[f] AMB on Facebook

Thoughts on the need for galvanic isolation.

gamma3, gamma24 (plug-in for the γ3)

Thoughts on the need for galvanic isolation.

Postby garbz » May 29th, 2017, 9:29 am

Hi all,

I'm about to start the process of getting and building 2 of the γ3 DACs with very different purposes. One effectively will be a USB only DAC, and the other effectively a standalone only without USB module.

Looking at the design of the γ3, a lot of very good work was put into the galvanic isolation on the γ3 focusing on the USB module and the controller. I'm looking at putting my own controller in both systems to fit them with some other equipment I already have, and I thought, is there any point in isolating a control module .... if its clock is stopped?

There's no question about USB, it's a huge source of noise. But the noise which is potentially produced by a microcontroller only happens in any switching state and when when you have something along the lines of an AVR with PowerDown mode then there should be no running clock sources. The only time it would run would be to start up the system, and then on any interrupt generated by manual intervention (i.e. pushing a button some where) during which any affect on quality of sound should be a none issue anyway.

If you do that is there any real point in galvanically isolating anymore? Maybe feed it from it's own regulator, but I can't see there being a whole lot of benefit in providing it its own powersupply.

jh4db536 likes this post.
garbz
 
Posts: 9
Likes: 0 post
Liked in: 4 posts
Joined: May 29th, 2017, 12:51 am
Location: Netherlands
Country: Netherlands (nl)

Re: Thoughts on the need for galvanic isolation.

Postby amb » May 29th, 2017, 2:01 pm

One major benefit of galvanic isolation is the total elimination of ground loop. Even if your own microcontroller is "stopped", there is still the potential of a ground loop via the AC cord and the I2S bus, even with its own power supply. The digital isolators will eliminate that.
User avatar
amb
Site Admin
 
Posts: 10378
Likes: 175 posts
Liked in: 482 posts
Joined: March 29th, 2010, 9:49 pm
Location: Sunnyvale, CA
Country: United States (us)

Re: Thoughts on the need for galvanic isolation.

Postby garbz » May 30th, 2017, 6:09 am

I don't understand where, but maybe I have it differently in my head. I was proposing to co-locate the microcontroller with the γ24 and power them from a common supply mounted on a common backplane. It would share the supply that comes into the γ24 board GVDD (or DVDD depending on which schematic you look at) and then split off another 3.3V line which we could effectively call +3.3D :)

There should be no looping anywhere as there is one common ground then within the chassis and all external sources are then completely isolated (optically or via transformer). Does my rambling make sense?
garbz
 
Posts: 9
Likes: 0 post
Liked in: 4 posts
Joined: May 29th, 2017, 12:51 am
Location: Netherlands
Country: Netherlands (nl)

Re: Thoughts on the need for galvanic isolation.

Postby amb » May 30th, 2017, 12:40 pm

Will your microcontroller be completely self-contained? I.e., no connection to any external device that may have a ground that ends up connecting to AC earth? If so, and if you use the γ24 board GVDD to power your microcontroller, then there won't be a ground loop.

EDIT: But we're only talking about the digital isolator U15 (on the γ3 backplane board). Isolation is still a good idea between the LCDuino-1 and γ24 (U11-U14). Also, I think maybe it would be better for you to power your microcontroller from LVDD instead, because that's the "really dirty" digital area whereas GVDD is a cleaner digital ground (and you should avoid polluting it with noise). The only issue is that LVDD is always on. For the ζ1 USB-I2S module, the power is switched via U20 on the backplane (synchronized with the power-up of GVDD via isolation). I don't know what your microcontroller's power requirement is, maybe you could use this, since you won't be using the ζ1.
User avatar
amb
Site Admin
 
Posts: 10378
Likes: 175 posts
Liked in: 482 posts
Joined: March 29th, 2010, 9:49 pm
Location: Sunnyvale, CA
Country: United States (us)

Re: Thoughts on the need for galvanic isolation.

Postby garbz » May 30th, 2017, 12:54 pm

Yes, I should have clarified what I meant to "fit with some other systems". It will be self contained and only controlling the DAC, but I'm writing my own software for it, have the same controller in other equipment and effectively want to be able to reuse a lot of C code on a platform I know rather than work with an Arduino. I do also have connections to the power on/power off of other devices, but I already isolate those.

Thanks for the input, expect an order for a few boards and bits soon :-)
garbz
 
Posts: 9
Likes: 0 post
Liked in: 4 posts
Joined: May 29th, 2017, 12:51 am
Location: Netherlands
Country: Netherlands (nl)

Re: Thoughts on the need for galvanic isolation.

Postby amb » May 30th, 2017, 12:56 pm

In case you haven't noticed, read the "EDIT" in my post above.
User avatar
amb
Site Admin
 
Posts: 10378
Likes: 175 posts
Liked in: 482 posts
Joined: March 29th, 2010, 9:49 pm
Location: Sunnyvale, CA
Country: United States (us)

Re: Thoughts on the need for galvanic isolation.

Postby garbz » June 1st, 2017, 3:25 am

Oh thanks I hadn't noticed :)

But in the case that I'm not using a ζ1, and not using an LCDuino, and considering any switching circuit on the microcontrollwer will be standby when I'm not doing something like touching buttons I was aiming to forgo the entire LVDD circuit.

As for which isolation we're talking about. U15 will be surplus since there's no ζ1 and no dirty USB input to isolate. And my case is U11-U14 should probably not be required providing the microcontroller and any associated parts (no display just a few LEDs) are actually in a steady state. I don't see a reason for isolation since the result would be no "dirty" power due to no switching or clocking operations in steady state.
garbz
 
Posts: 9
Likes: 0 post
Liked in: 4 posts
Joined: May 29th, 2017, 12:51 am
Location: Netherlands
Country: Netherlands (nl)

Thoughts on the need for galvanic isolation.

Postby amb » June 1st, 2017, 5:57 am

I am a bit confused. What purpose will your microcontroller serve? Since we're talking about I2S, your microcontroller must be active and running in order to output I2S data, right? So how could it be in standby? If I misunderstand you and your microcontroller does not handle digital audio stream, then it would make sense.

But where does the music come from? You mentioned two builds, one with USB and one without. For the first one, what's handling the USB? For the second one, are you using S/PDIF (coax, AES, optical)?
User avatar
amb
Site Admin
 
Posts: 10378
Likes: 175 posts
Liked in: 482 posts
Joined: March 29th, 2010, 9:49 pm
Location: Sunnyvale, CA
Country: United States (us)

Re: Thoughts on the need for galvanic isolation.

Postby garbz » June 1st, 2017, 12:48 pm

I2C. The purpose of my microcontroller is no different to the purpose of your LCDuino except on an even more basic level. Get the system up and running, change channel when I need to.

Sorry I confused it by talking about 2 builds. The one with the USB will effectively be a reference build.

The one with S/PDIF will be coax (pulse transformer isolated per your design) and optical (isolated by it's nature). This is the one I'm talking about. Since this will have no ζ1 module, and since I will be replacing the LCDuino with something that starts the system up and then powers itself off waking only on an asynchronous interrupt of someone hitting the channel up and down key, staying alive just long enough to send the signal to the receiver to change channel and then powering it off again, do you see any downside in this forgoing isolation completely?

That's the question I'm talking about. At its most basic level, forgetting ground loops and all that, when you software interface to the receiver / DAC via a microcontroller the source of noise contaminating the powersupply is logic switching noise. Given the very basic requirements and ability to shut down the clock completely in the uC (then current draw is only 500nA as well, is there any reason to go to the effort of isolating and building an isolated PSU for this part.

I think no, but I wanted to see if I'm not missing something obvious. Sorry if I'm explaining this poorly.
garbz
 
Posts: 9
Likes: 0 post
Liked in: 4 posts
Joined: May 29th, 2017, 12:51 am
Location: Netherlands
Country: Netherlands (nl)

Re: Thoughts on the need for galvanic isolation.

Postby amb » June 1st, 2017, 1:22 pm

OK, all clear now.
User avatar
amb
Site Admin
 
Posts: 10378
Likes: 175 posts
Liked in: 482 posts
Joined: March 29th, 2010, 9:49 pm
Location: Sunnyvale, CA
Country: United States (us)

Next

Return to γ3 high resolution DAC and γ24 DAC core module

Who is online

Users browsing this forum: No registered users and 1 guest

Quick Links

AMB audio DIY resources
AMB audio shop