5 rails power supply schematic check

[beatnik]

Hi everybody

Can someone check this schematic for a 5 rails power supply please?

 

[Samuel Groner]

Just two notes after a quick look:

* I'd replace D10 with a 30 V/5 W zener to protect the regulator from overvoltage if the output is shorted.
* It is IMO not clever to make the trimmers cover such a large voltage range for various reasons. Set the rough gain with a fixed resistor and reduce the trimmer value to what's actually needed to catch the regulator tolerance. In fact I'd just skip the trimmers--in very few applications you need more than 5% accuracy for the DC rails. Trimmers are expensive and can fail.

Samuel

 

[Harpo]

This has been covered many times before. Did you already try the SEARCH button?

Some remarks/questions:
- what is the transformers secondary voltage?
- what is the expected load current per rail?
- B1 and B2 rating is missing
- caps rating is missing
- SW1 should be replaced by a wire link or jumper, in your case probably always set for voltage doubler position, to prevent a toggling between doubler/trippler feed
- a single 470uF cap instead of your 500uF series caps C13+C16 with balancing resistors might be more efficient and less space consuming, depending on required rating
- IC5 should be a LM317-HV to give it a fighting chance to survive a shorted 48V rail, a TL783 might be a better fit. At least put a maybe 30V/2W zener in parallel to D10 to clamp it for the max.differential voltage between regulator i/o for a time limited -not permanent- short.
- maybe personal preference, but I'd exchange the trimmers A and E lead ends for a CW increase in output voltage
- trimmer R9 would better be a 1W rated part in this config
- if your target voltages +/-15V, +/-18V, +48V are nearly fixed (you probably never trim these down to 1.25V and fight the generated heat), I'd split the voltage setting resistors (your trimmers) to a fixed value resistor, setting the min.output voltage and a trimmer in series to this resistor for a variation in output voltage of maybe +/-10%. Your 5K trimmers for +/-15V and +/-18V should be sized down anyway to max.2K (giving out 22V if there is sufficient voltage in front of your regulators), or some 100R in series with prementioned fixed resistor. 

 

[beatnik]

i tried to search for 5 rails power supply but didn't find many informations. with such helpful replies i made a good choice starting a new thread!
(thanks  )

- the switch will be only a wire jumper, i used a switch in the schematic because it has a suitable package for the layout
- i forgot to tell that the main use for this psu will be for powering quad ssl9k preamp, that needs +48, +/-15 and +/-18
- so i was thinking about using a 2x18V secondary and setting the jumper as tripler.. am i right? about the current draw i'm not sure, but if i understood the datasheets correctly, the TL052 has about 5mA supply current, as the NN5534, but i don't know the current for the LM394.
I definitely need your help for understanding the expected load current, and in choosing the right power transformer VA rating.
- B1 and B2 are missing values because i don't know exactely what to use. I thinked about a 4A 200V just because i founded it on another project, but let me know tips for choosing the right one
- caps rating are 50V for 1000uF and 63V for 10uF... correct?
- i will change the trimmer configuration for fixed resistor + 100R trimmer, or maybe leave only the fixed resistor as both of you suggested
- would the zener make the LM317 safe enough, or is better to swap for a TL783 in any case?
- i didn't understand the C13/C16 modification, please can you explain?

 

[beatnik]

please don't leave me behind 

 

[signalflow]

Are you making this for the sole purpose of powering your ssl9k?
If so you may find it more cost effective to get a jlm powerstation.
It does exactly what you need.  Just a thought.

-Casey

 

[lernith]

I might be able to help a bit--- (?)

C13 and C16 are in series, which lets you use a lower voltage rating but it halves their combined capacitance.  There might be a good reason to use an arrangement like that at 48v but my limited and unsubtle knowledge couldn't say.  Even 450v caps are usually easy to get (depending on what land mass you're on).  I think Harpo was saying: instead of C13, C16, R13, and R14, just have a single 470uF cap.  Or you can go pi-crazy and have your 1000uF cap, a 22Ohm/1W resister in series, then your other 1000uF cap.

"caps rating are 50V for 1000uF and 63V for 10uF... correct?"

Your highest output voltage is 48v, your regulators will want a few volts to drop, so your unregulated DC will need to be at least 50v.  So your caps (all of them, pre-regulators) should be maybe 100v, just to account for voltage spikes on the grid.  This over specification should also give the caps a longer life.  Someone with more practical experience please refute if they see fit.
The other 10uF caps should be ~1.5-times the voltage that the regulator upstream is putting out.  This voltage should be more stable.

I'm not too cozy with the ssl9k....

Also, did you only search for "5 rails power supply"?  Because that's pretty specific.  Too specific for a web search.  Broken down, this is a "+/- 15v PSU" and a "+/- 18v PSU" and a "Phantom power PSU". 

You might want to use that 2x18v secondary transformer with the secondaries in parallel to cover the +/-15v and +/-18v regulators and figure out a voltage tripler (triple'er?!) for the 48v (the Green Pre META has plenty of tips and schematics on those).

Good luck, keep reading, keep creating.

 

[beatnik]

Are you making this for the sole purpose of powering your ssl9k?
If so you may find it more cost effective to get a jlm powerstation.
It does exactly what you need.  Just a thought.

-Casey

The first purpose is to power the ssl, i know a jlm works perfectly in this application, but i prefer to self etch my boards when possible. and put my layouts available for download. that's the spirit of diy, in my opinion...

I might be able to help a bit--- (?)

C13 and C16 are in series, which lets you use a lower voltage rating but it halves their combined capacitance.  There might be a good reason to use an arrangement like that at 48v but my limited and unsubtle knowledge couldn't say.  Even 450v caps are usually easy to get (depending on what land mass you're on).  I think Harpo was saying: instead of C13, C16, R13, and R14, just have a single 470uF cap.  Or you can go pi-crazy and have your 1000uF cap, a 22Ohm/1W resister in series, then your other 1000uF cap.

I'll try to put a new schematic to discuss this arrangement, i agree the 48V is the most critical part of this circuit

Ok for the capacitor voltage, you definitely answered my question.

I'll make more search and come up with new ideas for the circuit. stay tuned!

thank you all!

 

[helterbelter]

In case of the 9k pre:
I used zener diodes for the 15v rails. 2 per board. Works great. I used Keith's PSU, but you could also use other triple rail PSU's like Peter's green PSU.

But I do respect your wish to etch your own. (thumbs up!)