Hello folks,
here is something i have been working on for some time.
I am willing to offer the pcbs as a group guy.
let me know if there is interest.
Price $6,5 usd a piece+ paypal and Shipping&Handling.
1.6mm thick, fr-4 , red mask, double sided, 70um copper, white silk and HASL finish.
Here is 3D preview of this PSU how it looks when it is build. It gives at output 1mVAC ripple on 5A load, and on lower load output ripple is around 0,5mVAC.
Output BJTs are mounted on main heatsink to cool it down, this is using capacitance multiplier to supress AC ripple at output, and is not voltage regulator.
Input diodes are choosen to get temperature at 5A load is around 44C, and needs to mount at 10mm above pcb to cool it down.
So it is high quality psu for amplifiers with very low output ripple voltage.
The current GB list is as follows:
ronovar - 4pcs
here is something i have been working on for some time.
I am willing to offer the pcbs as a group guy.
let me know if there is interest.
Price $6,5 usd a piece+ paypal and Shipping&Handling.
1.6mm thick, fr-4 , red mask, double sided, 70um copper, white silk and HASL finish.
Here is 3D preview of this PSU how it looks when it is build. It gives at output 1mVAC ripple on 5A load, and on lower load output ripple is around 0,5mVAC.
Output BJTs are mounted on main heatsink to cool it down, this is using capacitance multiplier to supress AC ripple at output, and is not voltage regulator.
Input diodes are choosen to get temperature at 5A load is around 44C, and needs to mount at 10mm above pcb to cool it down.
So it is high quality psu for amplifiers with very low output ripple voltage.
The current GB list is as follows:
ronovar - 4pcs
Ronovar,
Looks good. What would be suitable input and output range, and drop accross the PSU/Cap multiplier?
Looks good. What would be suitable input and output range, and drop accross the PSU/Cap multiplier?
Thank you Francois G,
Input drop on diode is 2x0,87VDC, and Capacitive Multiplier voltage drop is about 2VDC at 5A load.
So voltage drop is 3,74VDC and i will say 4VDC total.
If you need +/-50VDC output voltage you need toroid of: 50VDC + 4VDC = 54VDC / 1,4142 = 38,20VAC toroid.
If you need higher output voltage you need to replace all 63VDC capacitors with higher voltage, for example 80V or 100V capacitors, but for most amplifiers we use +/-50VDC.
Input drop on diode is 2x0,87VDC, and Capacitive Multiplier voltage drop is about 2VDC at 5A load.
So voltage drop is 3,74VDC and i will say 4VDC total.
If you need +/-50VDC output voltage you need toroid of: 50VDC + 4VDC = 54VDC / 1,4142 = 38,20VAC toroid.
If you need higher output voltage you need to replace all 63VDC capacitors with higher voltage, for example 80V or 100V capacitors, but for most amplifiers we use +/-50VDC.
Cap Mx post CRC is an excellent way to get a low ripple PSU. I also used an active bridge to reduce heat diode switch noise, and I have done a similar PSU and achieved under 1mV rms ripple (rail was at 36V)for 4.7A current for a SE Class A amp that had low PSRR. One thing you need to make sure is to spec the CRC to stay within the ripple current rating of the cap. Unlike normal “bigger is better” cap bank linear PSU’s. The biggest size cap is not always the best and low ESR caps are also not the best. Too big and the ripple current increases. It’s a tradeoff. A simulation is the best way to spec the CRC values. I ended up at 15,000uF and 0.2ohm 10W resistors (two 0.1ohm 5W in series). Caps rated for 8A ripple current were spec’d based on oversize by 2x to handle the 4A ripple current.
Hi,
R3 and R4 with R7 and R8 form low pass filter with large capacitors and improve PSRR at 100Hz as lower ripple.
Another feauture is limiter for large inrush current when power is switched on.
In practice first pair of resistor will take more heat as second resistor pair.
Put into simulator rcrc filtering and meausure ac ripple at output. You will see that rcrc is more psrr and ripple lower that crc combination.
R3 and R4 with R7 and R8 form low pass filter with large capacitors and improve PSRR at 100Hz as lower ripple.
Another feauture is limiter for large inrush current when power is switched on.
In practice first pair of resistor will take more heat as second resistor pair.
Put into simulator rcrc filtering and meausure ac ripple at output. You will see that rcrc is more psrr and ripple lower that crc combination.
Using mosfets and LT4320 is great ideal bridge rectifier but i avoid it because it is hard to obtain from mouser and other large semiconductors store.
Think of that if people buy 200pcs of pcb for psu so you need 200pcs LT4320, so where you will get that much rectifiers?
This is reason why i avoid using LT4320, because components needs to be user friendly.
Think of that if people buy 200pcs of pcb for psu so you need 200pcs LT4320, so where you will get that much rectifiers?
This is reason why i avoid using LT4320, because components needs to be user friendly.
Yes, they are bottom.
No any interest in groupbuy this high quality psu with low ripple at output?
No any interest in groupbuy this high quality psu with low ripple at output?
Hi ronovar, very nice layout.
Could you explain to me the reasoning behind splitting the vcc and ve grounds between the supply caps?
It looks to me you have split the ground plane all the way to the output caps so it is essentially a star ground psu?
Could you explain to me the reasoning behind splitting the vcc and ve grounds between the supply caps?
It looks to me you have split the ground plane all the way to the output caps so it is essentially a star ground psu?
It is virtual ground, so that large capacitors is filtering virtual ground also, so we have clean V+, VGND, V- voltage.
Virtual Ground is using V- and V+ on dual secondary windings...this way all voltage polarity is clean and ready to go to capacitive multiplier.
Virtual Ground is using V- and V+ on dual secondary windings...this way all voltage polarity is clean and ready to go to capacitive multiplier.
Very interesting.
Do you not expect any benefits combining + and - ground? The impedance will be lowered between the psu caps and in respect to eachother they will be higher performing so itll be similar to having bigger caps. At least thats the way i see it. Ive fiddled with different psu ground schemes including one like yours and tho they all sound different i cant decide which one performs the best.
The kind we see most- without any slits and one big ground plane, has the best bass ime.
Do you not expect any benefits combining + and - ground? The impedance will be lowered between the psu caps and in respect to eachother they will be higher performing so itll be similar to having bigger caps. At least thats the way i see it. Ive fiddled with different psu ground schemes including one like yours and tho they all sound different i cant decide which one performs the best.
The kind we see most- without any slits and one big ground plane, has the best bass ime.
I don't like to use toroidals with single secondaries and big wire for ground for same reason as you describe above for Vgnd.
This is pcb projected for dual secondaries (so toroid have none big gnd wire), and vgnd is taken from output of capacitors, this way we have filtered all polarity and keep clean rails.
Clean power supply rails sound different, and how different it sound i will let too your ears to decide, because each person ears hear different. As some ears like amplifiers with high THD (darzeel amp) others like more ampliffiers with very low THD.
This is pcb projected for dual secondaries (so toroid have none big gnd wire), and vgnd is taken from output of capacitors, this way we have filtered all polarity and keep clean rails.
Clean power supply rails sound different, and how different it sound i will let too your ears to decide, because each person ears hear different. As some ears like amplifiers with high THD (darzeel amp) others like more ampliffiers with very low THD.
So am looking for power supply option for Wolverine amp. Transformer required is 500 va and 35 ac secondary. Will your option work for me?
If i im calculating right it is siggested to use output bjt that can handle 20A or more..because output bjt needs to be 3x larger that output currents needs to be....or if you are interested to buy multiple pcb i can redesign pcb with 2 output bjt pairs to handle 30Amps load divided by 3 it can handle amplifier that pulls maximum 10Amps of load.
500W toroid / 35VAC = 14,28/2 = 7,14Amperes maximum current per rail!
So one output par can hold maximum 5A of load.
How many pcb do you interest in buying if i redesign pcb to 2 output bjt?
500W toroid / 35VAC = 14,28/2 = 7,14Amperes maximum current per rail!
So one output par can hold maximum 5A of load.
How many pcb do you interest in buying if i redesign pcb to 2 output bjt?
I'll grab a set! But the PCBs should maybe be a little thicker due to the weight of capacitors in the long run.