Here's my Mouser part lists.
Two excel files, one fairly complete, the other does not include any of the output or headphone circuit, basically ending at the DAC.
I think the parts should match the board, I did check, but I also have not bought anything yet, so no guarantees.
I used polymer caps in the digital section, and mostly panny FC's elsewhere. I did use Nichicons in some of the analog section, but I am also not going to populate that circuit. If you are, you may want to consider upgrading.
I may actually order Sanyo Oscon's instead of the polymers, haven't decided yet.
I didn't find all of the connectors and some other stuff because I am not going to buy those parts right now.
For the main oscillator, I am thinking about using a surface mount Crystek instead of the suggested one, but this may not but a simple mod to make. I would use copper tape to connect it, to make sure I get nice, low impedance connections to the oscillator.
Randy
Two excel files, one fairly complete, the other does not include any of the output or headphone circuit, basically ending at the DAC.
I think the parts should match the board, I did check, but I also have not bought anything yet, so no guarantees.
I used polymer caps in the digital section, and mostly panny FC's elsewhere. I did use Nichicons in some of the analog section, but I am also not going to populate that circuit. If you are, you may want to consider upgrading.
I may actually order Sanyo Oscon's instead of the polymers, haven't decided yet.
I didn't find all of the connectors and some other stuff because I am not going to buy those parts right now.
For the main oscillator, I am thinking about using a surface mount Crystek instead of the suggested one, but this may not but a simple mod to make. I would use copper tape to connect it, to make sure I get nice, low impedance connections to the oscillator.
Randy
Randy great BOM, can you export this directly from mouser (if you've already got a project on the site?)
Hi
I didn't make a project at Mouser for this, just have all the links to the parts in the spreadsheet right now.
I am creating an order for my buy right now, and I can turn that into a project. I am ordering 1 or 2 extra of the small cheap parts, they are easy to lose when you are soldering. Also, I am ordering the no output version.
Randy
Randy,
Looking at the TI TCM2702PJT, there are two versions available from Mouser.
595-PCM2707PJT @$10.13
595-PCM2707PJTRG4 @ $6.80
The only difference I see is the package, T&R vs Tray.
Is there a reason to buy the more expensive one?
I believe they are charging more because of special handling.
Cut T&R is in an antistatic tape. Cut, bag, ship.
Tray has to be transferred from a tray to an antistatic bag. I believe you are paying $3.33 for handling.
Looking at the TI TCM2702PJT, there are two versions available from Mouser.
595-PCM2707PJT @$10.13
595-PCM2707PJTRG4 @ $6.80
The only difference I see is the package, T&R vs Tray.
Is there a reason to buy the more expensive one?
I believe they are charging more because of special handling.
Cut T&R is in an antistatic tape. Cut, bag, ship.
Tray has to be transferred from a tray to an antistatic bag. I believe you are paying $3.33 for handling.
I looked at the datasheet, and agree.
The part in the link in my excel file is $6.80, unless I am missing something?
This is different from the links I posted a few weeks ago, I saw your post about it, and checked the datasheet, and changed the part used.
Randy
Yes, when I get a chance, I will post what I plan to do, but basically I plan to power the oscillator from a different reg then the 8741. In the design, they are both powered from the first +5 reg.
Also, if you use the SPDIF input (I am not at first), it would be a good idea to give the hc04 it's own reg also.
Randy
Someone asked about a sony remote.
I found one that looked just like the one hifdino pictured on his webpage. Haven't even started building this yet so no garantees. It was called a "Sony Trinitron RM-757 TV Remote Control ".
On a side note anyone plan on not installing the upsampler? I understand there would be jitter reduction, but with the aurodino and an asynch usb I would want to chose upsampling 2x,4x, or none, with access to all 3x5 filters. with the upsampling chip you lose access to 10 filters, iow with the upsampling chip installed you are stuck with only 2x over sampling.
I found one that looked just like the one hifdino pictured on his webpage. Haven't even started building this yet so no garantees. It was called a "Sony Trinitron RM-757 TV Remote Control ".
On a side note anyone plan on not installing the upsampler? I understand there would be jitter reduction, but with the aurodino and an asynch usb I would want to chose upsampling 2x,4x, or none, with access to all 3x5 filters. with the upsampling chip you lose access to 10 filters, iow with the upsampling chip installed you are stuck with only 2x over sampling.
Saw that there is a new DAC Magic plus version released by Cambridge. For the price I started investigating just buying one of those instead of building this. It has hirez asynch usb integrated (Xmos) and dual mono (8740). Now for my price range over the years anytime I buy a DAC instead of building one I am always disapointed by something or a roadblock in the commercial product that makes modding to what I need near impossible.
Reading the wm8741 datasheet I was surprised to see that the THD is -100db for the WM8741 while -104 for the WM8740. While the SNR is a little better for the WM8741.
The big difference is the Magic is using a separate programmed 384khz upsampler, it allows selection of linear phase, minimum phase and steep of Camridge's own filter elements.
Here with this kit and Hifidino we can pick Wolfsons filters including apodizing and also with a little more code add a separate filter (the PMD100 was mentioned in the WM8741 datasheet which certainly caught my attention.)
For the asynch usb we can use the xmos reference board or the WaveIO using the same windows drivers as the Magic+. Driver support was a concern I had w.r.t the Xmos but it should be good now with the Magicplus sure to be a big seller.
And of course with this kit's little pcb it will be easier to build separate analog, digital, and clock powr supplies. Last with th kit we have flexibility w.r.t. the analog stage.
So I guess for my goals the kit pcbs will be my path foward rather than buying a DACMagic Plus.
Wrt to the ASRC chip, I've decided to leave it out. With an xmos board I think I can share the same clocks with the WM8741's, so no real need to install the ASRC.
Its interesting that Cambridge decided to use their own programs to make an 8x upsampler instead of using the wm8741's (with wm's internal filters.) Maybe the extra ~ $20 per unit for the 8741's played part in that desicion or their filters are that much better than Wolfsons?
As far as the SMD as long as you watch the heat you can't screw up so have the low temp solder on hand for safe removal. I've never had problems using this technique, sure a little gets mixed in but its not like I plan on doing any off-roading or boating with this DAC.
Planning on throwing in a "wire" for the headphone out.
The thing I need to research is the digital volume control, the datasheet of the wm8741 says it works with 32+32 wordlengths, but I don't know if the Xmos outputs in this format like most USB would or if it matters and the WM8741 operates the volume control from 32 bits regardless the incoming bit depth. The WM8740 makes no mention of 32 bit, so this could mean a much better digital volume control with the wm8741 or it it just truncates the 32 down to 24 ?
Well this is going to be an interesting project, I guess my goal is more flexibility and better sound than the Magic+, for less that $600, it will definately be more fun
Reading the wm8741 datasheet I was surprised to see that the THD is -100db for the WM8741 while -104 for the WM8740. While the SNR is a little better for the WM8741.
The big difference is the Magic is using a separate programmed 384khz upsampler, it allows selection of linear phase, minimum phase and steep of Camridge's own filter elements.
Here with this kit and Hifidino we can pick Wolfsons filters including apodizing and also with a little more code add a separate filter (the PMD100 was mentioned in the WM8741 datasheet which certainly caught my attention.)
For the asynch usb we can use the xmos reference board or the WaveIO using the same windows drivers as the Magic+. Driver support was a concern I had w.r.t the Xmos but it should be good now with the Magicplus sure to be a big seller.
And of course with this kit's little pcb it will be easier to build separate analog, digital, and clock powr supplies. Last with th kit we have flexibility w.r.t. the analog stage.
So I guess for my goals the kit pcbs will be my path foward rather than buying a DACMagic Plus.
Wrt to the ASRC chip, I've decided to leave it out. With an xmos board I think I can share the same clocks with the WM8741's, so no real need to install the ASRC.
Its interesting that Cambridge decided to use their own programs to make an 8x upsampler instead of using the wm8741's (with wm's internal filters.) Maybe the extra ~ $20 per unit for the 8741's played part in that desicion or their filters are that much better than Wolfsons?
As far as the SMD as long as you watch the heat you can't screw up so have the low temp solder on hand for safe removal. I've never had problems using this technique, sure a little gets mixed in but its not like I plan on doing any off-roading or boating with this DAC.
Planning on throwing in a "wire" for the headphone out.
The thing I need to research is the digital volume control, the datasheet of the wm8741 says it works with 32+32 wordlengths, but I don't know if the Xmos outputs in this format like most USB would or if it matters and the WM8741 operates the volume control from 32 bits regardless the incoming bit depth. The WM8740 makes no mention of 32 bit, so this could mean a much better digital volume control with the wm8741 or it it just truncates the 32 down to 24 ?
Well this is going to be an interesting project, I guess my goal is more flexibility and better sound than the Magic+, for less that $600, it will definately be more fun
Last edited:
OK,
I didn't see a reply and thought my original post had been missed (or I missed something in the data sheet).
Steven
Shared mouser project
Here's Randy's BOM as a shared project on mouser (without the headphone amp section.)
http://www.mouser.com:80/ProjectManager/ProjectDetail.aspx?AccessID=89c84f9dce
access id: 89c84f9dce
You should be able to log into mouser and import it into your own account. (it's missing component references)
Here's Randy's BOM as a shared project on mouser (without the headphone amp section.)
http://www.mouser.com:80/ProjectManager/ProjectDetail.aspx?AccessID=89c84f9dce
access id: 89c84f9dce
You should be able to log into mouser and import it into your own account. (it's missing component references)
Thanks for that
I think I made a couple mistakes.
I'll try to check it soon.
BTW, I decided to split my order, some stuff from Digikey, and most of it from Mouser.
My reason is that Digikey sells Elna Silmic II's. You need to search for "Elan RFS". Here is the catalog page
Silmic II
I am going to end up ordering the ferrites, a few IC's, and the Vregs from DK, to hit their $25 min.
While I'm sharing, you can get Oscons from here
oscon
They also carry poscap's, which are supposed to be good.
Randy
I think I made a couple mistakes.
I'll try to check it soon.
BTW, I decided to split my order, some stuff from Digikey, and most of it from Mouser.
My reason is that Digikey sells Elna Silmic II's. You need to search for "Elan RFS". Here is the catalog page
Silmic II
I am going to end up ordering the ferrites, a few IC's, and the Vregs from DK, to hit their $25 min.
While I'm sharing, you can get Oscons from here
oscon
They also carry poscap's, which are supposed to be good.
Randy
OPA2365 THD is specified at 600 Ohm load, however output voltage swing is specified at 10K Ohm load.
I'd be inclined to stick to a 10K (or greater) load.
WM8741 spec is 2Vrms out, so I would expect the buffer output to be limited to slightly less than that, depending on the value of the analog supply.
If you mean drive directly from the VOUTRP/VOUTTRN and VOUTLP/VOUTLN pins, from the WM8741, that is ill-advised. If you do, use diff-input transformers for coupling.
I'd be inclined to stick to a 10K (or greater) load.
WM8741 spec is 2Vrms out, so I would expect the buffer output to be limited to slightly less than that, depending on the value of the analog supply.
If you mean drive directly from the VOUTRP/VOUTTRN and VOUTLP/VOUTLN pins, from the WM8741, that is ill-advised. If you do, use diff-input transformers for coupling.
Last edited:
So I'm planning to use a transformer on the dac output, for a bal to se conversion, so I don't need the output stage.
I'm not sure if I have created confusion by creating a BOM without the output stage? You could decide to include the opamp, and related parts, so you don't have to worry about it. Or, just build the entire board, with all parts.
So, I am have just confused people?
FYI, I am planning to build a "customized" version. I only plan to use the USB input, and not planning to use the output stage. I am not going to install the parts I don't need, based on this.
Randy
I'm not sure if I have created confusion by creating a BOM without the output stage? You could decide to include the opamp, and related parts, so you don't have to worry about it. Or, just build the entire board, with all parts.
So, I am have just confused people?
FYI, I am planning to build a "customized" version. I only plan to use the USB input, and not planning to use the output stage. I am not going to install the parts I don't need, based on this.
Randy
I think the problem with driving the transformer directly from the DAC is that the DC resistance of the transformer is too low. The dc output of the analog signals is going to be one half AVDD. You would need to capacitivly couple the output(s) to the transformer to prevent excessive dc loading. One would need large (47uF min, probably 100uF) non-polarized caps for this purpose.
You need a buffer/level shifter before the transformer, and the OPA2365 with the capacitor coupling caps performs these functions. Level shifting is accomplished via the output cap C41. But now you have a single ended signal.
One possibility might be to bypass the OPA2365, and use the headphone amp position as unity gain amps to drive the transformer. It would require finding a better op-amp that was pin compatible, adding a V- supply to give more voltage swing, and capacitivly coupling between the D/A and the buffers.
I'm not sure how feasible this really is without compromising performance and s/n ratio.
Another option might be a longtail pair made of FETs (2SK170s?) with a current source in the tail, driving the transformer. (I was looking at doing this but with vacuum tubes instead of FETs.
You need a buffer/level shifter before the transformer, and the OPA2365 with the capacitor coupling caps performs these functions. Level shifting is accomplished via the output cap C41. But now you have a single ended signal.
One possibility might be to bypass the OPA2365, and use the headphone amp position as unity gain amps to drive the transformer. It would require finding a better op-amp that was pin compatible, adding a V- supply to give more voltage swing, and capacitivly coupling between the D/A and the buffers.
I'm not sure how feasible this really is without compromising performance and s/n ratio.
Another option might be a longtail pair made of FETs (2SK170s?) with a current source in the tail, driving the transformer. (I was looking at doing this but with vacuum tubes instead of FETs.
Last edited:
If you use the differential signals from the WM8741 , you would have one on each end of the of the two primaries in series . So the net DC across the primary would be zero, I don't see the need for a capacitor unless the primaries are poorly matched. What am I missing?
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.