Hi,
I have a power supply as shown in a valve pre-amp, and I am trying to understand what purpose the centre tap provides, and what would happen if it was no longer in place?
I also have some 800VAC caps and wondered why I could not put these across the whole voltage drop after the first centre tap capacitor pair?
Thanks for the education 🙂
I have a power supply as shown in a valve pre-amp, and I am trying to understand what purpose the centre tap provides, and what would happen if it was no longer in place?
I also have some 800VAC caps and wondered why I could not put these across the whole voltage drop after the first centre tap capacitor pair?
Thanks for the education 🙂
As you seem to realise, it allows you to use two reservoir caps in series to enable lower voltage parts to be used which are cheaper and more common. If it was not in place then the caps loose a 'fixed' reference point and may then share the total voltage drop unequally. The resistors are not needed as drawn but definitely are if the CT is omitted. Without a CT both caps must be identical to share equally, differences in leakage current would swing the 50:50 share one way or the other, hence the resistors to swamp that effect.
Mooly,
Thanks so much for yet more guidance. 🙂
I am finally taking my transformer out of my old ARC SP8 as you advised a while back to remove the last remnants of hum, and at the same time I got hold of some 40 uF 440VAC Shizuki oil filled caps I thought I could fit externally as they are to large to fit inside the SP8 Chassis.
I had intended on putting two across the centre tap along with some balance resistors and then a couple across the full DC voltage, and was just making sure there was nothing devious or clever beyond the obvious around the centre tap arrangement.
Thanks!
Thanks so much for yet more guidance. 🙂
I am finally taking my transformer out of my old ARC SP8 as you advised a while back to remove the last remnants of hum, and at the same time I got hold of some 40 uF 440VAC Shizuki oil filled caps I thought I could fit externally as they are to large to fit inside the SP8 Chassis.
I had intended on putting two across the centre tap along with some balance resistors and then a couple across the full DC voltage, and was just making sure there was nothing devious or clever beyond the obvious around the centre tap arrangement.
Thanks!
If you keep the centre tap then you don't need balance resistors, there is no possibility of imbalance. Each cap sees the rectified peak value of the AC as measured from CT to output. So 450 volt AC means each side of the CT is 225 volts AC and each cap sees just 225*1.4 as a DC voltage.
Those resistors in your diagram are probably just discharge resistors for safety.
If you do away with the CT then the resistors as well as the caps must be identical. The resistors need to pass several times more current than the maximum leakage current expected in the caps in order to swamp that effect. That makes the non CT version less appealing because you now have added dissipation. Any resistors MUST be correctly rated for the voltage as well as power dissipation to prevent failure and fire risk.
Those resistors in your diagram are probably just discharge resistors for safety.
If you do away with the CT then the resistors as well as the caps must be identical. The resistors need to pass several times more current than the maximum leakage current expected in the caps in order to swamp that effect. That makes the non CT version less appealing because you now have added dissipation. Any resistors MUST be correctly rated for the voltage as well as power dissipation to prevent failure and fire risk.
Thanks.
My intention is to keep the CT and use 2W 910K matched carbon resistors for discharge.
The other calculation I am not sure of is the amount of additional capacitance I can add to the circuit in order to ensure the increase in capacitor charging current is not too much.
The original circuit has two 100 uF in series across the 450-CT-0 a total of 50uF?
and I was imagining having two 45uF across 450-CT-0 and then two more 45uF across the whole 0-450, having a total of 112uF?
I tried to simulate this on PSUD2 but have no way of understanding the real load resistance of the voltage regulator circuit (other than by deduction on the target p2p of the ripple on the original circuit topology)
My intention is to keep the CT and use 2W 910K matched carbon resistors for discharge.
The other calculation I am not sure of is the amount of additional capacitance I can add to the circuit in order to ensure the increase in capacitor charging current is not too much.
The original circuit has two 100 uF in series across the 450-CT-0 a total of 50uF?
and I was imagining having two 45uF across 450-CT-0 and then two more 45uF across the whole 0-450, having a total of 112uF?
I tried to simulate this on PSUD2 but have no way of understanding the real load resistance of the voltage regulator circuit (other than by deduction on the target p2p of the ripple on the original circuit topology)
The wattage of the resistors is only half the problem, they must be specified to withstand the voltage they will see as well. All resistors have a maximum voltage rating irrespective of the power they will see dissipated.
The caps aren't particularly large and the relatively high resistance of the secondary windings will play a big part here in limiting currents. I can't see any problem with 112uF. The circuit I'm looking at shows 200uF? caps, 150uF in brackets and a parts list that says 100uF!
The caps aren't particularly large and the relatively high resistance of the secondary windings will play a big part here in limiting currents. I can't see any problem with 112uF. The circuit I'm looking at shows 200uF? caps, 150uF in brackets and a parts list that says 100uF!
yeah the 2W are Allen Bradley good for 750V HB grade I think.
If I use all the spare caps I have (6 all together) then that would add another 80uF making it just under 200uF.
If I use all the spare caps I have (6 all together) then that would add another 80uF making it just under 200uF.
if you intend on using the center tap to source half of B+ you can....i.e. a main B+ of 630vdc unloaded gives you 300vdc at the center tap...
if this is not your intention you can use equalization resistors of about 330k 2w anyway for safety...
depending on the size of your power transformer and your tube plate loadings, the raw b+ will drop some....
if you do not wish to use series caps, there are now 450vac filter caps in the 60ufd range that you can use....
https://www.amazon.com/CBB60-Motor-Capacitor-450Vac-Frequency/dp/B087X2TJZN
if this is not your intention you can use equalization resistors of about 330k 2w anyway for safety...
depending on the size of your power transformer and your tube plate loadings, the raw b+ will drop some....
if you do not wish to use series caps, there are now 450vac filter caps in the 60ufd range that you can use....
https://www.amazon.com/CBB60-Motor-Capacitor-450Vac-Frequency/dp/B087X2TJZN
Attachments
Last edited:
I have these for the smoothing caps, which are no problem for the voltage required.
440VAC 45uF, supposedly = 800VDC as quoted from some suppliers
My old original circuit used 270K balance resistors and 2 x 400V 100 uF caps in series, feeding a tube regulation circuit, producing 400V from the 630V input.
I had intended using two initially at 45uF in the place of the original 100uF in series across the centre tap, BUT I guess I don't actually need to with that voltage rating?
440VAC 45uF, supposedly = 800VDC as quoted from some suppliers
My old original circuit used 270K balance resistors and 2 x 400V 100 uF caps in series, feeding a tube regulation circuit, producing 400V from the 630V input.
I had intended using two initially at 45uF in the place of the original 100uF in series across the centre tap, BUT I guess I don't actually need to with that voltage rating?
If it uses the CT then they are not for balancing, they are just safety resistors to discharge the caps. No CT, you need balancing, with a CT and you don't.
Mooly, thanks - this makes sense.
I guess the original design was more belt'n'braces using both?
I guess the original design was more belt'n'braces using both?
Its just the logical way to do it tbh 🙂 No possibility of problems, the caps need no great safety margin to cater for imbalance and they will therefore be cheaper. Its sound engineering and sound commercial practice.
likely to be less noise with center tap.
and lowered voltage requirements for
capacitors
also an advantage for low voltage AC heater
and lowered voltage requirements for
capacitors
also an advantage for low voltage AC heater
I don’t understand the center tap and why it’s connected between the two caps.
If the center tap were disconnected, what difference would it make? Mooly explains in post 2, but aren’t series caps often used without a midpoint reference?
If the center tap were disconnected, what difference would it make? Mooly explains in post 2, but aren’t series caps often used without a midpoint reference?
The CT kinda enforces equal voltage sharing between C14 and C15. Without it, the sharing job would be handled by R31/32 but less effectively. Two caps in series will have voltages across them in inverse ratio of their relative capacitances - since 'lytics have fairly wide tolerances the voltage difference can be substantial.
OK, makes sense. Just never thought about it or knew that it was a problem! But if you have the CT, why not?
For the most part within simplicity sake.
Center tap power transformer usually more expensive
and generally used for lower noise/ higher quality designs.
in very generalized manner.
if high voltage and low voltage windings are center tap
leads to believe a lower noise/higher cost/quality design approach.
Center tap power transformer usually more expensive
and generally used for lower noise/ higher quality designs.
in very generalized manner.
if high voltage and low voltage windings are center tap
leads to believe a lower noise/higher cost/quality design approach.
Thanks everyone, I think I will use two of my 45uF caps between from CT to B+ and CT to GND, then add a couple from B+ to GND. I will add some reasonably high value 910K 2W 750V Allen Bradley resistors and it should work out well.
I can fit in 2 more 45uF in the case so I can try these as additional supporting across B+ to GND if they make any improvement to the Sound - it will be interesting to see 🙂
I can fit in 2 more 45uF in the case so I can try these as additional supporting across B+ to GND if they make any improvement to the Sound - it will be interesting to see 🙂
i have an issue with this, if the ct is not exactly located as in bifilar winding, then the imbalance can be a problem too..
diy'ers need to take care, while the center tap is not really required, make sure the voltages are equal on either legs...