Some pictures of my new creature:
I find this arrangement with circuits on both sides of a plate very convenient: you get most of the benefits of a double chassis without the problem of cabling.
D.
I find this arrangement with circuits on both sides of a plate very convenient: you get most of the benefits of a double chassis without the problem of cabling.
D.
Ryanj, thank you very much for posting! Really appreciated since I have just re-started my dusty TDA1541 project and like to try out the SEN i/v with 2SK170BL's which I still have a small stash.
Patrick, just want to say thanks a million for sharing such innovative design and great tips! Been following on the sideline and has just gotten motivation to restart my TDA1541 DAC project after many many years of absence! Thanks again!
You can see here how a current source may be constructed for the TDA1541, using a 2SK170 JFET :
Using the AD488 as common base zero feedback I/V stage, for current output dac's - DACs & Digital Source / Transport - StereoNET
http://www.stereo.net.au/forums/uploads/monthly_04_2013/post-106386-0-34640200-1367203422_thumb.gif
Patrick
Using the AD488 as common base zero feedback I/V stage, for current output dac's - DACs & Digital Source / Transport - StereoNET
http://www.stereo.net.au/forums/uploads/monthly_04_2013/post-106386-0-34640200-1367203422_thumb.gif
Patrick
Actually to make it clear you can delete the AD844 and everything connected to it.
The purpose of the CCS is to cancel the DC current that comes out of the DAC at zero signal. Let us say, for the sake of discussion, this is 1mA (may be different).
Then if there is signal, that 1mA will vary up or down depending on signal. At half max amplitude, for instance, it will vary between 0.5mA and 1.5mA.
If you put this in an I/V converter it is not a problem, unless you want to DC couple it, without a coupling cap.
In that case, you want to feed a constant 1mA into the DAC, so that without signal nothing comes out of it, and with the same signal as before it will vary between +0.5mA and -0.5mA, so a pure AC signal and no coupling cap needed.
So we need something that gives a constant current, independent of the voltage it is connected to. Look at the FET in the circuit Patrick linked to. At switch on, the FET will start to conduct, this will start to generate a voltage across the source R. The gate is at zero volts, so that current will grow until the voltage across the source R exactly cancels the required G-to-S voltage required for that current.
The pot is used to fiddle the effective source R so that the final point is exactly 1mA (in this discussion).
And once the current finds its set point, it remains constant, whatever the voltage is that will be at the DAC output. Voilà, a constant current source!
Jan
The purpose of the CCS is to cancel the DC current that comes out of the DAC at zero signal. Let us say, for the sake of discussion, this is 1mA (may be different).
Then if there is signal, that 1mA will vary up or down depending on signal. At half max amplitude, for instance, it will vary between 0.5mA and 1.5mA.
If you put this in an I/V converter it is not a problem, unless you want to DC couple it, without a coupling cap.
In that case, you want to feed a constant 1mA into the DAC, so that without signal nothing comes out of it, and with the same signal as before it will vary between +0.5mA and -0.5mA, so a pure AC signal and no coupling cap needed.
So we need something that gives a constant current, independent of the voltage it is connected to. Look at the FET in the circuit Patrick linked to. At switch on, the FET will start to conduct, this will start to generate a voltage across the source R. The gate is at zero volts, so that current will grow until the voltage across the source R exactly cancels the required G-to-S voltage required for that current.
The pot is used to fiddle the effective source R so that the final point is exactly 1mA (in this discussion).
And once the current finds its set point, it remains constant, whatever the voltage is that will be at the DAC output. Voilà, a constant current source!
Jan
Last edited:
Thanks Jan for the detailed explanation.
So the K216 is basically a source follower due to the it's biasing and varying the gate voltage with trimmer, similar to a pull up situation. While the 170 circuit is sourcing the DAC output which controls vgs until Iout finally settles to the required amount by trimmer/R source and finally remains constant.
So the K216 is basically a source follower due to the it's biasing and varying the gate voltage with trimmer, similar to a pull up situation. While the 170 circuit is sourcing the DAC output which controls vgs until Iout finally settles to the required amount by trimmer/R source and finally remains constant.
Not sure if you could call the circuit I posted a CCS or a constant voltage source follower, Jan would know a lot better than I, but I've tried the circuit in the link EUVL posted and it drifts quite a lot more compared to the one I use.
Ryan
Ryan
Hi Jan,
The circuit in the link Patrick posted but without the opamp. Also I used 5v instead of 15v. I mean the voltage drifted by something like +-10mV at the output of the dac.
The circuit in the link Patrick posted but without the opamp. Also I used 5v instead of 15v. I mean the voltage drifted by something like +-10mV at the output of the dac.
JFET current drifts with temperature.
It will drift a lot less when degenerated a lot.
So e.g. if you use 2SK369V and degenrated to 1~2mA, it will be a lot more stable,
then e.g. a 2SK170GR used in the same way.
2Sk209BL or 2SK117BL would also be good choice.
Patrick
It will drift a lot less when degenerated a lot.
So e.g. if you use 2SK369V and degenrated to 1~2mA, it will be a lot more stable,
then e.g. a 2SK170GR used in the same way.
2Sk209BL or 2SK117BL would also be good choice.
Patrick
Sorry I don't have time to chase up stuff somewhere in previous posts.
It could be the DAC that drifted. IIRC this is an archaic TDA15-something?
It could be drift in the I/V impedance.
Jan
Last edited:
Hi Jan,
Here is the link to the circuit in question:
-->CLICK HERE<--
Yes, the DAC we talk about is the TDA1541A which has a -2mA of a current offset at its output.
Patrick, sorry for my ignorance, but what's the definition of a degenrated JFET?
Ryan
A JFET with a source resistor which reduces the current to (much) less then Idss.
The bigger the difference between the actual CCS current and Idss, the more stable is the CCS current.
Try google :
http://www.colorado.edu/physics/phys3330/phys3330_fa11/pdfdocs/AN102FETbiasing.pdf
Patrick
The bigger the difference between the actual CCS current and Idss, the more stable is the CCS current.
Try google :
http://www.colorado.edu/physics/phys3330/phys3330_fa11/pdfdocs/AN102FETbiasing.pdf
Patrick
Thanks Patrick,
Yes I found that PDF before, I'll have a proper read of it.
Thanks for your explanation.
Ryan
Yes I found that PDF before, I'll have a proper read of it.
Thanks for your explanation.
Ryan
And of course you can also use a temperature compensated LM334 CCS.
Almost forgotten .....
http://www.diyaudio.com/forums/soli...ource-ccs-audio-applications.html#post5190090
Patrick
Almost forgotten .....
http://www.diyaudio.com/forums/soli...ource-ccs-audio-applications.html#post5190090
Patrick