyM20J]

11 minutes ago, Yetti said:

Could also be a screw dropped on the transistor board that is rolling around

Good idea...I'll pull the rheostat box down from the panel and take a look.  This is exactly the sort of thing that would do it!

[Yetti]

this is a good thread.   In the end you probably want to go PWM.    The PWM from Amazon and Spruce are the same except for the price.  

 

[FlyBoyM20J]

26 minutes ago, Yetti said:

this is a good thread.   In the end you probably want to go PWM.    The PWM from Amazon and Spruce are the same except for the price.  

 

Excellent!  I was just out on a walk considering how dead-end incandescent panel lighting is.  This would be *so simple* to diagnose if the loads weren't essentially inductors!

But that would require a PWM controller for a dimmer, even if one could swap out the post lights with LEDs in some way. 

So, I'm back from my walk and you've already answered that question!

But what about the dimmable lights inside radios and instruments?  We don't want to send them PWM, do we?

[Yetti]

1 hour ago, FlyBoyM20J said:

Excellent!  I was just out on a walk considering how dead-end incandescent panel lighting is.  This would be *so simple* to diagnose if the loads weren't essentially inductors!

But that would require a PWM controller for a dimmer, even if one could swap out the post lights with LEDs in some way. 

So, I'm back from my walk and you've already answered that question!

But what about the dimmable lights inside radios and instruments?  We don't want to send them PWM, do we?

PWM is used for running motors slower...   Pretty sure the filament in a incandescent bulb will do just fine.  

[FlyBoyM20J]

12 minutes ago, Yetti said:

PWM is used for running motors slower...   Pretty sure the filament in a incandescent bulb will do just fine.  

Yeah, motor coils are rather large inductors with a series resistance that's presumably higher than a few ohms, with a decently hefty tau.  Incandescent filaments are going to have a much smaller time constant.  I'd worry that this would amount to cycling the filaments full-scale at a high rate, though it all depends on the details (PWM frequency being one of them). 

But I'll defer to your experience.

[PT20J]

So, did I understand that you are using a fast blow fuse in a slow blow application? If so, does a slow blow fuse also blow? Have you tried seeing if it blows if you run the engine with the alternator off?

[FlyBoyM20J]

17 minutes ago, PT20J said:

So, did I understand that you are using a fast blow fuse in a slow blow application? If so, does a slow blow fuse also blow? Have you tried seeing if it blows if you run the engine with the alternator off?

Yes, what I have are Littlefuse 5A fast-blow parts.  But the fuse that blew the first time was the original slow-blow fuse.  I was flying the plane (actually, in rollout from a landing at night) and as I was exiting the runway, the panel lights flickered for a second or two and I heard static over the headset and then out they all went.

The landing wasn't rough at all but it must have shaken something loose since I'd been flying at night with these panel lights for upwards of a year with no issues and I was much more likely to have a bouncy landing last year when I first got the plane.

I don't believe the slow-blow/fast-blow nature of the fuse is at the crux of my problem.  Slow-blow might be good for initial power-on of the panel lights, especially if one rapidly turns up the rheostat.  But once we have them going, anything that makes the transistor pass > 5A is going to be a reason to break the circuit, regardless of how transitory it is, I would think.

All that said, the fast-blow fuses do not blow no matter how fast I turn up the rheostat.  They blow due to something else that tends to happen after the first few minutes of having the panel lights turned on while the engine is running.

Running the engine with alternator off would be an interesting test.  System voltage would be much lower.

 

[Yetti]

4 hours ago, FlyBoyM20J said:

Yeah, motor coils are rather large inductors with a series resistance that's presumably higher than a few ohms, with a decently hefty tau.  Incandescent filaments are going to have a much smaller time constant.  I'd worry that this would amount to cycling the filaments full-scale at a high rate, though it all depends on the details (PWM frequency being one of them). 

But I'll defer to your experience.

I am sure that an lectrical gineer will be along.  My understanding is the filament does not really cool down so it is not "cycling". 

[carusoam]

Are you equipped With a volt meter and resistance meter..?

Looking for a short using these devices has got to beat blowing fuses with the chance of random arcs melting something important...

All the lights are powered to one location near the pilots feet... the rheostat delivers power to a post... each light  power from the post...

One of them has no resistance under some shaking circumstance...

This is going to be tough if it won’t repeat itself easily...

Make a list of each lightbulb... and confirm the health of its wiring... one of them is going to be visually ugly...

PP thoughts only, from my M20C lighting experience...

Best regards,

-a-

[FlyBoyM20J]

7 hours ago, carusoam said:

Are you equipped With a volt meter and resistance meter..?

Looking for a short using these devices has got to beat blowing fuses with the chance of random arcs melting something important...

 

Yes, I am!  Unfortunately, the update time of my meter (and most that I've used) isn't good for finding transitory changes in impedance.

My plan, instead, is to remove the fuse, pull the panel lights CB for isolation, and apply +12V at that bus (or rather, at the fuse output point) using a bench power supply from work.  I'll set the bench supply to limit output current to about 1A.  And then start shaking things.  The CC trip on the bench supply is fairly quick and will amount to blowing a 1A fuse over and over again.  I think this will help me find a transitory short most efficiently.

[jcovington]

1 hour ago, FlyBoyM20J said:

My plan, instead, is to remove the fuse, pull the panel lights CB for isolation, and apply +12V at that bus (or rather, at the fuse output point) using a bench power supply from work.  I'll set the bench supply to limit output current to about 1A.  And then start shaking things.  The CC trip on the bench supply is fairly quick and will amount to blowing a 1A fuse over and over again.  I think this will help me find a transitory short most efficiently.

Something else to think about although it doesn't sound like exactly the same problem. A few years ago I helped a C172 owner diagnose some odd problems with his panel lighting. I don't recall the fuse blowing but the rheostat became very warm and the lights dimmed in random fashions. Turned out that someone performing a panel upgrade years earlier had soldered the new lights onto the rheostat rather than the light controller. Instead of having the rheostat supply a small current to control the transistors within the light controller it was attempting to dim the panel lights directly. Didn't work very well. The solution was to move the wire to the controller side and everything worked well afterwards.

Jim

[FlyBoyM20J]

I had a chance to pull out the dimmer box, open it up, verify the wiring, and that there wasn't something metal loose in there.  I also used a current-limiting bench supply to power the circuit loads and determine the operating power level: about 2.5A at 13V.  That is, I set the bench supply voltage to 13V and initially limited the current to 1A.  I slowly increased the current limit until the supply stopped being current-limited (thus, the load was running at full power, limited by voltage at 13V), and this current level was right at 2.5A.  The panel lights were nice and bright.

So, the 5A fuse gives plenty of headroom.

I tried to locate the radio lite bus strip but still can't find it!  Anyone how where it is exactly on an M20J?

Even though I couldn't find the bus, so couldn't follow individual wires to post lights (this would be tough anyhow given the density of wiring behind the panel), I did shake many a wire attempting to generate the transitory short condition while powering the panel light loads from this bench supply.  No luck, though.  I'm going to have to revisit this test next time I can take off the glare shield and do a more thorough shakedown.

 

[larrynimmo]

Divide and Conquer is my advice....disconnect areas of lighting put a fuse in and see if it blows...do so until they stop blowing.

[Bryan]

So you have a picture of the fuse location?

i am having similar problems with my 28v M20K panel lights. My plane is now in avionics for glass upgrade so I am hoping removing most of the instruments will solve my problem. I have not found a fuse but had my POT replaced and the controller in the avionics bay inspected, cleaned, and rebuilt.

[FlyBoyM20J]

2 hours ago, Bryan said:

So you have a picture of the fuse location?

i am having similar problems with my 28v M20K panel lights. My plane is now in avionics for glass upgrade so I am hoping removing most of the instruments will solve my problem. I have not found a fuse but had my POT replaced and the controller in the avionics bay inspected, cleaned, and rebuilt.

I don't have a picture handy (although there is one in the Service Manual Vol 1 for my M20J), but the fuses are screwed into little gray fuse holders on the bottom panel of the rheostat box.  I mean, if you can turn the right-hand pot that controls the transistor that limits current to the panel lights, then reach underneath the panel right there under that pot and you'll be touching the fuse holder for that circuit.

In my picture above, I've removed the pot box from the panel and it's hanging sideways, facing down with its cover removed.  You can see the pot shafts (sans knobs) there on the bottom, more or less, and there on the right-hand side of it is one of the gray screw-cap fuse holders still in place.  That's the one for the glare shield circuit.  I've removed the cover and fuse for the panel lights circuit.

Can only speak for my 1982 J, though.

I should note (as an edit here) that my bench supply test above proves to me that my rheostat controller circuit cannot be at fault.  I'm powering all the panel lights directly from a bench supply limited to 13V and they pull only 2.5A at full power.  The controller circuit is protected by a 5A fuse, but the only way that fuse can blow is if the controller circuit is turned up sufficiently to allow > 5A of current *and* that the loads pull > 5A of current.  So, in my case, the circuit is working just fine but something downstream has an occasional short to ground causing the fuse to blow, as it should.  Finding that short is going to take more work since it seems to happen only while the engine is running. 

My next test will probably be to do this same bench supply arrangement but while the engine is idling, and also in a position to pull individual wires off the panel lite bus to finally see which line has the short on it.  I'll wait for warmer weather before I do that.

 

[Yetti]

On 4/3/2019 at 7:08 PM, FlyBoyM20J said:

I tried to locate the radio lite bus strip but still can't find it!  Anyone how where it is exactly on an M20J?

 

 

 

I don't think there is a bus for the radio lights.   on the 75F.  The 75F is a testbed for the J.    there is a wire coming off the transistor board that is run to all the radios that need panel light input.   Seemed hokey.