LM27761: Can you share a PSRR verses frequency plot?

Hi John,

We don't have PSRR data for this IC (and it isn't an ultra-low noise LDO), but the LM27762 is a very similar device which does. It doesn't meet your PSRR spec. But do they have a 100 Hz signal on the input that they need to reject? If not, they should not be concerned with the PSRR at this operating point.

What Vin and Iout do they need? The TPS63710 is a very unique device--giving LDO-like performance with just a switcher. It gets very close to the aggressive 1 uV/root (Hz) at 10 Hz spec and this can be improved with a larger cap on the CAP pin.

Hey Chris,

I appreciate the super quick response.  I don't know the frequency of the signal they are monitoring but I do know they are using the TI ADS1299 at 500 SPS, so we can assume the frequency they are measuring is in the 250Hz or less range.  I believe that is what is driving this large PSRR requirement.

I have also asked them for more details on what is driving this PSRR requirement since the ADS1299 is showing a PSRR performance of over 100dB at 100Hz for a gain of 24 that they are using. So between the PSRR of the TPS63710 and the ADS1299 they may be okay.

They are using a +5V supply and need to generate ±2.5V rails for the ADS1299 (both positive and negative). These are low current rails needing 10-20mA max. It looks like the TPS63710 device is a 1A device which will be a bit of overkill for this application.

Last thing, one thing I was concerned about with both the LM27761 and LM27762 is that they both have a 2MHz switching freqeuncy. This is right at the modulation frequency of the ADS1299 and the decimation filter will have a gain right around 1 (will not reject this switching noise). See below.

Let me know if you have any questions but my main concerns were to find a inverting regulator that doesn"t switch near the sample frequency and doesn't switch near the modulation frequency.

Let me know what you think and thanks again,

John

Thanks for the details John. Yes, TPS63710 is overkill for this socket, but please still show it to them as it may be useful on other projects.

Both LM27761 and LM27762 have a PFM mode, so the noise will not be at 2 MHz but rather something lower. However, the TPS6040x device is a completely different inverting charge pump device with a much lower frequency.

Hi Chris,

Thanks again. It doesn't look like you can force either the LM27761 or LM27762 into PFM mode, can you? I also do not see a frequency verses load current plots in the datasheet so I cannot estimate if the device will transition to constant switching frequency mode at 10-20mA. 

I like both of these devices but I do not think that my customer will be willing to consider these unless we can show some data that says the device will not be in constant switching frequency mode with switching frequencies near 2MHz.

The TPS6040x devices might be the only option unless there is other data you can share with the LM2776x devices.

Thanks,

John

Will the TPS6040x work for them? Based on what we discussed, I don't see why it wouldn't. The non-TPS60400 part numbers do not operate with a power save mode, which is generally preferred in this type of application in order to give ripple at a defined frequency.