converting KPS1 (SSLTech's) PSU for +24, +48, w/no neg. rail

[kato]

Seems there are two projects here that use +24v with no negative rail - the Neve and Fetboy.

Normally I would use a bipolar supply like PeterC's, stuff all components, even for the negative rail, and simply don't hook up the negative side. Then I saw Fucanay's post Using Green PSU as Fetboy PSU and I like the idea of using a simplified design when negative rail is not needed. I'm wondering about modifying Keef's trusty PSU ±18v design to do 24v.

* Is it possible to use an adjustable regulator in place of the 7818?

* Can I just lob off the negative rail components?

* Will this cause smoke or splattered electrolyte?

I'm an untrained hobbiest hacking up a schematic with photoshop so I'd appreciate any comment from more experienced builders, especially if it will save me from blowing up some parts.

Original working schematic. ±18v, fixed regulator:

Hacked schematic. +24v, adjustable regulator:

[similar question]

 

[rodabod]

Alternatively, you could just use a 7924 regulator. Remember to select an appropriate transformer so that your rectified DC is always above the regulation voltage of 24V.

 

[3nity]

This is a good idea!
I think theres a lot of fetboys and 1290 out there but not a proper psu!
I think the best thing is a 2 x LM317 psu!

I dont really trust the 7XXX regulators!

I really like Matt (fucanay) idea about using a cheap 12-0-12V transformer for a 24V then regulating it to +24V and trippling with some Diodes to +48V.

But i would like to redesign a board for that purpose only!

I'm into it since i'm putting together 2 Fets and 2 Spectrasonics as well as 2 x 1290!!

Jorge.

 

[fucanay]

Let me clarify, that it was MattA, not me, that originally did that fetboy PSU and he shared that with the group a long time ago and I just brought it to the forefront because I wasn't understanding it and had problems with it.

That said, I don't think what you're trying to do Keith's board is all that different than what Matt did with Peters PSU board, except that with Peters, it was already designed for the LM317 and on Keith's you need a second bridge rectifier.

I too would like to see a ready made PSU board like this, but with some additions to it. Like for instance, a place to put in a large watt resistor to soak up some over voltage before the regulator if the transformer is a bit over spec. What rodabod said about selecting the right transformer is important, but in my case, it wasn't so easy because of every transformer I bought being over the stated voltage output. My 12-0-12 actually put out about 29 volts when wired in series. Rectify that and you have a little over 40VDC and then you have to regulate down 16 volts. I put 20ohms worth of resistors after the rectifier to soak that up, but it's wasted energy at that point.

Matt

 

[jdbakker]

Do not use the hacked schematic!

As drawn, the leftmost two diodes of the bottom diode bridge will short the negative half of the secondary winding (during each cycle) to its center tap! This is generally not what you want.

If you take SSLTech's schematic and just leave out those diodes and the negative regulator (ie only use the rightmost two diodes of the bottom bridge instead of the whole bridge) you should be set, although a 20-0-20V transformer might be a bit on the low side for +24V regulated out, depeiding on line conditions.

JD 'Danger, Will Robinson!' B.

 

[Harpo]

while you're at it, also change the connection of pin 1 and the 4368R voltage setting resistor in the +24V part, wired the same way as in the +48V part.

If you only need 0V/+24V/+48V and want to stick to a center tapped transformer, you may exchange the 7818 and 7918 with a 7824 and 7924 regulator, remove the +48V part and remove the labeling. The previous labled -18V out will become the 0V, previous 0V will become +24V and previous +18V will become +48V. Yust make shure, there is no connection from previously labled 0V to gnd or the case. Should work, but can lead to trouble if anybody has to fix it one day.

A maybe better approach could be a dual secondary transformer, using one winding for +24V regulation and the other/separate winding to a voltage doubler and 783 regulator (or 317 regulator with some additional zener/diode safety).
just my 2ct.

 

[kato]

Nice suggestions. I would like to not get my face blown off so I appreciate the "Danger, Will Robinson!"-type responses.

I was unsure of how to do the ground rail/0v connections and based it on my memory of this thread about converting a bipolar schematic to monopolar. It suddenly dawns on me that you would never use a center tap for a PSU without a negative side (after reading Harpo's comment.) Am I right about this? You would always wire the secondary in series or parallel. Referring to Mark Burnley's transformer hookup diagram.

[quote author="JD"]
As drawn, the leftmost two diodes of the bottom diode bridge will short the negative half of the secondary winding (during each cycle) to its center tap! This is generally not what you want.
[/quote]

Thanks!!


-----

[quote author="Harpo"]
while you're at it, also change the connection of pin 1 and the 4368R voltage setting resistor in the +24V part, wired the same way as in the +48V part.

If you only need 0V/+24V/+48V and want to stick to a center tapped transformer, you may exchange the 7818 and 7918 with a 7824 and 7924 regulator, remove the +48V part and remove the labeling. The previous labled -18V out will become the 0V, previous 0V will become +24V and previous +18V will become +48V. Yust make shure, there is no connection from previously labled 0V to gnd or the case. Should work, but can lead to trouble if anybody has to fix it one day.

A maybe better approach could be a dual secondary transformer, using one winding for +24V regulation and the other/separate winding to a voltage doubler and 783 regulator (or 317 regulator with some additional zener/diode safety).
just my 2ct.
[/quote]

Thanks for these suggestions - this gives me a lot of good suggestions, and got me using my brain about how the secondary should be wired.

[quote author="Fucanay"]
Let me clarify, that it was MattA, not me, that originally did that fetboy PSU and he shared that with the group a long time ago and I just brought it to the forefront because I wasn't understanding it and had problems with it.
[/quote]

Thanks for the clarification. I was previously referring to that PSU as the "Fucanay/PeterC Fetboy PSU." I missed the original post from MattA.

[quote author="Fucanay"]
That said, I don't think what you're trying to do Keith's board is all that different than what Matt did with Peters PSU board, except that with Peters, it was already designed for the LM317 and on Keith's you need a second bridge rectifier.
[/quote]

That's right, I'm trying to do the exact same thing. But I thought this would be easier because I was getting lost in the other schematic and wasn't sure it was a working schematic. But I suppose the LM317 and 1 less bridge could actually make the PeterC-based one simpler. Really I just think it's fun and interesting.


[quote author="Fucanay"]
I too would like to see a ready made PSU board like this, but with some additions to it. Like for instance, a place to put in a large watt resistor to soak up some over voltage before the regulator if the transformer is a bit over spec. What rodabod said about selecting the right transformer is important, but in my case, it wasn't so easy because of every transformer I bought being over the stated voltage output. My 12-0-12 actually put out about 29 volts when wired in series. Rectify that and you have a little over 40VDC and then you have to regulate down 16 volts. I put 20ohms worth of resistors after the rectifier to soak that up, but it's wasted energy at that point.
[/quote]

Agreed. I've encountered the same thing, as I noted in the other thread: overvoltage on many commercially available torroids.

I'll rework the schematic based on everyone's comments.
Is there any reason to keep the center tap for this application?
I'm going to remove it if there's no reason for it. Although I'm still confused about what "0V" on a power supply board would be without it.

 

[fucanay]

[quote author="kato"] Although I'm still confused about what "0V" on a power supply board would be without it.[/quote]

Re-read through the fetboy PSU thread. I didn't get it either, but Ptownkid filled in the gaps for me. Basically, with the CT transformer wired in series, it becomes 0V and +24V because there is the potential for +24V between the two wires.

Matt

 

[kato]

[quote author="fucanay"]

Re-read through the fetboy PSU thread. I didn't get it either, but Ptownkid filled in the gaps for me. Basically, with the CT transformer wired in series, it becomes 0V and +24V because there is the potential for +24V between the two wires.

[/quote]

OK, that makes sense.
I'm going to re-read the thread a few more times to let it sink in.

Version 2:
- removed center tap, moved LM317 adjusting resistors.
- still need to implement an optional voltage dropping resistor before the regulator, where? directly before?
Again, any comments are appreciated.

 

[3nity]

What about this?
http://w1.316.telia.com/~u31617586/API psu.jpg

Changing the 7818 to 7824!

Thanks!

 

[MartyMart]

Keith,s PSU schem became the 9k power board that Gustav supplies with
each single SSL 9k order.

I've used at least 5 of these now for other projects, by simply removing all
the 48v section I have either :

+/- 24v using 7824/7924
+/- 18v using 7818/7918
+/- 15v using 7815/7915

Also made a few single supplies which include the 48v but remove the
79** section for just a positive supply.

I have not had a single problem with this.
Obviously use the correct transformer for the voltage you require.

MM.

EDIT : Here's what I have on the board for a 48v and single +24v supply
REF: Gustav's 9k PSU board layout for parts :

Both BR's installed
Both 100uf 100v caps installed
Positive rail , 100n 63v poly cap/1000uf 35v cap/LM7824/final 100n 63v poly cap installed
48v rail , 220uf 100v cap/100n 63v poly cap/ 1N4006/LM317T/8k2/240 ohm and
final 100n 63v poly cap installed.

So on the board I'm only missing C2 100n/ C7 1000uf/ IC2 (79**) and C9 100n

 

[rodabod]

[quote author="rodabod"]Alternatively, you could just use a 7924 regulator. Remember to select an appropriate transformer so that your rectified DC is always above the regulation voltage of 24V.[/quote]

Excuse the typo - I thought you meant -24V. Should be 7824.

If you have the SSL9K board, then just go with it. Easy-peasy.

 

[kato]

[quote author="3nity"]What about this?
http://w1.316.telia.com/~u31617586/API psu.jpg

Changing the 7818 to 7824!
[/quote]

Interesting 48v section, from a wall-wart! Added to my collection of schematics to study...

[quote author="Rodabod"]
Alternatively, you could just use a 7924 regulator. Remember to select an appropriate transformer so that your rectified DC is always above the regulation voltage of 24V.
[/quote]

Good suggestion - that was my first impulse. But I like the adjustable reg to add a trimpot to dial in the voltage exactly - or change the voltage for other uses.

Version 3:
implementing JD's suggestion to remove half of diode bridge 1 - removing the center tap may not be enough because I still have the minus sides (-) of both bridges connected which may be a problem.

[quote author="MM"]
. . .
So on the board I'm only missing C2 100n/ C7 1000uf/ IC2 (79**) and C9 100n
[/quote]

Thanks MartyMart!!
That's an enormous help because it tells me (without C2, C7, or C9) the minus side of BR1 is not connected to anything - but it's still working for the positive voltages. I wonder if it's beneficial in some way to have the full bridge in there even if the negative side isn't connected to anything. (I suspect not.)

Version 4:
BR1 intact, minus side not connected to anything

I suspect #3 is the most correct. Thanks everyone for all the help.
I'll report back if anything explodes. Additional comments are always encouraged.

 

[MartyMart]

[quote author="kato"]
Thanks MartyMart!!
That's an enormous help because it tells me (without C2, C7, or C9) the minus side of BR1 is not connected to anything - but it's still working for the positive voltages. I wonder if it's beneficial in some way to have the full bridge in there even if the negative side isn't connected to anything. (I suspect not.)
[/quote]

The BR doesn't seem to mind that missing connection.
Perhaps it's technically NOT good form but someone much brighter would
have to confirm that !!
Kept the centre tap on all my versions BTW

MM.

 

[jdbakker]

[quote author="kato"]Version 3:
implementing JD's suggestion to remove half of diode bridge 1 - removing the center tap may not be enough because I still have the minus sides (-) of both bridges connected which may be a problem.

[/quote]

You'll still need the center tap here. If you don't want to use a transformer with center tap (or you don't have one handy) you can use the Bo Hansén schematic Trinity linked to, although having a half-wave rectifier it might have more ripple.

JDB.

EDIT: This link discusses some of the principles behind the +48V-generating circuit.

 

[kato]

I finally gave up on this - but not without designing, etching, and drilling a PCB based on ver. 4.

I didn't bother to stuff the board so as not to waste the $8 in parts. After re-reading JD's post's it's clear the center tap is an integral part of a bipolar supply and you can't realize the full power of the transformer by just eliminating the negative rail components.

Also, this thread adds further confusion:
http://www.groupdiy.com/index.php?topic=2564
(another person attempting the same mod)

It all makes me wonder, what if the 48v and 24v were on separate boards with their ground tied together?

For now, I've moved onto other things and will revisit this project when I gain more experience. Thanks again everyone for your contributions.

 

[kato]

A google search brought up Bill Bowden's adjustable 317 based PSU:

Obviously, two of these boards would work - one for 24v and one for 48v.
IF there were two separate transformers.

Would sharing a secondary between both supplies cause a problem - any more than running an extension cable for multiple 110v AC sources?

 

[Svart]

I have a schematic that I did some time ago that would work for you. It's something like a hash between all of these designs.

It voltage doubles and zener regulates like Bo Hansen's. It also uses zener regulation for the 24v too.

As with any partial rectification, the caps take a BEATING and will dissipate significant heat. This will dry them out and eventually they will fail, very similarly to how the bulk caps fail in SMPS units. Under load I simulated 1W+ coming from the bulk caps after rectification...

 

[Di_Sousa]

I finally gave up on this - but not without designing, etching, and drilling a PCB based on ver. 4.

For now, I've moved onto other things and will revisit this project when I gain more experience.  Thanks again everyone for your contributions.

Hi Kato,
you should be back on this project. It was a great idea in the first place, so it deserves probably a bit more effort from the forum.

those Gustav PSU boards are quite nice, being able to use them for other projects would be a practical and affordable solution.

best regards

N

 

[kato]

Well, you'll be happy to hear you can accomplish what you want using the KPS1 / SSL9k boards from Gustav.
(See the parts Marty left off above, and use the 7824 fixed regulator for +24 v.)

My pursuit was purely academic / learning experience rather than practical, because I like the idea of *fewer parts* i.e. no unpopulated spaces in a board. And also, using the full available current, without throwing away the negative rail (if I understood correctly.)

I just wanted to get up and running with my Neve and Fetboy projects without wasting any more time with a PSU beyond my level of understanding. So I just used the HT-DCPS I had onhand, stuffed all the parts, and left the negative side disconnected.

But I may come back to it one day just for fun, stuff my self-etched board, watch it explode, or maybe try a dual-version of the bill bowden schem I posted - except what's with the 22,000uF cap?? That seems insane. Good luck!