Hi,
I am wondering if someone can help me with a Second order low pass filer using transistor for PCM1704 DAC output. Will a passive filter work better than an active one?
I am wondering if someone can help me with a Second order low pass filer using transistor for PCM1704 DAC output. Will a passive filter work better than an active one?
Yes a passive filter will generally sound better than an active one as it won't be prone to IMD. Why do you want only 2nd order and what function is the transistor playing?
I am not sure if it need to be one with transistor ( or active) but no OPAMP based on is being considered.
There is no particular reason for selecting second order. That is what being used in other DAC models, active though.
There is no particular reason for selecting second order. That is what being used in other DAC models, active though.
I'd suggest going for a higher order filter if you're going to use the PCM1704 in NOS. I use 11th order currently (5 inductors, 6 caps per channel). If you want to run OS then you can use fewer components in the filter as it doesn't need to be as steep. However I reckon NOS and a steep filter sounds best.
I am using it as NOS configuration, I would appreciate if you please share the filter schematics/ details. Also curious to know how you made those inductors.I guess the filter is a Bessel-Butterworth?
Last edited:
Hi,
You will certainly need golden ears to hear the effect of an 11th order
supersonic filter over one that simply does the the job required on a DAC
output, which is typically 2nd order LP op-amp based I/V conversion.
Assuming NOS means no output stage and OS means output stage,
how on earth do you come up with a 11th order filter as sensible ?
rgds, sreten.
You will certainly need golden ears to hear the effect of an 11th order
supersonic filter over one that simply does the the job required on a DAC
output, which is typically 2nd order LP op-amp based I/V conversion.
Assuming NOS means no output stage and OS means output stage,
how on earth do you come up with a 11th order filter as sensible ?
rgds, sreten.
Last edited:
This is my DAC design, I do not want to make it too complicated. So any idea is welcome as long as it is practical and does not involve extreme measures. Sreten I have the I/V conversion taken care of from PCM datasheet. Not sure if I need a low pass filter or not after IV. A transistor based Low pass filter is the one I am looking for. Second or third order perhaps. But not sure which way to go
Attachments
Last edited:
Its on my blog, here - http://www.diyaudio.com/forums/blogs/abraxalito/1119-ozone-chebyshev-filter.html
This particular filter uses bought inductors (TDK) but in previous designs I've used pot cores to fashion inductors from. No, the filter's Chebyshev to get a steep slope before the imaging components kick in. In going NOS there's not much room (frequency-wise) between the top of the audio specturm and the start of the imaging products (24.1kHz). Which is why so many DACs have historically used oversampling so the steep filter's digital, rather than analog.
I'd dispense with the OPA627 doing I/V duty and employ it as a post-amplifier. Actually an OPA637 would do that job better as it needs at least 20dB of gain, depending on what output level you'd like. Use an LC filter between the DAC and the OPA. Don't skimp on decoupling - a low impedance power supply is called for given you've spent such a lot on the DAC chips.
You mean a whole DAC design (I'm not using PCM1704, rather TDA1387) ? There's the DAC and filter schematic up on a thread in Digital Source here - http://www.diyaudio.com/forums/digi...-good-tda1545a-dac-kits-pcbs.html#post3918486
Is it possible to get a schema of your IV stage ?I cant find it in the link, sorry. I do not want to damage those DAC chips ,since I do not know what I am doing.
The I/V stage is shown in the schematic on that thread - simply a resistor (50R) to GND. There's no post-amplifier stage shown on that design though. You could use a 50R resistor with the PCM1704, it would not damage it. You'd only get about 10mV output though so if you wanted 1VRMS at the output you'd need a 40dB gain stage - certainly easily implemented with an OPA637..
does this mean the current output of the DAC at 0dBfs is 0.2mA?
That seems incredibly low.
@Andrew - you have to factor in that the filter loses 6dB and that the peak current from the DAC corresponds to the peak-to-peak output voltage. 10mVRMS is about 30mV peak-to-peak and (from memory) the DAC's output current is 1.2mA. So if no filter loss, this current would give around 60mV in 50ohms.
@kinku - Mouser has them (albeit rather pricey, not nearly as pricey as PCM1704s though) - SLF7045T-681MR20-PF TDK | Mouser
@kinku - Mouser has them (albeit rather pricey, not nearly as pricey as PCM1704s though) - SLF7045T-681MR20-PF TDK | Mouser
good op amp I/V?
we do have better than 5534 today - if you accept the noise level abrax' circuits give you can even use ADA4899 - how about that -117 dB @ 1 MHz 2Vpp output spec? - hear any IMD now?
http://www.analog.com/en/high-speed...on-amplifiers/ada4899-1/products/product.html
use 2 to make the I/V and 1st biquad - with 3 audio anti image poles upfront with these extreme linearity op amps then any decent fet input audio op amps for the rest of the filter poles after will have no problem with the rolled off image and glitch components
use multiple feedback filters and the op amp inputs are at virtual gnd and avoid input Z modulation effects too - maybe OPA1644
Last edited:
- Status
- Not open for further replies.