About pulse tx connections:
Signal input to pin 1 - signal ouput pin 4
Ground input to pin 3 - ground ouput pin 6 (if needed to isolate also ground?
Datasheet pulse tx http://media.digikey.com/pdf/Data%20Sheets/Newava%20PDFs/S22160.pdf
Signal input to pin 1 - signal ouput pin 4
Ground input to pin 3 - ground ouput pin 6 (if needed to isolate also ground?
Datasheet pulse tx http://media.digikey.com/pdf/Data%20Sheets/Newava%20PDFs/S22160.pdf
back from a holiday I have asked qusp for a perfect quad of selected FETS and a heatsink...
(Expensive these FETS are pfff... but I just want them ;-) )
(Expensive these FETS are pfff... but I just want them ;-) )
Hi Merlin,
I have always had similar problems with the Buffalo II, but less so with the SPDIF input. I tried using the I2S input scheme the other day, and it doesn't work at all unless you pull out the controller. I guess TP optimized it for SPDIF input and as a result the I2S doesn't work correctly with some sources, which is a shame. So now I'm stuck using either SPDIF or losing my volume control and using I2S.
I also experienced dropouts on the SPDIF input of the Buffalo when my entire system was wired up with coax SPDIF. I introduced galvanic isolation between my music server and my Behringer DEQ2496, and the problem went away.
Anyhow... that TX you've chose looks like it should work well. Don't forget to include the resistor and the capacitor also needed for that section.
JBdV:
Good call on the matched fets. Tightly matching those parts will provide an actual tangible benefit in both channel matching and distortion performance, so I'd spend money there before I spent it anywhere else. When I was developing the circuit I kept noticing that the distortion would always be better in one channel than the other on my prototypes, and I finally tracked it down to device matching. I spent a good while matching the fets in my I/V and it definitely paid off.
Cheers,
Owen
I have always had similar problems with the Buffalo II, but less so with the SPDIF input. I tried using the I2S input scheme the other day, and it doesn't work at all unless you pull out the controller. I guess TP optimized it for SPDIF input and as a result the I2S doesn't work correctly with some sources, which is a shame. So now I'm stuck using either SPDIF or losing my volume control and using I2S.
I also experienced dropouts on the SPDIF input of the Buffalo when my entire system was wired up with coax SPDIF. I introduced galvanic isolation between my music server and my Behringer DEQ2496, and the problem went away.
Anyhow... that TX you've chose looks like it should work well. Don't forget to include the resistor and the capacitor also needed for that section.
JBdV:
Good call on the matched fets. Tightly matching those parts will provide an actual tangible benefit in both channel matching and distortion performance, so I'd spend money there before I spent it anywhere else. When I was developing the circuit I kept noticing that the distortion would always be better in one channel than the other on my prototypes, and I finally tracked it down to device matching. I spent a good while matching the fets in my I/V and it definitely paid off.
Cheers,
Owen
indeed, I only expressed that there may not be a huge benefit to matching the whole quad, its ideal of course, but as long as the 2 in each channel are matched you get the higher performance, the VERY slight difference in channels can be swamped by an attenuator in the next stage (if you still use that kind of thing) but hey we are audiophiles hehe.
strange, i've never experienced many dropouts like you guys are getting with my buff,2, whether with the volumite, or with spdif or i2s. ive had a couple though with other peoples coax sources. never optical and ive never had any problems at all with my ackodac, he has done a really good job with both i2s and spdif with termination and the impedance matched transformer coupled inputs with comparator.
opc, have you got the latest version of the buffalo 2 firmware? or still the first one, as i remember you got your b2 pretty early, i do read of a number of people having trouble with i2s at higher bitrates though.
basically guys, forget that stuff, usb->i2s is the way to go, i hardly ever use the other inputs except at audio meets.
strange, i've never experienced many dropouts like you guys are getting with my buff,2, whether with the volumite, or with spdif or i2s. ive had a couple though with other peoples coax sources. never optical and ive never had any problems at all with my ackodac, he has done a really good job with both i2s and spdif with termination and the impedance matched transformer coupled inputs with comparator.
opc, have you got the latest version of the buffalo 2 firmware? or still the first one, as i remember you got your b2 pretty early, i do read of a number of people having trouble with i2s at higher bitrates though.
basically guys, forget that stuff, usb->i2s is the way to go, i hardly ever use the other inputs except at audio meets.
Felipe, no, ground is part of the signal, whats the point of isolating only one? ground is the problem generally, not signal. connect both signal and ground
Ok I will isolate both: signal & ground.
After the pulse tx the cap is in series with the signal & the value is 0.01uF???
try both ways as i have seen it done both ways, but i have always done it isolating both. as i said mate, i dont use that part of the board, you have the bom. i dont have the board in front of me. you'll have to ask owen about the cap whether it goes on hot or ground, but either way its in series yes
Hi Felipe,
You will need the cap there to prevent the transformer from saturating (due to DC voltage). It should be a 0.01uf COG/NPO.
As for the resistance, it's up to you. Either 75 or 110 depending.
The only point in using the transformer is to provide galvanic isolation, so you must float both the signal and ground at the input of the transformer. If you connect either the signal or ground directly, you've eliminated the purpose of the transformer.
Let me know how it goes!
Cheers,
Owen
You will need the cap there to prevent the transformer from saturating (due to DC voltage). It should be a 0.01uf COG/NPO.
As for the resistance, it's up to you. Either 75 or 110 depending.
The only point in using the transformer is to provide galvanic isolation, so you must float both the signal and ground at the input of the transformer. If you connect either the signal or ground directly, you've eliminated the purpose of the transformer.
Let me know how it goes!
Cheers,
Owen
Hi Owen,
Now no dropouts, onlyt connected pulse tx with 75R terminated at the output tx.
I don't have on hand 0.01uF COG/NPO cap, I looked for at Mouser & only have 1pF COG/NPO.
I'm happy I can start my listening sessions.
Thanks again for your kind support Owen
Cheers,
Felipe
Now no dropouts, onlyt connected pulse tx with 75R terminated at the output tx.
I don't have on hand 0.01uF COG/NPO cap, I looked for at Mouser & only have 1pF COG/NPO.
I'm happy I can start my listening sessions.
Thanks again for your kind support Owen
Cheers,
Felipe
Zero Cool:
Sorry to leave you hanging on your question a few pages back... I got caught up trying to get Felipe's board up and running.
The final schematic and BOM is shown on the first page of this build thread. There is also a separate development thread that basically goes from start to finish in the development process. The initial schematic was basically a D1 with a B1 output buffer, but the buffer was found to be unnecessary, and the performance of the simple differential common-gate amplifier was optimized by picking the best FET and operating it at the ideal voltage and current. The result of that is the schematic you're seeing, which is about as simple as you can possibly get when it comes to the output of a DAC.
If you really want to know the entire story, read through the whole development thread here:
http://www.diyaudio.com/forums/digital-line-level/154866-new-take-classic-pass-labs-d1-ess-dac.html
Did you end up getting the board and kit from richluvsound?
Regards,
Owen
Sorry to leave you hanging on your question a few pages back... I got caught up trying to get Felipe's board up and running.
The final schematic and BOM is shown on the first page of this build thread. There is also a separate development thread that basically goes from start to finish in the development process. The initial schematic was basically a D1 with a B1 output buffer, but the buffer was found to be unnecessary, and the performance of the simple differential common-gate amplifier was optimized by picking the best FET and operating it at the ideal voltage and current. The result of that is the schematic you're seeing, which is about as simple as you can possibly get when it comes to the output of a DAC.
If you really want to know the entire story, read through the whole development thread here:
http://www.diyaudio.com/forums/digital-line-level/154866-new-take-classic-pass-labs-d1-ess-dac.html
Did you end up getting the board and kit from richluvsound?
Regards,
Owen
there are lots of 10nf np0 its a zero btw, not an O, but you might find they use the other name c0g also with a zero
thanks for the info on that cap btw opc, i hadnt realized it was for the dc, thought it was filter only
thanks for the info on that cap btw opc, i hadnt realized it was for the dc, thought it was filter only
Felipe
here is over 50 1206 size 10nf c0g/np0 caps at mouser
I prefer the TDK or murata
I find the 5% parts are a good middle ground, you dont need to go higher than that for this purpose, only for specific filtering purpose and cost can quadruple for 2%. if you dont need anything else at mouser, let me know, i'll pop one in the mail for you free of charge. better than spending too much at mouser for 1 item.
here is over 50 1206 size 10nf c0g/np0 caps at mouser
I prefer the TDK or murata
I find the 5% parts are a good middle ground, you dont need to go higher than that for this purpose, only for specific filtering purpose and cost can quadruple for 2%. if you dont need anything else at mouser, let me know, i'll pop one in the mail for you free of charge. better than spending too much at mouser for 1 item.
Zero Cool:
I just browsed through that thread (I suppose you could call it a trip down memory lane) and most of the good stuff happens from page 1 to page 22. I know reading all 75 pages is a little daunting, so focus on those. By page 22, the final schematic has been posted, and the rest is just board layout.
qusp:
You're correct about the 0/O thing... I'm just lazy and assume people know what I'm talking about.
That part is an 0805, but a quick search on DK shows at least two dozen parts in stock. It's certainly not a difficult cap to come by.
Those little high speed transformers are very sensitive to DC and will easily saturate. The cap helps prevent that and has no effect on the SPDIF signal.
Cheers,
Owen
I just browsed through that thread (I suppose you could call it a trip down memory lane) and most of the good stuff happens from page 1 to page 22. I know reading all 75 pages is a little daunting, so focus on those. By page 22, the final schematic has been posted, and the rest is just board layout.
qusp:
You're correct about the 0/O thing... I'm just lazy and assume people know what I'm talking about.
That part is an 0805, but a quick search on DK shows at least two dozen parts in stock. It's certainly not a difficult cap to come by.
Those little high speed transformers are very sensitive to DC and will easily saturate. The cap helps prevent that and has no effect on the SPDIF signal.
Cheers,
Owen
So I guess this isn't for my PCB? I'm confused... There's a spot on the PCB to install an 0805.
Great photos qusp! Looks like that is going to be a stellar setup, though I would not expect anything less from you. =)
Here's a real newbie question. I noticed you are using both a 4.7uF and 0.1uF film cap (in parallel I'm assuming?), for your coupling capacitors qusp. What is the benefit of using the 0.1uF cap as well?
Here's a real newbie question. I noticed you are using both a 4.7uF and 0.1uF film cap (in parallel I'm assuming?), for your coupling capacitors qusp. What is the benefit of using the 0.1uF cap as well?