Svart
Well-known member
ok another question.. Is it the general concensus that the IC is actually "good" or is it one of those flukes that sounds good but looks awful on paper?
Behringer ADA8000 pre bypass questions
- Thread starter asm
- Start date
Help Support GroupDIY Audio Forum:
asm
Well-known member
svart, id like to see your layout idea, ive never done one before, and its driving me nuts. even on something this simple.
SMD is fine, what about ease of solderbility? ive never done caps and resistors that small.
what about JLM's mod. looks to simple to be true? (if it works then your tha man, actually, your tha man besides that)
SMD is fine, what about ease of solderbility? ive never done caps and resistors that small.
what about JLM's mod. looks to simple to be true? (if it works then your tha man, actually, your tha man besides that)
Svart
Well-known member
though the Datasheet is vague I think he might be on to something, however there are always drawbacks to simplicity.
for the layout, do you want to try Bcarso's design? It might take me a few days due to workload and finding parts in the right packages..
SMD is easy. get yourself a good set of tweezers but not from the woman's drawer in the bathroom, they tend to get really really really mad when you do that.. also get a good set of tips for the iron and some really small solder. those three things can make SMD life great or simply Hell without them.
the only other thing is that the resistors are usually 5%. the small value caps can be NPO and panasonic FC caps are availiable in SMD versions too.. :green:
I would layout a simple one sided prototype just for the sake of seeing it work but if I were to do a real layout I would like to see it dual sided.. either way you could likely get all 8 channel inputs onto something the size of a business card easily, likely even smaller.
just some thoughts. :guinness:
for the layout, do you want to try Bcarso's design? It might take me a few days due to workload and finding parts in the right packages..
SMD is easy. get yourself a good set of tweezers but not from the woman's drawer in the bathroom, they tend to get really really really mad when you do that.. also get a good set of tips for the iron and some really small solder. those three things can make SMD life great or simply Hell without them.
the only other thing is that the resistors are usually 5%. the small value caps can be NPO and panasonic FC caps are availiable in SMD versions too.. :green:
I would layout a simple one sided prototype just for the sake of seeing it work but if I were to do a real layout I would like to see it dual sided.. either way you could likely get all 8 channel inputs onto something the size of a business card easily, likely even smaller.
just some thoughts. :guinness:
asm
Well-known member
actually, if we allready knew the parts (which look to be all common values) in normal size footprints, might be best to do a prototype board on that. the size might be a bit bigger, but if a few people want these boards then they could stuff them up too normally w/o any problems.
a business size card would be very cool if this layout works :razz:
just an idea.
a business size card would be very cool if this layout works :razz:
just an idea.
bcarso
Well-known member
The main feature of the suggested buffer is to intrinsically limit the signal swings so as to not exceed +5V, or go below ground, on the A/D, since many such parts will go into latchup under these conditions and self-destruct. Also, it provides +6dB of gain, owing to the complementary outputs, which may or may not be what you want.
Wavefront is not clear in their datasheet about the consequences of exceeding the rails. But their "single-ended input conditioning circuit" shows unspecified opamps. These could I suppose be rail-to-rail output swing low voltage parts run from +/-2.5V, but of course the B*hringer schematic from gyraf shows high voltage amps. Maybe the Wavefront part has protection against latchup. It is frankly odd, and suspicious, that there is no Absolute Max ratings table.
Note that even with r-to-r limited to 5V a coupling cap can produce voltage excursions with low-frequency square wave drive above and below the rails of the 1101. The buffer will do this too without the protection diodes. But that would be an extreme signal.
Certainly you can take advantage of the diff. inputs of the 1101 and drive them directly, if your source Z is reasonably low. The spec on input Z for the 1101 is 160k differential, and each input is internally biased to half Vcc, i.e., 2.5V.
I'm suspicious as well of the common A and D ground claim, although it is more feasible given their differential inputs.
The details of the clocking requirements are given in the datasheet ("2s complement serial data in a standard MSB-first format.")
Wavefront is not clear in their datasheet about the consequences of exceeding the rails. But their "single-ended input conditioning circuit" shows unspecified opamps. These could I suppose be rail-to-rail output swing low voltage parts run from +/-2.5V, but of course the B*hringer schematic from gyraf shows high voltage amps. Maybe the Wavefront part has protection against latchup. It is frankly odd, and suspicious, that there is no Absolute Max ratings table.
Note that even with r-to-r limited to 5V a coupling cap can produce voltage excursions with low-frequency square wave drive above and below the rails of the 1101. The buffer will do this too without the protection diodes. But that would be an extreme signal.
Certainly you can take advantage of the diff. inputs of the 1101 and drive them directly, if your source Z is reasonably low. The spec on input Z for the 1101 is 160k differential, and each input is internally biased to half Vcc, i.e., 2.5V.
I'm suspicious as well of the common A and D ground claim, although it is more feasible given their differential inputs.
The details of the clocking requirements are given in the datasheet ("2s complement serial data in a standard MSB-first format.")
bcarso
Well-known member
PS: 1% resistors are readily available in SMD.
bcarso
Well-known member
Also re JLM's remark: the request was for a single-ended input.
Svart
Well-known member
"PS: 1% resistors are readily available in SMD."
yeah, sorry, you are right, I keep forgetting that since I have nothing but drawers full of 5% resistors to work with. on another note though, I've yet to run into this "noise" that people get from using anything other than 1% metal films...
I can still lay this out and people can use whatever they want though..
I'll do a simple layout in the next few days, however I don't have one of these B*hringer units so I'll just leave header pads for wiring the rails and the I/O fair enough?
:thumb:
yeah, sorry, you are right, I keep forgetting that since I have nothing but drawers full of 5% resistors to work with. on another note though, I've yet to run into this "noise" that people get from using anything other than 1% metal films...
I can still lay this out and people can use whatever they want though..
I'll do a simple layout in the next few days, however I don't have one of these B*hringer units so I'll just leave header pads for wiring the rails and the I/O fair enough?
:thumb:
Mbira
Well-known member
Svart,
How much could we bribe you with to make a non-smd layout? As a home-etcher with my own drawers full of thru-hole stuff, smd is a nightmare. I'm probably not alone on that
. Not trying to take a mile or anything! :green:
Joel
How much could we bribe you with to make a non-smd layout? As a home-etcher with my own drawers full of thru-hole stuff, smd is a nightmare. I'm probably not alone on that
. Not trying to take a mile or anything! :green: Joel
asm
Well-known member
yeh, i agree. even though smd is alot more professional looking, it makes it a bit harder. especially for us beginners.
i guess the layout really whouldnt change much, same path, just different pad sizes and spacings.
i guess the layout really whouldnt change much, same path, just different pad sizes and spacings.
bcarso
Well-known member
Of course the ADC and DAC are only SMD, so you need the thing to be at least partially SMD.
Kev
Well-known member
[quote author="bcarso"]Also re JLM's remark: the request was for a single-ended input.[/quote]
just use pins 2 and 3
just use pins 2 and 3
Svart
Well-known member
so where do we want to start and stop the PCB? do want a simple 2 pin header so you can plug in what/where you want? Do you want to include the DAC/ADC on the PCB or do you want to only have the input section only on the pcb and just tie into the inputs on the DAC with small wires? I mean in truth we could so as much as layout a whole new board and copy the MCU code (if it has one..) and reprogram.. or just a tiny input section to try out.
let me know what everyone wants and i'll get on it.
:thumb:
let me know what everyone wants and i'll get on it.
:thumb:
Svart u seem well up on all this stuff :thumb:. I've been going in circles with all things ad/da for ages and still don't have anything to show yet, but am learning bit-by-bit...
As you mentioned MCU - would u know about implementing s/mux so higher sample rates can be acheived with adat? Perhaps not with these alesis chips (not sure they support sr's greater then 48khz) but with say some of the TI/akm chips out there...
Cheers :guinness:
As you mentioned MCU - would u know about implementing s/mux so higher sample rates can be acheived with adat? Perhaps not with these alesis chips (not sure they support sr's greater then 48khz) but with say some of the TI/akm chips out there...
Cheers :guinness:
Svart
Well-known member
I'm no digital engineer but some chips just need to see a higher clock rate. as for the other stuff, layout is not a specific science, just gotta be careful of how you run your traces.
I'm going to do a layout of Bcarso's input section first, then we can move on from there if so desired.
:thumb:
I'm going to do a layout of Bcarso's input section first, then we can move on from there if so desired.
:thumb:
asm
Well-known member
svart,
awesome man.
i think the best bet, and easiest, would just have header pads on input and out. then we can wire a switch or something, so if we want to use the original line going thru the pre we can.
awesome man.
i think the best bet, and easiest, would just have header pads on input and out. then we can wire a switch or something, so if we want to use the original line going thru the pre we can.
Svart
Well-known member
ok got a quick schemo, netlist and board done in circuitmaker for one channel. let me know where i can send it.
I just looked over it quickly and *think* that it's ok but haven't etched one to check. please look over it and see if there are any issues. I didn't include the protection diodes due to needing another wire to run to the logic rail, I figure if we need it I can add it later.
I used through hole parts and larger pads for everyone. the layout is fairly tight and I can do other things like laying pads to stand the resistors up and using smaller footprint caps for a smaller board.
and again this was just a quick and dirty board to get people prototyping quickly.
opinions are greatly appreciated especially if we decide to to a multi channel version.
:thumb:
I just looked over it quickly and *think* that it's ok but haven't etched one to check. please look over it and see if there are any issues. I didn't include the protection diodes due to needing another wire to run to the logic rail, I figure if we need it I can add it later.
I used through hole parts and larger pads for everyone. the layout is fairly tight and I can do other things like laying pads to stand the resistors up and using smaller footprint caps for a smaller board.
and again this was just a quick and dirty board to get people prototyping quickly.
opinions are greatly appreciated especially if we decide to to a multi channel version.
:thumb:
bcarso
Well-known member
But...you know, much as I would like to see the circuit function, unless you really need the 6dB Kev and JLM are right---if the diff front end of the ADC is high enough performance you could manage just running in directly. For single-ended sources more-or-less referenced already to the same ground system you ought to be ok.
For quasi-floating sources with significant mains-coupled current, like a source powered by a wallwart with tens of pF of interwinding capacitance, you might want to provide for a moderate value R from the cold side of the input to system ground, so you aren't asking the front end of the 1101 to reject large common-mode voltages. The impedance differentially is 160k, which probably means that there are 80k Rs to each input from the internal half-supply. At low frequencies, if mains coupling was pure sinusoid, 80k would be enough to keep you within the range. If there is a lot of hash at much higher frequencies, that coupled with the likely falloff in common-mode rejection might lead to trouble.
I would definitely use the clamp diodes from each input to +5 and ground though in this case, not knowing what kind of stuff might be hitting the frnt end. Also, although through hole is generally fine it apparently helps per Wavefront's notes to put the lowpass differential cap as close to the package as possible, and this would be facilitated by an SMD device there.
For quasi-floating sources with significant mains-coupled current, like a source powered by a wallwart with tens of pF of interwinding capacitance, you might want to provide for a moderate value R from the cold side of the input to system ground, so you aren't asking the front end of the 1101 to reject large common-mode voltages. The impedance differentially is 160k, which probably means that there are 80k Rs to each input from the internal half-supply. At low frequencies, if mains coupling was pure sinusoid, 80k would be enough to keep you within the range. If there is a lot of hash at much higher frequencies, that coupled with the likely falloff in common-mode rejection might lead to trouble.
I would definitely use the clamp diodes from each input to +5 and ground though in this case, not knowing what kind of stuff might be hitting the frnt end. Also, although through hole is generally fine it apparently helps per Wavefront's notes to put the lowpass differential cap as close to the package as possible, and this would be facilitated by an SMD device there.
Svart
Well-known member
or soldering a cap right to the pins.. done that many times. :green:
It's interesting to say the least that the adc can take such a high input signal. oh well, at least I spent a slow rainyday at work doing things for the united common front of DIY rebellion!
or at least not falling asleep.. :thumb:
It's interesting to say the least that the adc can take such a high input signal. oh well, at least I spent a slow rainyday at work doing things for the united common front of DIY rebellion!
or at least not falling asleep.. :thumb:
bcarso
Well-known member
"oh well, at least I spent a slow rainyday at work doing things for the united common front of DIY rebellion!"
And I finally contributed a circuit with pretty good performance out of two transistors per channel. Maybe it is useful somewhere.
BTW the claim to fame of the Wavefront, emblazoned across their datasheet ("ClockEZ ADC') is that the internal PLL and clock generation circuitry makes it easier to use since it eliminates a lot of high frequency stuff running all over the board. Perhaps this is another reason why they insist a single ground plane suffices.
I may give these folks a call to get some additional answers, especially what really happens to the device when you overdrive the inputs.
And I finally contributed a circuit with pretty good performance out of two transistors per channel. Maybe it is useful somewhere.
BTW the claim to fame of the Wavefront, emblazoned across their datasheet ("ClockEZ ADC') is that the internal PLL and clock generation circuitry makes it easier to use since it eliminates a lot of high frequency stuff running all over the board. Perhaps this is another reason why they insist a single ground plane suffices.
I may give these folks a call to get some additional answers, especially what really happens to the device when you overdrive the inputs.
