ADP7142 for use with ADAU1701

Hello chmed,

I am not comfortable with this but not knowing anything else about your design I could be too cautious.

The rough max power for the part is going to be around 600mW. I calculated around 584mW but the IOVDD is not really included in this so if you are talking to other parts or driving an LED this would go up. So with 600mW that would be 182ma off of a 3.3V power supply. That is kind of close to the max and that is not counting any other parts in the system. I will agree that the max numbers are over all temperatures and silicon skews running a maximum program but still... not much margin.

You can ask this question to the linear guys but you should provide them a more complete accounting of the power requirements for your design and see what they say. I was only looking at the ADAU1701.

Dave T

Hmm, I'm curious where your 600mw figure came from. The 1701 datasheet talks about 85ma - analog and 60ma - digital for max currents. That doesn't include the I/O as you say but I'm driving high impedance inputs from the analog outputs and only planning to use one gpio connected to a switch. The other current drains are a bank of 5 eeproms (3ma each according to datasheet and that's only for writing) and a reset manager like on the dev board. Does the 1.8v regulator draw current that is not accounted for?

Hello Chmed,

85ma * 3.3V = 280mW

60ma * 1.8V = 108mW

Oops, now I see my mistake in trying to get an answer too quickly. I counted the reset currents...

So that is 388mW without IOVDD.

388mW / 3.3V = 118 ma without IOVDD.

I am also not counting the waste current from the DVDD voltage regulator. That is one thing our datasheets are not clear on is if the current going into the pass transistor is included in the tables. I am fairly certain it is not. So that would be 1.5V differential for the regulator times the DVDD current would give you 90mW dissipated by the pass transistor. So that is an additional 90/3.3 = 27ma of current from the main regulator.

So the IOVDD may be a little higher than you expect but you are not taxing it too much. it might be 10 or 20 ma. Then you say to add 15ma for EEPROMs. The reset generator is very little. So you end up less than the 200ma budget but it is not a lot of margin. Unfortunately, we did not make it easy to measure the current draw of the part on the Mini eval board. The design goal was small and cheap so features had to go.

Anyhow, that is my 2-cents on your question. Perhaps others may chime in on this. We are so trained here to look at the maximums and design for the worst case over all temperatures that it is difficult to look at a more realistic use case that is not under such environmental extremes and where one failure out of 100,000 units is not acceptable! So keep that in mind.

Dave T

Thanks Dave. This feedback helps. Unfortunately, after closer examination it looks like my hunch was right about using these tiny regulators to drop down this far. Although I think the total current is acceptable, anything more than a few volts of differential will yield an unacceptable junction temp. I think I'll just go with a 7805 drop-in style dc-dc convertor to get the voltage down to an acceptable range and then use either a 3336 or 7142 (cheaper) to finish it off. My next concern is if the switching noise of the dc-dc will be a problem. I think I can position it at least 1.5 inches away from the 1701 chip physically but....

Hello chmed,

Check the switching frequency of the converter. If it is lower than something like 300Khz then I would not use it. Up in the MHz would be good. This way the noise is easily filtered out and is not likely to beat with the audio or other clocks to end up in the audio band.

Then make sure you have a good ground plane and power plane. Place plenty of bypass caps between the planes and around the edges of the plane. Then pay attention to where the currents will flow for the input and output of the regulator to make sure you do not cross into sensitive analog signal traces. All this will give you a chance at a quiet PCB implementation. No guarantees!

Dave T