Help with my power supply schematic...

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

T-Dogg

Well-known member
Joined
Jun 11, 2004
Messages
81
Location
NJ, US of A
Can someone take a look at the link below? It's a quick schematic I drew up, based on (probably misinterpreted) info from the first 50 pages of the Art of Electronics... LOL. Truth is, I know it's not right, but I thought it would help me understand things a bit better if I showed some effort and took a stab at it, in hopes that some of you folks could tell me where I'm going wrong...

I'm trying to wind up with the following voltages, to be used as a power supply for two Calrec PQ1347's:

+16v
-16v
0v
+5v
+8v
+48v

I'm not even worrying about the phantom yet, and I'm going under the assumption that +/-18v will work instead of +/-16(?)... I'm trying to get the lower voltages using simple voltage dividers -- but gotta admit I'm not really sure how it affects the the higher voltage taps I'm taking them off of. So, if y'all don't mind... learn me!!! Tell me what I should do and why. Thanks!!!

http://redbuoy12.com/Images/CalrecPowerSupply.bmp
 
My 2 cents...

I'd use something larger than 1000 uF for the filter cap, and bypass them with a decent film or NPO ceramic capacitor.

Unless the current "draw" for the 8 and 5 VDC rails is tiny (and constant), passive dropping resistors won't work correctly.

Bri
 
On the regulated (output) side of the 3 terminal regulators, install .1 and 1 uf caps to ground. Keeps em from oscillating.

Also, use a 7808 and 7805 downstream from the 7818 unless, like bri says, its super low current. I would anyway.
 
Hiya Larry! Good point re. the output bypass caps on the regulators...I missed that point.

Dogg...look at the "app notes" for the regulators, and also note the reverse-biased diode commonly suggested (from output to input) typically shown in the app notes.

For the "derived" 8 and 5 Volt rails, I strongly suggest adding a "crowbar" overvoltage protection circuit as well.

Bri
 
[quote author="Larrchild"]On the regulated (output) side of the 3 terminal regulators, install .1 and 1 uf caps to ground. Keeps em from oscillating.[/quote]
Or it may make them oscillate. They are less likely to oscillate if you only put electrolytics at the inputs and outputs. Anything from 10-47µF is fine at the output.

I don't remember the exact reason, but somebody made a very throrough analysis of the stability, and came to the above conclusion. It has worked fine for me...

Best regards,

Mikkel C. Simonsen
 
[quote author="mcs"][quote author="Larrchild"]On the regulated (output) side of the 3 terminal regulators, install .1 and 1 uf caps to ground. Keeps em from oscillating.[/quote]
Or it may make them oscillate. They are less likely to oscillate if you only put electrolytics at the inputs and outputs. Anything from 10-47µF is fine at the output.

I don't remember the exact reason, but somebody made a very throrough analysis of the stability, and came to the above conclusion. It has worked fine for me...

Best regards,

Mikkel C. Simonsen[/quote]

The standard three-term fixed regs are usually pretty tolerant in my experience. But as mentioned elsewhere some of the "low dropout" ones are very prone to oscillation if the cap is either too low-ESR or too high-ESR. Also, the TL431 shunt voltage regulator has an elaborate chart of permissible shunt capacitance vs current and stability---quite tricky.
 
:grin: Thanks everyone!!! Very helpful, I'll try to implement everyone's input tonight. But, now I have even more questions!!! I'll try to research some o' this tonight too, but if anyone feel's like jumpin in:

--With regard to larger filter caps, is there a general rule for determining what is appropriate?

--So to avoid oscillation, it's possible I could bypass with a cap after every regulator -- or I could use on cap to ground right before the end of each voltage tap based on what MCS said? Maybe I'll try both when I breadboard it up. If I bypass every reg, why use two caps (.1uf and 1uf)? Do these values change depending on current or anything?

--Brian, you said to bypass the filter caps? Like, with another cap to ground right after? What does this do (I believe ya, just curious!) Is value important?

--Won't be a need for a "crowbar circuit" provided I use voltage regs on the smaller voltage taps, right? Also, do I want to re-adjust my voltage dividers to feed each regulator with slightly more than the voltage I'll be getting out of them? (thought I read something about this somewhere)

--What exactly is the net result of using the +18v tap to generate my 8v and 5v taps? I just pulls more current through? Doesn't pull the voltage on the 18v tap down or anything?
 
yeah, as usual, these guys are right, not me, lol...Bcarso was talking once about a national 3 terminal reg that needed just the right ESR on that cap. Too hi or low and she took off like an F-16.

Ditch the .1 idea and use 10-47uf like the grownups reccomend. And that backwards diode across in and out is good too.

i'm out.
 
> filter caps, is there a general rule

Start with 1,000 uFd per Amp of load current. In DIY, at transistor voltages, there is rarely an excuse to go less than 1,000uFd even if the load is 0.010A (10mA): caps are too inexpensive to cheap-out on.

More is generally better. I've used 4,700uFd for a 60mA load. (But this was without active regulation: three 4,700uFd with some resistors gave me DC clean enough to eat off of.)

With plenty-big caps, the rectifiers MUST be rated for the full transformer current, even if you are not actually using all that current, because the peaks are limited only by transformer losses. Since a blown diode can cause a cascade of catastrophe, double the tranny rating is wise.
 
doubling the current capability of the trafo will cut certain types of noise too. you can tweak a lot of things in the PSU.. check some of the other threads for details.
 
Thanks for the info.

Just a quick question - Are there any good tricks for determining if a regualtor is oscilating? I am assuming that the oscillation is usually above the range of the standard scope used for audio DIY (mine's 20MHZ).

Happy Thanksgiving!
 
Ok, I've finally digested all this damn turkey and can move again... I'd like to redraw that power supply schematic and post it later with additions(same thing, just to get some critiquing from the group...)

One thing I've done is replace all the passive voltage dividers I was using with regulators for the 8v and 5v taps... Looking at data sheets for these 78XX regulators, it seems like they can all accept 35v-40v MAX for V_in... So I've simply taken parallel feeds from my rectified +18v line (right after the FWB, before the LM7818) and used that as the input to all my regulators... Is it that easy? No need to calculate damn passive voltage dividers to get V_in closer to my target V_out?
 
I've posted a newer version of that power supply schematic... Really, I'm curious to see if the regulators will function properly (assuming they're rated for 35v-40v maximum input) as I have it set up... I've been foolin around with simulator software -- seems like it'll work although my understanding of sim software is even worse than my understanding of electronics... I tried to incorporate all everyone's feedback, so if ya get a chance check out the link below and let me know if ya think it's a solid working power supply. Thanks again for all the help!

...next I gotta figure out a 48v solution...

http://redbuoy12.com/Images/calrecPS2.bmp

Thanx guys!
 
hey that is pretty much a *regular* differential PSU arrangement. Will work fine for most everything.

Check SSLtech's PSU for the 9k preamps for more ideas. his arrangement is similar with 48v output also.
 
Thanks Svart-- I'll check out the G9k power supply to see how the 48v is generated. This is cool... If this is the basic idea, I think I understand enough to build up a power supply for any low voltage solid state pre, comp, eq, etc... At its simplest implementation, its not too hard after all. I know it's basic stuff, but the sh'it I've learned on this forum just amazes me. Now I have a GSSL, a working Ampeg amp, and soon some racked Calrec's, all because of the Lab.

I went out and picked up all the parts on Tuesday, and should have the PSU finished tonight. I used a pre-drilled Eurocard, and linked everything with wires and solder links on the solder side of the board. Did something stupid with the layout, mounting the regulators in the wrong direction, which complicated my wiring traces a bit. It'll work, but I'd like to refine the way I'm doing this in general... The wiring I used was a bit too thick, and for real small links the wire insulation seems to heat too easily... Probably should've used all hard links, I just thought the insulation might be an extra safegaurd against shorting... Used molex connectors for the voltage tap outs and for the incoming AC, but probably should've used something more heavy duty for the AC off the tranny... Ah, live and learn!!!

Without a pre-fabbed PCB, what could I use instead of one of those general pre-drilled eurocard things? Anything cheaper? Perfboard doesn't hold onto the components so well...
 
It might be time to invest in some photoetching equipment! It makes higher quality PCBs very easy to make!

Not to blow your mind just yet but as you get more into PSUs you will start to see more interesting things like using adjustable regulators and discrete regulators which reduce rail noise and so on.

getting a solid PSU is KEY to having a good working system and unfortunately too many people tend to overlook this in the production industry. Hell I have a Tascam MX2424 that must have the bare minimum switching PSU because if i add more than 1 drive to it it tends to drag the rails down under heavy use and eventually locks up..

There are a few really good threads on PSU design and improvements floating around. do a search, there is more than enough info there to keep you reading for days on end!

:thumb:
 
[quote author="T-Dogg"]I've posted a newer version of that power supply schematic... Really, I'm curious to see if the regulators will function properly (assuming they're rated for 35v-40v maximum input) as I have it set up... I've been foolin around with simulator software -- seems like it'll work although my understanding of sim software is even worse than my understanding of electronics... I tried to incorporate all everyone's feedback, so if ya get a chance check out the link below and let me know if ya think it's a solid working power supply. Thanks again for all the help!

...next I gotta figure out a 48v solution...

http://redbuoy12.com/Images/calrecPS2.bmp

Thanx guys![/quote]

Polarity of C7 is backwards in the schemo.

I would SERIOUSLY suggest adding a crowbar overvoltage protection circuit for the 8 and 5 V rails.

Bri
 
Oh shiziznit. Thanks for C7 Brian, missed that! I was just about to fire this up today, and figured I'd check in before doing so. Glad I did.

I left out the crowbar circuits thinkin the regulators had some kinda internal protection circuitry built in. But I'll put them in there. Gotta read up on it first -- zener diode with breakdown at 9v (or 6v) plus a SCR... What do I want to short when the gate triggers? The regulated rail output to ground?
 
Tim, 7xxx regulators have short protection, but NO overvoltage limiters.

I've seen crowbars connected in a number of ways. My GUT instinct for best reliability would be to install it ahead of the regulator (with a fuse on the rail line, of course).

You have to size the SCR (and the accompanying series resistor) to handle the DUMPING of the charge in the main filter lytic when the SCR is triggered. But, SCRs with large current capacity are typically inexpensive compared to the downstream components in the actual gear which will fry with an excessive overvoltage.

That's what concerns me about your design, since the 5 and 8 V rails are fed from a MUCH higher "raw" rail. If a regulator fails, that "raw" DC will appear on the 5 or 8 V line and cause Magic Smoke to be released inside of the gear.

Bri
 

Latest posts

Back
Top