Hi
I would like help with voltage regulators that I am building ( trying to build ) for a couple of preamps. I am having trouble finding the right location and value for the capacitors that go close to the regulators.
Both pre amps have remote power supply transformers rectified to provide 24-0-24 to the preamps via low reistance cable. The pre amps work off 18-0-18 volt supplies and I am planning to put the voltage regulators in the pre amp cases. The pre amps supplies are for line stage sections only so I was not making a complicated voltage regulator. I was planning to use Lm317/LM337 or LM7818/LM7918 although I am aware that some would use more sophisticated regulators.
I found that without a capacitor between the input and ground of the regulator there is a big loss of treble and some confusion to the sound. In fact the sound is better without the regulator. I read the national semiconductor datasheets. They indicate that a capacitor will help between input and ground but are vague on the value and location. I experimented by adding different capacitors.
It seems that my best sound is with 1000uf soldered across the input and ground pins of the regulator. Is this normal or am I doing something wrong.
By the way I already have a capacitor between the output and ground pin but that seemed easy to fix in size and location and does not seem to have any effect on the capacitor across ground and input.
All help appreciated.
Don
I would like help with voltage regulators that I am building ( trying to build ) for a couple of preamps. I am having trouble finding the right location and value for the capacitors that go close to the regulators.
Both pre amps have remote power supply transformers rectified to provide 24-0-24 to the preamps via low reistance cable. The pre amps work off 18-0-18 volt supplies and I am planning to put the voltage regulators in the pre amp cases. The pre amps supplies are for line stage sections only so I was not making a complicated voltage regulator. I was planning to use Lm317/LM337 or LM7818/LM7918 although I am aware that some would use more sophisticated regulators.
I found that without a capacitor between the input and ground of the regulator there is a big loss of treble and some confusion to the sound. In fact the sound is better without the regulator. I read the national semiconductor datasheets. They indicate that a capacitor will help between input and ground but are vague on the value and location. I experimented by adding different capacitors.
It seems that my best sound is with 1000uf soldered across the input and ground pins of the regulator. Is this normal or am I doing something wrong.
By the way I already have a capacitor between the output and ground pin but that seemed easy to fix in size and location and does not seem to have any effect on the capacitor across ground and input.
All help appreciated.
Don
Hi,
input and output caps are compulsory for 78xx, 79xx, 317 & 337 regulators.
The ST data sheet specifies 330nF in to ground and 100nF out to ground for the fixed regs.
Nat semi specifies the adjustable to need caps from in to ground and out to ground and their performance is significantly inferior to the fixed if the adjust to ground cap is omitted. They also mention stability issues if the wrong type of cap is used.
BTW,
I interpret their instructions to mean caps directly on or adjacent to the relevant regulator pins.
input and output caps are compulsory for 78xx, 79xx, 317 & 337 regulators.
The ST data sheet specifies 330nF in to ground and 100nF out to ground for the fixed regs.
Nat semi specifies the adjustable to need caps from in to ground and out to ground and their performance is significantly inferior to the fixed if the adjust to ground cap is omitted. They also mention stability issues if the wrong type of cap is used.
BTW,
I interpret their instructions to mean caps directly on or adjacent to the relevant regulator pins.
in any case, the adjustable regulators are quieter than the fixed voltage units, much better still if you bypass the adjust pin with an electrolytic.
when Audio Amateur (or was it AX?) had an article on the Adcom GFP565 modification all the respondents stated that cleaning up the power supply with a high quality regulator had the most dramatic improvative effect on the sound. the circuit boards are still available from Old Colony Sound and are easy to implement if you have the real estate.
when Audio Amateur (or was it AX?) had an article on the Adcom GFP565 modification all the respondents stated that cleaning up the power supply with a high quality regulator had the most dramatic improvative effect on the sound. the circuit boards are still available from Old Colony Sound and are easy to implement if you have the real estate.
Hi Andrew
Yes; I find that soldered right on to the pin seems to work best. It just looks such a mess.
However, I am surprised that if it is so important to solder onto the pins I would have expected the datasheet to say that. I just assumed that I was doing something else wrong that made it necessary to solder the caps directly onto the pins . I am probably still doing something wrong.
Everyone who replied
Thanks everyone for the comments and help.
Don
Yes; I find that soldered right on to the pin seems to work best. It just looks such a mess.
However, I am surprised that if it is so important to solder onto the pins I would have expected the datasheet to say that. I just assumed that I was doing something else wrong that made it necessary to solder the caps directly onto the pins . I am probably still doing something wrong.
Everyone who replied
Thanks everyone for the comments and help.
Don
sugestion
make shure that capacitor of filter in the input is close to regulator in some datashets is requiered a 100n with it if cap is located far away from bridge ....
also capacitors used to be located in parallel with diodes in many audio and or other application ..... makes rectifing faster
but best of all try after the 100n of the output and close to regulator to add a tantalium 10 mfd
this will change also a lot
make shure that capacitor of filter in the input is close to regulator in some datashets is requiered a 100n with it if cap is located far away from bridge ....
also capacitors used to be located in parallel with diodes in many audio and or other application ..... makes rectifing faster
but best of all try after the 100n of the output and close to regulator to add a tantalium 10 mfd
this will change also a lot
Actually I don't know if a super regulator say with a jfet input like the ad8610 _without the pre reg_ (as this is unfair comparison) could beat a 7815, 7805 at psrr in the Mhz+ region.
Maybe Jan Didden could help us out on that. But i think there definately better than adjustables Lm317 types.
Maybe Jan Didden could help us out on that. But i think there definately better than adjustables Lm317 types.
Hi, Fanuc,
The values for LM317 at 100khz = 25dB-40dB (depending on Cadj).
For LM2940 at 100khz it's 65dB (but the curve is strange, at 1Mhz it becomes 70dB)
LM2940 is non-adjustable, maybe is this what you mean by comparison of adjustable and non-adjustable (where non-adjustable is better)?
The values for LM317 at 100khz = 25dB-40dB (depending on Cadj).
For LM2940 at 100khz it's 65dB (but the curve is strange, at 1Mhz it becomes 70dB)
LM2940 is non-adjustable, maybe is this what you mean by comparison of adjustable and non-adjustable (where non-adjustable is better)?
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.