[cobaltmute]

I was just looking at R6 and R7, the through hole resistors that are part of the output filter.

There is a couple of options that you can use here:
- The spec'ed RN50.
- You can use CMF50 in the same place which is the same resistor in non-Military grade.
- CMF55 (and of course RN55) fit in the position. They are tight, but they do fit.
- Xicon MF-RC 270 series
- Panasonic EROS2

and probably some others.

I'd stick with a good quality option here. These are directly in the signal path and you want something decent here.

 

[keyid]

would a tombstoned 100ohm carbon film work well for r6/r7?

 

[cobaltmute]

I would think you can get away with it. R6/7 are a simple low pass filter so it shouldn't be too sensitive to orientation.

As for carbon versus metal film, that would seem to be personal preference. The only reason I didn't do SMD for these positions was some negative comments to the nickel barriers I've read and the not nickel ones were $17 a resistor.

 

[stwspoon]

I have a few observations based on my build and eval of the HPDAC last year.
Looking at your schematic, PSEL on the PCM270x is grounded, On the HPDAC it is pulled up to indicate 500 ma power to the host. Maybe have a few smt resistor pads or jumpers to select pull up or pull down based on how the user will power the grubdac.

HOST is tied to Vdd on the GrubDac and is tied to VBus (through a ferrite) on the HPDac. This would be useful so that the PCM270x can properly detect when the usb is inserted when the grubdac is not USB powered.

Also could the SSPND pin be brought out to a pad? It may be useful in reducing inrush currents by keeping the DC/DC converter on a Carrie disabled until a few milliseconds after the USB is inserted ( see page 26 of the PCM270x data sheet for the power up timing)

Definitely add me to your list for your next round of boards.

 

[cobaltmute]

Quote:

Originally Posted by stwspoon /img/forum/go_quote.gif I have a few observations based on my build and eval of the HPDAC last year. Looking at your schematic, PSEL on the PCM270x is grounded, On the HPDAC it is pulled up to indicate 500 ma power to the host. Maybe have a few smt resistor pads or jumpers to select pull up or pull down based on how the user will power the grubdac.

I suggest you re-read the meaning of PSEL. It does not indicate the power consumption of the device - it specifies how the PCM2707 is getting its power.

Quote:

HOST is tied to Vdd on the GrubDac and is tied to VBus (through a ferrite) on the HPDac. This would be useful so that the PCM270x can properly detect when the usb is inserted when the grubdac is not USB powered.

Again check the datasheet for what HOST does. The options for what you want to do with PSEL/HOST don't work as a combination for Power/Bus Detection.

I'll have to consider the not-powered by USB options. It was not something that was an initial design consideration. I have run these settings with external power on my other DAC design and it did work and function, but probably breaks a bunch of USB standards.

Quote:

Also could the SSPND pin be brought out to a pad? It may be useful in reducing inrush currents by keeping the DC/DC converter on a Carrie disabled until a few milliseconds after the USB is inserted ( see page 26 of the PCM270x data sheet for the power up timing)

Again, something I'll consider but possible a good use for that pin. The only problem is to latch on the DC/DC once power is established as SSPND will go low again after power on.

Quote:

Definitely add me to your list for your next round of boards.

There is still a lot of the first round boards available.

 

[cobaltmute]

I've added an alternate J1 USB connector to the online BOM as the original is out of stock.

I have listed both, but you only need one. When the cheaper option is back in stock, I will remove the more expensive one.

 

[stwspoon]

Quote:

Originally Posted by cobaltmute /img/forum/go_quote.gif I suggest you re-read the meaning of PSEL. It does not indicate the power consumption of the device - it specifies how the PCM2707 is getting its power.

That is right. PSEL specifies how the PCM2707 is getting power TO THE HOST via the configuration descriptor( p 16). What shuts down the onboard regulator is connecting Vbus to Vcc. In bus powered mode, HOST just toggles the 100ma /500 ma flag in the configuration descriptor.

Quote:

Originally Posted by cobaltmute /img/forum/go_quote.gif Again check the datasheet for what HOST does. The options for what you want to do with PSEL/HOST don't work as a combination for Power/Bus Detection. I'll have to consider the not-powered by USB options. It was not something that was an initial design consideration. I have run these settings with external power on my other DAC design and it did work and function, but probably breaks a bunch of USB standards.

Yea, what meets USB spec and what mostly works is quite different. DAMHINT. I learned a lot wrestling with the TPS61170 boost regulator on the HPDAC. The DCP design in Carrie is much better. Most USB host devices will let you draw ~ 500 ma (900 ma for USB 3.0

) without looking at the config descriptor, but not all.

Quote:

Originally Posted by cobaltmute /img/forum/go_quote.gif Again, something I'll consider but possible a good use for that pin. The only problem is to latch on the DC/DC once power is established as SSPND will go low again after power on. There is still a lot of the first round boards available.

Great! I thought they were all gone. I must have gotten confused with the Carrie boards as that is what I want to pair it with.

 

[cobaltmute]

Quote:

Originally Posted by stwspoon /img/forum/go_quote.gif That is right. PSEL specifies how the PCM2707 is getting power TO THE HOST via the configuration descriptor( p 16). What shuts down the onboard regulator is connecting Vbus to Vcc. In bus powered mode, HOST just toggles the 100ma /500 ma flag in the configuration descriptor.

I'd really like something to backup the comment shutting down the regs. If that is true, I can see where you are going and it looks like it would be somewhat easy to implement. The only issue (compared to what is done now) is to separate out the power to the on-board regulator from Vbus.

 

[nightanole]

Its true that vcc to host and psel to ground shuts down the internal regs, its the way the gamma dac does it.

 

[cobaltmute]

I'm well aware of what the gamma1 does.

But the concept is that Vbus or Vdd to Vcc alone shuts down regulators is a good idea. This allows us to put PSEL high, indicating to the PC that the DAC is bus-powered. This allows us to use the HOST pin to tell the PC that we are going to pull upto 500mA. In the case where the Carrie is attached, it helps us ensure the proper power is available

The way I've done it, which is the same as the gamma1, tells the PC that the DAC is self powered, which by USB standards only allows 100mA draw. As amb says, a number of stupid devices always pull more so do we worry about it?

 

[nightanole]

Ya my light up snow men and fans use the 5v from usb and the pc could care less. I wonder if this whole 100ma-500ma key is for older usb 1.0 buses.

 

[stwspoon]

Here is where I got my info from:
ecp.cc

Specifically the notes around VR3 near the bottom of the page.

Also diyaudio has several long threads on getting the PCM270x to work
Another PCM2707 problem... - diyAudio
USB powered NOS USB DAC with built-in headphone amp - diyAudio

On not backfeeding the USB bus in self powered mode
USB interface NOS DAC?? - Page 2 - diyAudio

 

[stwspoon]

Desktop PCs usually don't care about USB power, but laptops and bus powered USB hubs can be more particular. They all current limit as part of the spec is a test where each USB pin is shorted to the others for 24 hrs to simulate a cable being smashed/shredded/cut/etc. The port doesn't have to work afterwards, but it can't catch fire

I have a docking station that I can't plug my USB fan into, but it works fine in the laptop.

 

[cobaltmute]

I've added a note to the Carrie thread about my issues with power and startup. This contains important information if you are going to integrate the GrubDAC with something else and draw all power from the USB bus.

It appears the stwspoon is correct about the pin settings and power draw when having the DAC integrated with another device. If fact, I just Googled this tidbit:USB-IF Compliance Updates

In r1b of the board, PSEL with be set high which tells the host it is bus-powered. HOST is already high (indicating 500mA) and since all power comes from the USB bus, I'm not worried about disconnect/reconnect issue.

This issue will not affect stand-alone GrubDACs and may or may not affect GrubDAC + Carrie setups depending upon your PC hardware.

EDIT: I've updated the Carrie thread with an easier fix to this issue.

 

[amc]

If you have boards left I would like 2. Sent PM