Have this preamp i really like after some mods but the transformer seems way underdimensioned. After replacing op amps from NE5532 to opa1612 i noticed temp rises.. now Around 45-48 degres celcius or 113-118 F. After some 1-2 h playing music.
Think they saved some money here and logic section (volume, relays and all) take its power from same. See schematic.
I mesured voltage and with load its 27.5V ac and 30.5v ac without load.
My idea was to use only the orginal one for logic and add a toroid or something similar to feed audiocurcuit separatly.
But i can’t find a toroid that has centertap and 27.5v, i find many 2x25 and some 2x30v (Noratel/ multicomp) maybe some 20-30va? Some 1 amp rating would be enough. I’m not sure how to calculate current rating here. But the whole preamp sucka about 13.5w from the wall in standby with no music.
Can i use 2x25v for both (if i wanna change to toroid for both for example?)
The 7824 doesn’t need some min 26.5v to regulate well?
And with my limited knowledge a 2x25v can easily be connected with a CT just connect the 0 together from both windings?
Also is it possible when using 2 transformers to let the centertap be to same ground?
Anything i should think about regarding fuses? Now there is 2x800mA.
Hope to learn something new and same time improve the sound a bit.
Regards
Think they saved some money here and logic section (volume, relays and all) take its power from same. See schematic.
I mesured voltage and with load its 27.5V ac and 30.5v ac without load.
My idea was to use only the orginal one for logic and add a toroid or something similar to feed audiocurcuit separatly.
But i can’t find a toroid that has centertap and 27.5v, i find many 2x25 and some 2x30v (Noratel/ multicomp) maybe some 20-30va? Some 1 amp rating would be enough. I’m not sure how to calculate current rating here. But the whole preamp sucka about 13.5w from the wall in standby with no music.
Can i use 2x25v for both (if i wanna change to toroid for both for example?)
The 7824 doesn’t need some min 26.5v to regulate well?
And with my limited knowledge a 2x25v can easily be connected with a CT just connect the 0 together from both windings?
Also is it possible when using 2 transformers to let the centertap be to same ground?
Anything i should think about regarding fuses? Now there is 2x800mA.
Hope to learn something new and same time improve the sound a bit.
Regards
Attachments
If you look at the datasheets for the regulators you will see that they will operate with inputs up to a maximum of 37V.
You are trying to obtain +/- 24V after the LM317/LM337 regulators.
So yes you will need at least 27V after the rectifiers.
One way to do this is to buy an 18-0-18V toroidal transformer and add a few turns to it. You don't have to take it apart, just wind the extra windings around the core and add them in series with the production windings.
You might be able to find a 24-0-24V transformer which will do the job.
You are trying to obtain +/- 24V after the LM317/LM337 regulators.
So yes you will need at least 27V after the rectifiers.
One way to do this is to buy an 18-0-18V toroidal transformer and add a few turns to it. You don't have to take it apart, just wind the extra windings around the core and add them in series with the production windings.
You might be able to find a 24-0-24V transformer which will do the job.
If your mods included fitting oversized reservoir caps then that could contribute to higher core temperatures as the charging current per cycle will now be higher although spread over shorter time intervals (each cycle). That can push the transformer closer to saturation the effect of which will manifest as heat.
Whatabout this one? Biggest i can fit into the chassie.
https://uk.farnell.com/multicomp/mc...oidal-2-x-25v-30va/dp/9530339?st=25v toroidal
NE5532s need 8-9 mA a pop. OPA1612 is rated 3.6 mA a channel, i.e. 7.2 mA. An actual increase in power consumption may be indicative of oscillation here. You better hunt that down first - if fingers are too insensitive, cover suspect parts in alcohol and watch evaporation speed. OPA1612 is a substantially faster, very low voltage noise (exposed input capacitance!) part. Misimprovement is a very real possibility. (Why do people think that mods always result in improvements?)
Looking at the NAD 116 board layout, I don't see any local rail decoupling at the opamps, which is Not Good in this context - that would have to be added for fancy opamps (and may even be a good idea when sticking with NE5532s, honestly).
Moreover, most opamps are seeing impedances of at least 1 kOhm, so you actually just need a medium-low-noise part, not a super-low noise one - think LME49860, OPA1622 or maybe OPA1656. Actually, noise performance may be of little importance altogether, as the opamps are preceded by the main +16.5 dB amp stage with what is likely to be >20 nV/√(Hz) of noise output.
What I don't like is that the +/-18 V supply is right on the edge of spec for all the OPAs. It is, however, likely to benefit main amp performance.
Honestly, for what is ultimately an output buffer you don't need a fancy opamp. Are the original 5532s Signetics parts? I would suggest putting those back while adding rail decoupling - one cap rail-to-rail on the underside of each opamp at least, and if you want some rail-to-ground caps as well you'll have to run an extra wire for power ground following V+/V- on the board.
Instead of trying to replace the audio power supply xfmr (should it really be necessary), I would be looking at the logic supply instead. A transformer with a ca. 24 V secondary should be much more readily available. Also check how much current is flowing over LK34, as a 5 V supply derived from something like 30 V unregulated seems very inefficient to say the least. I'd say anything over ca. 40 mA would warrant a dedicated transformer to feed the +5V reg.
Looking at the NAD 116 board layout, I don't see any local rail decoupling at the opamps, which is Not Good in this context - that would have to be added for fancy opamps (and may even be a good idea when sticking with NE5532s, honestly).
Moreover, most opamps are seeing impedances of at least 1 kOhm, so you actually just need a medium-low-noise part, not a super-low noise one - think LME49860, OPA1622 or maybe OPA1656. Actually, noise performance may be of little importance altogether, as the opamps are preceded by the main +16.5 dB amp stage with what is likely to be >20 nV/√(Hz) of noise output.
What I don't like is that the +/-18 V supply is right on the edge of spec for all the OPAs. It is, however, likely to benefit main amp performance.
Honestly, for what is ultimately an output buffer you don't need a fancy opamp. Are the original 5532s Signetics parts? I would suggest putting those back while adding rail decoupling - one cap rail-to-rail on the underside of each opamp at least, and if you want some rail-to-ground caps as well you'll have to run an extra wire for power ground following V+/V- on the board.
Instead of trying to replace the audio power supply xfmr (should it really be necessary), I would be looking at the logic supply instead. A transformer with a ca. 24 V secondary should be much more readily available. Also check how much current is flowing over LK34, as a 5 V supply derived from something like 30 V unregulated seems very inefficient to say the least. I'd say anything over ca. 40 mA would warrant a dedicated transformer to feed the +5V reg.
So if its ocillation, the op anp would show that by being warm?
Orginal op amps were made by Philips thailand, same thought. No decoupling.
I only have some 0,1uf wima/ ero mkt caps. Would that do? Right on the pins.
I put the ops in sockets so i can easily change.
Orginal op amps were made by Philips thailand, same thought. No decoupling.
I only have some 0,1uf wima/ ero mkt caps. Would that do? Right on the pins.
I put the ops in sockets so i can easily change.
Basically if we are now calculating 1.414 times 25v thats about 35v.
But then why i need that high voltage? I could then get away with a 2x18v to feed audio section. And a 25v to feed logic. Why the need for so high voltage? 27.5v (like now is 38.8v) isn’t that pushing everything to the limit? The 7824/7805 can take max 40 or 37v what i can see?
But then why i need that high voltage? I could then get away with a 2x18v to feed audio section. And a 25v to feed logic. Why the need for so high voltage? 27.5v (like now is 38.8v) isn’t that pushing everything to the limit? The 7824/7805 can take max 40 or 37v what i can see?
When I replaced 1 v/usec slew rate op amps with 7 v/usec ones, I got oscillation. I only had to put one .1 uf disc capacitor from rail to rail within an inch of two packages, to get rid of it. Plus 33 pf disk capacitors parallel the feedback resistors on every circuit.
I don't have a working scope; I detected the oscillation with an analog VOM with a 10 VAC scale. You can buy such a VOM for about $20 on e-bay. You have to put a capacitor series the negative input to prevent DC from reading on the AC scale. One uses .047 uF for audio use. A 1 mhz oscillation will also read through a 390 pf cap. Audio signals will not. So that is how I detect ultrasonic oscillation without a device (used oscilloscope) containing 80 expired electrolytic caps.
DVM AC scales only read accurately 50-60 hz, except the $180 RMS models. & the RMS model dVMs will miss ultrasonics, their upper limit of response is 7 khz.
I don't have a working scope; I detected the oscillation with an analog VOM with a 10 VAC scale. You can buy such a VOM for about $20 on e-bay. You have to put a capacitor series the negative input to prevent DC from reading on the AC scale. One uses .047 uF for audio use. A 1 mhz oscillation will also read through a 390 pf cap. Audio signals will not. So that is how I detect ultrasonic oscillation without a device (used oscilloscope) containing 80 expired electrolytic caps.
DVM AC scales only read accurately 50-60 hz, except the $180 RMS models. & the RMS model dVMs will miss ultrasonics, their upper limit of response is 7 khz.
I didn’t check temp on the 1612 before adding decoupling underneath the board but now original transformer is less warm, about 35-40 instead of 48 and opamps are around 40 degress. Little bit of alcohol vaporize withing a few seconds.
I’m thinking if i use the 1612 i have to move decoupling up ”on top” (not underneath) because now with socket and all it’s still say 2-3cm between opamp and 1612. I will tey the ne5532 again also.
I added an extra pt for logics and what i can understand this is a half wave rectifier curcuit for logic supply but is this correct how to implent a second transformer. The whole mainboard share the same “ground” and the audio section has its centertap to “ground” aka the mainboard but will there be issues if i use the same “ground” for the new pt? See my drawing please.
I’ve also added an extra fuse on that new pt but unsure what right fuse will be. Current draw over the 2.2R fusible res was max 230ma.
Added a mainfuse also to both pt and its now about 250ma. There was only those 2x800ma on the secondary. I’m guessing too that they are too big now when the load isn’t that big over that part of the curcuit anymore.
So please some help
Have a nice day
I’ve also added an extra fuse on that new pt but unsure what right fuse will be. Current draw over the 2.2R fusible res was max 230ma.
Added a mainfuse also to both pt and its now about 250ma. There was only those 2x800ma on the secondary. I’m guessing too that they are too big now when the load isn’t that big over that part of the curcuit anymore.
So please some help
Have a nice day
Attachments
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.