LM317: LM317 datasheet following pcn : legacy vs new chip

Hello Sebastien,

Im not sure what you mean by your first question. Could you elaborate or provide screenshots of what you are referring to? What do mean that Rthja seems to be inverted?

For the transient response concern, im not sure why we measured the line and load transients at different parameter and if you want I can ask about this. In terms of performance, looking at the 5.5 Electrical Characteristics (Both Legacy and New Chip) table in the datasheet, the performance is relatively the same for line transient and slightly different on the load transient:

In regards to purchasing old die, I can take a look into seeing if we are able to ship old die (legacy chip) but we are transitioning all old material to new chip material. If this is possible, what are the volumes are you thinking of for old material for LM317MQ or LM317, specifically what OPNs? 

Hope this helps

Jorge Del Rio

Product Marketing Engineer | Mid-Wide LDOs | Industrial Sector

Dear Jorge,

RTHJA:

For the LM317, thermal impedance is 67°C/W for legacy chip and 60°C/W for new chip (so less for new chip). While in the LM317MQ, the Rth for the legacy is 60°C/W and 78°C/W for the new chip (so higher for new chip). I don't understand why the evolution of Rth is not in the same way....

TRANSIENT RESPONSE:

OK for the line and load regulation, but it is not the transient case. Compare figure 7.3 and 7.4 of the LM317MQ (p17) > Vout = 10V for legacy device while Vout = 1.25V in new chip. Compare also figure 7.5 and 7.6, same case for Vout... If we keep the same Vout condition as before, is it the same results ? I need to justify that the new die has no impact on current design without testing > same condition in datasheet....

I repeat but it seems that the shrinking die in the new RFAB factory has an impact on die and performance... Is it possible to buy old factory die (SFAB I guess) ? What is the process please ?

Reagrds

Sébastien

Hello Sebastien, 

For LM317:

When surface-mount packages are used (SOT-223), the thermal resistance is reduced by increasing the PCB copper area that is thermally connected to the package. This is mentioned under the Thermal Information (New Chip) table. Look at the 8.5.1.1.1 Heat Sink Requirements in the datasheet for specifics. Typically, larger die size will cause better thermals and so will increasing the pass element size. Sometimes it’s because the thermal pad has been grown.

For Transient:

Transient response looks fairly similar but slightly better on the new device. In practice, this is dominated by the phase margin and therefore loosely corresponds to PSRR, which is also pretty similar in the >10kHz region (The PSRR spec change is for 120Hz). This means output capacitor selection will affect it significantly more than legacy/new.

Buying old material (legacy chip):

Unfortunately, when purchasing this device, you would be getting a mix of both new and old material until we completely flush out the old material. Once all old material is depleted, all devices would be using new chip material. 

Hello Jorge

LM317:
I know that the Rthja is influenced by copper area or die size but my question is not here. My question is why there is difference in the 2 datasheet (LM317 and LLM317MQ) as we know that the Rthja is always given with the minimum copper area and as we know that the LM317 is the same die than LM317MQ (I guess) ?
So let me ask you my question once again: For the LM317, thermal impedance is 67°C/W for legacy chip and 60°C/W for new chip (so less for new chip). While in the LM317MQ, the Rth for the legacy is 60°C/W and 78°C/W for the new chip (so higher for new chip). I don't understand why the evolution of Rth is not in the same way.... Which datasheet is telling the truth ?

Transient:
You can not say that the response is similar while it is not tsted in the same condition. So if it is not tested in the same condition, we can not say that the behavior is similar... In addition, I dont see the link between output voltage transient response regarding output current transient and PSRR. PSRR is regarding input voltage AC voltage, not fast output current transient. So have you transient simulation results or tests results in the same condition than before (legacy) > figure 7.4 & 7.6 (New chip) with Vout = 10V or 3.3V ?

PURCHASE:
Is it not possible to buy SFAB die in a specific purchase as I have done some weeks ago for an issue with SN74HC gates ???

Regards,
Sébastien