Re: Re: Re: Re: Re: The AD1864NK good than AD1865
Maybe a pot beside each 1.5K resitor will help zero the output.
Usualy, at least from the members of this forum, we preffer to add a cap at the output instead of using it at the input of the following stage. And I see also that most commercial gear follow the same pattern excep that in some cases they add a cap at both ends.
Would you hesitate to connect your DAC to a line amp or an amp that has no input cap?
kello said:
Maybe a pot beside each 1.5K resitor will help zero the output.
Usualy, at least from the members of this forum, we preffer to add a cap at the output instead of using it at the input of the following stage. And I see also that most commercial gear follow the same pattern excep that in some cases they add a cap at both ends.
Would you hesitate to connect your DAC to a line amp or an amp that has no input cap?
Hi!
the value of DC out -1/+3mV, is max value.so using in normal system I think is OK. cause mean time the DC out is <1mV.
using E-cap for DC couple,I can't easy full manage the sound Quality,and than good E-cap the price is very high. so in this design I remove it and without any problem.
of course,I would like the DC is zero in output,but I think is system,the DC output to speaker not higher 30mV<max > is OK<normal time <10mV),and in system using one e-Cap for DC couple is enough.
for different system or application, the design idea is different .I see too.
in commercial gear more think application to other system and safe reason.
I think you are right too,but I keep my idea in my mind.
the value of DC out -1/+3mV, is max value.so using in normal system I think is OK. cause mean time the DC out is <1mV.
using E-cap for DC couple,I can't easy full manage the sound Quality,and than good E-cap the price is very high. so in this design I remove it and without any problem.
of course,I would like the DC is zero in output,but I think is system,the DC output to speaker not higher 30mV<max > is OK<normal time <10mV),and in system using one e-Cap for DC couple is enough.
for different system or application, the design idea is different .I see too.
in commercial gear more think application to other system and safe reason.
I think you are right too,but I keep my idea in my mind.
I/V Conversion for AD1865N-K
Can anyone provide a simple answer?
My current project:
I have been reading this thread and I have also been working on a project- with a single AD1865N-K. So far I have a made a prototype using 2 x AD1851N-J mono DACs, but now wish to use the single AD1865N-K which is supposed to have a better sound (and a better THD figure). I am currently using the internal I/V conversion of the AD1851N-J s and direct coupling the output. I am using +/-5V Ni-cad supplies- these work fine, sound good and I wish to keep these for simplicity in the overall circuit design.
My question:
For my intended AD1865-K project could I use 2 x OPA627 as the I/V conversion to gain improved performance?
or
are the internal op-amps of the AD1865 somehow specially balanced or matched or better for +/-5V supply I have available?
Can anyone provide a simple answer?
My current project:
I have been reading this thread and I have also been working on a project- with a single AD1865N-K. So far I have a made a prototype using 2 x AD1851N-J mono DACs, but now wish to use the single AD1865N-K which is supposed to have a better sound (and a better THD figure). I am currently using the internal I/V conversion of the AD1851N-J s and direct coupling the output. I am using +/-5V Ni-cad supplies- these work fine, sound good and I wish to keep these for simplicity in the overall circuit design.
My question:
For my intended AD1865-K project could I use 2 x OPA627 as the I/V conversion to gain improved performance?
or
are the internal op-amps of the AD1865 somehow specially balanced or matched or better for +/-5V supply I have available?
IV-Converter for AD1865
Hi Oli,
The OPA627 can be used as a IV-converter for the AD1865. Just connect the current output to the inverting input of the opamp and the non-inverting input to ground. You can use the on-chip feedback resistor or a separate resistor form opamp output to inverting input. I also experimented with current feedback opamps like AD811 and OPA603. Back in time when I was using the AD1865 I employed a discrete opamp. Nowdays I would perhaps use the OPA604 or AD8610. Beware the latter does not work off +/-15 V supplies! Use +/-12V!
Passive IV-conversion is also possible. This is done by Audionote in the DAC1.1 & 1.2.
The internal opamps of the AD1865 are not bad but external opamps are better, I feel.
Hi Oli,
The OPA627 can be used as a IV-converter for the AD1865. Just connect the current output to the inverting input of the opamp and the non-inverting input to ground. You can use the on-chip feedback resistor or a separate resistor form opamp output to inverting input. I also experimented with current feedback opamps like AD811 and OPA603. Back in time when I was using the AD1865 I employed a discrete opamp. Nowdays I would perhaps use the OPA604 or AD8610. Beware the latter does not work off +/-15 V supplies! Use +/-12V!
Passive IV-conversion is also possible. This is done by Audionote in the DAC1.1 & 1.2.
The internal opamps of the AD1865 are not bad but external opamps are better, I feel.
Hi Oli,
you can use AD844 like suggested in this post ad844 i/v . That's what I use. It sounds very good and it doesn't need any additional buffering because the buffer is already in the IC. I tried passive I/V with transformer and fet buffer, but didn't like it . Perhaps the problem was in transformer (solid quality Beyerdinamc mic xformer), but the sound was lifeless and it lacked details. I also tried passive I/V without transformer and with fet buffer, but it was not good either. Maybe Audio Note DACs sound better with tube amplification stage. Anyway; I prefere TDA1541, but AD1865 is very good.
you can use AD844 like suggested in this post ad844 i/v . That's what I use. It sounds very good and it doesn't need any additional buffering because the buffer is already in the IC. I tried passive I/V with transformer and fet buffer, but didn't like it . Perhaps the problem was in transformer (solid quality Beyerdinamc mic xformer), but the sound was lifeless and it lacked details. I also tried passive I/V without transformer and with fet buffer, but it was not good either. Maybe Audio Note DACs sound better with tube amplification stage. Anyway; I prefere TDA1541, but AD1865 is very good.
Is +/- 5 volts okay for OPA627
Thanks for the ideas regarding I/V conversion for AD1865-K guys!
I would like to try the OPA627 option, but will my +/- 5 (Five) Volt supply be okay?
Elso, when you worked with the AD1865 did you use a capacitor in the feedback loop as a first order output filter to remove excessive utrasonic garbage that might make your power amplifier unstable?
P.S. asynchronous reclocking sounds great!
Thanks for the ideas regarding I/V conversion for AD1865-K guys!
I would like to try the OPA627 option, but will my +/- 5 (Five) Volt supply be okay?
Elso, when you worked with the AD1865 did you use a capacitor in the feedback loop as a first order output filter to remove excessive utrasonic garbage that might make your power amplifier unstable?
P.S. asynchronous reclocking sounds great!
Ad1865 Dac
Hi Oli,
I am not sure the OPA627 can work off +/-5V supplies. All specs in the datasheet are at +/-15V.
Yes you can put the first pole of the analog low-pass filter in the IV-converter. I simulated a MFB third order lowpass filter with FilterPro from www.ti.com. Schematic is attached. ??? is connected to ground.
If the crossoverfrequency is too low for you you can choose any frequency with the filterprogram as welll as filtertype: Bessel, Butterworth, Chebyshev.
Hey, I did not know you have built my Asynchronous Reclocker. My impression is it gives more improvement on the AD1865 than on the TDA1543 f.a. Glad it works to your satisfaction.
Hi Oli,
I am not sure the OPA627 can work off +/-5V supplies. All specs in the datasheet are at +/-15V.
Yes you can put the first pole of the analog low-pass filter in the IV-converter. I simulated a MFB third order lowpass filter with FilterPro from www.ti.com. Schematic is attached. ??? is connected to ground.
If the crossoverfrequency is too low for you you can choose any frequency with the filterprogram as welll as filtertype: Bessel, Butterworth, Chebyshev.
Hey, I did not know you have built my Asynchronous Reclocker. My impression is it gives more improvement on the AD1865 than on the TDA1543 f.a. Glad it works to your satisfaction.
Attachments
Is any filtering required?
Thanks Elso for the Schematic and the link for Filterpro!
I do not want too high an order of active filter as I feel it negates the purpose of the zero oversampling 'filterless' philosophy.
Would it be better to avoid the filter all together?
BTW I believe I am currently performing the filter operation with my AD1861N-Js incorrectly. I am simply connecting a 470pF polystyrene capacitor at the Iout terminal and shunting ultrasonics to ground. The output impedance of the Iout terminal is used as the resistor in this RC circuit. I am not so sure that this is such good practice and it might be better to connect the capacitor in the feedback loop of the I/V conversion op-amp as you have shown in your diagram. Is this because the latter arrangement presents a very high output impedance to the resistor ladder- this is desirable, and the feedback resistor is at a stable value - unlike the output impedance of the Iout pin?
Following your thoughts on op-amp choice the AD8610 looks very good- Swings to withinin 1.2V of rail voltage (ideal for my +/- 5V supplies), consumes a mere 3mA and offers superb load driving ability- thus requiring no bufffer stage. Many specifications of this Op-Amp better the OP627.
Your thoughts?
Thanks Elso for the Schematic and the link for Filterpro!
I do not want too high an order of active filter as I feel it negates the purpose of the zero oversampling 'filterless' philosophy.
Would it be better to avoid the filter all together?
BTW I believe I am currently performing the filter operation with my AD1861N-Js incorrectly. I am simply connecting a 470pF polystyrene capacitor at the Iout terminal and shunting ultrasonics to ground. The output impedance of the Iout terminal is used as the resistor in this RC circuit. I am not so sure that this is such good practice and it might be better to connect the capacitor in the feedback loop of the I/V conversion op-amp as you have shown in your diagram. Is this because the latter arrangement presents a very high output impedance to the resistor ladder- this is desirable, and the feedback resistor is at a stable value - unlike the output impedance of the Iout pin?
Following your thoughts on op-amp choice the AD8610 looks very good- Swings to withinin 1.2V of rail voltage (ideal for my +/- 5V supplies), consumes a mere 3mA and offers superb load driving ability- thus requiring no bufffer stage. Many specifications of this Op-Amp better the OP627.
Your thoughts?
Non-os Dac
Hi Oli,
The reason for analog low-pass filtering is I want a smooth 3150 sine wave, not the one with the staircase on it. And with filtering applied it sounded better.
I made some posts in the past on this subject:
http://www.diyaudio.com/forums/showthread.php?postid=72484#post72484
and Jean-Pauls post (my reply om the next page of that thread):
http://www.diyaudio.com/forums/showthread.php?postid=142850#post142850
There are more but I can't find it back with that lousy search engine here.
The inverting input of the opamp presents a virtual ground to the I-output of the DAC.
AD8610 seems better suited to low supply voltages than OPA627. If it sounds better I don't know, have not tried it yet.
Hi Oli,
The reason for analog low-pass filtering is I want a smooth 3150 sine wave, not the one with the staircase on it. And with filtering applied it sounded better.
I made some posts in the past on this subject:
http://www.diyaudio.com/forums/showthread.php?postid=72484#post72484
and Jean-Pauls post (my reply om the next page of that thread):
http://www.diyaudio.com/forums/showthread.php?postid=142850#post142850
There are more but I can't find it back with that lousy search engine here.
The inverting input of the opamp presents a virtual ground to the I-output of the DAC.
AD8610 seems better suited to low supply voltages than OPA627. If it sounds better I don't know, have not tried it yet.
HF attenuation capacitor outside the opamp feedback loop?
Thanks for your thoughts Elso. It has just struck me that you are using third order bessel filtering at only 10kHz! Do you hear any loss of HF?
I am only using first order filtering at 200kHZ! and I am currently placing a 470pF capacitor between Iout and Agnd, thus preventing HF from entering the feedback loop of the opamp inside my AD1861s. Yesterday I placed this capacitor inside the feedback loop (providing a slightly lower HF roll off frequency) I preferred the original arrangement- more gentle and dynamic sounding when required- Why do people not adopt this technique- something similar to Kusunuki's original output stage, with capacitors shorting HF noise directly from Iout straight to ground?
Thanks for your thoughts Elso. It has just struck me that you are using third order bessel filtering at only 10kHz! Do you hear any loss of HF?
I am only using first order filtering at 200kHZ! and I am currently placing a 470pF capacitor between Iout and Agnd, thus preventing HF from entering the feedback loop of the opamp inside my AD1861s. Yesterday I placed this capacitor inside the feedback loop (providing a slightly lower HF roll off frequency) I preferred the original arrangement- more gentle and dynamic sounding when required- Why do people not adopt this technique- something similar to Kusunuki's original output stage, with capacitors shorting HF noise directly from Iout straight to ground?
Re: Caps dorectly at I-out
Hi Oli,
I have tried a 2nF cap directly at the I-out to ground as in some Analog Devices datasheet. It did not have any effect, neither sonically nor on the scope.
Yes there is a slight loss of extreme highs with my filter.
Only a steep filter will avoid that but that will have as much ringing as a digital filter and bad sound.
As said before I can not bend the laws of physics......
This is the compromise I chose.
Oli said:Thanks for your thoughts Elso. It has just struck me that you are using third order bessel filtering at only 10kHz! Do you hear any loss of HF?
I am only using first order filtering at 200kHZ! and I am currently placing a 470pF capacitor between Iout and Agnd, thus preventing HF from entering the feedback loop of the opamp inside my AD1861s. Yesterday I placed this capacitor inside the feedback loop (providing a slightly lower HF roll off frequency) I preferred the original arrangement- more gentle and dynamic sounding when required- Why do people not adopt this technique- something similar to Kusunuki's original output stage, with capacitors shorting HF noise directly from Iout straight to ground?
Hi Oli,
I have tried a 2nF cap directly at the I-out to ground as in some Analog Devices datasheet. It did not have any effect, neither sonically nor on the scope.
Yes there is a slight loss of extreme highs with my filter.
Only a steep filter will avoid that but that will have as much ringing as a digital filter and bad sound.
As said before I can not bend the laws of physics......
This is the compromise I chose.
No capacitor at Iout
Can you clarify your statement Elso- Using a capacitor at Iout did not perform any measurable filtering at all?
I confess that I tried computer modelling this arrangement and found this did not work as a filter. It is most likely that I am running my system wide open with no filtering at all! (except for the sinx/x roll off incurred by the staircase shaped waveforms, internal filtering in my power amplifiers and inductance of my tweeters)
I believe that a capacitor in the feedback loop of the AD1861 opamp is causing this component to perform with unacceptably poor slew rate. Hence the reason for my preferred choice- a better sound is achieved with the capacitor out of the feedback loop of this opamp.
I may as well remove the capacitor since it serves no purpose and improve my opamp stability. Does this seem reasonable?
Can you clarify your statement Elso- Using a capacitor at Iout did not perform any measurable filtering at all?
I confess that I tried computer modelling this arrangement and found this did not work as a filter. It is most likely that I am running my system wide open with no filtering at all! (except for the sinx/x roll off incurred by the staircase shaped waveforms, internal filtering in my power amplifiers and inductance of my tweeters)
I believe that a capacitor in the feedback loop of the AD1861 opamp is causing this component to perform with unacceptably poor slew rate. Hence the reason for my preferred choice- a better sound is achieved with the capacitor out of the feedback loop of this opamp.
I may as well remove the capacitor since it serves no purpose and improve my opamp stability. Does this seem reasonable?
How do I calculate the compensation capacitor?
I have been reading the data sheet you mentioned. Perhaps my computer model showed no effect as I was using an ideal op-amp model (not a realistic situation!)
I would like to retain the capacitor at Iout as it does obviously have an effect. I would like to calculate a suitable compensation capacitor. The data sheet gives confusing information about the choice of compensation capacitor (The formua makes no sense). Is there a formula for choosing an appropriate capacitor at Iout and a compensation capacitor for a given opamp and desired transfer function.
I have been reading the data sheet you mentioned. Perhaps my computer model showed no effect as I was using an ideal op-amp model (not a realistic situation!)
I would like to retain the capacitor at Iout as it does obviously have an effect. I would like to calculate a suitable compensation capacitor. The data sheet gives confusing information about the choice of compensation capacitor (The formua makes no sense). Is there a formula for choosing an appropriate capacitor at Iout and a compensation capacitor for a given opamp and desired transfer function.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.