will this work? 48/24 dc PSU

Here is a bitmap of a design I am not sure if this transformer wiring will work. The Xformer is 27-0-27. The ground reference is on -27 and the 48V side is tapped +27 to -27. The 24v side is from 0 to -27. Some of the resistor values are not there but will be calculated for the right "fit".

Thanks


http://guavatone.com/images/psu.jpg


PS forgive me for posting another PSU like this. it somewhat like the post on modifying SSL Tech's PSU.

It looks like the voltage doubler's charging currents are being pulled through the lower voltage power supply, which is a bad idea.

Use only half of the lower bridge (leave the "-" terminal open---you would use it in other situations as an output for a negative regulator's bulk cap). Make the low voltage reg common (your "PE" ground symbol) the 0 volt (center tap) of the transformer. Make the "-" terminal of the upper bridge rectifier connect to the transformer center tap as well.

Feed the top connection of the lower bridge from "+27" instead of "0".

Showing the actual transformer would be helpful---when one sees things like +27 and -27 the assumption is that we're dealing with d.c.

Draw a new schematic. Wear safety glasses.

Sorry, the Transformer is 27-0-27 secondary. Doesn't this mean 54v total?

This is the post I found that had this type of setup

http://www.groupdiy.com/index.php?topic=2564&postdays=0&postorder=asc&start=30

They said to use 2 27 taps for 48v and 0-27 for 24v.

Do you mean to do this? I redrew schematic.

http://guavatone.com/images/psu2.jpg

Also, where is the doubler? I thought the top was just a bridge.

Maybe the caps make a charge-pump doubler?

The caps are essential. A normal half-wave doubler would use the upper half of the upper bridge and one side of the transformer relative to the center. This circuit gives full-wave doubling. It's better for a variety of reasons.

The lower section is a full-wave center-tapped bridge, if the other two diodes in the bridge were used for the negative supply. Your new schematic is still full-wave for the low voltage positive unreg supply.

Thanks,
Is the new schematic better for this application?

The new schematic looks fine. Although I just got back from dinner, so you may want a second opinion.

Good thing I wore my safety glasses (or walked into the other room as I pluged in AC.

SMOKINGGGGGGGGGGG Cap after the top bridge rect. hmmm...

it was a 1000 uF/50V cap. All the 1000uF caps are 50V BTW.

Was it too small?

everything was wired as schematic.

Anyway my 24VDC is stable and at 700 mA which is fine for the modules I am powering.

Anyone?


I am sorry to rehash this 24/48V topic but my powerone linear is getting fixed and I have a session. I can post PCBs if anyone want to smoke their Elcaps.

Yeah the cap's voltage was quite a bit small. Check your math on the rectified AC to get your cap's minimum voltage then round up to the next easily available voltage rating. HINT: double their voltage on the caps before the regulators. Check the voltages on the other caps too or you'll have more fishy smell in your lab.

:green:

EDIT:

Ok I was going to leave you to figure it out but I'll continue the lesson :thumb:

your rectified AC will be the total AC voltage swing * 1.35 = DC output.

so assuming your trafo is putting out 54vac then:

54*1.35=73vdc.

Too big for 63v caps, next easily available cap is 100v

Sorry I missed that detail.

Also, with light loading many transformers are considerably higher voltage out than their specs say, so it makes the problem even worse. Maybe your 27VAC was already a measured value though.

100V rated cap is the way to go as suggested by Svart.

Unloaded I have 67 VAC on the full swing of the xformer. Rectified DC on the bottom leg is 45.5 DC. This seems very inefficeint but I will make do until I get a propper xformer and a propper circuit.

Should be able to use the 50V caps on the bottom?

Svart, are you tlking about the top leg doubbler at 73vdc?


Also, once I get into 100v caps the highest is 680uF that I see at mouser. I thought I would want at least 1000uF x2.

647-UVZ2A102MHD

Nichicon VZ 1000uf 100v @ mouser

You *could* use the 50v caps if the DC is 45.5 as caps work a little better near their max voltage. you could add some zener protection for the caps that are running close to max if you wish. As a rule of thumb that I use, if the cap voltage is within 10v of the normal DC level, then move to the next voltage value up. this way if you have any serious fluctuations you won't blow your caps. However, you could likely get away with using all 100v caps if buying them in bulk will be cheaper in the long run..

if not then:

45.5VDC on the bottom portion of the circuit would like to see a 63v cap.

the upper portion around 73 VDC needs a 100V cap.

after the regulators the 48VDC will likely be more stable but without zener protection, you should still use a 63v cap. The regulated 24vdc would use a 35v cap.


:thumb:

Youch! 67V rms means as much as 93.5VDC out of the doubler, without any load. You might take note of your a.c. line voltage in as well, and then budget for a 15% high line condition.

It's true that it doesn't take much of a load to lower the doubler unreg voltage, but why play with fire.

Most 'lytics will shrug off a slight overvoltage, but it's a bad idea to count on that.* I think I would use 63V parts below and 150V above for safety. And, as you say, eventually get a more suitable transformer.



*I did once and it was a train wreck when the purchasing agent could't get the Rubycons that had been assumed and got some dirt-floor jobs instead. After about 10-20 minutes of operation, PAAFFFF! and vapor coming out of the little hole for the on-off switch on the powered speaker, for a fair number of the units. How do you spell crisis?

The question is do I really need a doubler on the top?

maybe a 76V zener?

You won't get to 48V without a doubler, unfortunately.

Should I splurge for Panasonic TS-UP/HA caps. These are the few large caps I see at digikey. I think they are made for PSUs. I can go as high as 4700 to 3300 and use one cap on each side rather than 2.

"67V rms means as much as 93.5VDC"

wow, I didn't catch that, was reading too fast.

yeah you'll need to up those caps to 150v or.. maybe the addition of some kind of led/bulb/resistor power LED circuit to load the circuit down a little to tame it a bit?

enlarging the capacitance isn't going to affect it's voltage rating(in case you didn't know), unless of course you are operating at the max DC level and any spikes could put you over... the voltage rating is the sole value that we are concerned with right now due to obvious fire hazards.

The capacitance is only going to affect the ability to smooth the ripple of the DC and smooth any current/voltage sags created by large(current) drains in/on the system. Usually you see 1000uf for every amp or so of expected draw.

Thaks for everyone's help. I just did a 10 hour mix session with the 24V side of my PSU ( I removed the inout caps for the doubler) running power for 4 Audix EQ/Preamp. Everything seemed to work fine with the 50V filter caps on the 24V Rail. I measured 23.9VDC before the session and I believe its putting out 700mA or so. I thought that the headroom on the modules were a little low but I may have been hitting the input too hard.

Anyway, after this week its back to the drawing board to try to get phantom to these as well. most likely I'll have to get another Xformer, my next design will probably be 48V phantom and maybe a lower voltage tap for relay switches and LEDs.

I was talking cap size because I thought I would want the most filtering I could get.

Where would I put the LED in the circuit to tame the beast?
on

OK these lm317s are hot. I dropped the 48v side. Can I just use half the transformer safely?


I don't know what could happen but right now Something is making these modules behave strangely.

Hot as in REALLY hot? heatsinks? are all of them getting hot or just the one for the phantom? IF you were really getting around 94v out of the doubler into your floating Vreg for the phantom then you are regulating down to 48v meaning UP to 46v dissipation as heat unless the current draw pulls the voltage to the input of the Vreg down considerably.

Here's what we need to know.. we need to know how much current you expect to draw from the 48v, then we need to know how much that load brings the input voltage down on the input of the Vreg. we can then figure your exact dissipation in heat.

as for the 24v line, does it sags considerably with the modules hooked up? Hook up a voltmeter and watch the voltage as you feed various signals through the channels. watch to see if the current draw causes fluctuations in the voltage.

you can't reliably use just one side of the secondary wiring. you'll need to hook those in parallel if you intend to use just one output.