Hello people,
I am going to build a power supply to power some guitar effects up.
The voltages that I am going to need are:
So I need 6 fixed supplies, and two that I will be able to alter depending on my needs. Note that I don't need the adjustment to be controlled easily, without opening the chassis etc. No external pots or switches needed, to be accurate.
I was leaning towards using LM317 voltage regulators, with which I am familiar. Actually I have designed the power supply based on LM317's (and LM350's for the variable supplies), but here I am seeking for a better alternative.
So, I have come across some transistor regulators appearing to have better noise specs than ordinary LM317 designs (how about ripple rejection?). I am mainly referring to this site, which should be known to many people in here, I suppose: Simple Voltage Regulators Part 1: Noise - [English]
But I am not familiar with this circuitry, and it seems to me that using zener diodes with fixed reverse voltage characteristics does not leave much room for flexibility.
So, should anyone be able to explain to me how to modify these circuits to produce 9V,0.5A outputs or even better to turn them into variable regulators, I would be glad to know.
Is there really any cheap and simple circuit out there to beat the ordinary LM317 design? That could be the thread's title in the first place, actually.
I am going to build a power supply to power some guitar effects up.
The voltages that I am going to need are:
- 6x9V,0.5A
- 2x9 to 30V,1.5A
So I need 6 fixed supplies, and two that I will be able to alter depending on my needs. Note that I don't need the adjustment to be controlled easily, without opening the chassis etc. No external pots or switches needed, to be accurate.
I was leaning towards using LM317 voltage regulators, with which I am familiar. Actually I have designed the power supply based on LM317's (and LM350's for the variable supplies), but here I am seeking for a better alternative.
So, I have come across some transistor regulators appearing to have better noise specs than ordinary LM317 designs (how about ripple rejection?). I am mainly referring to this site, which should be known to many people in here, I suppose: Simple Voltage Regulators Part 1: Noise - [English]
But I am not familiar with this circuitry, and it seems to me that using zener diodes with fixed reverse voltage characteristics does not leave much room for flexibility.
So, should anyone be able to explain to me how to modify these circuits to produce 9V,0.5A outputs or even better to turn them into variable regulators, I would be glad to know.
Is there really any cheap and simple circuit out there to beat the ordinary LM317 design? That could be the thread's title in the first place, actually.
Last edited:
I am sorry but I did not understand! You mean to use L7809 exclusively for the 9V supplies or to use LM317 first and after that the L7809?
Are L(M)7809 considered to be any better than LM317?
well a transformer that puts out 12v AC at 1A and 24v AC at 3 A would be good , you can use a couple lm7809 regs for the 2x9v DC from the 12v AC tap and a LM317 or lm338 for the 9v -30v DC using the pot on the adjust leg of the reg off the 24v AC tap .... Using a TX with 2 seperste taps will give you some isolation between the 9v dc and 30v dc supplies ......
Cheers
PS if you want to go super simple and cheap you could use a single 12v DC Xformer and rectify the 12v ac to 17v dc and regulate down to 9v dc and then use a voltage doubler off the 12v ac to get your 34v DC and regulate down to 30 v and adjust with a trim pot on the adjust leg of the regulator ......
Cheers
PS if you want to go super simple and cheap you could use a single 12v DC Xformer and rectify the 12v ac to 17v dc and regulate down to 9v dc and then use a voltage doubler off the 12v ac to get your 34v DC and regulate down to 30 v and adjust with a trim pot on the adjust leg of the regulator ......
Please tell me your opinion on this: I was planning to use LM350T's for the 1.5A supplies - do you think it is better to do so, or is it better to use a LM317 stage in addition with a transistor or mosfet to provide the extra current? I know LM317 is specified at 1.5A load, but I don't want to push it that far.
Symmetrical power supply
I am wondering whether the following two options are equally good:
(1) creating a +/-16V supply using a LM317 plus one LM337 with separate rectifiers and secondaries
(2) creating a +/-16V supply using two LM317's with separate rectifiers and secondaries, by means of connecting the output of the second one to ground - thus, it could provide the negative rail
I think the second option has the following advantages:
Other than these assets, which drawbacks do you assume option (2) could have? Or in what sense could the circuit behave differently?
I am wondering whether the following two options are equally good:
(1) creating a +/-16V supply using a LM317 plus one LM337 with separate rectifiers and secondaries
(2) creating a +/-16V supply using two LM317's with separate rectifiers and secondaries, by means of connecting the output of the second one to ground - thus, it could provide the negative rail
I think the second option has the following advantages:
- easier layout for the pcb construction (ok a bit subjective)
- LM317 has better noise characteristics (maybe not audible, but this is what I got from a site that performed measurements)
- I can be more familiar with LM317's behaviour, since there is a lot of information about it everywhere
- costs - well that is no issue speaking of one power supply, but being able to order only one IC drops the price, so one can invest in buying better LM317's
Other than these assets, which drawbacks do you assume option (2) could have? Or in what sense could the circuit behave differently?
Buy a bag of LM317 regs, it will cost you few €. Then check this link: Using 3-pin regulators off-piste: part 4 and go to the bottom, last picture is tracking preregulator with LED. Cheap, easy to build, good PSSR, better than single LM317.
Yes, there are better regulators, but they are more comlicated or more expensive or...
And yes, IMHO option (2) is better than (1).
Personally I have found that ST parts don't always meet published specs. Other than that, it shouldn't matter for a 3 terminal regulator.
Well yes, I am going to buy a small batch and see what to do: either one of the options on the site, or a classic one-LM317 regulator.
About the filtering before feeding the regulator, I was thinking of a C-R-C with values of 4700u - 2R2 - 2200u, for a maximum load of 200-300 mADC. Do you think it is adequate for both implementations?
edit: Seeing that you two also seem to prefer the option of 2 positive regulators wired to give symmetrical rails, why are negative regulators produced?
About the filtering before feeding the regulator, I was thinking of a C-R-C with values of 4700u - 2R2 - 2200u, for a maximum load of 200-300 mADC. Do you think it is adequate for both implementations?
edit: Seeing that you two also seem to prefer the option of 2 positive regulators wired to give symmetrical rails, why are negative regulators produced?
Last edited:
for centre tapped transformer..............
edit: Seeing that you two also seem to prefer the option of 2 positive regulators wired to give symmetrical rails, why are negative regulators produced?
Hmm, and option 2 could not be implemented in that case?
I have calculated that I need a 12VAC secondary for a single LM317 build, so it will always operate with Vin-Vout>4V. Now if two are cascaded, will I also need to apply the Vin-Vout>4V rule for the first one again? If so, I figure I will need 3VAC more (2.85 to be precise) per supply.
Hmm, and option 2 could not be implemented in that case?
I have calculated that I need a 12VAC secondary for a single LM317 build, so it will always operate with Vin-Vout>4V. Now if two are cascaded, will I also need to apply the Vin-Vout>4V rule for the first one again? If so, I figure I will need 3VAC more (2.85 to be precise) per supply.
No, using a center-tapped transformer, the ground for the negative and positive supplies are connected together at the transformer and you must use a negative rail regulator. For your option 2 (two + regulators) you need two separate windings or two transformers; i.e. the two supplies must be isolated from each other
Option 2 has other advantages, such as being able to use the two supplies as two positive supplies.
Of course, I think it has many advantages! And yes, I will use two separate windings (one transformer) and bridges.
About the AC's needed? Is my assumption that I will need 3VAC more if I build the pre-regulator valid?
Yes, if you cascase two regulators, then you need additional voltage overhead, since there will be some drop between input and output of the first stage. The absolute minimum amount of input-output difference is referred to as the dropout voltage. If the input falls to less than this (output + dropout) then the output falls and becomes unregulated. Keep in mind the amount of ripple in the input voltage, since you need to consider the minimum voltage at the input, not the average voltage.
An extra 3 VAC would be the minimum that I would use.
The pre-regulator is a good idea since you want a wide range adjustable output at high current. This way you split the voltage drop and power dissipation between two active devices. Otherwise you may exceed the power dissipation of a single regulator, when operating at high current and low voltage output.
An extra 3 VAC would be the minimum that I would use.
The pre-regulator is a good idea since you want a wide range adjustable output at high current. This way you split the voltage drop and power dissipation between two active devices. Otherwise you may exceed the power dissipation of a single regulator, when operating at high current and low voltage output.
No. For this application you want all the 9V supplies to be galvanicaly isolated.
I assume each 9 supply power one effects pedal. The best practice is to use a use a transformer with many taps like the one linked to below. Other wise you risk ground loops
https://taweber.powweb.com/store/wpdlxfmr2sch.jpg
If you buy a good commercial made pedal board power suppliy this is the transformer you will find inside. It's not that expensive.
Let the power supplies "float" and not create a ground loop through the guitar cable shields.
As for "noise". Those three terminal LM7809 parts are about the best, even compared to batteries. I would NOT bother with a "pre regulator" but you can use a CRC filter rather then the more common C only type filter. This does three things (1) the "R" helps drop some volts and takes some heat off the 7809 and you get about (2) a few more db of ripple rejections and (3) because the R limits in-rush current yo can make the second C really huge.
But the most reduction in noise will come from breaking the ground loops. A "loop" is easy to understand. Any time you can traces a circle with grounds you have a potential problem. Ideally you want all those pedals to be tied to ground ONLY through the audio cables.
If you don'tfloat the grounds then you will forever have to use plug in this power supply to the same AC mains outlet as the amplifiers and venue's PA mixer. With home stereo this is a non-issue because people tend to put all the gear in one place and use really short cables. Not true of live sound.
Last edited:
Of course, I am planning to use completely separate secondaries, rated at 12VAC, 0,4A. It will be a 80VA transformer, or to keep sizes a bit down (and maybe costs a little up
) I will use two separate 50VA transformers.No earth connections, everything floating: only one at the amplifier, each pedal will be connected to it through signal cables.
I will include soft start circuit - so inrush would not be a problem. Yes, I was thinking of a 4700u - 2R2 - 4700u (2200u) filter. Maybe I will put identical caps to buy them on a scale and save some money.
I will also use the C-R-C filter to attenuate high frequency noise, not only ripple. As I have seen on some sites, the regulators are not fully capable of doing this well on their own.
And I was planning to use LM317's - I have not considered LM7809. Have you encountered any significant advantages of the LM7809 compared to LM317, which is by the way a well discussed regulator?
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.