Hi,
I try to output a symetric +/- 20 V with those 2 low drop programmable regs (in symetric to output +/gng/-)
I choosed R1 : 12 K ohms & R2 180 K ohms to programm something around 20 V output
the symetric supply after the symetric caps are +/- 32 V unloaded with a 18V traffo !
When I connect to the load I have +12 v at the output of the first caps before the LT1963A and -10.5 V at the output of the smoothing caps before the LT3015 ! (where i measured the +/- 32 V unloaded) !
At the output the BC550/560 (6 in diamond buffer arrangement are more than hot : it smells !)
What do I do wrong please ? bad R2 choice ? ( I don't understood to find the correct value !)
Thanks a lot in advance for your help
I try to output a symetric +/- 20 V with those 2 low drop programmable regs (in symetric to output +/gng/-)
I choosed R1 : 12 K ohms & R2 180 K ohms to programm something around 20 V output
the symetric supply after the symetric caps are +/- 32 V unloaded with a 18V traffo !
When I connect to the load I have +12 v at the output of the first caps before the LT1963A and -10.5 V at the output of the smoothing caps before the LT3015 ! (where i measured the +/- 32 V unloaded) !
At the output the BC550/560 (6 in diamond buffer arrangement are more than hot : it smells !)
What do I do wrong please ? bad R2 choice ? ( I don't understood to find the correct value !)
Thanks a lot in advance for your help
Eldam,
Have a look at the datasheet.
The LT1963 has an absolute maximum input voltage with respect to ground of 20V. A pity.
The LT3015 is somewhat more tolerant, but with mains fluctuations, you've probably fried it too.
To stay on the safe side with the same transformer, you should use pre-regulators (LM317/337) with less than 20V output to feed your low noise regs.
The final output voltage will be something around +/-18V wich should be enough to feed your buffer.
Francis
Have a look at the datasheet.
The LT1963 has an absolute maximum input voltage with respect to ground of 20V. A pity.
The LT3015 is somewhat more tolerant, but with mains fluctuations, you've probably fried it too.
To stay on the safe side with the same transformer, you should use pre-regulators (LM317/337) with less than 20V output to feed your low noise regs.
The final output voltage will be something around +/-18V wich should be enough to feed your buffer.
Francis
Thanks a lot for your help Francis
Problems is I have already bought two of each (already populated on the pcbs) for two symetrical supply for a Diamond buffer made with discrete BC550/560 for an AD1862 DAC by Painkiller member ! (two mono board so 2 symetric PSU with for each one LT1963A & LT3015 with Gnd in common to output +/-20 V (I could live with less : 19V also !)
I saw the LT1963A is 20 V max at input and output, but the datasheet just give a formula for programming the output ! I'm lost in the "virgules" of the second part of the formula ! 😱
At least not sure I'm Under 20V at the input : I have a 18V R-core from Selectronic with a common 22 R /2W after the diodes bridge, I measure 31.9 V after the smoothing capacitor unloaded ! If I'm right I shoulf have 20.3 V loaded looking at PSUII designer soft: maybe too much for the LT1963A ???? I should reduce the 0.3 V with a Pi filter before !
With the negativ LT3015 I should be nearer than 19V with R1 : 12.1K and R2 180K... problem is I took the same value for LT1963A... they advise here not to go further than R1=4.15K for more precision while in theory 12.1 should work !
Here I both don't understand if the 18V traffo voltage value is good enough ! Basicly I beleive I'm wrong to use the programming formula of both regs !🙁
Any connaissor of those two regulators around here please ?
I keep also the idea to use a pre-reg to input a safer <20V , not having LM317/337 on hands though !
Problems is I have already bought two of each (already populated on the pcbs) for two symetrical supply for a Diamond buffer made with discrete BC550/560 for an AD1862 DAC by Painkiller member ! (two mono board so 2 symetric PSU with for each one LT1963A & LT3015 with Gnd in common to output +/-20 V (I could live with less : 19V also !)
I saw the LT1963A is 20 V max at input and output, but the datasheet just give a formula for programming the output ! I'm lost in the "virgules" of the second part of the formula ! 😱
At least not sure I'm Under 20V at the input : I have a 18V R-core from Selectronic with a common 22 R /2W after the diodes bridge, I measure 31.9 V after the smoothing capacitor unloaded ! If I'm right I shoulf have 20.3 V loaded looking at PSUII designer soft: maybe too much for the LT1963A ???? I should reduce the 0.3 V with a Pi filter before !
With the negativ LT3015 I should be nearer than 19V with R1 : 12.1K and R2 180K... problem is I took the same value for LT1963A... they advise here not to go further than R1=4.15K for more precision while in theory 12.1 should work !
Here I both don't understand if the 18V traffo voltage value is good enough ! Basicly I beleive I'm wrong to use the programming formula of both regs !🙁
Any connaissor of those two regulators around here please ?
I keep also the idea to use a pre-reg to input a safer <20V , not having LM317/337 on hands though !
Last edited:
Relying on loading of the xformer and smoothing caps to reduce input voltage of the reg is not good practice.
At power on, with large capacitive loading of the reg, it can enter in current limit and even for a short time, input voltage may be well above what you expected.
Those LDO's have need for output capacitance close to them or oscillations may occur, it could also explain over heating.
When it comes to the resistor divider, you don't have to stick to such highish values.
The LT1963 has two orders of magnitude higher bias current than its counterpart, as a result with R2=180k, a difference in output voltage of 0.5V !
Using R1=1.2K and R2=18K (or even lower) will work well and the second part of the formula is now out of the party, with still around 20V output.
I would say the LT1963 is not suited to the task like you plan to use it.
Good luck with your experiment,
Francis
At power on, with large capacitive loading of the reg, it can enter in current limit and even for a short time, input voltage may be well above what you expected.
Those LDO's have need for output capacitance close to them or oscillations may occur, it could also explain over heating.
When it comes to the resistor divider, you don't have to stick to such highish values.
The LT1963 has two orders of magnitude higher bias current than its counterpart, as a result with R2=180k, a difference in output voltage of 0.5V !
Using R1=1.2K and R2=18K (or even lower) will work well and the second part of the formula is now out of the party, with still around 20V output.
I would say the LT1963 is not suited to the task like you plan to use it.
Good luck with your experiment,
Francis
Thanks to have read the datasheet.
The first resistor between the smoothing cap is more a technic I saw from Pedja Rogic in the first AYA2 dacs, I surmise to attenuate the commutation peaks of the diodes ?! Certainly 22R is too huge ! ANd it will be better with a Pi after the first cap as well !
There is less than 2 cm between the regs and the load with 50 uF between, certainbly not a question of capacitance of the output cap and distance to the load.
Maybe it's just my PS was bigger than the 20 V max of the LT1963A, and because the bad choice of the programmating resistors (12.1 K R1 & 180K R2) as well it fried the reg(s) !
they output 1.5 A and are certainly strong enough but as it's a diamond buffer I need exactly the same symetric voltage for + & - ... so not understanding the formula of the datasheets and their different needs per reg reference and the simple rules of making 20V (or a little less) from a 18V traffo is a problem for sure 😱
Hopping I have not fried the discrete diamond buffer but it smelled burned !
!
Ok : desoldering pump !
Thank you for the good luck ! The first experiment was more a transcient towards the death of the Power Stage !
The first resistor between the smoothing cap is more a technic I saw from Pedja Rogic in the first AYA2 dacs, I surmise to attenuate the commutation peaks of the diodes ?! Certainly 22R is too huge ! ANd it will be better with a Pi after the first cap as well !
There is less than 2 cm between the regs and the load with 50 uF between, certainbly not a question of capacitance of the output cap and distance to the load.
Maybe it's just my PS was bigger than the 20 V max of the LT1963A, and because the bad choice of the programmating resistors (12.1 K R1 & 180K R2) as well it fried the reg(s) !
they output 1.5 A and are certainly strong enough but as it's a diamond buffer I need exactly the same symetric voltage for + & - ... so not understanding the formula of the datasheets and their different needs per reg reference and the simple rules of making 20V (or a little less) from a 18V traffo is a problem for sure 😱
Hopping I have not fried the discrete diamond buffer but it smelled burned !
! Ok : desoldering pump !
Thank you for the good luck ! The first experiment was more a transcient towards the death of the Power Stage !
- Status
- Not open for further replies.