Hi,
When looking for a good but cheap supply for digital circuits, there is a lot to find on this site. Too much is a better word, lots of schematics and even more opinions. However this seemed like a good startingplace for a cheap but good supply for the digital parts of my dac. When i come to feed the DACs itself, i'll try the more advanced stuff. This is for the digital circuits and maybe the xtal.
So i remembered a small trick to use a fet to drive a led over a big voltagerange (also mentioned somewhere in this post, allthough i connected gate to drain ??) Well i made it as attached. The result is a 5.31V supply.
However, i have no good measuring equipment (old tube-scope.. with 50Hz without signal) and no DAC yet to test this by ear. So it would be nice if somebody would try this and report back :->
I find the LT1086 too expensive for my needs (need 3 supplies for the dig. part alone), and the other stuff too complicated to make three. Other options, zener+transitor, shunt?
Would there be an improvement by using a 3.8V ref instead of the 2 LED's (other than getting closer to 5V)?
I could try red LEDs to get the voltage down a bit.
No caps attached yet, that is for later, on in and output terminal and accros the LEDs.
Greetings,
GuidoB
When looking for a good but cheap supply for digital circuits, there is a lot to find on this site. Too much is a better word, lots of schematics and even more opinions. However this seemed like a good startingplace for a cheap but good supply for the digital parts of my dac. When i come to feed the DACs itself, i'll try the more advanced stuff. This is for the digital circuits and maybe the xtal.
So i remembered a small trick to use a fet to drive a led over a big voltagerange (also mentioned somewhere in this post, allthough i connected gate to drain ??) Well i made it as attached. The result is a 5.31V supply.
However, i have no good measuring equipment (old tube-scope.. with 50Hz without signal) and no DAC yet to test this by ear. So it would be nice if somebody would try this and report back :->
I find the LT1086 too expensive for my needs (need 3 supplies for the dig. part alone), and the other stuff too complicated to make three. Other options, zener+transitor, shunt?
Would there be an improvement by using a 3.8V ref instead of the 2 LED's (other than getting closer to 5V)?
I could try red LEDs to get the voltage down a bit.
No caps attached yet, that is for later, on in and output terminal and accros the LEDs.
Greetings,
GuidoB
Ok,
But does it matter? I am not interested in a supply which is exact 5.000V, but more in one which keeps the voltage as stable as possible.
Doesn't matter then if it varies a bit then from supply to supply.
So my question is, would the fet give a supply with a better regulation (noise, load) than the resistor.
Or does the fet vary more on load or temp or something else then than a resistor ?? Would think it is not as depending on Vout as with the resistor.
Thanx,
GuidoB
But does it matter? I am not interested in a supply which is exact 5.000V, but more in one which keeps the voltage as stable as possible.
Doesn't matter then if it varies a bit then from supply to supply.
So my question is, would the fet give a supply with a better regulation (noise, load) than the resistor.
Or does the fet vary more on load or temp or something else then than a resistor ?? Would think it is not as depending on Vout as with the resistor.
Thanx,
GuidoB
Data sheets for BS170:
http://www.fairchildsemi.com/ds/BS/BS170.pdf
http://www.futureai.com/electronics/datasheets/fet/BS170-D.pdf
http://www.semiconductors.philips.com/acrobat/datasheets/BS170_CNV_2.pdf
http://katalogi.iele.polsl.gliwice....-all--&ile=50&start=0&szukaj.x=16&szukaj.y=14
The typical Vgs threshold voltage is 2 volts for 1 ma. You should have about 1 to 3 ma through the LEDs. You might have to sort quite a few to get one with a turn on voltage around 1.25 volts. The drain to source impedance should be much higher than a resistor picked for 1 ma. Read the data sheet for the LM317 closely to see how the adjustment terminal works. You could add a diode biased by a current source to get your nominal 2 volts and not have to sort through as many BS170s. It is a good circuit if get the right fet Vgs. Temperature stability should be fine. Try it, its a pretty simple circuit. I would add a 1k gate resistor to the MOSFET as they are prone to oscillate without one.
http://www.fairchildsemi.com/ds/BS/BS170.pdf
http://www.futureai.com/electronics/datasheets/fet/BS170-D.pdf
http://www.semiconductors.philips.com/acrobat/datasheets/BS170_CNV_2.pdf
http://katalogi.iele.polsl.gliwice....-all--&ile=50&start=0&szukaj.x=16&szukaj.y=14
The typical Vgs threshold voltage is 2 volts for 1 ma. You should have about 1 to 3 ma through the LEDs. You might have to sort quite a few to get one with a turn on voltage around 1.25 volts. The drain to source impedance should be much higher than a resistor picked for 1 ma. Read the data sheet for the LM317 closely to see how the adjustment terminal works. You could add a diode biased by a current source to get your nominal 2 volts and not have to sort through as many BS170s. It is a good circuit if get the right fet Vgs. Temperature stability should be fine. Try it, its a pretty simple circuit. I would add a 1k gate resistor to the MOSFET as they are prone to oscillate without one.
I've used the LM334 with the LM317.
It is connected between output and adjust and feeds a string of leds. The 1.25V is within the operating range of this ccs.
With respect to the LM334's output impedance; the datasheet only shows a graph till 10kHz; I guess it's impedance at higher frequency will drop quickly and not be as good as a fet or mosfet used as ccs.
It is connected between output and adjust and feeds a string of leds. The 1.25V is within the operating range of this ccs.
With respect to the LM334's output impedance; the datasheet only shows a graph till 10kHz; I guess it's impedance at higher frequency will drop quickly and not be as good as a fet or mosfet used as ccs.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.