Some questions regarding Pedja's RPSU:
1. Can I use MUR680 instead of SB560? (I want to use BrainGT PCB only without C1/C2 4.7uF).
2. C3/C4 100uF-470uF, how to determine the value ? Are those caps instead of 1000uF/1500uF on BrainGT pcb?
3. What voltage will I get in case of using 18V TX rails ?
thanks in advance.
Yuval
1. Can I use MUR680 instead of SB560? (I want to use BrainGT PCB only without C1/C2 4.7uF).
2. C3/C4 100uF-470uF, how to determine the value ? Are those caps instead of 1000uF/1500uF on BrainGT pcb?
3. What voltage will I get in case of using 18V TX rails ?
thanks in advance.
Yuval
Attachments
1. Can I use MUR680 instead of SB560? (I want to use BrainGT PCB only without C1/C2 4.7uF).
YES
2. C3/C4 100uF-470uF, how to determine the value ?
What sounds best
Are those caps instead of 1000uF/1500uF on BrainGT pcb?
no, the 1000/1500uf on the pcb's you have would be the 4700uf called for in the diagram you attached
3. What voltage will I get in case of using 18V TX rails ?
something less probably you will need to drop 2-3 volts with the regulator you are using so maybe 21-22 at the most
YES
2. C3/C4 100uF-470uF, how to determine the value ?
What sounds best
Are those caps instead of 1000uF/1500uF on BrainGT pcb?
no, the 1000/1500uf on the pcb's you have would be the 4700uf called for in the diagram you attached
3. What voltage will I get in case of using 18V TX rails ?
something less probably you will need to drop 2-3 volts with the regulator you are using so maybe 21-22 at the most
Hello Yuval,
Pedja
Yes and no. The job done by the usual 1000uF caps in the unregulated GC is in the case of the regulated supply shared between input and output caps and the reg itself. Anyway, C3 and C4 from the schematic should be mounted as close as possible to the LM3875. I've finished using 100uF here (not going lower than that).yuval said:C3/C4 100uF-470uF, how to determine the value ? Are those caps instead of 1000uF/1500uF on BrainGT pcb?
As officeboy said, not much and likely even less (reservoir capacitors yet can't hold the nominal voltage unconditionally constant), maybe it is good not to count on more than 19-20V.What voltage will I get in case of using 18V TX rails ?
Pedja
thanks both for your answers.
The 1500uF caps on BrainGT's pcb are mounted close the chip, I just need to swap it with 100uF. True or False ?
Just to be sure I get it right, the 1.414 factor (which makes 18V -> 25V under load on the chip) with unregulated PSU doesn't work for the regulated PSU ?
I plan to get something between 30W to 40W @ 6 ohm which means the chip should get 24V - 27V. In that case I'de better go for 24V rails, right ?
C3 and C4 from the schematic should be mounted as close as possible to the LM3875.
The 1500uF caps on BrainGT's pcb are mounted close the chip, I just need to swap it with 100uF. True or False ?
Just to be sure I get it right, the 1.414 factor (which makes 18V -> 25V under load on the chip) with unregulated PSU doesn't work for the regulated PSU ?
I plan to get something between 30W to 40W @ 6 ohm which means the chip should get 24V - 27V. In that case I'de better go for 24V rails, right ?
True.yuval said:The 1500uF caps on BrainGT's pcb are mounted close the chip, I just need to swap it with 100uF. True or False ?
It works always considering the rectified voltage after diodes, but you should also take into account a forward voltage drop of the used diodes (0.67V at 5A for SB560, 1.5V at 8A for MUR860). When you regulate it, you should also take into account a dropout voltage of the regulator (3V maximum for LM338), and the fact you will still have certain ripple on the input caps that will rise at the higher loads (so count a few additional Volts for this).Just to be sure I get it right, the 1.414 factor (which makes 18V -> 25V under load on the chip) with unregulated PSU doesn't work for the regulated PSU ?
Right.I plan to get something between 30W to 40W @ 6 ohm which means the chip should get 24V - 27V. In that case I'de better go for 24V rails, right ?
Pedja
BTW
You can not get 35W @ 6 Ohm with +/-25V supply as shown in the LM3875 datasheet using nothing but 1000uF per rail. You will have considerable voltage sagging under higher loads. In fact, the diagram shown by National worth only for regulated (stable voltage) supply.yuval said:Just to be sure I get it right, the 1.414 factor (which makes 18V -> 25V under load on the chip) with unregulated PSU doesn't work for the regulated PSU ?
I plan to get something between 30W to 40W @ 6 ohm which means the chip should get 24V - 27V. In that case I'de better go for 24V rails, right ?
Hi Pedja,
I see your point, I assume my speakers/room size could work with less than 35W (it's 88db / 6 ohm small room), but anyway I've decided to build the Regulated PS like you explaind using 2*24V rails - monoblocks/dual mono configuration.
Many thaks for your help.
Yuval
I see your point, I assume my speakers/room size could work with less than 35W (it's 88db / 6 ohm small room), but anyway I've decided to build the Regulated PS like you explaind using 2*24V rails - monoblocks/dual mono configuration.
Many thaks for your help.
Yuval
Yuval, I have done just that, 26v rails gives me 35W into 6Ohm at 1KHZ and 50W at 10KHZ 🙂 I also measured 35W at 50HZ into the same load.
Mains transformer is 25-0 0-25v and gives 36.8v before the 338 regulators.
Mains transformer is 25-0 0-25v and gives 36.8v before the 338 regulators.
hi Neal, what is voltage mesure after the Lm338 ?
35W @ 6 ohm means 27v regulated supply voltage, is this what u have ? (36.8V before LM338 should be at least 33.8v after which is more than chip can get 🙄 ).
35W @ 6 ohm means 27v regulated supply voltage, is this what u have ? (36.8V before LM338 should be at least 33.8v after which is more than chip can get 🙄 ).
Hi Yuval,
35v is the max *differential* voltage between input and output.
I am using 8TQ100 Sckottky rectifiers and measure 36.8v into the chip. My fixed resistor combination sets o/p voltage to 25.6v for each rail.
I aimed for 35W into 6Ohm load. A good compromise I thought as a large number of box 'speakers measure 6Ohm nominal.
Equation 5 from the data sheet: sqr root of (2xRLxPo) = sqr root (2x6x35) = 20.49v plus 5v dropout for the LM3875 = 25.49v supply rails.
Pk current = 3.41 Amp
Voltage measurment into a 6.8R load at 1KHZ just before clip was 15.2v or 34W close enough!
35v is the max *differential* voltage between input and output.
I am using 8TQ100 Sckottky rectifiers and measure 36.8v into the chip. My fixed resistor combination sets o/p voltage to 25.6v for each rail.
I aimed for 35W into 6Ohm load. A good compromise I thought as a large number of box 'speakers measure 6Ohm nominal.
Equation 5 from the data sheet: sqr root of (2xRLxPo) = sqr root (2x6x35) = 20.49v plus 5v dropout for the LM3875 = 25.49v supply rails.
Pk current = 3.41 Amp
Voltage measurment into a 6.8R load at 1KHZ just before clip was 15.2v or 34W close enough!
. My fixed resistor combination sets o/p voltage to 25.6v for each rail.
Could you explain what resistor is it? maybe you could add the circuit?
Yuval, refer back to Pedja's diagram you first posted.
Instead of the 120R/2K2 resistor combination I use a 240R sense resistor and a 4K7 adjustment resistor. From the 338 data sheet, ((4700 / 240) +1) * 1.24 set o/p voltage to 25.5v or thearabouts. You could substitute a 120R sense resistor and a 5K multi turn pot for the 4K7 to get precise adjustment of o/p voltage.
Instead of the 120R/2K2 resistor combination I use a 240R sense resistor and a 4K7 adjustment resistor. From the 338 data sheet, ((4700 / 240) +1) * 1.24 set o/p voltage to 25.5v or thearabouts. You could substitute a 120R sense resistor and a 5K multi turn pot for the 4K7 to get precise adjustment of o/p voltage.
- Status
- Not open for further replies.
- Home
- Amplifiers
- Chip Amps
- Pedja's Regulated PSU