Thanks again for your very good advice, I will try first with two 9V alkaline batteries first.
If I understood well your schematic: I need to connect a 10R + 100pF polystyrol between each DAC output (10R) + (100p) to AGnd ES9018 (Agnd is gnd in the AVcc TP module?)
First I need to learn what is a horse, thanks for support Patrick.
Felipe
If I understood well your schematic: I need to connect a 10R + 100pF polystyrol between each DAC output (10R) + (100p) to AGnd ES9018 (Agnd is gnd in the AVcc TP module?)
First I need to learn what is a horse, thanks for support Patrick.
Felipe
> I will try first with two 9V alkaline batteries first.
You need 2x 9V battery for one single ended SEN.
So for 2 channel balanced, you need 8x 9V battery altogether. Still less than 10 Euros.
Once working you can decide whether to go rechargeable, or mains based floating.
> I need to connect a 10R + 100pF polystyrol between each DAC output
> (10R) + (100p) to AGnd ES9018 (Agnd is gnd in the AVcc TP module?)
Correct. You may also use 100p silver mica if you wish.
And you should do this close to the ES9018 PCB from whatever supplier.
The purpose is to shunt some of the HF components of the i_out of the DAC.
Wish Nic will show you some photos soon.
Patrick
You need 2x 9V battery for one single ended SEN.
So for 2 channel balanced, you need 8x 9V battery altogether. Still less than 10 Euros.
Once working you can decide whether to go rechargeable, or mains based floating.
> I need to connect a 10R + 100pF polystyrol between each DAC output
> (10R) + (100p) to AGnd ES9018 (Agnd is gnd in the AVcc TP module?)
Correct. You may also use 100p silver mica if you wish.
And you should do this close to the ES9018 PCB from whatever supplier.
The purpose is to shunt some of the HF components of the i_out of the DAC.
Wish Nic will show you some photos soon.
Patrick
Here the schematics again properly drawn.
Unless I made some careless mistakes, they should be the same as that of post#593.
I have removed the 10R-100p at the current outputs of the ES9018.
They are most likely to be not very effective there due to the low Zin of the SEN IV V18.
You should instead experiment with C_iv in parallel with R_iv, starting with 100p.
Charger circuit later in the week.
Patrick
Unless I made some careless mistakes, they should be the same as that of post#593.
I have removed the 10R-100p at the current outputs of the ES9018.
They are most likely to be not very effective there due to the low Zin of the SEN IV V18.
You should instead experiment with C_iv in parallel with R_iv, starting with 100p.
Charger circuit later in the week.
Patrick
Attachments
You can probably see something very similar if you look for the O2 portable headamp in the Forum
Firstly a general remark :
I hope you guys would understand that this circuit was published for the experienced DIYers in mind, for them to try out the idea behind. And I also hope that you would appreciate I cannot possibly have the time to studyand then publish circuits for every DAC ICs around. Even then, it would be against my own discipline of publishing anything without proving in circuit. The only DAC IC I intend to try with this is PCM1704 and AD1865. I only coached Nic because there are a whole group of people who wanted to use the ES9018, and they (as well as I) have been waiting too long for someone to build and inform them. In the end, I decided to take things in my own hand, with a lot of help and donkey work from Nic.
There are already two variants covered, one with an offset current (TDA1541), and another with a offset voltage (ES9018). If you think you understand the similarity or difference between the DAC you choose and the two mentioned above, I encourage you to experiment. If you do not feel confident enough, then probably you should wait till someone else does. Because although this IV circuit is simple and elegant, it is complicated at the same time (in that it is not easy to understand, and requires floating supplies, etc.)
To answer jambul's specific question :
The PCM1794 has a bias current of 6.2mA and a dynamic current of 7.8mA p-p. If I understand the datasheet correctly, you need to have a constant current sink of 6.2mA at each Iout of the DAC, and use a R_iv of 680R. So you need something similar to but also different from the TDA1541 circuit published earlier,
http://www.diyaudio.com/forums/digi...minimalistic-iv-converter-12.html#post2935106
and certainly not that for the ES9018. But I do not have that IC, nor do I intend to try in the near future. Maybe Ryanj can give you some hints.
Thank you for your interest in any case,
Patrick
I hope you guys would understand that this circuit was published for the experienced DIYers in mind, for them to try out the idea behind. And I also hope that you would appreciate I cannot possibly have the time to studyand then publish circuits for every DAC ICs around. Even then, it would be against my own discipline of publishing anything without proving in circuit. The only DAC IC I intend to try with this is PCM1704 and AD1865. I only coached Nic because there are a whole group of people who wanted to use the ES9018, and they (as well as I) have been waiting too long for someone to build and inform them. In the end, I decided to take things in my own hand, with a lot of help and donkey work from Nic.
There are already two variants covered, one with an offset current (TDA1541), and another with a offset voltage (ES9018). If you think you understand the similarity or difference between the DAC you choose and the two mentioned above, I encourage you to experiment. If you do not feel confident enough, then probably you should wait till someone else does. Because although this IV circuit is simple and elegant, it is complicated at the same time (in that it is not easy to understand, and requires floating supplies, etc.)
To answer jambul's specific question :
The PCM1794 has a bias current of 6.2mA and a dynamic current of 7.8mA p-p. If I understand the datasheet correctly, you need to have a constant current sink of 6.2mA at each Iout of the DAC, and use a R_iv of 680R. So you need something similar to but also different from the TDA1541 circuit published earlier,
http://www.diyaudio.com/forums/digi...minimalistic-iv-converter-12.html#post2935106
and certainly not that for the ES9018. But I do not have that IC, nor do I intend to try in the near future. Maybe Ryanj can give you some hints.
Thank you for your interest in any case,
Patrick
xen-audio Minimalistic NiMH charger
This is a simple, minimalisitic charger that we use ourselves for charging NiMH battery packs. We normally leave it always on (hence the low triggle charge current of C/40), and it will take about 8 hours for a full charge.
Open the spreadsheet as read only, and enter input values in the two cells with green background. The rest will be calculated automatically for you.
But there are many modern charger ICs for NiMH, so by all means look around for something more sophisticated.
Patrick
This is a simple, minimalisitic charger that we use ourselves for charging NiMH battery packs. We normally leave it always on (hence the low triggle charge current of C/40), and it will take about 8 hours for a full charge.
Open the spreadsheet as read only, and enter input values in the two cells with green background. The rest will be calculated automatically for you.
But there are many modern charger ICs for NiMH, so by all means look around for something more sophisticated.
Patrick
Attachments
Last edited:
An ES9018 with SEN wings
A few weeks ago Patrick was kind enough to come up with a simple schematics describing, in a way that even I could understand it, how the SEN circuit could be used with the ES9018 DAC chip.
As I have the CEN/SEN evaluation pack (essentially PCB's and heatsinks), and recently Idss matched a fair amount of 2SK369V's, I decided to grab the offered tutoring and try the circuit out.
To be honest, I expected it to be the first build of several as much more complex solutions with constant current sources utilizing high precision voltage references have been deemed required to keep DC-offset to tolerable levels.
After getting hold of a few missing components I assembled the attendant circuitry on a small perf-board and connected it to the the SEN and my ES9018 DAC (tweaked Buffalo II from Twisted Pear Audio) exactly as per Patrick's schematics.
To my pleasant surprise the DC-offset was easily dialed down to less than 1 mV and remained stable for several hours. At cold-start I did not see offset exceeding 5 mV and it rapidly settled to below 1 mV. This is more that stable enough for me to have a good listen.....
OK - so how did it sound? Well considering that I had only build one balanced channel it sounded very, very good. Seriously good.....
Having learned some layout issues during the first build I decided to redo it with decent components and to make the second channel at the same time.
Considering that I have only listened for a few hours, I can only say that what I predicted from my one channel listening was confirmed - it is truly outstanding....
Cheers,
Nic
A few weeks ago Patrick was kind enough to come up with a simple schematics describing, in a way that even I could understand it, how the SEN circuit could be used with the ES9018 DAC chip.
As I have the CEN/SEN evaluation pack (essentially PCB's and heatsinks), and recently Idss matched a fair amount of 2SK369V's, I decided to grab the offered tutoring and try the circuit out.
To be honest, I expected it to be the first build of several as much more complex solutions with constant current sources utilizing high precision voltage references have been deemed required to keep DC-offset to tolerable levels.
After getting hold of a few missing components I assembled the attendant circuitry on a small perf-board and connected it to the the SEN and my ES9018 DAC (tweaked Buffalo II from Twisted Pear Audio) exactly as per Patrick's schematics.
To my pleasant surprise the DC-offset was easily dialed down to less than 1 mV and remained stable for several hours. At cold-start I did not see offset exceeding 5 mV and it rapidly settled to below 1 mV. This is more that stable enough for me to have a good listen.....
OK - so how did it sound? Well considering that I had only build one balanced channel it sounded very, very good. Seriously good.....
Having learned some layout issues during the first build I decided to redo it with decent components and to make the second channel at the same time.
Considering that I have only listened for a few hours, I can only say that what I predicted from my one channel listening was confirmed - it is truly outstanding....
Cheers,
Nic
Attachments
Nic,
Thank you very much for sharing the experience with us, and the many photos.
Very impressive "quick & dirty" test build !!
1mV DC offset is perfect. You won't get much better even with a servo.
Even 5mV at cold start is also more than OK. Good FET matching and gluing.
I presume you use one Vref to supply 4 singled-ended SEN ?
Also suggest you disconnect the 10R-100p at the ES9018, and move the 100p to parallel the 680R R_iv.
Should improve smoothness further.
Patrick
Thank you very much for sharing the experience with us, and the many photos.
Very impressive "quick & dirty" test build !!
1mV DC offset is perfect. You won't get much better even with a servo.
Even 5mV at cold start is also more than OK. Good FET matching and gluing.
I presume you use one Vref to supply 4 singled-ended SEN ?
Also suggest you disconnect the 10R-100p at the ES9018, and move the 100p to parallel the 680R R_iv.
Should improve smoothness further.
Patrick
One more suggestion.
Since you want to use Tx and floating regulators, I suggest you use separate transformers for the left and right channel. Will improve imaging. Supplying the positive and negative phase of each channel from a common Tx is OK.
Also, in theory at least, R-Core transformers should be better than toroids in this position.
They have less inter-winding capacitances, and hence will couple less back to the primaries.
Patrick
Since you want to use Tx and floating regulators, I suggest you use separate transformers for the left and right channel. Will improve imaging. Supplying the positive and negative phase of each channel from a common Tx is OK.
Also, in theory at least, R-Core transformers should be better than toroids in this position.
They have less inter-winding capacitances, and hence will couple less back to the primaries.
Patrick
> 1mV DC offset is perfect. You won't get much better even with a servo.
Even 5mV at cold start is also more than OK. Good FET matching and gluing.
The funny thing is that I get the low offset at exactly Vref/2 (i.e. when the voltage divider resistors are 1.000K and 1.000K (and no pot).
>I presume you use one Vref to supply 4 singled-ended SEN ?
No it is actually one Vref/channel (i.e. per two single ended SENs). Taking from the dual-shunt supply sitting on top of the ES9018 on the Buffalo.
>Also suggest you disconnect the 10R-100p at the ES9018, and move the 100p to parallel the 680R R_iv.
Will I not be feeding the SEN a lot of HF trash that the FETs have to deal with if I remove this filter? They are at the moment running at 45-50 degree C.
Even 5mV at cold start is also more than OK. Good FET matching and gluing.
The funny thing is that I get the low offset at exactly Vref/2 (i.e. when the voltage divider resistors are 1.000K and 1.000K (and no pot).
>I presume you use one Vref to supply 4 singled-ended SEN ?
No it is actually one Vref/channel (i.e. per two single ended SENs). Taking from the dual-shunt supply sitting on top of the ES9018 on the Buffalo.
>Also suggest you disconnect the 10R-100p at the ES9018, and move the 100p to parallel the 680R R_iv.
Will I not be feeding the SEN a lot of HF trash that the FETs have to deal with if I remove this filter? They are at the moment running at 45-50 degree C.
Last edited:
> Since you want to use Tx and floating regulators, I suggest you use separate transformers for the left and right channel. Will improve imaging. Supplying the positive and negative phase of each channel from a common Tx is OK.
I will consider this the next time I order custom made transformers. A lot of secondaries are needed (the one in the build has 14). I "sacrificed" one channel of a dual-mono build to test this out so I actually used to have one TX per channel.
Fact is that the DAC in its current form has absolutely no problem with imaging
Thanks for a great circuit,
Nic
I will consider this the next time I order custom made transformers. A lot of secondaries are needed (the one in the build has 14). I "sacrificed" one channel of a dual-mono build to test this out so I actually used to have one TX per channel.
Fact is that the DAC in its current form has absolutely no problem with imaging
Thanks for a great circuit,
Nic
nice work Nic! I had a feeling youde beet me to the punch, hopefully my circuit will still be of interest to people as truth be told and typical of me i allowed myself to get caught up in the details and variations on the reference circuit when as youve shown a simple 2 resistor divider is all thats needed to get started if your AVCC supplies are steady.
Doing it with AC is no mean feat, but its more convenient than batteries for me at least. a charging circuit for NIMH could be fairly straight forward, by harnessing the signal lock LED to switch DPDT relays so that when music is playing its disconnected but when its not the charger connects.
Doing it with AC is no mean feat, but its more convenient than batteries for me at least. a charging circuit for NIMH could be fairly straight forward, by harnessing the signal lock LED to switch DPDT relays so that when music is playing its disconnected but when its not the charger connects.
Last edited:
> Will I not be feeding the SEN a lot of HF trash that the FETs have to deal with if I remove this filter?
It is at a position which is not verz effective, as there is no voltage swing there.
A cap needs voltage swing before it can sink current.
Try it and then decide.
> They are at the moment running at 45-50 degree C.
That is just fine.
> it is actually one Vref/channel (i.e. per two single ended SENs).
> Taking from the dual-shunt supply sitting on top of the ES9018 on the Buffalo.
Are there two AVcc supplying the 9018, one per channel ?
Patrick
It is at a position which is not verz effective, as there is no voltage swing there.
A cap needs voltage swing before it can sink current.
Try it and then decide.
> They are at the moment running at 45-50 degree C.
That is just fine.
> it is actually one Vref/channel (i.e. per two single ended SENs).
> Taking from the dual-shunt supply sitting on top of the ES9018 on the Buffalo.
Are there two AVcc supplying the 9018, one per channel ?
Patrick