ADAU1466 operating current and clock speed variation

Hello archerator,

The DVDD current draw has two basic variables, well three if you could silicon skews.

The two major factors are length of the program and temperature.

The length of the program and the sample rate will change the DVDD current draw a lot. When a sample period starts the core starts processing and draws a lot of current. Then when the core is done processing it will idle for the rest of the sample period. So the longer the program the longer the core is drawing a lot of current.

Then the temperature is a big factor. The higher the temperature the more leakage current in the die. So that is also a major factor. This is why the thermal pad specifications are so important with the array of vias.

Then if the silicon process for that lot happens to be a little faster then it will draw a little more current. This is not as significant since the process variables are typically held to close tolerances but we need to plan for the worst case.

Then this part is automotive qualified which means we have to add some extra guard band on the specifications.

Regarding the core clock speed. Yes, you can run the clock way down but not quite half speed. So this will throw off all the sample rates stated in the register settings but I think you will know this and adapt. For modules in SigmaStudio that require to know the sample rate then you have to set it to as close as you can to the actual sample rate.

So for the most part we don't recommend you move the core rate too far from nominal since the PLL has to lock to the clock input and if it goes too far it will not lock.

You can use the 150MHz versions of the Sigma300, that would be the ADAU1450. If you need more memory then stay tuned and you will see some solutions coming soon.

This part does draw a lot of power so it will be tough for a battery operated device unless it is large enough to have a large battery.

Can you tell me more details about your application?

Dave T

Hi Dave,

Thanks for the quick response, I appreciate it!

The part that is most confusing to me about the current spec is that the 233-495mA spec is already listed as the "Maximum Program" spec, which I interpret to mean that core processing is taking up ALL of the time allotment, or at least as close as possible to all of it. With that limitation I wouldn't expect the current range to be so wide except from temp variation, which is why I was wondering if there was a current draw vs temp graph available for max program.

The application is a belt clip guitar effects module / headphone amp that will be configurable from a smartphone via bluetooth. We actually already used the ADAU1452 in a similar AC adapter powered floor pedal version of the same thing.

We would like to keep the DSP power to 0.5W if possible, and if spec stayed closer to the 233mA end of the range, we would be able to meet that if we ran the 1.2V from a switching supply, which we are planning on doing. I have considerable experience with designing good switching supplies so efficiency and switching noise will not be an issue.

When we first opened this project we were looking at much lower power AD devices such as the ADAU1761, which almost won the ticket until we decided we really needed more "oompfh" to meet our original configurability goal. The ADAU1450 was considered next but that wasn't really much better for us since it has so little RAM. I am MOST curious about your suggestion to stay tuned for more solutions coming soon, so I was hoping you could elaborate on that if possible!

This is a bit off-topic, so I apologize in advance... We would actually prefer to write our own algorithms but we have found it difficult to find DSPs that meet our specs that we can also acquire the SDK for. For example, for the floor pedal version we would have much preferred to use something like the ADSP-SC571, but we simply don't have the budget at this time for licenses for CrossCore Embedded Studio. We have also looked at competitor parts of course but there's always one problem or another that turns us off to them.

Hello archerator,

I cannot comment on a public forum about parts that are not released yet.

What have you wanted to do that the current library did not support?

Where in the USA are you located? What city?

Dave T

Dave,

I do like how Sigma Studio operates, but there always seems to be one or more algorithms that are available in one part but not another, and vice versa for the 'other' part. That's a bit abstract, so I'll give an example. The larger devices such as the 145x and 146x series have a Reverb effect, but don't have a Chorus effect. The Chorus effect is, however, available on smaller parts.

Other algorithms that smaller parts have that I wish the larger ones had is the Sub-Harmonic Synth, the Graphic EQ (although we can accomplish basically the same with other filters), and Mono-to-Stereo synth.

Also, even if the effect is available, some of them aren't as configurable as I would hope. Reverb (as you can probably tell now this is an important one for us) is missing quite a lot of options, particularly reflection and diffusion adjustments. Also, it would be nice if the 3 adjustments that are available (Reverb Time, HF Damping, and Bass Reverb Gain) had adjustable frequency bands.

The Chorus effect has similar problems for us. In general, like I said above, we do like the interface and the ease of use, which definitely makes up for a lot of this, but when it comes down to it, for us it's all about configurability.

I should clarify that I'm not surprised that the Reverb in particular isn't available on the small parts since that is a very resource intensive algorithm generally, but I am rather surprised about the absence of many algorithms from the larger parts.

I'm located in Rochester NY.