G'day Guys,
I am currently designing a PCB for a fairly bog standard phono stage based on the NE5532. Nothing fancy or important here.
The typical way I would handle the power supply would be something like some 7815/7915 regulators on an off board power supply pcb.
Is there a good reason not to put the 7815/7915 regulators right beside the OPamps power supply pins?
Is there a technical reason not to do this?
I can think of cost reasons but it seems to me like putting the regulator right beside the OPamp with the regulator output cap functioning as the OPamp decoupling cap to be superior.
I am currently designing a PCB for a fairly bog standard phono stage based on the NE5532. Nothing fancy or important here.
The typical way I would handle the power supply would be something like some 7815/7915 regulators on an off board power supply pcb.
Is there a good reason not to put the 7815/7915 regulators right beside the OPamps power supply pins?
Is there a technical reason not to do this?
I can think of cost reasons but it seems to me like putting the regulator right beside the OPamp with the regulator output cap functioning as the OPamp decoupling cap to be superior.
One issue I just thought of would be that OPamps typically like a 100nF ceramic capacitor beside them for decoupling.
Off of the top of my head I don't think 3 terminal regulators like ceramic caps at their output?
Off of the top of my head I don't think 3 terminal regulators like ceramic caps at their output?
Local regulation is technically the best and today it is even standard. About all new regulators are based on this "small current loop" principle and therefor high current regulators are not popular anymore (pity). DC power distribution and/or preregulation with 7818/7918 or just unregulated DC and then per section local ultra low noise (LDO) regulators. Or 7815/7915 and +/-12V with local ultra low noise regulators close to the load. For a phono preamp the "ultra low noise" will not exactly be fancy but a necessity 🙂 It is 2024 and things can be executed to level "excellent" with today's parts. Why put time, money and effort in "good enough"?
You are right that then 1 cap per rail is adequate. Also pay attention to input filtering in today's polluted environments. CLC/CRC before the pre regulators also improves matters.
You might want to try out the less bog standard but stellar performing OPA828. Posts here triggered me to try it out and it is a winner.
You are right that then 1 cap per rail is adequate. Also pay attention to input filtering in today's polluted environments. CLC/CRC before the pre regulators also improves matters.
You might want to try out the less bog standard but stellar performing OPA828. Posts here triggered me to try it out and it is a winner.
Last edited:
Thanks.
My evil plan is to use a Cap Mx for the 'main' psu.
I will start with NE5532 and probably upgrade to nicer opamps later on. I have quite a liking for the LME47920.
This project is a commission for a friend so he can spend the extra money on nicer opamps later if he wants.
I get the chance to play with some new concepts in the mean time like local regulation, cap mx etc.
My evil plan is to use a Cap Mx for the 'main' psu.
I will start with NE5532 and probably upgrade to nicer opamps later on. I have quite a liking for the LME47920.
This project is a commission for a friend so he can spend the extra money on nicer opamps later if he wants.
I get the chance to play with some new concepts in the mean time like local regulation, cap mx etc.
Choose wisely and use SMD. Opamp rolling is useless when using one of the best straight away. Cheaper and more effective too. Shortest connections/PCB tracks, no IC sockets .... Give laymen the choice of opamp and cap rolling and they will do this on a weekly basis 🙂 One day tower like structures in bright colors with fancy names will make that covers can not be closed anymore. D cell like capacitors will be used with precious metals and the finest of oils for uplifting veils and blacker than black backgrounds meanwhile resulting in a ruined device. Time, money and effort that could have been spent in making a well designed and outstanding device with adequate shielding against nasties.
I have many fancy DIL8 opamps but will only use OPA828, OPA1642 as of now.
I have many fancy DIL8 opamps but will only use OPA828, OPA1642 as of now.
Last edited:
Would a standard 10uF electro be ideal or a smaller film/ceramic cap?
eg: 100nF ceramic or a 1uF film cap.
Depends on the regulator. Modern ones have no issue with ceramic SMD caps close to the pins. Use X7R as minimum standard.
When choosing TH 1 to 10 uF electrolytic caps are OK for both classic and modern regulators but why would you use parts that wear out?
My choice would be to add PCB pads for both types 🙂
When choosing TH 1 to 10 uF electrolytic caps are OK for both classic and modern regulators but why would you use parts that wear out?
My choice would be to add PCB pads for both types 🙂
Reading the data sheets of the 78/79xx regulators they all same the same thing word for word:
"An input bypass capacitor should be selected to provide good high−frequency characteristics to insure stable operation under all load conditions. A 0.33 F or larger tantalum, mylar, or other capacitor having low internal impedance at high frequencies should be chosen."
Given the low current consumption, I don't think the typicall 2200uF input cap I see on most designs is necessary.
I am thinking of using a 1uF polyester cap on the input and output which should save a lot of pcb space and allow me to keep things very close to both the opamp and regulator pins.
Does this sound like a good idea?
"An input bypass capacitor should be selected to provide good high−frequency characteristics to insure stable operation under all load conditions. A 0.33 F or larger tantalum, mylar, or other capacitor having low internal impedance at high frequencies should be chosen."
Given the low current consumption, I don't think the typicall 2200uF input cap I see on most designs is necessary.
I am thinking of using a 1uF polyester cap on the input and output which should save a lot of pcb space and allow me to keep things very close to both the opamp and regulator pins.
Does this sound like a good idea?
No.
1. IF you use 78xx/79xx don't use too low ESR caps like film caps at the outputs*.
2. A 5 mm pitch 1 µF film cap at the inputs is OK.
3. A 2200 µF input cap is nonsense, probably the bulk filter cap is meant.
4. almost every possible regulator needs both an input and output bypass cap but the latter needs special attention depending on the regulator. If you want to experiment add PCB pads for both 0805 SMD caps and 2.5 mm electrolytic caps for freedom of choice. From experience I know that a 100 nF film cap works out OK at outputs of 78xx/79xx.
5. some design tips: make the design as compact as possible. One board with least possible wiring and separate PSU board. Distribute DC instead of AC. Have inputs and outputs not in each others vicinity. Design inputs and outputs at the side where the connectors will be and/or use PCB mount RCA connectors for no wiring at all. Try to make the design RF proof.
6. Tip: check the current the opamp is using as 79xx needs a minimum current for sure. A remedy can be to include the power on LED as load for the 79xx.
*Using too low ESR caps at the outputs of many classic regulators is asking for oscillation.
1. IF you use 78xx/79xx don't use too low ESR caps like film caps at the outputs*.
2. A 5 mm pitch 1 µF film cap at the inputs is OK.
3. A 2200 µF input cap is nonsense, probably the bulk filter cap is meant.
4. almost every possible regulator needs both an input and output bypass cap but the latter needs special attention depending on the regulator. If you want to experiment add PCB pads for both 0805 SMD caps and 2.5 mm electrolytic caps for freedom of choice. From experience I know that a 100 nF film cap works out OK at outputs of 78xx/79xx.
5. some design tips: make the design as compact as possible. One board with least possible wiring and separate PSU board. Distribute DC instead of AC. Have inputs and outputs not in each others vicinity. Design inputs and outputs at the side where the connectors will be and/or use PCB mount RCA connectors for no wiring at all. Try to make the design RF proof.
6. Tip: check the current the opamp is using as 79xx needs a minimum current for sure. A remedy can be to include the power on LED as load for the 79xx.
*Using too low ESR caps at the outputs of many classic regulators is asking for oscillation.
Last edited:
Have a look at L78L15 in SOT-89. This looks like a nice part to place on a small PCB and will deliver 100mA, which is heaps for a pair of opamps. The 5532 is a couple of dB noisier than the 5534A, which is my choice in this application. This opamp will get you a noise floor 10dB below the noise on the disk, if properly implemented. The high cost OPA828 will only gain you a few tenths of a dB of real SNR. You will never hear it in practice.
Oh oh can of worms 🙂 OPA637 anyone?! It is not all about noise. If that would be the number 1 priority the system itself would be rejected.
Last edited:
These days music goes through a dozen opamps if that, and some DACs and ADCs... Hundreds of opamps is from a bygone age, and in fact a lot of TL072's would be involved too.
805 size SMD film bypass caps don't have too low ESR (at least for some regulators I have tried them with) and are more linear than X7R. That said, it might not hurt to try some different caps, as well as try resistor loading of the regulator outputs so as to bias the pass transistor into a reasonably high gain region.
I've been playing with packages over the last couple of evenings.
The L78L15/79L15 in the TO92 package would be lovely but the appear to be EOL so I don't want to base a design around them.
The sot-89 package is pretty small but I think I could manage the challenge.
I am going for the DO213AB package for the diodes. SM4004. I think this will be easier to layout the protection diodes around the regulator compared to the much larger TH footprints.
I am toying with the idea of moving to SMD for the resistors too. I've done 0605 packages once before so I know its doable. Once again it possibly makes layout easier around larger items like film caps which depending on what I end up going with might be 15mm pitch PP types.
I think I will stick to through hole for the regulator input and output caps since they are cheap and easy. I had a bit of a look at the SMD options but I couldn't find easily accessible ESR data so I will stick with the electros which give me the info up front.
I will include a footprint for SOIC-8 for future proofing. Maybe on the underside of the board so that I could use a socket on the top for the initial use of NE5532 and then have the option of upgrading to an SOIC chip later without having to remake the whole board.
The L78L15/79L15 in the TO92 package would be lovely but the appear to be EOL so I don't want to base a design around them.
The sot-89 package is pretty small but I think I could manage the challenge.
I am going for the DO213AB package for the diodes. SM4004. I think this will be easier to layout the protection diodes around the regulator compared to the much larger TH footprints.
I am toying with the idea of moving to SMD for the resistors too. I've done 0605 packages once before so I know its doable. Once again it possibly makes layout easier around larger items like film caps which depending on what I end up going with might be 15mm pitch PP types.
I think I will stick to through hole for the regulator input and output caps since they are cheap and easy. I had a bit of a look at the SMD options but I couldn't find easily accessible ESR data so I will stick with the electros which give me the info up front.
I will include a footprint for SOIC-8 for future proofing. Maybe on the underside of the board so that I could use a socket on the top for the initial use of NE5532 and then have the option of upgrading to an SOIC chip later without having to remake the whole board.
Last edited:
SOIC is easy to hand solder, and easy to obtain. I wouldn't bother with DIP ICs these days in a new design.
That does make sense. But I know that the individual that will try this out first loves to upgrade things.
I may as well build him an upgrade path into it.