Don't know why it should matter, but at the moment it happens to be: Old Acer Travelmate 5720 with Windows-10 64-bit -> USB2 ->
Plugable (brand) USB2 2.0 4-Port Hub BC 1.2 Charging (Model:USB2-HUB4BC) -> USB2 -> Amanero(or XMOS) -> I2S -> , etc.
Last edited:
You could use, say, one of the low cost $30 AK4137 boards. They have I2S inputs and outputs, so you should be able to patch it in between RPi I2S outputs and ES9038Q2M board I2S inputs. Basically: RPi -> I2S -> AK4137 -> I2S -> ES9038Q2M
Normally the AK4137 board would get some signals from the Amanero(or XMOS) board to tell it if the incoming data is PCM or DSD, and at least for DSD what the sample rate is. If those signals are missing from RPi (DSDON, F1, F2), then you can set them manually with the buttons on AK4137 board.
If you want to try what I am working on to try to make Q2M work as a master mode dac with AK4137 output clocking as slaved, we could talk about that some more.
That one looks like hacking into it for master mode dac use would be more difficult than for the AK4137 board I am using. AK4137 I2S/DSD Sample Rate Conversion Board Supports PCM/DSD Interconversion DOP | eBay
And, for only a few dollars more you get a display, too
EDIT: This one has the 768k PCM option available: AK4137 I2S/DSD Sample Rate Conversion Board Supports PCM/DSD Interchange | eBay
Last edited:
@ Markw4
i have a ak4137 board and spend a lot of time reading the dataset. we have the driver for 9038q2m on master mode, i'd like to get a best way to control the ak4137 and q2m via i2c commands from pi ( volumio ) so we could change fomats, filters or bypass ak4137.
any suggestions?
i have a ak4137 board and spend a lot of time reading the dataset. we have the driver for 9038q2m on master mode, i'd like to get a best way to control the ak4137 and q2m via i2c commands from pi ( volumio ) so we could change fomats, filters or bypass ak4137.
any suggestions?
Last edited:
@eslei,
As one can see from looking at AK4137 data sheet, some things are configured over I2C bus, and some things by dedicated pins on the chip. I haven't added up the total number of pins that would need to be controlled but it could be quite a few. If RPi has enough spare GPIO pins to run I2C bus and control all the other AK4137 pins, then it would only be a matter of wiring them up together. However, it seems likely AK4137 might need too many GPIO pins to do it that way. In that case one could use an intermediate chip to control all or some of the pins, like say for example, an Arduino. In that case RPi would talk to Arduino, and Arduino would talk to AK4137. That could be a lot of work to implement though.
I think we will need to count up all the AK4137 pins we might need to control, then decide if RPi can do it directly, or if an intermediary chip such as Arduino would be needed. Also, if you just want to make one to prove it works, it might make more sense to limit the number of AK4137 things to be controlled by RPi to a minimum. That would mean some things might still have to set by pushing buttons on an AK4137 board. It would probably be a good idea to make up a list of 'must have' things to control, and another list of 'nice to have' things, then see how much work would be involved to implement each list.
For example, I don't know if the filters would need to be software controlled. Probably one would find there is a preferred filter for PCM and another one for DSD. In that case the existing AK4137 board can remember the preference for each one. For other things like setting input sample rate for AK4137 that might need to be under RPi software control in order to allow playlists of mixed file types.
The way I see it at this point, a basic test of AK4137 would not take any RPi control. Even master mode PCM would not require it if using the AK4137 board I have. However, master mode DSD would require at least RPi control of I2C bus registers for AK4137.
So, I think you should think about how important having various features under software control are, and how much work you are willing to do to make it happen. There is always a way to give you full control of everything if you are willing to do enough work.
As one can see from looking at AK4137 data sheet, some things are configured over I2C bus, and some things by dedicated pins on the chip. I haven't added up the total number of pins that would need to be controlled but it could be quite a few. If RPi has enough spare GPIO pins to run I2C bus and control all the other AK4137 pins, then it would only be a matter of wiring them up together. However, it seems likely AK4137 might need too many GPIO pins to do it that way. In that case one could use an intermediate chip to control all or some of the pins, like say for example, an Arduino. In that case RPi would talk to Arduino, and Arduino would talk to AK4137. That could be a lot of work to implement though.
I think we will need to count up all the AK4137 pins we might need to control, then decide if RPi can do it directly, or if an intermediary chip such as Arduino would be needed. Also, if you just want to make one to prove it works, it might make more sense to limit the number of AK4137 things to be controlled by RPi to a minimum. That would mean some things might still have to set by pushing buttons on an AK4137 board. It would probably be a good idea to make up a list of 'must have' things to control, and another list of 'nice to have' things, then see how much work would be involved to implement each list.
For example, I don't know if the filters would need to be software controlled. Probably one would find there is a preferred filter for PCM and another one for DSD. In that case the existing AK4137 board can remember the preference for each one. For other things like setting input sample rate for AK4137 that might need to be under RPi software control in order to allow playlists of mixed file types.
The way I see it at this point, a basic test of AK4137 would not take any RPi control. Even master mode PCM would not require it if using the AK4137 board I have. However, master mode DSD would require at least RPi control of I2C bus registers for AK4137.
So, I think you should think about how important having various features under software control are, and how much work you are willing to do to make it happen. There is always a way to give you full control of everything if you are willing to do enough work.
Last edited:
I'm considering using a LT3035 driving a LiFe battery in a ghetto setup into AVCC. To keep the parts list to a minimum I don't want to disconnect the battery. The digital side will be powered by the LT3042 regulators I already have. So what would happen if the AVCC has power and the digital side of the chip looses power? Does the chip just turn off?
Last edited:
Okay, thank you. Everyone in this thread should probably be aware if you aren't already that ESS, maker of ES9038Q2M, treats all data sheets as confidential information only available under NDA. However, they seem to be happy to provide them to people willing to sign an NDA in order to have access. Signing an NDA limits what can be talked about publicly and or with people who have not signed an NDA. Only information that has been legally made public can be discussed openly.
Now that we have that out of the way, regarding LiPO4 batteries and AVCC, some people have suggested trying it so one time I did. It didn't sound right to me at all. There was obviously some modulation of the voltage due to battery and or interconnection impedances. I guess with big enough batteries then internal impedance should be able to be very low, but then there are contacts and wiring between batteries and dac chip and then from dac board ground back to batteries. Often I see battery setups with skinny wires running around and its hard to see how that could be better than a very high quality voltage regulator located right at the dac pins, or at least with sense leads right at the dac chip to compensate for wire resistance.
Regarding questions about actual AVCC currents, some info was published on Dimdim's site, IIRC. Also, IIRC, Iancanada posted some numbers for ES9038Q2M vs ES9028k2m. Turns out 28k2m runs at higher output currents than 38Q2M. Maybe they felt k2m ran at too much current for a mobile part. Anyway, ESS gives current draw numbers under very specific chip operating conditions most of us here won't be using. So whatever they say is wrong, which they must certainly know. They seem to leave it up to each designer to measure current draws under the conditions they intend to operate the chips. We do know that an LME49720 can easily power 38Q2M AVCC for one channel, and so that kind of puts a limit on what the current could be.
Now that we have that out of the way, regarding LiPO4 batteries and AVCC, some people have suggested trying it so one time I did. It didn't sound right to me at all. There was obviously some modulation of the voltage due to battery and or interconnection impedances. I guess with big enough batteries then internal impedance should be able to be very low, but then there are contacts and wiring between batteries and dac chip and then from dac board ground back to batteries. Often I see battery setups with skinny wires running around and its hard to see how that could be better than a very high quality voltage regulator located right at the dac pins, or at least with sense leads right at the dac chip to compensate for wire resistance.
Regarding questions about actual AVCC currents, some info was published on Dimdim's site, IIRC. Also, IIRC, Iancanada posted some numbers for ES9038Q2M vs ES9028k2m. Turns out 28k2m runs at higher output currents than 38Q2M. Maybe they felt k2m ran at too much current for a mobile part. Anyway, ESS gives current draw numbers under very specific chip operating conditions most of us here won't be using. So whatever they say is wrong, which they must certainly know. They seem to leave it up to each designer to measure current draws under the conditions they intend to operate the chips. We do know that an LME49720 can easily power 38Q2M AVCC for one channel, and so that kind of puts a limit on what the current could be.
Last edited:
Thanks Mark. I plan to implement the battery with super short thick soldered leads directly to battery and onto the to chip leg on positive. The 26650 is capable of 35A which indicates very low internal resistance. I dont know the capacitance of the battery unfortunately and I've read here on the forum that the regulator isn't fond of running super caps. That's why I'm thinking of using a battery instead. Battery would have to go back to regulator ground pin like star ground perhaps. I see that the regulator have current limit, so if for some reason the battery is empty, the regulator will only supply current at a rate that the current limit dictates I guess. If I never turned of the dac I think this solution would work for some time before running into some issues. If by some miracle the dac chip would switch of by loosing power to the other inputs, this way of powering the chip could provide good bass and still be hassle free.
impulse60,
May I ask if you have tried a high quality AVCC supply like we use here and found some problem with it? If not, I don't understand why to go batteries first. Also, if you don't have a high quality regulator solution to compare with, how would you know if the battery supply even works as well the high quality regulator solution does?
I know some people try LDOs and aren't happy, so they go to super caps, or batteries. But, they don't seem to try comparing with a type of regulator that is already known to work very well for AVCC, such as the opamp type we use. The only LDOs I see people using seriously for AVCC is ADM7150 or ADM7151, IRRC, not LT3045 or LT3042. Thing is, noise is not the only issue, output impedance across the audio band is a big one too, and not something every LDO seems to be equally good at.
May I ask if you have tried a high quality AVCC supply like we use here and found some problem with it? If not, I don't understand why to go batteries first. Also, if you don't have a high quality regulator solution to compare with, how would you know if the battery supply even works as well the high quality regulator solution does?
I know some people try LDOs and aren't happy, so they go to super caps, or batteries. But, they don't seem to try comparing with a type of regulator that is already known to work very well for AVCC, such as the opamp type we use. The only LDOs I see people using seriously for AVCC is ADM7150 or ADM7151, IRRC, not LT3045 or LT3042. Thing is, noise is not the only issue, output impedance across the audio band is a big one too, and not something every LDO seems to be equally good at.
Last edited:
I simply wan't to try it
Fair enough, can't argue with that.
If you would like to buy a solution rather than build, TP offers little ADM7151 based AVCC boards that you could also use, if you ever wanted to, for an ES9038PRO dac. They should probably also be usable whenever you need one or two 600mA, 3.3v LDOs for anything. Maybe not a bad investment to have around. I should try one myself one of these days to see how they sound (presumably just fine, but want to see for myself if its something I would recommend to folks).
Last edited:
In other news, Arduino software for dac master mode testing is about ready. Basic functionality for accessing AK4137 I2C registers seems to be working and was added to the existing Arduino dac control program. Want to write out a little test plan so I don't accidentally end up with Q2M and AK both trying to drive the same clock lines at the same time. Should find out if it works before too long. Very interesting part of the project, seems to me anyway.
TP?Fair enough, can't argue with that.
If you would like to buy a solution rather than build, TP offers little ADM7151 based AVCC boards that you could also use, if you ever wanted to, for an ES9038PRO dac. They should probably also be usable whenever you need one or two 600mA, 3.3v LDOs for anything. Maybe not a bad investment to have around. I should try one myself one of these days to see how they sound (presumably just fine, but want to see for myself if its something I would recommend to folks).
Looking at TP Trident-SR again, I see they actually set for 3.6v, pushing AVCC towards the limits in order to get a little more free output level out of the dac. Probably okay for AVCC, but not necessarily for other nominally 3.3v loads. If contemplated for other than AVCC use, please check data sheets for voltage limits of whatever parts you might be considering using it with.
Hi
I am waiting for my ver. 1.06 board from china. Meanwhile I want to prepare for the fun. I do have 2 pcs permaloy 1:4 signal transformer. input inductance is 2.2H and 190 Ohm. Could you please advise whether I should solder that directly, previously removing the 6,2 Ohm resistors?
Reading this very long post I have understood that this board performs better in current mode and worse in voltage mode, can this be achieved with output transformers?
Thanks
I am waiting for my ver. 1.06 board from china. Meanwhile I want to prepare for the fun. I do have 2 pcs permaloy 1:4 signal transformer. input inductance is 2.2H and 190 Ohm. Could you please advise whether I should solder that directly, previously removing the 6,2 Ohm resistors?
Reading this very long post I have understood that this board performs better in current mode and worse in voltage mode, can this be achieved with output transformers?
Thanks