diyAudio

diyAudio (http://www.diyaudio.com/forums/)
-   Digital Line Level (http://www.diyaudio.com/forums/digital-line-level/)
-   -   I/V Stage for Raindrop_hui's DAC (http://www.diyaudio.com/forums/digital-line-level/218775-i-v-stage-raindrop_huis-dac.html)

R4s3tsu 30th August 2012 07:25 AM

I/V Stage for Raindrop_hui's DAC
 
1 Attachment(s)
Hello there. Thats my first post lol.
Some time ago I've bought this dac
(SUMMER SALE) DAC TDA1541+WM8805 1.0UI multi-input | eBay
http://img.auctiva.com/imgdata/1/1/3...73111586_o.jpg

The DAC is not arrived yet, but I'v got a question:
The description says it's NON-OS.Without any analog stages. So I need some tips for this.
I've been looking for good I/V stage for this board and bought some kind of "Pedja Rogic's Buffer"
The schematic is in attachements.
So can I use this I/V stage with this DAC? Cos some people say it's can't work with TDA1541 NOS dacs. So, will it work or not?

Ceglar 30th August 2012 08:46 AM

you paid for that?

good trolling, bad fishing.

bye

UV101 30th August 2012 09:05 AM

That looks way too busy!

I would (do) use a single Opamp I/V stage. Provided you are not driving a long cable into a difficult preamp it will be very good if you use decent components. I've also used this formation in several dacs. You can either use good caps to block the dc on the output or add the components in the 2nd circuit to set the dc offset to zero.

With respect to nos, that's Non over sampling. This means there is no digital filter before the dac. This will have no effect on the current output of the dac and the type of output stage it can drive. The only information you need is that it's a tda1541 and you need an output stage.

http://i497.photobucket.com/albums/r.../CDPoutput.jpg

ryanj 30th August 2012 10:24 AM

Try the Cen out by EUVL. Worked well for me.

Post #578

Ryan

Calvin 30th August 2012 12:34 PM

Hi,

as I understand You already bought the IV-converter described in the attachment?
The circuit will work with a current output DAC like the TDA1541. The AD844 is a different kind of beast, as it is no classic OPamp. It consist of a current-conveyor input stage and a (here unused) buffer stage. The current conveyor part is made up from a grounded base stage feeding into current mirrors. This is a really smart way to achieve low THD without relying on global feedback and hence omitting with all the flaws of global feedback in this application. DA1,2 and DA3,4 are just paralleled AD844s, probabely to achieve a lower input impedance, which otherwise would be a bit high at ~50Ohms. Additionally the two AD844 share the signal current. So the current swing each AD844 has to deal with is halved, which reduces THD.
R1,2 and the associated C15,16 are the I/V conversion parts. The Rs define the output voltage which is Idac x R1. So, +-2mA from the TDA1541 result in 2.11Vrms. The caps C15,16 limit the bandwidth and filter aliasing artefacts. R3/C19 and R4/C20 form a additional lowpass filter. With NOS-DACs heavy post-filtering is required to deal with aliasing artefacts. A simple -6dB/oct filter is not sufficient. This is a major drawback of the circuit in #3, besides that it relys on a globally fed back OPamp.
The CEN is a very simple current conveyor using JFETs and probabely the best sounding circuit of the three and probabely even giving lowest THD of all three circuits.
DA5, DA6 and associated parts are AD844 configured as operational Buffers/Amplifiers. Here its made use of the AD844īs integrated Buffer stage (output taken from Pin 6). Since the requirements are by far less stringent to this stage, the rather costly AD844 is an overkill. A much cheaper NE5534 would suffice fully. DA5 and DA6 are configured in noninverting mode. Hence their gain is >1. If the complete circuit has too much gain, it may be reduced by changing R9,10, or by lowering the value of R1,2.

Thereīs a minor drawback which simulation of the current conveyor DA1,DA2 revealed. The AD844s THD rises the lower the DACs output impedance is and the larger the input cap C1 is. Best is if Zout of the DAC remains >100kOhm. Then THD is mostly due to C1. With C1=2.2nF as in the attachment THD@1kHz is 0.088%, -61dB, mainly odd harmonics. The same with C1=47p gave 0.0035%, -89dB (THD-figures at the input pin of R3), quite a difference.
The figures alone are still ok, though nothing spectacular. The distortion spectrum shows a clean decay towards higher harmonics, but is dominated by K3 and K5. Unfortunately the datasheet I have doesnīt give numbers for the TDA1541s output impedance.

jauu
Calvin

ps. some similar to the AD844 discrete I/Vs are discussed in "dac I/V convertion with very low distortion". See #251

R4s3tsu 30th August 2012 05:19 PM

Thank you guys for reply. I want to try my stage. But as far I understood I should do something with DAC's outputs, then connect it to stage?

abraxalito 31st August 2012 02:33 AM

Quote:

Originally Posted by Calvin (Post 3145463)
Thereīs a minor drawback which simulation of the current conveyor DA1,DA2 revealed. The AD844s THD rises the lower the DACs output impedance is and the larger the input cap C1 is. Best is if Zout of the DAC remains >100kOhm. Then THD is mostly due to C1. With C1=2.2nF as in the attachment THD@1kHz is 0.088%, -61dB, mainly odd harmonics. The same with C1=47p gave 0.0035%, -89dB (THD-figures at the input pin of R3), quite a difference.

Caps on the output of current output DACs is one of my pet subjects, so this is very interesting. I've found lower sized caps sound better with the DACs I've tried but I doubt that the SQ difference is down to THD. Do you have any explanation to suggest for what might cause the distortion to rise with larger values? Also were you simulating with a sine wave input or with steps?

R4s3tsu 31st August 2012 05:16 AM

1 Attachment(s)
UV101,
Should I do this circuits on the DAC\s outs and then connect it to stage?
Also here is the photo of that stage. Any tips for component improvement/optimization?
Some people recommended to replace LT1363 to BUF03 voltage buffer.

Calvin 31st August 2012 08:29 AM

Hi,

I can only guess as to why larger caps introduce more distortion.
A larger cap requires larger charging currents. Compare to the charging current of big Electrolyts in a power amp supply. Now this applies to charging from a voltage source and wouldnīt apply to a current source.
But as far as Iīve seen figures most audio DACs are rather low in output impedance and are by far no ideal current sources. PCM63 was specced at ~700Ohms, The PCM179x series probabely around 2kOhms, the ESS were resported to be as low as 300Ohms.
Since the DAC outputs waveform is stepped with very steep flanks (iirc a rise time of <200ns for FS of PCM179x and AD1955), and the output current of the DAC (i.e. charging current of the cap) is limited, and the input impedance of the I/V-stage beeing not 0, there will be voltage modulations at the I/Vs input.
I havenīt seen such drastical increase in THD with other grounded base circuits that have much lower input impedance and hence lower voltage modulation. The AD844 is specced at 50Ohms, two in parallel giving 25Ohms input impedance of the afore mentioned circuit. This leads to a voltage swing at the I/Vs input of ~20mVrms, while it ideally should be 0V.
I simm with sine as well as pulse.
The LT1363 is a fast OPamp and probabely not really cheap. In this application, such a fast OPamp is not needed at all. It was probabely chosen because it is speecced to be able to drive
rather low load impedances. If low impedance capability is no requirement a NE5534 would suffice and be much less problematic to the OPamps surroundings like layout and PS. If the stage works and sounds to Your satisfaction donīt bother tweaking. Against most claims of vast improvements rather the opposite is the case.
The BUF03 is marked as obsolete part on ADIs website.
jauu
Calvin

R4s3tsu 31st August 2012 09:24 PM

Calvin,
So it's better to use a smaller cap values then bigger?
As far as I know the purpose of LT1363 in this circuit is the follower function? Anyway even if BUF03 is obsolete it's still easy to find it on markets. I'll definitely will try it.


All times are GMT. The time now is 06:18 PM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright Đ1999-2014 diyAudio


Content Relevant URLs by vBSEO 3.3.2