PSU OK .. but opamps get hot.. is it normal?

[pablobolche]

I am in thee process of finishing a piece of gear and I am doubtful of the capacitors for my etched power supply.
I have successfully made psu for preamps and other low consumption gear (basically copied and adapted JLM`s ACDC schematic)
My power requirements for this one is higher :
- 480mA for a 12v rail for 36 relays when all are on.. checked with my multimeter..
- 400-800mA for +-15v rails for 24 ne5532´s this is from the datasheet saying max supply current is 16mA per opamp all though I don´t know if it is per rail so that why I swing between 400 and 800 mA.

My doubt is regarding the two 10uf output caps per rail (don´t know what´s the name for these)..
Would 10uf for these caps be suffice as in JLM´s schematic or would I need to pump them up?
I have attached a jpeg with the caps marked in red.
Any comments are welcome..and hope there´s no problem with posting part of JLM´s acdc schematic here.
thank you

***Title Edit***

 

[JohnRoberts]

The last time I looked at this on the bench was a few decades ago, but I made empirical output impedance/regulation measurements by cap coupling an AC current of a several mA into the output and watching what happened (note: the Vreg needs to be loaded as it will be in typical use since AC current more than the DC current load will give an unrealistic result).

From memory I think I found around 1,000 uF gave me the best result when using a common three terminal 78/7915 vreg. The old 78/79xx used 741 speed opamp technology, so the newer 317 may deliver lower active source impedance and not benefit as much as the lower vreg.

Short answer is you can answer this for yourself with some simple experiments. The Vreg data sheet will also offer some good starting point advice, they have surely tested the part in many common applications.

JR

 

[abechap024]

http://www.ti.com/lit/ds/symlink/lm117.pdf

That will tell you But to summarize, there isn't much improvement increasing those values above 10uf. But experiment, cause depending on the circuit and how far the power supply is from the circuit, I've experienced improvements by bumping it up.

If your worried about it why not just stick 100uf in there? The price isn't much more and space isn't at a premium if your designing it yourself.

 

[PRR]

You need a good-quality ~10uFd per reg so they stay stable. That's for any current, assuming it flows "out".

If you have spikes injected IN to the reg, and they do harm, you might want more. In this case you don't: the relays may kick but they are on their own dedicated supply and a little kick does them no harm.

If you want to keep impedance AT the chip pins _low_, you need a 0.1uFd ceramic for each chip and a 100uFd electro for every handful of chips. See any large vintage channel strip. With that much bypassing ON the audio board, no extra capacitance is needed on the regulator output.

 

[pablobolche]

Ok the PSu works great regulatotrs have good heat sink  and keeps pretty cool..
the thing is that  ne5532 at +-15v... get real hot... can't take a measurment now but way hottar then the regulator..
read the spec sheet and they say they work at 70 celcius..
are there heat sinks for opamps?? is it a good idea.. or is this over kill?
Maybe I am doing some thing funny on the psu ...  with all 24 opamps hooked up my regulation only achieves +-14v.. is this normal?? I am running off a 12v dual 2Amp transformer maybe?? although regulator keeps cool.. I have a dual 18v tranformers that I am going to try to see if I get steady +-15v rail now..


anybody want to give me a hint?
thanks..

 

[gemini86]

I've never had an opamp run more than warm, except in a case of oscillation.

Have you listened to it yet?

 

[pablobolche]

I am using Audiox sum-12 arquitecture.. now changing decoupling caps to see any differences.. trying with Wima and the cheapest 100nf ceramic caps I found in a local electronic store..
sounds ok.. almost as the original but can't identify it as of yet..
but its ok.. no noise.. will post later about the caps..

the temp issue is weird.. though..
also its summer here as 35-36 celcious... and as a smoker I work with open windows so it hot here..
tried with the psu set to +-12v and lower overall temp on the opamps but still hot..

thanx

 

[Harpo]

I am using Audiox sum-12 arquitecture..
the temp issue is weird.. though..

Only reason I could imagine is a (mistakenly) nearly shorted -output (U2B/R32) to gnd instead of leaving this pin unconnected for an unbalanced output connection.

 

[pablobolche]

I´ll check the my schematics and pcb for this... .. maybe got some solder blues  ..
thanks

 

[pablobolche]

Still working on this...

I have breadboarded the the summing circuit with no socket for the IC's.. and temperature seems fine.....is it possible that the IC sockets are acting funny? not enough thermal trasmision?

Another funky thing now about the PSU.... I have a regular multimeter and you know when you put it in beep mode.. to check continuity..ok  ..I noticed that when I turn on the power.. I get a faint continuity between my GND and the -V.. not the full strong beep.. versus when power off.. there is no continuity between GND and -V. Is this normal?

I have selfetched a couple of PSU using JLM ACDC schematic and never noticed this...

thing is this Faint GND also contaminates my audio inputs and output of the summing.. is this weird or the way it is??
could this be the reason for the high temp on the opamps

Please advise..

 

[Rob Flinn]

It is normal to get a beep when you put a meter on continuity across a cap, which is what you are doing it just happens to be in circuit.  Try it with a cap out of circuit.

 

[CJ]

did you put the opamps in backwards?

pin 4 becomes 8 and the people stop dancing,

 

[pablobolche]

Rob-- the multimeter beep yeah... a newbie thing.. no problem... there is no continuity  between my voltages and gnd nor my audio signals and gnd... great..

CJ -- the opamps are in correct orientation.. I thought about this and rechecked with Audiox pcb layout and I am good..

The opamp heat is what troubles me first of all because I am new to these things and I don't know what to expect and second of all I have been working on this piece for almost  10 months on and off and as this piece of gear is for live audio and I don't want it to go bad live.

I have measured temp after power on for 15 minutes.. not passing audio.. the opamps go from 39.9 ºC to 41.5ºc.. ambiente temp is around 26ºC.. this is normal?? operating temp for ne5532 is upto 70ºc son it might be ok  I just don't know..

I will hook up audio and force these guys to working conditions and measure again..

I want to repeat my question on opamp sockets... could these make opamps work weird or hotter(the opamps don't have a really tight fit although they pass audio and it seems to work ok.. in fact it sounds very nice).

thanks

 

[pablobolche]

Ok found my answer here on the forum... sockets aren't very good for heat dissipation..

The datasheet of OPA604 advises to solder the IC directly to achieve best heat dissipation

I guess I will keep the sockets for now but will solder the chips directly to the PCB after the audio company uses  the machine for a while in real world ..

Thanks