Prune's volume control is OK. It will only work with a balanced signal, but that's OK as well.
This simple fet follower has many qualities, and some drawbacks. The positive qualities are: simplicity, class A operation, bias stability, and very low DC offset (with matched parts). The limitation is low tolerance for low resistance loads, below 2K or so. My associate, Bob Crump, successfully used this follower in a preamp for several years.
The next level is the complementary fet follower. This has the advantage of better current drive, and cancellation of even order distortion. It has a problem of being more difficult to control offset. It is also not as self limiting as the simple fet follower, but it will usually self limit, before it breaks.
This simple fet follower has many qualities, and some drawbacks. The positive qualities are: simplicity, class A operation, bias stability, and very low DC offset (with matched parts). The limitation is low tolerance for low resistance loads, below 2K or so. My associate, Bob Crump, successfully used this follower in a preamp for several years.
The next level is the complementary fet follower. This has the advantage of better current drive, and cancellation of even order distortion. It has a problem of being more difficult to control offset. It is also not as self limiting as the simple fet follower, but it will usually self limit, before it breaks.
My DAC won't drive well loads below 10K. The PCB version of the Aleph-X already has 10K input-to-ground resistor, so the pot would either be lower (Peter Daniel used 500 ohm) which I can't drive, or, as someone else suggested, a linear 100K, but that can create a lowpass filter when combined with the input capacitance of the Aleph-X. There's another thread about this here. There was a suggestion to put a pot in place of the ground shunting resistor, but hifizen told me that's not a good idea (though I forgot why...)
The only thing ahead of the buffer in the DAC is the DAC chip itself, a CS43122 (voltage output), and someone suggested that the capacitance of a pot is too high for the CS43122 to drive well. The DAC is jwb's, schematic is here:
http://atari.saturn5.com/~jwb/43122.png
In the datasheet it said 100pF max load capacitance, but for testing they used only 10pF to get their good results. Also there's DC offset so I'm not sure how to wire the pot.
I'm sorry, but I only got into DIY a year ago so I'm relying on what others tell me, so I apologize for the stupid questions.
What about this guy's suggestion: http://www.diyaudio.com/forums/showthread.php?postid=332245#post332245 ?
http://atari.saturn5.com/~jwb/43122.png
In the datasheet it said 100pF max load capacitance, but for testing they used only 10pF to get their good results. Also there's DC offset so I'm not sure how to wire the pot.
I'm sorry, but I only got into DIY a year ago so I'm relying on what others tell me, so I apologize for the stupid questions.
What about this guy's suggestion: http://www.diyaudio.com/forums/showthread.php?postid=332245#post332245 ?
Prune said:
So, the input capacitance due to the MOSFETs in the Aleph-X is too high to drive the thing from a passive attenuator unless it's a 500 ohm attenuator, that big-*** MOSFET buffer you've got on the output of your DAC can't drive a 500 ohm load, and your DAC chip can't handle the capacitance of a pot?
What a mess.
I'd go with the pot after the DAC chip.
se
The DAC buffer has about -1V DC offset, and I'm limited in the size of DC blocking film caps I can fit in there; lower loads with these limited size caps result in too much attenuation of the lower frequencies.
I was thinking of using a Bourns 81 or 82 series cermet, as the AKSA guy recommends. I didn't find any capacitance data in the datasheet.
I was thinking of using a Bourns 81 or 82 series cermet, as the AKSA guy recommends. I didn't find any capacitance data in the datasheet.
Well, I'm sure someone knows enough to answer this for me!
I contacted International Rectifier, and they just told me to look at some application notes (937 and 944), which I did, but they are all written from the point of view that the HEXFET will be used in a switching application, and as a newbie I couldn't figure it out...
Figure 6 in the datasheet shows V_gs vs gate charge, but I'm still not sure how to determine from that if it falls within the ability of the CS43122 to drive it in a follower configuration.
Again, the DAC schematic is here, and the CS43122 specifies a load of 1Kohm min. and 100pF max. The chip's outputs swing between 0.9V and 3.7V. So how do I calculate this?
I contacted International Rectifier, and they just told me to look at some application notes (937 and 944), which I did, but they are all written from the point of view that the HEXFET will be used in a switching application, and as a newbie I couldn't figure it out...
Figure 6 in the datasheet shows V_gs vs gate charge, but I'm still not sure how to determine from that if it falls within the ability of the CS43122 to drive it in a follower configuration.
Again, the DAC schematic is here, and the CS43122 specifies a load of 1Kohm min. and 100pF max. The chip's outputs swing between 0.9V and 3.7V. So how do I calculate this?
The data sheets for FETs typically show both Ciss and Crss. Crss is another way of saying Cgd, while Ciss is the same as Cgd + Cgs. So if you need to know Cgs, simply subtract Crss from Ciss. For a source follower as shown in your circuit, there is no contribution from Cgs, since there is (essentially) no AC signal between the gate and the source.
The data sheet for the IRF610 specifies Crss as 15 pF, but this is at 25 volts. If we look at a graph shown later in the data sheet, we can see that at 10 volts (this looks to be what the device will see in your circuit), the Crss is actually around 38 pF.
This is within your specified requirements, but if you look at this graph you will see that the input capacitance is quite non-linear. For this application, with only 10 volts across the part and biased around 5 mA, I see no reason at all to use a MOSFET. If you use a JFET instead, the circuit will be quieter, more linear, and smaller. Plus there will not be the 4 volt bias offset required for the MOSFET.
For example a 2SK170 will have about 6 pF Cgd at 10 volts D-S, and the required bias voltage is only about 0.2 V or so.
Hope this helps!
The data sheet for the IRF610 specifies Crss as 15 pF, but this is at 25 volts. If we look at a graph shown later in the data sheet, we can see that at 10 volts (this looks to be what the device will see in your circuit), the Crss is actually around 38 pF.
This is within your specified requirements, but if you look at this graph you will see that the input capacitance is quite non-linear. For this application, with only 10 volts across the part and biased around 5 mA, I see no reason at all to use a MOSFET. If you use a JFET instead, the circuit will be quieter, more linear, and smaller. Plus there will not be the 4 volt bias offset required for the MOSFET.
For example a 2SK170 will have about 6 pF Cgd at 10 volts D-S, and the required bias voltage is only about 0.2 V or so.
Hope this helps!
Thanks a lot.
As part of the analog filter, there is a 300 ohm resistor between the DAC chip and IRF610 gate. Does this change anything? In another thread, greyhorse said that the DAC datasheet load capacitance specification only applies for directly driving such a capacitance, not through a resistor. So does that resistor make the situation better or worse, or no change?
And I'm still wondering about putting a pot for volume control before the buffer.
As part of the analog filter, there is a 300 ohm resistor between the DAC chip and IRF610 gate. Does this change anything? In another thread, greyhorse said that the DAC datasheet load capacitance specification only applies for directly driving such a capacitance, not through a resistor. So does that resistor make the situation better or worse, or no change?
And I'm still wondering about putting a pot for volume control before the buffer.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.