Digital circuits are always a challenge for linear regulators. Even a few inch of wiring sort of isolates the reg from the load for those fast edges.
The best way to handle that is to use good quality local decoupling, as close to the chips as you can get (SMD anyone?) to support the regulator.
Another option would be a shunt reg, IF you can get that close enough to the chips. A shunt an inch away defeats the purpose.
Jan
The best way to handle that is to use good quality local decoupling, as close to the chips as you can get (SMD anyone?) to support the regulator.
Another option would be a shunt reg, IF you can get that close enough to the chips. A shunt an inch away defeats the purpose.
Jan
Capacitance is an excellent source of current for amplifiers.
They are quiet, if not microphonic.
They are fast, if selected appropriately.
A capacitor that approaches perfection is the gap between two adjacent layers of a multilayer PCB. Some PCB manufacturers have patented their methods to create these near perfect capacitors.
Digital layout designers use these adjacent layers as their VHF decoupling. Two layer boards do not get close to the capacitances in 8layer to 12layer PCBs.
SMD caps with "designed in locations" of their vias, are the HF decoupling.
Leaded, or smd, electrolytics in strategic locations are the MF decoupling.
All the high speed current ON and current OFF, come from these three sets of "designed decoupling".
The regulator is simply a way to recharge the decoupling after the event/s
Except, where the regulator output is designed onto the current consumer device pins.
They are quiet, if not microphonic.
They are fast, if selected appropriately.
A capacitor that approaches perfection is the gap between two adjacent layers of a multilayer PCB. Some PCB manufacturers have patented their methods to create these near perfect capacitors.
Digital layout designers use these adjacent layers as their VHF decoupling. Two layer boards do not get close to the capacitances in 8layer to 12layer PCBs.
SMD caps with "designed in locations" of their vias, are the HF decoupling.
Leaded, or smd, electrolytics in strategic locations are the MF decoupling.
All the high speed current ON and current OFF, come from these three sets of "designed decoupling".
The regulator is simply a way to recharge the decoupling after the event/s
Except, where the regulator output is designed onto the current consumer device pins.
Last edited:
Filter cap capacitance?
I see that some are using quite small filter capacitors in CRC configuration before the regulator. I've read about using 2200uF or even 1000uF caps.
I was planning to use two 4700uF capacitors in CRC configuration or maybe even two 10000uF caps in CRC configuration. Would this be OK, or would this just be overkill?
I'm planning to power a B1 from Pass and some other buffer stages / pre amps with it in the near future. Just to see (or hear) which I like best.
I see that some are using quite small filter capacitors in CRC configuration before the regulator. I've read about using 2200uF or even 1000uF caps.
I was planning to use two 4700uF capacitors in CRC configuration or maybe even two 10000uF caps in CRC configuration. Would this be OK, or would this just be overkill?
I'm planning to power a B1 from Pass and some other buffer stages / pre amps with it in the near future. Just to see (or hear) which I like best.
Last edited:
I see that some are using quite small filter capacitors in CRC configuration before the regulator. I've read about using 2200uF or even 1000uF caps.
I was planning to use two 4700uF capacitors in CRC configuration or maybe even two 10000uF caps in CRC configuration. Would this be OK, or would this just be overkill?
I'm planning to power a B1 from Pass and some other buffer stages / pre amps with it in the near future. Just to see (or hear) which I like best.
I am using two 2,200 uf caps per rail. I don't think 4,700uf is too much. I've seen a lot of preamps with 20,000+uf per rail. I think it's more dependent on the load and circuit you are using. Maybe Jan will have better info about the particulars of the super reg. But the caps on the reg are not meant to be a capacitor bank...I think you should have some caps before the reg.
I've read that the Pass B1 is not really that power supply dependent...sounds good with anything, even switch mode PSUs. I think the best regulator for one application is not always the best for another...but I am enjoying the super reg in my Pass Ba-3 preamp.
Intuitively one would think that more capacitance gives less ripple amplitude, which is true, but there's another factor. If the cap increases, the current pulses from the rectifier to charge them also get larger and narrower. meaning they contain more and more high frequency components. You can see that on a scope: the ripple gets more and more sharp corners, with lower capacitance the corners are more rounded.
Those higher harmonic components may get into your circuit through several ways, so its better not to overdo it. That said, its not easy to give a clear advice, but in my experience anything above 10,000uF is wasted money for a pre-amp type supply and may actually degrade overall performance. And it is better to use 2 x 5,000uF with a resistor between, than a single 10,000uF.
Jan
The general guidance for a long time has been:
use 2200uF for each ampere of current.
If you know the circuit uses a quiescent current of 50mA and a peak transient current of 250mA then you can either:
a.) use 0.05*2200uF to suit the quiescent current, i.e. 110uF (100uF, or 120uF, or 150uF)
or
b.) use 0.25*2200uF to suit the peak transient current, i.e. 550uF (470uF, or 1mF)
This would be after a regulator if used.
If a CRC was being used instead of a regulator then I would use the b.) option after the R.
Before the R, the maximum I would use would be 50% of the second C.
But much smaller is probably OK, I would be very tempted to use the quiescent current capacitance value as a guide. i.e. 100uF to 150uF.
So the CRC ends up being: 150uF : R : 1mF, (R would be determined by acceptable voltage drop and desired RC time constant for effective filtering).
If you want, you can now tack a super regulator on the output of that CRC.
But you still need the transient peak current capacitor supply AFTER the regulator. And some of this will be MF and HF decoupling right at the main current consuming devices.
use 2200uF for each ampere of current.
If you know the circuit uses a quiescent current of 50mA and a peak transient current of 250mA then you can either:
a.) use 0.05*2200uF to suit the quiescent current, i.e. 110uF (100uF, or 120uF, or 150uF)
or
b.) use 0.25*2200uF to suit the peak transient current, i.e. 550uF (470uF, or 1mF)
This would be after a regulator if used.
If a CRC was being used instead of a regulator then I would use the b.) option after the R.
Before the R, the maximum I would use would be 50% of the second C.
But much smaller is probably OK, I would be very tempted to use the quiescent current capacitance value as a guide. i.e. 100uF to 150uF.
So the CRC ends up being: 150uF : R : 1mF, (R would be determined by acceptable voltage drop and desired RC time constant for effective filtering).
If you want, you can now tack a super regulator on the output of that CRC.
But you still need the transient peak current capacitor supply AFTER the regulator. And some of this will be MF and HF decoupling right at the main current consuming devices.
Last edited:
Thanks! I'll try some different values in the CRC. Values in the range between 1000uF and 4700uF and figure out the best or most acceptable value for R. I'll use a scope to measure the output of the CRC before connecting it to the regulator so I can feed the reg with the best found configuration of the CRC.
idle current draw
I tried to find the answer in all the articles and the thread, but I did not succeed in it. So here's another question:
What is the idle current draw of the regulator? I want to calculate the resistor value for the CRC. I know what the B1's current draw is, but I still need to know the current draw of the reg.
I tried to find the answer in all the articles and the thread, but I did not succeed in it. So here's another question:
What is the idle current draw of the regulator? I want to calculate the resistor value for the CRC. I know what the B1's current draw is, but I still need to know the current draw of the reg.
Hello,
I Just got 3 superreg boards.
I need to make a linear power supply with +12(1A), -12(0.3A), +5v (1A) outputs.
What parts will I need, and which transformers would be recommended? Can I use two torroids or do I need three?
Also, just saw this type of "GreenForce" ps.
GreenForce power supply Technical Details
is anyone familiar with this type of ps?
https://www.google.com/patents/US20...a=X&ei=ZAMyU6KdJ6OWyAG2koGQAQ&ved=0CD4Q6AEwAQ
regards,
Herman
I Just got 3 superreg boards.
I need to make a linear power supply with +12(1A), -12(0.3A), +5v (1A) outputs.
What parts will I need, and which transformers would be recommended? Can I use two torroids or do I need three?
Also, just saw this type of "GreenForce" ps.
GreenForce power supply Technical Details
is anyone familiar with this type of ps?
https://www.google.com/patents/US20...a=X&ei=ZAMyU6KdJ6OWyAG2koGQAQ&ved=0CD4Q6AEwAQ
regards,
Herman
Last edited:
The equation to use is
The advantage of describing it this way is, it applies to ALL Super Regulator schematics, regardless of the component labels applied to the resistors, zener diodes, IC voltage references, and so forth.
- Vref = Vout * {1 / [1 + (Rupper/Rlower)]}
The advantage of describing it this way is, it applies to ALL Super Regulator schematics, regardless of the component labels applied to the resistors, zener diodes, IC voltage references, and so forth.