For the record:
Yeah, I also get annoyed by guys who want us to do all their design work for them. Help is one thing. Expecting us to do all the work is another. Sometimes you will have to take our comments and seek out the end point in the direction we are steering you in. You know the old saying about teaching a man to fish...............
Sometimes the trick is to figure out which tangent to follow: not all of them will converge.
Jocko
Yeah, I also get annoyed by guys who want us to do all their design work for them. Help is one thing. Expecting us to do all the work is another. Sometimes you will have to take our comments and seek out the end point in the direction we are steering you in. You know the old saying about teaching a man to fish...............
Sometimes the trick is to figure out which tangent to follow: not all of them will converge.
Jocko
Ulas said:
This one was of course not a typo. It was a genuine error.
Some people function and work in different way then you are. Did you really lost the faith in other people to think that there is a hidden agenda behind everuthing other people do.
Yes of course it is better but what more can I do than thank you like I did in posts #73 #77 #91. No one here questioned your experitze and we would (are) look forward to them if your manners were better. Do you think you would be in a good health talking to someone in person and telling him he is a "moron"? I do not think so. Maybe a "moron" has changed a menaing since I learned english and has a new slang meaning. The point I wanted to make is nobody here was ever questioning your desire to help, but the way of doing it.
There you did it again. Raise your manners and respect towards other people to the level of your digital design knowledge. I like to think, more than you can imagine. I admitted here that analog is more interesting for me, that is why we approached like we approached this desigh. I wanted, cann't speak for Kim, to do a modular DAC design, reusing other people work and with help from others. I was couple of times thrilled here that several threads started with the thing closed to what I wanted. But they all have failed. So with Kim we made a thread. Look at the title, we are asking for help. That is why we started. I thought that diyaudio.com exists for the things like this. This is my hobby not my life.
Well believe it or not I do belive you. If you wan't, do the schematics and layout and I buy a PCB from you if it is not a sky high price.
Well I expendid the effort. When you pointed out fot the counter problem I also started to look for the other solution. Other people helped but as you can see I have found '161 and '163. Maybe not much in your eyes.
This is typo I would like to use VHC74 component. You must understand that I am using a protel for the design and trying to reuse a component symbol and layout as ever possible. At the end I will change the names so that everuthing looks nice and as it should be. I used flipflops for the inverter since I didnt want to use a whole chip just for one inverter. After I started to dig and found out that there is whole bunch of small inverters that I did not know of. Also I found out that vhc components are also not easy to come by so I might change this as well...
Happy Christmass to You too.
My remark is aimed at guys who want the "experts" and "pseudo-experts" to give them enough info for them to make a working unit that they can turn into a commercial venture. You said you aren't........so no problem.
(Actually, I haven't been paying much attention to what you are building. But my inbox gets flooded everytime someone make a horse's butt of himself.)
But some guys just like to complain and throw stones. Waiting for one of them to make you a PCB..........well, you get the idea.
Jocko
(Actually, I haven't been paying much attention to what you are building. But my inbox gets flooded everytime someone make a horse's butt of himself.)
But some guys just like to complain and throw stones. Waiting for one of them to make you a PCB..........well, you get the idea.
Jocko
Replacing the flipflop:
I've used a flipflop 74vhc74, as I mentioned, since I did not wanted to use just one inverter from hex inverter that I was familiar with. I was last couple of days "fooling" with the layout and cann't find satisfactory way to route this part.
I've looked around and found that there is a number of small inverters in SOT-23 package, small 5pin package containig just one of them. The names are unkown to me so I wonder if someone has an experinece with these parts, like what is more of an standard and recomended to use. My candidates from Farnell catalogue are
NC7SZ04M5 Fairchild (5ns propagation)
SN74AHCT1G04DBVR TI and Philips (around 10ns)
First one is far superior part and I wonder if it is true. 10ns is 100MHz operation so it should be fine. Both have the same pinning so they should be interchangable.
Question for Ulas (if you are still willing to answer): I read that you double the maximal AC characteristics stated from the datasheet. Is it from the experience or is just putting a big margin to be on the safe side?
Pred
PS
Still not sure if I should invert LRCK signal. Searched everywhere, could not find a thing. Any pointers for usefull reading?
I've used a flipflop 74vhc74, as I mentioned, since I did not wanted to use just one inverter from hex inverter that I was familiar with. I was last couple of days "fooling" with the layout and cann't find satisfactory way to route this part.
I've looked around and found that there is a number of small inverters in SOT-23 package, small 5pin package containig just one of them. The names are unkown to me so I wonder if someone has an experinece with these parts, like what is more of an standard and recomended to use. My candidates from Farnell catalogue are
NC7SZ04M5 Fairchild (5ns propagation)
SN74AHCT1G04DBVR TI and Philips (around 10ns)
First one is far superior part and I wonder if it is true. 10ns is 100MHz operation so it should be fine. Both have the same pinning so they should be interchangable.
Question for Ulas (if you are still willing to answer): I read that you double the maximal AC characteristics stated from the datasheet. Is it from the experience or is just putting a big margin to be on the safe side?
Pred
PS
Still not sure if I should invert LRCK signal. Searched everywhere, could not find a thing. Any pointers for usefull reading?
rfbrw said:
I want to use I2S format. Someone told me I should invert the signal when I generate them for the 192kHz operation. I still do not get why?
Anyway, I am sending a pcb layout. This is a first draft and just for those interested, to get an idea. There is a lot of things to clean and compact. I am open for comments, sugestions and whishes. For easy following left is CS, in the middle on top are the oscilators, on the bottom AD and at the output ADCs.
pred said:
Then follow the rules of the I2S bus format.
http://www.semiconductors.philips.com/acrobat_download/various/I2SBUS.pdf
Having a glance over just the AD1896 part of the schematic...
- Do you know why you're using 22uF 'compensation' capacitors on the TL431 regulator circuits? None of the literature I have for that part from TI or Linear describes adding a capacitor here, and curves from both companies indicate that the part will be stable without it. I'd also add a shunt C between the ferrite bead and the series resistor.
Also, make sure you calculate what the thermal dissipation in the series R and the TL431 itself will be. TO-92 packages can't dissipate a lot, and if the series R stays at 10 ohms I think you'll have thermal problems.
- Why use an OSCON for a reset time delay? and also, what happens when you remove power from the circuit and the AD1896 /RESET voltage - held by that OSCON - goes higher than VCC_IO? You might want to use an actual power on reset chip, they're a lot safer.
- Add ceramic decoupling caps to your I2S generation and retiming circuitry - OSCONs are low ESR, but not low impedance at 27MHz + harmonics. And again, I'd toss the 7416x/7474 combination and use a single VHC393 chip - with the two halves cascaded and clocked with 49.152 or 24.576 MHz, it will produce proper I2S clock and frame sync signals. I have used this chip for this application in professional practice.
- Why are there voltage dividers on the /EN input of your oscillators? enable/disable control?
- Decide on which digital audio format you're using on the AD1896 chip output and the PCM179x input, stick with it, and lose the configuration jumpers. It'll save board space and help with a nice layout.
And finally, don't get ahead of yourself - make sure your schematic is correct before jumping into PCB design. Ulas was correct on one thing; read the datasheets for everything you're using and make sure you're operating it within spec. Let me re-iterate this in a nice way.
- Do you know why you're using 22uF 'compensation' capacitors on the TL431 regulator circuits? None of the literature I have for that part from TI or Linear describes adding a capacitor here, and curves from both companies indicate that the part will be stable without it. I'd also add a shunt C between the ferrite bead and the series resistor.
Also, make sure you calculate what the thermal dissipation in the series R and the TL431 itself will be. TO-92 packages can't dissipate a lot, and if the series R stays at 10 ohms I think you'll have thermal problems.
- Why use an OSCON for a reset time delay? and also, what happens when you remove power from the circuit and the AD1896 /RESET voltage - held by that OSCON - goes higher than VCC_IO? You might want to use an actual power on reset chip, they're a lot safer.
- Add ceramic decoupling caps to your I2S generation and retiming circuitry - OSCONs are low ESR, but not low impedance at 27MHz + harmonics. And again, I'd toss the 7416x/7474 combination and use a single VHC393 chip - with the two halves cascaded and clocked with 49.152 or 24.576 MHz, it will produce proper I2S clock and frame sync signals. I have used this chip for this application in professional practice.
- Why are there voltage dividers on the /EN input of your oscillators? enable/disable control?
- Decide on which digital audio format you're using on the AD1896 chip output and the PCM179x input, stick with it, and lose the configuration jumpers. It'll save board space and help with a nice layout.
And finally, don't get ahead of yourself - make sure your schematic is correct before jumping into PCB design. Ulas was correct on one thing; read the datasheets for everything you're using and make sure you're operating it within spec. Let me re-iterate this in a nice way.
gmarsh said:
First thanks a lot for an indepth help and suggestions
The reason is to reduce the AC gain and therefore decrease noise and increase Power supply rejection. For the 3.3V it is around 3dB more in PSR and change of 180 to 140nV/rtHz in LT spice simulator. It is an option, some people use it some not, won't hurt to leave the space for it
For TLs there is no problem since we will calculate the current (that we use as a supply) and leave like 5-10mA flowing trough TL. Resistor is another matter and you are right. Guido sugested it will be nice to put the resistor. Since it needs to be a big component I will put it on the PS board.
For this one I do not have a reason, we started like that and it sticked. I look at the reset parts and they are so small that it is a better solution than capacitor. I will put something like STM810TWX6F from STM. Maxim parts are soooo more expensive...
This is an omition, an overlook. Thanks for seeing it.
I already have some '161 but I will look into it. I maybe need to order some parts soon and will see what can I do about this one. Since you are here. I think it is you that sugested in thread "AD1896 in slave mode how" that the SCLK has to be inverted to create a proper phase relation. I saw while looking that I did a mistake by using your post wrongly and inverted the LRCLK. I saw in datasheet that the change in LRCK is always on falling edge of the SCLK that agrees with your proposition. On the other hand I did not see that in the AD1896 eval board pdf or elsewhere. Is it necessary? I will go in detail through I2S specification...
I can swear I saw some oscillators that enable high and others are on low so I wanted to have both but not to use a jumper, or one of the resistors is like a jumper low oms one.
We decided for the I2S all the way through. The jumpers that stayed are for choosing the slave or master mode for the AD and for mono/stereo slow/fast filter on the DAC. You know, I always say I will experiment on the equipment I make and than when I see that I have to resolder and desolder I kinda am put off buy it. Except some times I tested some capacitors... For this one I really decided I will go changing clocks, mono stereo, I/V etc. I have one 4xTDA1543 made and I like it, so it will be always there if I want to listen, until I get tired playing with this one and put it as a master DAC
And finally, don't get ahead of yourself - make sure your schematic is correct before jumping into PCB design. Ulas was correct on one thing; read the datasheets for everything you're using and make sure you're operating it within spec. Let me re-iterate this in a nice way.
I really did not have an idea how big it will all be and I was curious. I have some trouble sleeping lately, maybe it is holiday or maybe it is Ulas, so I sat one night and played with layout knowing that I will redo it later. I hate routing of clock signals that I did and sence I will have problems if I have so much vias for 24MHz signals....
I will go and read and come back in a couple of weeks with hopefully better circuit and design. I'll be still reading the thred to check if someone has some more comments.
Thank you again for very helpfull suggestions!
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.