I prefer to put something passive (i.e. a filter) nowadays between the DAC and any active stage because the challenges are just too great for semiconductors unaided so I've moved away from active I/V. If you'd like to learn a bit more about passive filtering you could have a butchas at my thread over here : Digital that sounds like analog
As regards the TPA6120 yes its a good choice - I have whole tubes of ADI's equivalent to that which is AD8016 and planning to build a lowish-power poweramp with several of them paralleled
As regards the TPA6120 yes its a good choice - I have whole tubes of ADI's equivalent to that which is AD8016 and planning to build a lowish-power poweramp with several of them paralleled
Last edited:
No, none on your list are CFB - none of the ones I've so far mentioned are true CFB either since they'd not work in the current circuit without mods. A particular point is that a true CFB amp can't have a feedback capacitor or it goes wild
No, don't know any FET ones which are CFB but I agree that would be an interesting avenue to explore if anyone's up for designing a discrete stage, looks like it could be good...
already done, Patrick aka EUVL/Papa/John Curl and their SEN/CEN variants are current conveyers/transimpedance stages with passive IV on the output. simple circuits, I just wish it didnt necessitate floating supplies/Batteries.
the thing for the ESS in stereo mode though is the input needs to be high transconductance/low impedance, low noise, as well as having the ability to handle the above average current of 16ma that you see from the 9018.
because of the above, jfets are difficult to choose, mosfets tend to need to be high current gain power devices... and the best candidates objectively end up being bipolar transistors.
this thread has some very interesting discrete circuits for ES901X too including some CFB types, one of which is a discrete AD844 variant.
Last edited:
A current conveyor isn't the same thing as a CFB amp. Current conveyors operate open loop from my (limited) understanding.
Yes what you say about ESS's requirements are reminding me of an earlier discussion we had where I said I was just too lazy to go there, coz ESS has made it too hard NOS is sooo much easier, but by no means a walk in the park all the same.
Yes what you say about ESS's requirements are reminding me of an earlier discussion we had where I said I was just too lazy to go there, coz ESS has made it too hard
NOS is sooo much easier, but by no means a walk in the park all the same.
yeah fair call on the actual SEN, but check the development thread there is a number of simple CFB designs discussed. also in the link I forgot to add is added now has a number of options including CFB
yep I think current conveyer must be open loop and gain depends on the device characteristics like transconductance etc but it seems a little bit of a grey area, but transimpedence amps dont have to be open loop of course
interesting and topical paper at ADi
^^ CFB amp
yep I think current conveyer must be open loop and gain depends on the device characteristics like transconductance etc but it seems a little bit of a grey area, but transimpedence amps dont have to be open loop of course
interesting and topical paper at ADi
An externally hosted image should be here but it was not working when we last tested it.
^^ CFB amp
Last edited:
3.8 is safe as shown here: ESS Support Downloads in the analog power consumption chart.
If using slower clocks, then you can probably use the higher voltages. Apparently, higher voltages result in better performance.
The white paper says:
"With 4.0V power supply greater than 133dB can be achieved"
There is also some indication that slower clocks could give you better performance:
http://www.diyaudio.com/forums/digi...re-reference-dac-8-channel-4.html#post1441042
If using slower clocks, then you can probably use the higher voltages. Apparently, higher voltages result in better performance.
The white paper says:
"With 4.0V power supply greater than 133dB can be achieved"
There is also some indication that slower clocks could give you better performance:
http://www.diyaudio.com/forums/digi...re-reference-dac-8-channel-4.html#post1441042
its not safe at the speeds I run and potentially shortening the life of the dac chip for an extra 1-2db max (they cant even tell you how much) when your chances of seeing the full standard 133db are already nil? doesnt seem worthwhile to me.
of course it potentially give better performance, its the analogue reference for the output, meaning you get more swing, higher DNR and the the output will of course have higher DC offset to match. AVCC/2
I tried 4v, found no audible difference vs 3.6, but settled on 3.65. and I certainly wouldnt be recommending people here send it 4v with a shunt reg, generally shunt regs do not have very stable voltage output vs temperature due to the tempco on the voltage references used, part of the reasons I pulled back back then. you could potentially send it 4.5v....
running at 4v is like running the clock at 125MHz.... well actually at least running at 4v has a marginal measurable benefit.
of course it potentially give better performance, its the analogue reference for the output, meaning you get more swing, higher DNR and the the output will of course have higher DC offset to match. AVCC/2
I tried 4v, found no audible difference vs 3.6, but settled on 3.65. and I certainly wouldnt be recommending people here send it 4v with a shunt reg, generally shunt regs do not have very stable voltage output vs temperature due to the tempco on the voltage references used, part of the reasons I pulled back back then. you could potentially send it 4.5v....
running at 4v is like running the clock at 125MHz.... well actually at least running at 4v has a marginal measurable benefit.
On another thread I wondered aloud about SSM2141
http://www.analog.com/static/imported-files/data_sheets/SSM2141.pdf
is that device even worth any chatter? What with all the high-spec op amps being considered? {I'm not likely to use any op amps after the I->V anyway}.
I see it deployed on boards on eBay:
WZ 21B SSM2141 Balanced Unbalanced Board Balance Input Output Unbalance | eBay
and for various reasons I have some interest.
[modding Behringer DCX2496 which has AKM AK4393 differential out DACs]. SSM2141 reminds me of the Broskie cathode follower:
http://tubecad.com/Product_PDFs/BCF 9-Pin.pdf
though I have no idea what SSM2141 or the tube circuit would spec out at vs ESS reference design [or the better stuff being discussed HERE]. Anyone seen it used anywhere in real life?
http://www.analog.com/static/imported-files/data_sheets/SSM2141.pdf
is that device even worth any chatter? What with all the high-spec op amps being considered? {I'm not likely to use any op amps after the I->V anyway}.
I see it deployed on boards on eBay:
WZ 21B SSM2141 Balanced Unbalanced Board Balance Input Output Unbalance | eBay
and for various reasons I have some interest.
[modding Behringer DCX2496 which has AKM AK4393 differential out DACs]. SSM2141 reminds me of the Broskie cathode follower:
http://tubecad.com/Product_PDFs/BCF 9-Pin.pdf
though I have no idea what SSM2141 or the tube circuit would spec out at vs ESS reference design [or the better stuff being discussed HERE]. Anyone seen it used anywhere in real life?
Hey glt,
I was wondering the same thing - what about you ?
To get her running asap, I'll use a TL1963a 3.3V reg for the analog, and then when she's running and burned in, I'll try an AMB Sigma 11 with three IR leds as the Vref. They have a forward voltage of 1.14V each, so 3 will be 3.42V, to which I can add some gain to try higher voltages, or swap the leds for two red for 3.6V.... anyway, Sigma 11 is low impedance, low noise, high ripple rejection. A really good series reg imo.
The σ11 Regulated Power Supply
The TL1963a is a very nice sounding 3 pin reg - good punch and seems to have linear output impedance. Noise is not so good at 38uV. The DCY package fits on the board quite easily.
Linear Regulator (LDO) - Single Channel LDO - TL1963A - TI.com
So far I fitted both an adp151 3.3V for the digital and the TL1963a. Still waiting for parts. But so far :
I was wondering the same thing - what about you ?
To get her running asap, I'll use a TL1963a 3.3V reg for the analog, and then when she's running and burned in, I'll try an AMB Sigma 11 with three IR leds as the Vref. They have a forward voltage of 1.14V each, so 3 will be 3.42V, to which I can add some gain to try higher voltages, or swap the leds for two red for 3.6V.... anyway, Sigma 11 is low impedance, low noise, high ripple rejection. A really good series reg imo.
The σ11 Regulated Power Supply
The TL1963a is a very nice sounding 3 pin reg - good punch and seems to have linear output impedance. Noise is not so good at 38uV. The DCY package fits on the board quite easily.
Linear Regulator (LDO) - Single Channel LDO - TL1963A - TI.com
So far I fitted both an adp151 3.3V for the digital and the TL1963a. Still waiting for parts. But so far :
An externally hosted image should be here but it was not working when we last tested it.
Last edited:
Are you going to use a balanced output ? What are you planning to use as the filter ?
I am still using single ended so I'm going to use something like LME49990 in the output buffer/filter,.....
.... but I was wondering, since we know the current and voltage output of the DAC, could we do something "nos" with the I/V like a simple shunt resistor ?
Another option would be to use the LME49713 and change some of the resistor values to suit it ?
http://www.ti.com/product/lme49713
Last edited:
From a quick glance at the data, no - its a slow (10V/uS, 3MHz) opamp and some closely matched resistors. Will contribute a lot of IMD in this application.
The part I was referring to earlier with the extra-linear input stage was this one: http://www.analog.com/static/imported-files/data_sheets/AD8099.pdf
Its SOP8 (no DIL) and 12V max operation. But the noise spec will suit those looking to get close to the max possible from the DAC chip.
Hi
I can't get AD8099 soic samples to try
so I'll give that one a miss.For the LME49713, it seems easy enough - remove the 100pF cap (C1 in the datasheet circuit) and change the 680R (R1) to 1.2KR.
For pure CFB amps like LME49713 there are other tricks you can play which are outlined in this appnote from TI : http://www.ti.com/lit/an/slyt099/slyt099.pdf
Using the ferrite bead in series with the -ve input isn't something I've tried myself, but I sure plan to
Using the ferrite bead in series with the -ve input isn't something I've tried myself, but I sure plan to
That's a very interesting read. It seems I could keep the 680R for better noise and use a ferrite, AG or HG grade with low Q ( BLM18AG601SN1 ) to connect pin2 of the op amp. It might even be possible to simply short C1, which would create a feedback of pure resistance around 87R and add the ferrite with an impedance around 600R @100Mhz ?
And cheap too :
http://hk.element14.com/jsp/search/...ions=false&ref=globalsearch&_requestid=105100
And cheap too :
http://hk.element14.com/jsp/search/...ions=false&ref=globalsearch&_requestid=105100
Last edited:
Those prices look rather expensive to me - I'm used to paying a maximum of 0.1rmb for a bead, but then I do usually buy a whole reel of them at one time as I seem to use them by the bucketful
These are my current flavour of 'cooking' bead, chosen for having highest inductance at lower freqs - I use MPZ2012S102 : http://www.tdk.co.jp/tefe02/e9413_mpz.pdf
I do put a resistor to shunt them though, depending on application as it looks from comparing the impedance graph with my measurements that the inductance is somewhat non-linear at lower freqs. I've used them to decouple capacitive loads and got instability which was fixed by adding the shunt resistor.
These are my current flavour of 'cooking' bead, chosen for having highest inductance at lower freqs - I use MPZ2012S102 : http://www.tdk.co.jp/tefe02/e9413_mpz.pdf
I do put a resistor to shunt them though, depending on application as it looks from comparing the impedance graph with my measurements that the inductance is somewhat non-linear at lower freqs. I've used them to decouple capacitive loads and got instability which was fixed by adding the shunt resistor.
Want to try the Salas 1.1 for analog and the TPS7A47 eval board for the digital. But got to order the 1.2 V regulator first...
Just a quick update on those ferrites - found the Murata AG grade in 0805 size and 600R here, US$2 for 25 :
https://hongkong01.rs-online.com/web/p/ferrite-beads/7241494/
https://hongkong01.rs-online.com/web/p/ferrite-beads/7241494/
Still waiting for parts, so in the meantime I fitted a socket for the oscillator. Positioning is tricky but it's doable. I made it low-profile by cutting off the legs, soldering bridges, and filing down so it sat flat on the board. Put solder on one pad and positioned the socket.
The angle of this photo is not so good - looks like C1 connects to the socket with solder but that's just the angle.
The other angle :
Before filing down, with a Crystek that's grown legs :
For the xo decoupling cap C1, I soldered on a 0603 15pF to the board's pads and then a 1uF on top of it.
There's quite a lot of low esr ceramic caps on the outputs of the adp's. The datasheet says match the input and output caps, which I will do, but my experience with LM317s oscillating still makes me a little uncomfortable about this. What do you guys think ?
The angle of this photo is not so good - looks like C1 connects to the socket with solder but that's just the angle.
An externally hosted image should be here but it was not working when we last tested it.
The other angle :
An externally hosted image should be here but it was not working when we last tested it.
Before filing down, with a Crystek that's grown legs :
An externally hosted image should be here but it was not working when we last tested it.
For the xo decoupling cap C1, I soldered on a 0603 15pF to the board's pads and then a 1uF on top of it.
There's quite a lot of low esr ceramic caps on the outputs of the adp's. The datasheet says match the input and output caps, which I will do, but my experience with LM317s oscillating still makes me a little uncomfortable about this. What do you guys think ?
Last edited:
- Status
- Not open for further replies.