Merlin if you use 9140 instead of 140, it goes without saying. Those are N Mosfets you need in the negative side.
Yes, but better take out the Mosfets and cooked resistors from the wrong side, to make it non conducting, so not to catch fire.
queez
Salas - why did you guys use LEDs for Vref duty on these shunt PSs?
Reason I'm asking: I am also inclined to use LED for this, as I do believe they're quieter than most Vrefs, and as quiet as the most expensive ones (which as far as I know go as low as a few uV). LEDs also I believe have noise in the range of 6-8uV.
I do not know what their wideband performance is in this respect (I - totally - assume those 6-8uV to be in the tens of Hz). I could never find any good literature or specs about this, as nobody seems to be really interested in studying LEDs as voltage references, and even less in terms of audio specifics.
I'd be very interested in what you know about this.
Thank you,
Radu.
(P.S. I am getting close to prototyping two boards, very similar to each other, for your shunt PS with current/remote sense with high versatility and availability for customization. do you think there might be interest in putting this into a group buy/order or something along these lines?)
Salas - why did you guys use LEDs for Vref duty on these shunt PSs?
Reason I'm asking: I am also inclined to use LED for this, as I do believe they're quieter than most Vrefs, and as quiet as the most expensive ones (which as far as I know go as low as a few uV). LEDs also I believe have noise in the range of 6-8uV.
I do not know what their wideband performance is in this respect (I - totally - assume those 6-8uV to be in the tens of Hz). I could never find any good literature or specs about this, as nobody seems to be really interested in studying LEDs as voltage references, and even less in terms of audio specifics.
I'd be very interested in what you know about this.
Thank you,
Radu.
(P.S. I am getting close to prototyping two boards, very similar to each other, for your shunt PS with current/remote sense with high versatility and availability for customization. do you think there might be interest in putting this into a group buy/order or something along these lines?)
Because they are really low noise, low impedance, and stiff to their CCS. One generic random red fed by a high -Vp jfet, beat an LM329 in a test I asked syn08 to run on his HP analyzer in JC's Blowtorch thread once. Was something like 35nVsqHz density over audio as a total system, with lowest hum pick up due to close impedance to its current source. Noise does not correlate linearly with the more you use. But multiplying under square root. So there is a final population number to consider against other solutions given the Vout level and Vref you need.
About your other thoughts, PM me.
About your other thoughts, PM me.
In the Lightspeed Attenuator thread, it is recommended that two series regulators (12v and 5v) in series are used to reduce fluctuation. Can we use the Hypnotize in series to do that? In general, is there any issues in connecting two shunt regulators in series?
Some they already used the ''baby shunt'' I proposed to them some time ago, after the 3 legged series pre reg. It was only a series-parallel Leds matrix fed by an LM317 set up as CCS. Very small and shinny.
If you take care that the pre reg shunt has its CCS set for enough current so to feed the next with enough remaining not to shut off its own shunt element, then if you set the next one's Vout lower, they will work. They had this thing in DAC End recently where they feed multiple TL431 TO-92 local shunts fed from 3 V1.0 Simplistic branches. In the 3rd one they did not test the current draw, were chocking the main CCS and could not set it right for Vout. After they gave it ample current all worked.
If you take care that the pre reg shunt has its CCS set for enough current so to feed the next with enough remaining not to shut off its own shunt element, then if you set the next one's Vout lower, they will work. They had this thing in DAC End recently where they feed multiple TL431 TO-92 local shunts fed from 3 V1.0 Simplistic branches. In the 3rd one they did not test the current draw, were chocking the main CCS and could not set it right for Vout. After they gave it ample current all worked.
Thanks Salas. Even though we can use the shunt regulators for the Lightspeed Attenuator, is there any advantages in doing so? In this case, the load seen by the regulator is nearly constant, once we stop adjusting the volume control. We are not worried about drift in output. What we want is very little dynamic fluctuation in the regulated output. Does the shunt regulator offer any advantage over the series regulator here?
You must ask in the LSPD thread opinion from a few that used the ''baby''. Maybe they could discern a different tone or not. Strictly technically such a client circuit isn't demanding indeed. Maybe the Zo differences of the regs and the CCS isolation or not could be detectable? Don't know, haven't made it my self because I lack the application for a low Z volume controller yet.
I will ask about others' experience regarding the sound in the LSPD thread. I asked the question here more for the theoretical explanation. When I get my hand on one of these boards, I will try it out.
Strict Vrefs and good ripple kill can be found in series too. As I wrote, there are differences bcs of CCS isolation, Zo and its curve shape for this reg. Can go 5V straight with a couple of leds and a small resistor, try it some time.
Measurements +-24V
1st of all I want to know if the measurements mus be done with the load resistor 500R 5 W (I used 33R//33R 1W per side instead the 68R//68R 1W per side following the kind advice of Salas to get a current of 130mA, when after the reg. going on well I will change for 22R//22R 1W to get the 200mA) if yes these are mine measurements to get +-24V
Voltage across Rs = 60mV
Vout = 5.72V
V across test resistor (220R) = 1.27V
Base voltage = 4.45V (5.72V-1.27V)
V out target 24V - Base voltage 4.45V = 19.57V
60mV / 10R = 6mA
19.57V / 6mA = 3K262 so 3K3 the standard value
So with this resistor in both rails I will get +-24V 130mA, right?
How to know Jfets 0-3mV or 5mV?
1st of all I want to know if the measurements mus be done with the load resistor 500R 5 W (I used 33R//33R 1W per side instead the 68R//68R 1W per side following the kind advice of Salas to get a current of 130mA, when after the reg. going on well I will change for 22R//22R 1W to get the 200mA) if yes these are mine measurements to get +-24V
Voltage across Rs = 60mV
Vout = 5.72V
V across test resistor (220R) = 1.27V
Base voltage = 4.45V (5.72V-1.27V)
V out target 24V - Base voltage 4.45V = 19.57V
60mV / 10R = 6mA
19.57V / 6mA = 3K262 so 3K3 the standard value
So with this resistor in both rails I will get +-24V 130mA, right?
How to know Jfets 0-3mV or 5mV?
Yes 3k3 will take you to 24V. If your measurements are dead accurate, 24.25V.🙂
A load resistor drawing 50mA will not make much difference.
You mention Rs 60mV you mean the drop across the 10 Ohm source resistor of the 2SK170 in Vref.
What is the drop across 33R//33R?
A load resistor drawing 50mA will not make much difference.
You mention Rs 60mV you mean the drop across the 10 Ohm source resistor of the 2SK170 in Vref.
What is the drop across 33R//33R?
Measure 23.5V, how can increase till 24V, reducing 3k3 to 3k2 or 3K?
How can do to know Jfets 0-3mV or 5mV?
Last edited:
That question about jfets I don't get. What you mean exactly? Also for what application you make it?
Merlin - the JFET is conducting at IDSS, so what you need to know is the voltage dropped across LED's+Rset when they're conducting whatever IDSS is. IDSS is 6mA apparently, so 3.3kohms will drop 19.8 volts. More resistance will drop more volts 🙂
- Status
- Not open for further replies.