Jim Brown (who is very much into audio) writes about Henry W. Ott:
The author, and our entire industry, are greatly indebted to Henry Ott, whose excellent workshop on EMC I attended in 2004, and whose book, cited above, should be on everyone’s shelf. It would be difficult to conceive of an issue related to EMC that Henry hasn’t thought through in great detail, considering every possible ramification, from the micro to the macro, and from circuit performance to manufacturability to user-friendliness to where things will be in 20 years. A side-comment in his book (first published in 1976, updated in 1988 and again in 2009) makes it clear that he knew about
SCIN at least 20 years ago.
&
An absolutely essential book. Henry Ott nailed it,in this definitive text that ties together both theory and practice in EMC.
If you disagree with a single word in this book, you’re wrong!
The author, and our entire industry, are greatly indebted to Henry Ott, whose excellent workshop on EMC I attended in 2004, and whose book, cited above, should be on everyone’s shelf. It would be difficult to conceive of an issue related to EMC that Henry hasn’t thought through in great detail, considering every possible ramification, from the micro to the macro, and from circuit performance to manufacturability to user-friendliness to where things will be in 20 years. A side-comment in his book (first published in 1976, updated in 1988 and again in 2009) makes it clear that he knew about
SCIN at least 20 years ago.
&
An absolutely essential book. Henry Ott nailed it,in this definitive text that ties together both theory and practice in EMC.
If you disagree with a single word in this book, you’re wrong!
Headphone output ?
Hi,
I have a question about changing the EVM circuit to use a current feedback op amp, the LME49713HA. It has +/-100mA output and can "easily drive 150R loads" and it seems to have pretty low distortion and noise too, and is able to drive some capacitance. So I think it should be able to drive headphones like Senn HD650 directly, if I add some gain.
I am planning to remove C1 and add a 600z ferrite in series after R1, or I could keep C1 and cut the trace into the neg op amp input and put the ferrite there. This follows Ti's advice about using cfb op amps, that Abrax posted in the es9018 thread. So, without having sim software, which one seems like the best idea ? What about C2 ? I assume it can stay since it is not in the neg f'back loop so it won't cause instability.
Then I need to change the resistor values for some gain, say +3, so change R2 and R3.
I was also thinking of reducing R5 to see if it would improve dynamics. Did I understand this right ?
Hi,
I have a question about changing the EVM circuit to use a current feedback op amp, the LME49713HA. It has +/-100mA output and can "easily drive 150R loads" and it seems to have pretty low distortion and noise too, and is able to drive some capacitance. So I think it should be able to drive headphones like Senn HD650 directly, if I add some gain.
I am planning to remove C1 and add a 600z ferrite in series after R1, or I could keep C1 and cut the trace into the neg op amp input and put the ferrite there. This follows Ti's advice about using cfb op amps, that Abrax posted in the es9018 thread. So, without having sim software, which one seems like the best idea ? What about C2 ? I assume it can stay since it is not in the neg f'back loop so it won't cause instability.
Then I need to change the resistor values for some gain, say +3, so change R2 and R3.
I was also thinking of reducing R5 to see if it would improve dynamics. Did I understand this right ?
Attachments
You got me, I wasn't aware that it was - after all I don't have a groundplane, rather I use star earthing.
Yep, I agree. This is EMC101.
I agree - do you think I have a problem then?
This tends to be what happens in digital systems - the common-mode currents meander through the plane in ways that the designer hasn't thought about. I could point to any number of examples - one that comes to mind is the Metrum Octave which I think I commented on in a thread last year. The guys who know how to do this right at those at Berkeley Audio Design. Take a look at the Alpha DAC's internal PCB.
Well the requirements are different - in instrumentation the aim isn't to please the ears. But I'm more than willing to be shown wrong - it would result in even better SQ.
@marce - its not really that my views are whacky, rather that my practices are. I just report here what I do, what works for me.
Last edited:
I cannot get this DAC to play music.
I have checked all power points, signal connections, data connections, etc. There are no shorts and all connections are good, with all the right pins getting the voltages, gnd, signal, etc they need. I can get it to buzz, hum, crackle, pop, etc but not decode pcm.
I have all the dif0-2 high for 32 bit I2S (which is my source, tested on other DACs too) and I have tried tying dif2 low for 24 bit i2s. I have also tried tying dem1 low to set de-emphasis to off, just in case.
There is no dc offset or any ACV on the output either.
It seems to me that it is not an analgoue problem because I an get continuous noise/hum by putting a finger on the wire near the dac output. The op amaps and circuitry are working.
With the op amps on, if I turn off the DACs power and then on again, I can get a thump.
So, I think it must be the i2s. I checked the board has been labelled correctly, and i checked and re-checked my source and the connecting cable.
There doesn't seem to be anything left to check, so that's why I'm posting this. My first failure.
I have checked all power points, signal connections, data connections, etc. There are no shorts and all connections are good, with all the right pins getting the voltages, gnd, signal, etc they need. I can get it to buzz, hum, crackle, pop, etc but not decode pcm.
I have all the dif0-2 high for 32 bit I2S (which is my source, tested on other DACs too) and I have tried tying dif2 low for 24 bit i2s. I have also tried tying dem1 low to set de-emphasis to off, just in case.
There is no dc offset or any ACV on the output either.
It seems to me that it is not an analgoue problem because I an get continuous noise/hum by putting a finger on the wire near the dac output. The op amaps and circuitry are working.
With the op amps on, if I turn off the DACs power and then on again, I can get a thump.
So, I think it must be the i2s. I checked the board has been labelled correctly, and i checked and re-checked my source and the connecting cable.
There doesn't seem to be anything left to check, so that's why I'm posting this. My first failure.
Wah, very cool - great skills ! What size are they ? Only those four pins ? Any tips for me about building this ? Thanks.
and Kung Hei Fat Choi !
1) zero distance decouple cap is a must, without it, my ear will complain in 5 minutes.
2) use solder to fill all via
3) direct solder the opamp to the PCB,
4) use fet/jfet input opamp/ opa627/ad8610br etc.
Diyer can try one by one and should find the sound is improving more and more
1) check the 10uf and 20k near the reset, and 20k near the mute is all soldered ok.
2) try to short the reset after power on, if it sound, your 10uf and 20k near the reset should have problem.
3) check the fb is soldered
4) no need to install the switch SW1 if connected from our cm6631/a pcb, just make sure all 20k is soldered
1) all good - 5V accross 10uF cap
2) no sound/pop/etc, actually cuts noise when noise is present - Noise is sometimes present when the i2s gnd is disconnected, and very rarely there is distorted music hidden very quiet under the noise.
3) already checked and re-checked
4) had to try - tried everything.
The i2s sources are cm6631 and wolfson wm8805. both the same problem.
What voltage should there be on vcml and vcmr ? I have 0V. All power supply pins are correct.
cheers
2) no sound/pop/etc, actually cuts noise when noise is present - Noise is sometimes present when the i2s gnd is disconnected, and very rarely there is distorted music hidden very quiet under the noise.
3) already checked and re-checked
4) had to try - tried everything.
The i2s sources are cm6631 and wolfson wm8805. both the same problem.
What voltage should there be on vcml and vcmr ? I have 0V. All power supply pins are correct.
cheers
Last edited:
Anyone know what vcml/vcmr should be ? The datasheet says "normally connected to vss" and it says that vss1-4 are connected to gnd, so vcml/r should be 0V ? There's only one reference to it that I can find.
The fact I can get noise of various kinds indicates to me that the analogue stage is working, and that it is the digital side that has a fault. Nonetheless, I've been over everything on this DAC several times now and found no fault.
I've spent over 10 hours debugging it and I can do no more. I have no clue why it doesn't work. There are hi-res photos on my blog - better than this site will allow - just click on the photo. www.myl8test.wordpress.com/audio/ak4399-dac/ The photo looks like pin1 and 2 are bridged with solder but actually that is not the case - funny angle or lighting or something.
The fact I can get noise of various kinds indicates to me that the analogue stage is working, and that it is the digital side that has a fault. Nonetheless, I've been over everything on this DAC several times now and found no fault.
I've spent over 10 hours debugging it and I can do no more. I have no clue why it doesn't work. There are hi-res photos on my blog - better than this site will allow - just click on the photo. www.myl8test.wordpress.com/audio/ak4399-dac/ The photo looks like pin1 and 2 are bridged with solder but actually that is not the case - funny angle or lighting or something.
Last edited:
Hi,
It's working - what a relief.
Problem 1 - my CM6631 i2s has stopped working (but the spdif still works ?!). Something else to fix.
Problem 2 - the wm8805 only works if the AKM board is reset every time. I did not try the reset with this connected before. I have dif2 low for this too.
....so many hours wasted because of these two problems, but the AKM works. Thank you for your help ! It is very much appreciated.
The Lme47710ha sound horrible. LT1028 are much better but the treble has thd. OpA627 are "normal". I'll try more another time.
I had fitted 220R instead of 620R. This increases gain in that stage to ~3.5 and the output is high enough for hd650 at an excessive level.
Thanks again for your help !
It's working - what a relief.
Problem 1 - my CM6631 i2s has stopped working (but the spdif still works ?!). Something else to fix.
Problem 2 - the wm8805 only works if the AKM board is reset every time. I did not try the reset with this connected before. I have dif2 low for this too.
....so many hours wasted because of these two problems, but the AKM works. Thank you for your help ! It is very much appreciated.
The Lme47710ha sound horrible. LT1028 are much better but the treble has thd. OpA627 are "normal". I'll try more another time.
I had fitted 220R instead of 620R. This increases gain in that stage to ~3.5 and the output is high enough for hd650 at an excessive level.
Thanks again for your help !
PS My best effort at local decoupling. Only the digital is on the pin but the rest are quite close. Is it worth adding 100nF smd across the 10uF on vcml/vcmr ?
An externally hosted image should be here but it was not working when we last tested it.
Hi,
This DAC requires I2S or MSB or LSB or DSD input. I2S is the most common. If you want usb only, then I can recommend this as the best value for money (from the same seller, and no, I don't know him :
24bit 192kHz USB to I2S / SPDIF PCB CM6631 with ultralow noise 6.5uV regulator | eBay
and if you want both then this
Multi-input/converter module for DAC 8805X-J | eBay
although raindrop_Hui sells some real crap, its not too bad.
This DAC requires I2S or MSB or LSB or DSD input. I2S is the most common. If you want usb only, then I can recommend this as the best value for money (from the same seller, and no, I don't know him
:24bit 192kHz USB to I2S / SPDIF PCB CM6631 with ultralow noise 6.5uV regulator | eBay
and if you want both then this
Multi-input/converter module for DAC 8805X-J | eBay
although raindrop_Hui sells some real crap, its not too bad.
The old reset trick!
I am glad you had it working. Thanks for being the pioneer on these diyinhk boards
BTW, I checked your blog. Good start.
I am glad you had it working. Thanks for being the pioneer on these diyinhk boards
BTW, I checked your blog. Good start.
Haha... yeah, that old trick was new to me ! Every DAC I've played with just started playing music as soon as it got the i2s connected.
I found my cat6 cable is the fault, not the cm6631. I've never had a cat cable fail on me before either.
Also, with the cm6631, no reset is necessary, it works immediately, so only the wolfson wants a reset. I'll have to look into why.
The DAC sounds good, but as Diyinhk says, the mids can be a little strident with the wrong op amps. So, op amp rolling once again. Since I'm going to use this to drive headphones, I have LM7321 in the second stage, and at the mo, I'm listening to ad8597 in the first stage. The trick is to get sets that suit the headphones - HD650 and PXC250 - rather than amps and speakers.
Thanks for the kind words about the blog - man, its time consuming getting it set up but now it is easy enough to navigate and add extra stuff too so, yeah... a start ! I'm now even more impressed with yours - that's got to have taken ages to get it looking that good.
cheers
I found my cat6 cable is the fault, not the cm6631. I've never had a cat cable fail on me before either.
Also, with the cm6631, no reset is necessary, it works immediately, so only the wolfson wants a reset. I'll have to look into why.
The DAC sounds good, but as Diyinhk says, the mids can be a little strident with the wrong op amps. So, op amp rolling once again. Since I'm going to use this to drive headphones, I have LM7321 in the second stage, and at the mo, I'm listening to ad8597 in the first stage. The trick is to get sets that suit the headphones - HD650 and PXC250 - rather than amps and speakers.
Thanks for the kind words about the blog - man, its time consuming getting it set up but now it is easy enough to navigate and add extra stuff too so, yeah... a start ! I'm now even more impressed with yours - that's got to have taken ages to get it looking that good.
cheers
Another update is on the blog, but briefly, I tried adding a ferrite and using cfb. It either didn't work (oscillation) or worked but didn't sound as good as LM7321. I use opa827 in the first stage, but NE5534 are surprisingly nice too, and soooo cheap. This could be a real bargain DAC with the right choice of parts.
The last thing to do is solder the vias, and add a receiver. I found an SRC4392 so I thought I'd give it a try. It looks mediocre but the output I2S lines haven't been messed with so it could sound okay.
An externally hosted image should be here but it was not working when we last tested it.
The last thing to do is solder the vias, and add a receiver. I found an SRC4392 so I thought I'd give it a try. It looks mediocre but the output I2S lines haven't been messed with so it could sound okay.
- Status
- Not open for further replies.