Transformer math

[Axtman]

How can I determine the ac center-tapped voltages from the transformer secondaries that will produce the dc voltages shown in this schematic?

[mawyatt]

C and D need to be swapped voltage wise!!

Best

[Axtman]

Wow! I never saw that. You have great observational skills!

[Benta]

Well, the theoretical peak voltage is nominal secondary voltage x sqrt(2) which is around 1.41.
Adding in losses, 1.35 will be about right.
I'd choose 45-0-45 V for the upper secondary and 60-0-60 V for the lower one.
It's unregulated anyway, so it only has to be ballpark, no need for great precision.

[ledtester]

There is this design guidance from Hammond:

https://www.hammfg.com/electronics/transformers/rectifier

[Simmed]

hmmm
so its suppose to be 60v rail ?
but feeds a 63v capacitor ?
only 3v margin ?
 
then series regulation down to 15v ?
 

[Konkedout]

Well that is not all that is WRONG  Amended..see below.

Q1 and Q2 are connected as an "Oh, My Darling!!" ton.  :-)  So far, so good.

How are Q3 and Q4 connected?  I think it is supposed to be another Darlington, complementary to the other one.  But...I think it is more insultory than complementary.  Look at it!!

EDIT:  Well Q3 and Q4 are NOT a Darlington, but I guess it might work. 

1) Be careful about oscillation.

2) As for transformer turns, this is tricky and probably works best to simulate.  Transformer winding resistance and leakage inductance will have a big effect on the available DC voltage.

3) 1N966 can probably work OK; I like TL431 adjustable for better accuracy.  Yes this is MORE prone to oscillation but that can be avoided with proper compensation.

[Ian.M]

Well that is not all that is WRONG  Amended..see below.

Q1 and Q2 are connected as an "Oh, My Darling!!" ton.  :-)  So far, so good.

How are Q3 and Q4 connected?  I think it is supposed to be another Darlington, complementary to the other one.  But...I think it is more insultory than complementary.  Look at it!!

EDIT:  Well Q3 and Q4 are NOT a Darlington, but I guess it might work. 

Q3 and Q4 are a Sziklai pair, avoiding the need for a power PNP.

[MrAl]

How can I determine the ac center-tapped voltages from the transformer secondaries that will produce the dc voltages shown in this schematic?

Hi,

I do not see any ratings for the output voltages of the transformer.  We have to know that too.
We also have to know the current rating or VA rating of the transformer.

The peak output with a capacitor filter is usually 1.4142*Vac, so a 10vac winding produces about 14.1v DC peak.  With two diodes in series, it can fall to 12v or even a little lower with load because a load causes higher peak voltages across each diode.

Sometimes the ripple is estimated also, as:
dV=I/(f*C)
where
dV is the change in voltage (ripple),
I is the load current,
f is the frequency (twice the line frequency),
C is the capacitance in Farads.
This can help calculate the lowest voltage output.

Another thing to consider is the "low line" voltage.  This is typically 10 percent lower than nominal, so a 120vac line can go down to 108vac.  It some locations it can go down even more.  I think it is the same in the UK but you have to check the nominal voltage for any place first.

[Terry Bites]

I see 100V caps on the 60V rail?

[themadhippy]

I see 100V caps on the 60V rail?

time to go to specsavers,  im sure it says 63v by c5-c8

[Davecbyp]

You cant, theres no load. The resulting voltages sre a proportion of transformer impedance and load resistance.

The question is improper because those are REGULATED outputs. You must first find the minimum ' dropout voltages' for the regulatir circuits, at the low end, the high end is determined by the pass regulstor heat dissipation and breakdown currents. Again not possible without load info.

[Refrigerator]

Well that is not all that is WRONG  Amended..see below.

Q1 and Q2 are connected as an "Oh, My Darling!!" ton.  :-)  So far, so good.

How are Q3 and Q4 connected?  I think it is supposed to be another Darlington, complementary to the other one.  But...I think it is more insultory than complementary.  Look at it!!

EDIT:  Well Q3 and Q4 are NOT a Darlington, but I guess it might work. 

Q3 and Q4 are a Sziklai pair, avoiding the need for a power PNP.

Also known as the quasi-complementary pair