How to test the LM338K

[TK]

I think I have a bad LM338K on a vintage test equipment.  The +5V rail goes only up to 4.7V using the adjustment Pot.  I purchased a new LM338K from Jameco for $10 and it gives even lower output voltage.  I found somewhere that the LM338K, when ADJ pin is connected to GND, it should give 1.25V (The Reference Voltage) in the output pin.  When I do this test on the original LM338K that I suspect is bad, I get 1.75V.  On the new regulator purchased from Jameco, it gives only 0.9V!!!  Is this a valid test?  Input voltage is 12V DC.

[digsys]

Yes to what you say. As long as you have at least a 20mA load connected, and the I/O filter caps are per spec, then your numbers are correct.
1st note: There a LOT of FAKE linear regs out there. I've even found LM317s jammed in a LM338 case, so you may have a fake one.
LM338s were EOL ages ago, so fakes are rampant.

[David Hess]

LM338s were EOL ages ago, so fakes are rampant.

TI still makes LM338Ts and apparently LM338Ks but the later are really expensive.

[TK]

LM338s were EOL ages ago, so fakes are rampant.

TI still makes LM338Ts and apparently LM338Ks but the later are really expensive.

Yes, they are $60 a piece.

[rsjsouza]

A LM338T (on TO-220, not on TO-3) is only US$0.81 in volume, which is probably well worth trying to adapt to this equipment.

http://www.ti.com/product/LM338/samplebuy

The only LM338K I found at TI's website is the MIL version:

http://www.ti.com/product/LM338-MIL/samplebuy

[TK]

I will order the LM338T TO-220.  Thanks

[floobydust]

The LT338 is also very good, if you can find them.

[David Hess]

The LT338 is also very good, if you can find them.

I don't think Linear Technology made those for very long.  Even in my old LT databooks, it was already marked obsolete.

They do have the LT1085 (3A), LT1084 (5A) and LT1083 (7.5A) with the first two being improved replacements for the LM350T and LM338T.  These used to be available in TO-3 and TO-3P packages but now are only made in the TO-220 and surface mount TO-220 with the tap cut off.  These are low dropout regulators so a small 10uF solid tantalum or 47uF aluminum electrolytic needs to be added close to the output and perhaps a 10uF capacitor at the input; check the datasheet for details.

[BravoV]

I don't think Linear Technology made those for very long.  Even in my old LT databooks, it was already marked obsolete.

They do have the LT1085 (3A), LT1084 (5A) and LT1083 (7.5A) with the first two being improved replacements for the LM350T and LM338T.  These used to be available in TO-3 and TO-3P packages but now are only made in the TO-220 and surface mount TO-220 with the tap cut off.  These are low dropout regulators so a small 10uF solid tantalum or 47uF aluminum electrolytic needs to be added close to the output and perhaps a 10uF capacitor at the input; check the datasheet for details.

Even the LT108x series is NRND by Analog now.

[David Hess]

I don't think Linear Technology made those for very long.  Even in my old LT databooks, it was already marked obsolete.

They do have the LT1085 (3A), LT1084 (5A) and LT1083 (7.5A) with the first two being improved replacements for the LM350T and LM338T.  These used to be available in TO-3 and TO-3P packages but now are only made in the TO-220 and surface mount TO-220 with the tap cut off.  These are low dropout regulators so a small 10uF solid tantalum or 47uF aluminum electrolytic needs to be added close to the output and perhaps a 10uF capacitor at the input; check the datasheet for details.

Even the LT108x series is NRND by Analog now.

Hmm, their online database shows it still available but it would not surprise me if they basically discontinue most of Linear Technology's products.  Some things have already vanished.  It will be a shame to see some of their irreplaceable parts go like the low input bias current LT1008/LT1012/LT1097 series and the LT1043 switched capacitor building block.

[BravoV]

Hmm, their online database shows it still available but it would not surprise me if they basically discontinue most of Linear Technology's products.  Some things have already vanished.

Yeah, quite sad. 

I still have probably a dozen of LT1083s in TO-3P that I treat like treasure. 

[rsjsouza]

I don't think Linear Technology made those for very long.  Even in my old LT databooks, it was already marked obsolete.

They do have the LT1085 (3A), LT1084 (5A) and LT1083 (7.5A) with the first two being improved replacements for the LM350T and LM338T.  These used to be available in TO-3 and TO-3P packages but now are only made in the TO-220 and surface mount TO-220 with the tap cut off.  These are low dropout regulators so a small 10uF solid tantalum or 47uF aluminum electrolytic needs to be added close to the output and perhaps a 10uF capacitor at the input; check the datasheet for details.

Even the LT108x series is NRND by Analog now.

Hmm, their online database shows it still available but it would not surprise me if they basically discontinue most of Linear Technology's products.  Some things have already vanished.  It will be a shame to see some of their irreplaceable parts go like the low input bias current LT1008/LT1012/LT1097 series and the LT1043 switched capacitor building block.

Interesting; TI has the LM1085 (3A) and LM1084 (5A) still in active state - all TO-220. IIRC I have a few units of each purchased about a year ago. 

[David Hess]

Interesting; TI has the LM1085 (3A) and LM1084 (5A) still in active state - all TO-220. IIRC I have a few units of each purchased about a year ago.

Oh, I did not even check TI but they second source a number of LT products.  The price and availability is not bad either.

[rsjsouza]

Interesting; TI has the LM1085 (3A) and LM1084 (5A) still in active state - all TO-220. IIRC I have a few units of each purchased about a year ago.

Oh, I did not even check TI but they second source a number of LT products.  The price and availability is not bad either.

It was most probably NSC that second sourced these - the prefix LMxxxx is a dead giveaway.

[David Hess]

Oh, I did not even check TI but they second source a number of LT products.  The price and availability is not bad either.

It was most probably NSC that second sourced these - the prefix LMxxxx is a dead giveaway.

Oh, that explains why the TI second source for the LT1013/LT1014 is the LT1013/LT1014 and the second source originally produced by National of the LT1084 is the LM1084.

[Alex Eisenhut]

I think I have a bad LM338K on a vintage test equipment. 

Well if you're sure that's the trouble, I can send you a Linear Technology LM338K from my private reserve... Date code 9213.
For sure not a fake.
Just pay shipping.

[TK]

I think I have a bad LM338K on a vintage test equipment. 

Well if you're sure that's the trouble, I can send you a Linear Technology LM338K from my private reserve... Date code 9213.
For sure not a fake.
Just pay shipping.

Thank you for your kind offer!  I already ordered some LM338T.  I will also test the suspect LM338K in a test circuit, as well as the Jameco version that I purchased for $10 and post the results.  And will consider your offer as a last resort.

[TK]

I soldered the LM338T and it fixed the problem.  Thank you all for the suggestions and recommendations.