Hey guys and gal's,
I need to figure out a adjustable regulated bipolar supply with enough juice to drive 32 high current op amps ( ~100mA each ) and leaving me some headroom.
So...
32x 100mA = 3.2A + 0.8A ( 25% Headroom ) = 4A
I looked around and didnt find to many audiophile design ideas for bipolar supplies
I even went as far as looking at data sheets for ideas on a reference design with some luck but larger ratings were always shown for positive rails and not for negative
So, my thought was to use LM117/337 driving a larger TIP35/36 for each rail.
Any input or ideas are welcome.
I need to figure out a adjustable regulated bipolar supply with enough juice to drive 32 high current op amps ( ~100mA each ) and leaving me some headroom.
So...
32x 100mA = 3.2A + 0.8A ( 25% Headroom ) = 4A
I looked around and didnt find to many audiophile design ideas for bipolar supplies
I even went as far as looking at data sheets for ideas on a reference design with some luck but larger ratings were always shown for positive rails and not for negative
So, my thought was to use LM117/337 driving a larger TIP35/36 for each rail.
Any input or ideas are welcome.
You can use an LM338 or an LM1084. If you parallel the LM337 with a Transistor keep in mind that overload protection may not work anymore.
Regards,
Udo
Regards,
Udo
Yes, LM338 or LT/LM1084 should be fine. However, remember to calculate the power dissipation before you decide on heat sinking. You see quite a lot of these regulators on ebay boards etc. with very small heatsinks and then you'd never get anywhere near full output before the regulator shuts down.
/U.
/U.
Thanks guys.
Now the big question is through hole or smd
I mean both have their pros and cons, but SMD seems to becoming more of a standard, however larger SMD components are required for larger power devices.
Also with SMD components i can mount a heatsink below the board a couple it with VIA's for better heat transfer as well as adding a larger copper place to act as a passive radiator.
Now the big question is through hole or smd
I mean both have their pros and cons, but SMD seems to becoming more of a standard, however larger SMD components are required for larger power devices.
Also with SMD components i can mount a heatsink below the board a couple it with VIA's for better heat transfer as well as adding a larger copper place to act as a passive radiator.
I'd say through-hole is the only way you can get enough heatsinking if you plan on drawing 4 amps continuously (otherwise you need a massive PCB).
/U.
/U.
I stumbled across this last night:
Power Supply 1-30V 10A with LM317
I much more prefer this as I can do a mirror image configuration and employ a split configuration ( regulator board and power board ). This way the driver will be SMD as there is very little heat will be produced, and would allow me to mount the power devices on a suitable heatsink. To further reduce the stress on the devices only 2 TIP41/2 outputs are going to be used as even 1 would suffice but for better thermal transfer, 2 devices would run cooler.
An externally hosted image should be here but it was not working when we last tested it.
Power Supply 1-30V 10A with LM317
I much more prefer this as I can do a mirror image configuration and employ a split configuration ( regulator board and power board ). This way the driver will be SMD as there is very little heat will be produced, and would allow me to mount the power devices on a suitable heatsink. To further reduce the stress on the devices only 2 TIP41/2 outputs are going to be used as even 1 would suffice but for better thermal transfer, 2 devices would run cooler.
This is a crude power supply. In this supply cou can substitute a zener diode instead of the LM317 and get similar results.
The LM317 does not regulate the output voltage but regulates the base of the output transistors.
Also you loose the overcurrent and overheating protection of the LM317.
I would not recommend it.
Regards,
Udo
The LM317 does not regulate the output voltage but regulates the base of the output transistors.
Also you loose the overcurrent and overheating protection of the LM317.
I would not recommend it.
Regards,
Udo
Have you calculated the dissipation of the power device(s)?
In any case, that design leaves a lot to be desired.
I just noticed that!
The the ADJ terminal should have been tied to the output of the TIP41/2 ( PAD1 ) that way voltage tracking is there. As for the Over Temperature protection if its mounted on the same heatsink as the output devices it should be OK as it would read the temperature of the heatsink.
The overcurrent protection doesn't matter much as the toroid is rated at 5A for each secondary and VIN is 21VDC ( 15VAC into bridge rectifier ) and there will be a fuse on the A/C input to the rectifier.
Also each device is rated at 6A or 10A peak at 65W Max dissipation.
At the maximum output of the supply ( 5A ) each will see roughly 2.5A across 2 devices OR 1.67A across 3 devices.
Dissipation is would roughly be 26.7W per device with 2 output transistors OR 17.6W per device with 3 output transistors.
Last edited:
I am not totally convinced "audiophile" components would add anything to the quality of a power supply.
This design have the advantage of being a non switched regulator, so the only consideration is to avoid ground loops and have good mains frequency filtering.
- Bosse
This design have the advantage of being a non switched regulator, so the only consideration is to avoid ground loops and have good mains frequency filtering.
- Bosse
I see, you don't want the easy LM338 or LM1084 single chip solution😉
Another suggestion: Stay with the LM317 and use more than one for different section of your curcuit. This may even improve crosstalk and make cooling easier.
The reason behind the 317 / 338 pair is for actual proper regulation of the positive and negative rails rather than trying to make due with 2 positive regulators and hoping for the best noise rejection.
Also I don't like the idea of using the center tap as 0V or GND...
Much rather would use 2 bridge rectifiers and use the positive from one and the negative from the other to create Common / 0V / GND this way AC ripple and noise has less chance to make it through to the output.
As far as the Audio Grade caps go...
They were designed for maximum filtering in the audio band ( 10Hz -25KHz ) and low ESR. yes, some may say its complete overkill, but I DO think they make a huge difference. I was reading through the debate last year and I did some tests of my own and drew my own conclusions. Too bad my scope was damaged last year when my basement flooded.
Either way tests were simple:
A/C -Bridge Rectifier - Main Cap - Filter/ Decoupling Cap ( film )
Used 4 different grades and checked with scope.
The differences between your Generic and High End Audio Cap were Immense !
At idle the "out of band" ( 30KHz + ) noise was much much less as well as under load, but what was impressive is how quickly the cap recharged without any noticeable ( 0.1-0.3V for the Audio Grade VS 0.5 - 0.9V for the General Purpose Cap ) and was able to provide higher current on tap as well as faster recharge time.
So yes they do make a huge difference !
Working on it as we speak...
However its time to take the dog for a walk, as he's getting restless here.
Will post it when I get it done when i get back.
However its time to take the dog for a walk, as he's getting restless here.
Will post it when I get it done when i get back.
High Current 13.8V Power Supply is a pretty straightforward high current build. And adaptable to your needs.
And this is something I built a while back...http://www.diyaudio.com/forums/powe...h-supply-class-regulated-psu-based-lm317.html
Although I did end up building a dual bridge reg version - I could find that I'd you like...
And this is something I built a while back...http://www.diyaudio.com/forums/powe...h-supply-class-regulated-psu-based-lm317.html
Although I did end up building a dual bridge reg version - I could find that I'd you like...
- Status
- Not open for further replies.