difference LM338K-MIL and LM338K-STEEL P+

[PrograMedia]

I have made a small power supply for filament heating from a KT44 audio tube.

I use a LM338K with attached schematic.
R1=110 0,1%
R2=240 1% with bypass 18uF
C1=2200uF 10V Rubycon
C2=47uF 16V Jafflicon

Vin=7V from a regulated PSU
Output is exact 4V unloaded
Cooling is a 1mm copperplate 20x30cm

The load is 2Amp with a peak current when the filament is cold (cold resistance about 0,6) I measured 3Amps for a second...

I build the circuit twice.
With LM338K-MIL it works fine, with STEEL P+ voltage drops to 2V and I hear the filament making noise, perhaps the regulator is oscillating

Both LM338K are from National Semiconductors. I replaced STEEL P+ already but it still doesn't work.

Also tested Vin up to 10V but still STEEL P+ does not work.


Any ideas???

[singa]

Hi, The LM338k/MIL are the same except that the MIL part has gone through hours of burn in to survive and pass tests according to miltary specs.Your circuit seem to have missing 2 protection diodes if you are bypassing R2.So for the moment ,remove bypass cap 18uF.Add or bypass C1=1000uF with 0.1uF.

With R1=110, R2=240 should give you 4V output.Also note that LM338K require 3V headroom to function.I.E. 3V +4V =7V minimum input.By all calculation the circuit will work unless your regulated 7V cannot supply enough current.The filament noise you hear is probably filament expanding/cooling off.

edit- C1 at 10V use 16V better safety margin.

[PrograMedia]

Hello, thanks for your information.

I replaced the steel p+ version with ST Microelectronics version and it works also! No problems...

What is wrong with the steel p+ version??
It looks like the steel p+ is a fake component. It limits current at about 1 amps at any input voltage causing a voltage drop. I could easily wipe of the label, very unusual for to-3 components I think

Anyone who has worked with the LM 338k Steel p+ ??(See image) What are the results?