LM317 Output Drops to Zero

[DudeGibbs]

Hello,
Trying to light the 6.3v filament to a 6SL7 vacuum tube. Using the attached LM317 circuit to drop down 36v dc to 6.3v. Filament requires 300mA current. When I attach the output of the lm317 to the filament the output drops to zero. I read zero volts on the filament. Removing the lm317  from the filament I get 6.3v on the output again. Why does it drop to zero when I apply to the filament but goes back to 6.3 when I remove it?
Thank you for reading.

[David_AVD]

Using a linear regulator to drop 36V to 6.3V @ 0.3A will result in nearly 9 Watts of dissipation in the regulator !

I'm guessing that the regulator outputs the voltage for a brief time, heats up then shuts down ?

[DudeGibbs]

I'm guessing that the regulator outputs the voltage for a brief time, heats up then shuts down ?

Dropping to zero seems instant. There is  no heat from the regulator. I'm going to try a lower input voltage.

[David_AVD]

A quick photo of the circuit may help.   What package type is the regulator ?

[Brumby]

From a TI datasheet:
"The LM117 and LM317-N offer full overload protection such as current limit, thermal overload (LM117) protection and safe area protection."


Does the chip know it's going to fry if it tries to do what is asked of it?

[Audioguru]

I betcha your LM317 is missing a heatsink capable of taking almost 10W of heat away from the metal tab. Then its tiny chip inside heats so quickly to its maximum allowed temperature then it shuts down before its outside feels warm.

[DudeGibbs]

Thank you for responding

I think I'm exceeding the max current of the lm317. the filament is 3.6ohms and if I apply 6.3 volts I'll draw atleast 1.75 amps right? The regulator may be shutting itself off when I apply the voltage.

[MK14]

Hello,
Trying to light the 6.3v filament to a 6SL7 vacuum tube. Using the attached LM317 circuit to drop down 36v dc to 6.3v. Filament requires 300mA current. When I attach the output of the lm317 to the filament the output drops to zero. I read zero volts on the filament. Removing the lm317  from the filament I get 6.3v on the output again. Why does it drop to zero when I apply to the filament but goes back to 6.3 when I remove it?
Thank you for reading.

There seem to be a number of possible issues here.

Firstly the valve/tube heater (usually tungsten), will normally have a considerably lower resistance (just like a filament light bulb, which is also usually tungsten), when it is cold.
Maybe as much as a factor of 10x, or at least some big number. (Disclaimer: Some valve/tubes can have complications, which could change what I am saying here. But I'm ignoring them for now.).

Also unlike a light bulb which "instantly" turns on. Valves/tubes take considerably longer, like up to a minute, to heat up (because it is heating up a lot more, than just the filament...).

So let's say it was 1 Amp to 3 Amps (depending on the exact multiplication factor. Further confused because valves/tubes heaters run at a lower temperature, than bulbs).

So this huge initial current, combined with the huge voltage across the regulator (36V is relatively HUGE, a 30V+ differential), also I'm NOT clear about any heatsink arrangements. Sorry if I missed the details.

So I'd actually be much more surprised if it actually worked, SORRY!
In short, it has a wide range of protection features, built in. You are probably activating them ALL at the same time ? (partial joke, SORRY).

Any chance you could give it a more reasonable incoming supply voltage ? (If not there are other solutions).

Also there are higher current rated versions of the 317, which may be a better choice here, as well. E.g. LM317K.

Thank you for responding

I think I'm exceeding the max current of the lm317. the filament is 3.6ohms and if I apply 6.3 volts I'll draw atleast 1.75 amps right? The regulator may be shutting itself off when I apply the voltage.

Is that the COLD resistance reading with a meter, and did you subtract the meters probes short circuit resistance from the reading ?

[DudeGibbs]

I successfully lit the filament using the same lm317. This time I placed a 2n3055 transistor in parallel with the lm317(base connected to lm317 output, collector connected to 36v, emitter as new output reading 6.3v). both the 317 and 3055 have considerable heat sinks attached.

The voltage drops to around 4volts when applied to the filament. How can I manage this? Should I just raise the output voltage to 6.3v with the filament load?

Thank you all for your responses. I was getting annoyed at this lm317

[MK14]

I successfully lit the filament using the same lm317. This time I placed a 2n3055 transistor in parallel with the lm317(base connected to lm317 output, collector connected to 36v, emitter as new output reading 6.3v). both the 317 and 3055 have considerable heat sinks attached.

The voltage drops to around 4volts when applied to the filament. How can I manage this? Should I just raise the output voltage to 6.3v with the filament load?

Thank you all for your responses. I was getting annoyed at this lm317

Congratulations. It sounds like you got the 2N3055, to take much of the current, saving the LM317 from shutting down (if that is what was happening before).

I don't understand the bolded text, above. Is it possible you could re-explain it.
What drops to 4 volts ?
Manage what ?
How did you intend to raise it to 6.3V ?

[Tac Eht Xilef]

I think I'm exceeding the max current of the lm317. the filament is 3.6ohms and if I apply 6.3 volts I'll draw atleast 1.75 amps right?

No, that's the power the load is drawing (at cold - the filament resistance will increase as it heats up).

Right now, you should be thinking "so where's the other ~30v @ 300mA going?"...

[DudeGibbs]

The voltage drops to around 4volts when applied to the filament. How can I manage this? Should I just raise the output voltage to 6.3v with the filament load?

I don't understand the bolded text, above. Is it possible you could re-explain it.
What drops to 4 volts ?
Manage what ?
How did you intend to raise it to 6.3V ?

I adjust the lm317 to give 6.3v output. When I apply that voltage to the filament it drops to around 4volts. This is expected i guess because the filament is drawing 300mA. Im just wondering about any special way to remedy this. For now I will raise the lm317 voltage high enough so when the output voltage is applied to the filament it will drop to the necessary 6.3v.

i reckon the lm317 kept shutting itself down because the filament is basically a short circuit? So the 3055 acts as a buffer between the filament and lm317.

[MK14]

The voltage drops to around 4volts when applied to the filament. How can I manage this? Should I just raise the output voltage to 6.3v with the filament load?

I don't understand the bolded text, above. Is it possible you could re-explain it.
What drops to 4 volts ?
Manage what ?
How did you intend to raise it to 6.3V ?

I adjust the lm317 to give 6.3v output. When I apply that voltage to the filament it drops to around 4volts. This is expected i guess because the filament is drawing 300mA. Im just wondering about any special way to remedy this. For now I will raise the lm317 voltage high enough so when the output voltage is applied to the filament it will drop to the necessary 6.3v.

i reckon the lm317 kept shutting itself down because the filament is basically a short circuit? So the 3055 acts as a buffer between the filament and lm317.

The voltage should be very precisely controlled, within tight margins, by the LM317, automatically.

Since this does not appear to be happening (4V vs 6.3V), one or more things are likely to be wrong with your setup.

When you connected the 2N3055, did you get the schematic (for how to do this properly), from some kind of source ?
Or could it need improving ?

If it is correctly connected up (I'm NOT sure, without seeing a new schematic), then something is wrong, such as some type of power/current/voltage rating, forcing the LM317 into some type of partial shut down, or something.

But if it is simply not connected up the best way, it may just need to be connected, better.

Or it could be something else.

Best to find out what is really wrong, rather than trying to bodge it by forcing the LM317 to produce much higher voltages. Which could lead to unreliability in the near future, if something is wrong.

Also the 36V sounds like rather a lot of voltage to be giving the 2N3055 and LM317, as they will have to waste a lot of heat, which could also lead to reduced life expectancy of the parts.
It can work like that, but it is not a good idea. 36V may be too near the upper voltage limit of the LM317, as well. E.g. Mains supply voltage changes could lead to it breaking.

[Agent24]

Can we get any photos of your setup?
Can you show the new schematic with added 2N3055?

[DudeGibbs]

Attached is the lm317 circuit I've built. My circuit only uses one 3055 instead of the three shown. It has a small heatsink on it now but it still gets very hot. I'm only drawing 300mA Why all the heat?

 My circuit also lacks the 6a4 diode. Could this be the issue?

[David_AVD]

Really?   Did you even read any of the replies?  I told you about the 9 Watts of wasted energy in the first reply!

[Skimask]

Ok, you're not getting it.

In this particular situation, to regulate voltage, you have to dump it off somewhere...HEAT...which equates to WATTs.
Again, in this particular situation, it's a bit like regulating water pressure.  If you've got 100 PSI coming in and you only need 5 PSI, this circuit might dump the extra 95 PSI on the ground and leave you with 5 PSI.  (no, water pressure regulators don't work like that, it's just an example)

Since WATTs = Volts x Amps, if you've got 36 volts coming in, and you want 6 volts coming out, you have to dump those extra 30 volts somewhere...again, HEAT.  So, 30 volts at 300mA = (30 x .3) = 9 WATTs.
If you started off with only 9 volts and regulated that down to 6 volts, you'd only have to dump off the extra 3 volts.  And 3 volts at 300mA = (3 x .3) = .9 WATTs.

So, the key is to NOT dump off so much extra voltage, or get rid of the heat faster.  In this case, if you are still trying to put 36 volts in, you aren't going to dump 9 watts thru a heatsink.  I don't care how big it is.  The heat just won't move fast enough from the regulator to the heatsink.  That's all there is to it.

Even if you rebuilt that circuit using those '3055's, even if those '3055's are in a TO-3 package, good luck dropping 9 watts across ONE of them.  You might get away with dropping it across 3 of them...but why?

In short, you aren't listening to the advise given, you aren't paying attention, you don't know that you don't know what you don't know, and/or don't have a grasp on basic electronics yet.
You might get away with using three of those circuit back-to-back, with each one dropping the input voltage a bit, say the first one drops from 36 down to 24, the second one from 24 down to 12, and the last one from 12 to 6.3.  maybe...

[Agent24]

I have not seen that kind of design before, but I assume the diode is necessary. Yes, omitting it may be the cause of your voltage error.

You may only be drawing 300mA, but you are dropping 30 volts, so it's still going to get hot. Maybe your heatsink is not big enough and its still running into thermal limiting or something.

Photos really help so we can see what you are actually doing and if you have made any silly mistakes that may be totally obvious to us but not to you. It's much harder to troubleshoot something when we can't see exactly what it looks like.

[ludzinc]

Right there, first pic in the post, Vin max = 28V.

You're trying 36V.

Out if spec, you need to do something else.

[MK14]

Attached is the lm317 circuit I've built. My circuit only uses one 3055 instead of the three shown. It has a small heatsink on it now but it still gets very hot. I'm only drawing 300mA Why all the heat?

 My circuit also lacks the 6a4 diode. Could this be the issue?

Is the 220R resistor in your circuit in the right place ?
Because in the diagram (R1 in diagram below), it is connected differently.

[DudeGibbs]

Thanks to everyone for their input. I'm still learning the limits of these ic regulators. I didn't think 9watts would make that much heat but the lesson is learned. I'm going to use a different power supply that can give me 12vdc for the lm317's input.

Thanks again.

[MK14]

Thanks to everyone for their input. I'm still learning the limits of these ic regulators. I didn't think 9watts would make that much heat but the lesson is learned. I'm going to use a different power supply that can give me 12vdc for the lm317's input.

Thanks again.

Cool (excuse the pun).

That will be much better. The LM317 was also probably getting too much voltage (or was too near the limit), which can be problematic, as well.