Bauman PSU

Hi all, need someone to check work re: Bauman's +-18/48v/5v PSU, for 3*1*2 etc boards. This is my first DIY attempt, trying to grind out the math so concepts will sink in.

Schem...
http://www.thediypill.phx.com.br/forumfiles/psu312sch_v1.pdf

Trying to calculate value of R1 and find proper V for smooth DC at REG+
i'm using 3300uf for C1/C2
Fabio said in a post 'R1 = 22R if C1/C2 = 2200uf'
...also he wrote if i use 4700uf, R1 would be something like 10-12R

Toroid spec: Amveco 62085
22@ 2.27a, 44@ 1.136a 50Va
No Load current 8.0ma No Load V: 2x25.4 V

Vpeak after D1/D2 = 42.6V (rect drop) at 120hz,
Vavg = 27V, Veff = 29.2V
Iavg= 723ma, Ieff = 792ma
ripple freq = 120hz

Then i read with 'RC filter' after full-wave rectifier, multiply Vavg x 1.4? Is this right? That makes 37.2V..??? ..also there is Vripple equation..

can't figure why i would even want R1 there... then i see RC time constant. Okay, R1 is to make the caps discharge on the same cycle as DC ripple, for max smoothing. Do i have this right?

But what value for t? so i cheated and plugged in Fabio's numbers from before.
t= r x c = 99ms
...then with C1/C2 = 6600uf
r = .099 / .0066 = 14.66 = 15R
This seems to fall in range, given the 2 values above. But why 99ms?

cap reactance (.159/fc) = .20
impedance = 9
ohm's law- circuits with reactance and impedance:
I = E/Z =27/9 = 3a ???? dunno if this formula applies

which value for I to check V at REG+ ?,
plug in 15R with I to get V at REG+... I = lm317 max current = 1.5a
V at reg+ = 15 x 1.5a = 22.5V
that looks good, but wrong i suspect. Iavg x 1.4 = 1.01a
15 x 1.01 = 15.15V.... not so good.
or V= 15 x 3a (ohms law with resist and reactance) = 45V ???
or add impedance and R, 15 + 9= 24 x Iavg = 24.24V

Could someone clarify which values of I and V i should do these equations with? The Ac to Dc transition is throwing me off. Missing a crucial formula? if/why t=99ms? so many questions... this is probably 101 for most of you, but i gotta start somewhere...

...also, on schem, are C11/C12 there to show no load V to toroid? so 48V rail gets (in my case) 50.8V ?

I'm bad at supplies, and I've got Fabio's boards, so I'll throw my question in with this thread.

How would i get down to +/- 15v? I've got some JH 990a 15v amps that were available too cheap to pass up, and I assume since the 18v is a different version, I'd be pushing these a bit hard. Do I just select a different regulator or something. How about moving the voltage up to hear the 992's I have in their full glory?

(All of Fabio's six-pin opamp cards are happiest at about 18v, correct?)

Bear

Bear, from my, as you can see above, limited understanding...

i think you can set the gain of the 317/337s by adjusting the ADJ_POS and ADJ_NEG trimpots. More details on LM317 app sheet... my brand is:
http://www.national.com/ds/LM/LM117HV.pdf
top of p.6, fig3 'lm317 w protection diodes'
...your brand might be (slightly) different... pretty close though, i looked at a few of these...

.. the power trafo should be enough to give you +-18V at REG+in (to allow for some ripple error) which LM317 to give regulated 15V at REGOUT when you set the trimpot.

i'm in the little-knowledge=dangerous stage of learning though... a look at the app sheet might clear it up...

hope this helps

Yes to get +/- 15v instead of +/-18v just set ADJ_POS & ADJ_NEG pots to the voltage you want.

AS for R1 & R2. If your highest peak DC input voltage is under 40v they can be set to 0ohm as the 4400uF per rail is enough smoothing for the LM317 & LM337 to work perfectly. If you are above 40vdc then will need to make R1 & R2 large enough to drop the volts at C3 & C4 to 40vdc at the minimum load you will have on the power supply. The 10 to 22R values that Fabio mentions will give extra smoothing since he is only using less than 40vdc as the C1 to C4 are only 35v types. So your 2 x 25vac no load transformer will be fine.

The tolerance of R7 & R8 even at 1% will move the 48v rail up and down by a volt or two. If you want exactly 48v change R8 to a 10k 25turn trim pot.

The critical resistor to get right is R10 so at minimum load the voltage at the Vin pin of 78** is 35v or less. There should also be a cap at this pin to ground as the 78** will oscillate if this value gets to big. And if the only load is the led while no relays are switched on this resistor will need to be big in value. The 78** should also have the protection diodes like D13 & D14 across it as well for reliability. This Reg is also going to run very hot with any load due to the voltage dropped across it. Because the 78** is run from the pump charge rail which is only smoothed by 200uF the 48v smoothness will be greatly affected by the load on the 78**. It would be better to connect R10 to the join of R1 & C1 as R10 can be smaller in value and the 78** will not affect the 48v rail performance and it will not run as hot.

Not sure what you are going to run from the power supply. So I do not no the minimum load the +/-18v & 78** rails will have on them to give you a better idea on the values of R10, R1 & R2.

Joe
www.jlmaudio.com

It would be better to connect R10 to the join of R1 & C1 as R10 can be smaller in value and the 78** will not affect the 48v rail performance and it will not run as hot.

Click to expand...

Thanks for all the comments Joe! I was not sure about how good the C-R-C could be in comparison to a just C... so I decide to put it in.

About the relay regulator, that was not a so clever solution as you point out, I admit. I´ll connect that to R1 & C1 to keep less heat at it....
I did it also to try some double regulation for some single ended amps,
like Neve...

:guinness:
Fabio

Joe, thanks for the reply.... so i was way off with RC time constant? I wondered why Fabio suggested those specific values...

The +/-18v side of supply will run 4 x melcor/ap pres, in/out trafos each channel. it's gone from my notes but IIRC 150ma per pre..?... Tried for a link just now, Fab's server is down, it's on diypill dot com same as PS schem.
....my C1-C4 are 3300uf...

so for Vdrop across R1 etc, i figure approx. load current and use that for these equations... duhh.....

The 10 to 22R values that Fabio mentions will give extra smoothing since he is only using less than 40vdc as the C1 to C4 are only 35v types. If your highest peak DC input voltage is under 40v they can be set to 0ohm as the 4400uF per rail is enough smoothing for the LM317 & LM337 to work perfectly

Click to expand...

Which formula(s) to figure smoothing effect of caps on rectified DC? Like when you say 'enough smoothing' i would like to calculate that effect. My understanding was that RC network gives higher Vavg after bridge rectifier than just Vpeak x .637. only RC explanation i found said multiply Vavg by 1.4... but this was in context of AC, so don't know if it applies here. suppose cap reactance (i calculated .20 for f=120hz and c = .0066) comes in here somehow..??
also supposing to myself i need to know Vripple in order to check that the lowV end of ripple will still exceed 20V for REG+in, with a given value of R1. That's why the bigger caps for C1-2, right?

... yep, 48V to 5Vrail is sometimes only LED_SW LR. Fabio suggested 470R/5w for R10 a while back. R11 will need to be at least 470R also..? i'm not using the relays on his boards, so current will always be low (12 30ma leds max, maybe 5-6 avg ).

REG48: I'm guessing C11 and C12 are to block DC and show no-load V to toroid, so that D9-12 get 50.8... from there diode drop, reduce ripple a little and try to get at least 49v? or so to REG48... maybe not much room to adjust. (typed that so someone can say otherwise.)

'connect to 5v supply to C1/R1'... i see what you mean, smaller Vdrop, but i think the idea is to keep the 5v sw/led/relay rail off the opamp power rails...? see signature

Fabio, it's you... must have posted while i was writing

anything else i screwed up?

...hmmm... i was so way off... maybe start over...

since toroid 'sees' load at D1/D2, so 44V to D1/D2, then -1.4V diode drop, then take Vavg? This will never be above 40V... even No-load V after diodes (49.4V) x .637 is only 33V. so why R1?
... unless i am oversimplifying for Vin=44V and Vavg ?

which leads to ...info about RC network (i was wrong, it is with regard to DC) :

Note that smoothing significantly increases the average DC voltage to almost the peak value (1.4 × RMS value). For example 6V RMS AC is rectified to full wave DC of about 4.6V RMS (1.4V is lost in the bridge rectifier), with smoothing this increases to almost the peak value giving 1.4 × 4.6 = 6.4V smooth DC.
Smoothing is not perfect due to the capacitor voltage falling a little as it discharges, giving a small ripple voltage. For many circuits a ripple which is 10% of the supply voltage is satisfactory and the equation below gives the required value for the smoothing capacitor. A larger capacitor will give less ripple. The capacitor value must be doubled when smoothing half-wave DC.
Smoothing capacitor for 10% ripple, C = 5 × Io /Vs × f
Io = output current from the supply
Vs = supply voltage (peak value of unsmoothed DC)
f = frequency of the AC supply (hz)

Click to expand...

f=120hz after rectifier
could someone confirm this info? if it's correct i can calculate R1 and correct Vripple value to check max heat dissipation of REG+... and promise to never clutter the boards with this drivel again

thanks Joe, i see now to use current past REG+ for I, so figure equiv circuit.

opamp ma draw varies w/ out ohms and V, estimate 300ma max draw for 4ch. (2ch melcor 1731/ 400-17-2 out trafos, 2 25twentyBC / profile 4804). 2*5*2*0 sheet:
http://www.waltzingbear.com/Schematics/API/API_2520.htm

i know you all are short on DIY time, maybe just a yes or no to RC-avg equation? 1000 thanks.

since toroid 'sees' load at D1/D2, so 44V to D1/D2, then -1.4V diode drop, then take Vavg? This will never be above 40V... even No-load V after diodes (49.4V) x .637 is only 33V. so why R1?

Click to expand...

Yes 33v volts unregulated is what you will get. R1 in this case is not needed to get the volts below 40v and can be set to 0ohm but can still be used to smooth the supply for the reg more and as extra protection for the reg if the reg output is shorted. Since you need 18v and the the reg will need 4 to 5v to keep regulating we can waste at full load of 300mA.

In a simplified form

33V - (18V + 5V) = 10volts
10V/.3A = 33.33ohm
10v * .3A = 3W

So a 33ohm 3w to 5w resistor for R1 & R2 will work fine and give a smoother supply at C3 & C4 for the regulators to work from.

Joe
www.jlmaudio.com

thanks Joe, you make all that beautiful gear and still have time to straighten me out... so simple but i have 8 pages of scribbles here.

anyway i really appreciate it.

Very good info here...

I built this supply a couple weeks ago, but still have not connected any load to it. Luckily, my 48VDC was right on just using the 1% resistors. I think I will also connect R10 to the R1/C1 junction before loading for reasons already mentioned.

Cool Greg, glad the thread helped someone besides me.

After looking at it last night, I'm gonna try the same R10 to C1/R1 mod on at least one of the PS boards (have 2 here). R10 trace seems small/easy enough to remove...
If convenient, could you send me a pic of your mod when done? i'll do the same if you're interested.

I doubt I'll have to cut any traces. I used the 5W resistor for R10, so the legs are rather stiff. I'm probably just going lift the leg of R10 and connect that to the C1/R1 junction via wire. Or you could cut the trace and put a wire from C1/R1 junction to the R10 solder joint.

Either way, I'll post a picture when I'm done. I don't think I'll get a chance to work on this until the weekend... I have a GSSL to finish up :thumb:.

cool Greg, thanks.

Or you could cut the trace and put a wire from C1/R1 junction to the R10 solder joint.

Click to expand...

...that's what i'll try...will post pic also...

Hi
I have one 20VA toroid with 2x21,3V voltages off-load.
Can I use this for Bauman PSU?
Duka

Should work fine Duka! I always use 24V AC when I need regulated 24V DC so 21V AC for 18V DC will be fine, think the phantom will work fine to!

[quote author="Tekay"]Should work fine Duka! I always use 24V AC when I need regulated 24V DC so 21V AC for 18V DC will be fine, think the phantom will work fine to![/quote]
Thanks Thomas
Did you used 24VAC off-load or on-load?
Duka

Neve check the voltage unloaded, probably little higher.
Connect a 24V/10w lightbulb to one of the outputs and then measure the voltage, must be the easiest way to do that?

I moved R10 to the C1 junction... didn't cut any traces. I just lifted on leg and used a wire... worked great !!! :thumb:

[quote author="Tekay"]Neve check the voltage unloaded, probably little higher.
Connect a 24V/10w lightbulb to one of the outputs and then measure the voltage, must be the easiest way to do that?[/quote]
Thanks!
Is it enought for transformer 24V than be about 10-15% higher voltage off-load?
Duka