Hot for a component is usually specified on the datasheet.
Having the heatsinks installed doesn't mean correctly installed. Heat has to transfer to the heatsink, and the whole has to be at the same temperature. If the component is hotter than the heatsink, then the transfer is not happening as it should.
If the whole is at the same temperature and still hot to touch, then you need a larger heatsink. And so on.
But the current we are talking on this project are very low, so parts, particularly the SMDs on the board, shouldn't be hot at all. Perhaps you do have some sort of oscillation going on, though I doubt it might be because of low esr capacitors, except on some border case. Change the capacitors at the output of the regulators and see what happens, or leave just a film bypass cap.
You're welcome
Later I'll search and post something about it
The problem is not with input capacitors but with output ones...
Well, i do measure 55*C on the chip and roughly 52*C on the heatsink just now - i think i've smeared enough high-quality thermal compound in there (too much isn't good after all...). The 78XX are specified up to 120*C(!) absolute max, so ~50-60*C should be way in the safe range, right? I'm more worried about the AK @ 45*C... ;/
So you propose to completely remove the (1117-5v?) electrolytics (which would be C31, right?), or also maybe also the big buffers for the 7812 (C38?), correct? (i'm a little worried i could break something here, so i'd better try asking some dumb questions before i'll break something...).
Thanks a bunch! Regards
- Dario
Probably an obvious question, but I have to do it: did you check that the heatsink is isolated from all pins, particularly the middle one?
Where do I say remove the regulator 1117-5v? I say remove the large electrolytic capacitor at the output of 78XX and leave the film cap. This is just to see if it's the capacitor which is making any problem.
The SMD regulator should be just warm to touch.
In any case, 55 degrees C is not that high. Tighten up the screw holding the heatsink to the regulator. Temperature should be the same on both if transfer is OK.
Martin Clark of Acoustica,com has a 4 part tutorial on the use of 3 pin regulators, particularly focusing on the LM317/337 adjustable types. This tutorial is worth reading, especially part 3 which contains a discussion on the merits of the ESR value of the output cap. Even though Martin is discussing the adjustable regulator, a lot of what he says about output capacitor ESR is relevant to the fixed regulators.
Martin argues that Arbitrarily-low ESR is not necessarily a good thing and suggests that a good quality electrolytic output capacitor with an ESR of about 0.3R will be "critically damped" and suitable for the LM317. But you also need to read the notes.
If you go to Martin's website at Just one thing about music - when it hits you feel no pain, you can scroll down the menu on the left side of the page and click on "using 3-pin regulators. part 3", or try this url Using 3-pin regulators off-piste: part 3
Hey,
thanks again - i wouldn't have removed the 1117-5v, i asked whether i should remove the 100uF cap buffering the 1117-5v supply in addition to the large 2200uF of the 7812. Sorry if i got it wrong, i know my english can suck sometimes
.Thanks @ jnewbold too, will read later...
Regards
-Dario
*EDIT: I know 55*C isn't hot for a 7812/1117, what i'm worried about is the AK @ ~45*C - this chip has a maximum rating of 85*C and its cold in my room in winter, so ... i just wonder whether this is ok? on the other hand typical is 60mA@5V + ~10mA@3.3V? it *should* get a little warm from that...?)
Last edited:
From LM1117 (the 5V one) datasheet:
1.3 Output Capacitor
The output capacitor is critical in maintaining regulator stability,
and must meet the required conditions for both minimum
amount of capacitance and ESR (Equivalent Series
Resistance). The minimum output capacitance required by
the LM1117 is 10μF, if a tantalum capacitor is used. Any
increase of the output capacitance will merely improve the
loop stability and transient response. The ESR of the output
capacitor should range between 0.3Ω - 22Ω. In the case of
the adjustable regulator, when the CADJ is used, a larger
output capacitance (22μf tantalum) is required.
This NatSemi appnote explain very well why LDO regulators (like LM1117) don't like low-esr caps.
Standard voltage regulators (like LM78xx) are (much) less affected by the problem but, regarding heat, it must be taken in account that a big value low-esr cap require much more current to maintain its charge, thus increasing regulator's heat generation.
So my recommendation was right (better a non low-esr cap at the output of regulator) for both LDO and standard regulators but for the latters stability it's not a problem while overheating remain a problem.
Hey,
i've just toyed around with the 2200uF caps of the "first level" regulators, the 7812, 7912 and 7809 a little. Taking them out has no effect whatsoever. I haven't dared to get rid of the 1117-5v, the datasheet vetoed...
There was no effect measurable. After ~30min listening my 7812 was on ~50*C again, the 1117-5v and AK4393 were at ~45*C. A friend also found a neat paper which is basically about those AKs - http://www.scheme.ro/files/diverse/diverse pdf/sigma_delta_mutibit.pdf. Judging from this the temperature is totally normal. As when putting my power-meter serial in trafo -> dmm -> dac-board for every trafo winding i observer 30mA to the 7812, 25mA to the 7912 and something very low to the 7809. This also correlates to the power estimates given by the datasheets of the 4 sinks (receiver, dac, opa and 74hc04). So i guess i'm not really having a problem here after all? From what i understand at least?
Nevertheless i'd be really grateful if someone could measure the AKs surface temperature ...
Kind Regards, and thanks for all the reading! I'm still in the process of understanding it ;/
- NebuK
i've just toyed around with the 2200uF caps of the "first level" regulators, the 7812, 7912 and 7809 a little. Taking them out has no effect whatsoever. I haven't dared to get rid of the 1117-5v, the datasheet vetoed...
There was no effect measurable. After ~30min listening my 7812 was on ~50*C again, the 1117-5v and AK4393 were at ~45*C. A friend also found a neat paper which is basically about those AKs - http://www.scheme.ro/files/diverse/diverse pdf/sigma_delta_mutibit.pdf. Judging from this the temperature is totally normal. As when putting my power-meter serial in trafo -> dmm -> dac-board for every trafo winding i observer 30mA to the 7812, 25mA to the 7912 and something very low to the 7809. This also correlates to the power estimates given by the datasheets of the 4 sinks (receiver, dac, opa and 74hc04). So i guess i'm not really having a problem here after all? From what i understand at least?
Nevertheless i'd be really grateful if someone could measure the AKs surface temperature ...
Kind Regards, and thanks for all the reading! I'm still in the process of understanding it ;/
- NebuK
DMM's amp meter is not accurate, a better method is to use a resistor in series and measure the voltage drop across the resistor. Since you are measuring the current drawn from the transformer, it is AC, make sure your DMM can measure RMS. A 10R resistor is good in your situation, easy for calculation.
While I am only running 12-0-12, I still get over 17VDC after the first filter cap. Therefore, my 7812 and 7912 need to drop 5+V. However, they are only lurk warm without heat sink. Did you use isopropyl to clean up any possible stray flux? The stary flux sometimes can create a semi-short;hence, draw excessive current.
While I am only running 12-0-12, I still get over 17VDC after the first filter cap. Therefore, my 7812 and 7912 need to drop 5+V. However, they are only lurk warm without heat sink. Did you use isopropyl to clean up any possible stray flux? The stary flux sometimes can create a semi-short;hence, draw excessive current.
Thanks Clave + Erin for your efforts with this board. I have made most of the changes recommended, and really impressive performance.
The last change i have made was C40-43, 1500pF Wimas. Although this solved the raspiness at the treble, it also took away a lot of the realism and attach of the piano and other vocals.
Does this make sense? Any suggestions gratefully received.
The last change i have made was C40-43, 1500pF Wimas. Although this solved the raspiness at the treble, it also took away a lot of the realism and attach of the piano and other vocals.
Does this make sense? Any suggestions gratefully received.
Hey,
i've cleaned the LMs a little, not that much. I've soldered with no-clean flux, i've measured the resistance of the LM pins to each other, everything is kinda unmeasurably high for my DMM.
Audio is coming out nicely, there is no hearable distortion on both headphones (directly on the dac) or speakers. I could start to measure ...
The caps after the 1117 are same as also FCs, 47uF.
I'm still not sure what to make of this ... If this i really a ESR issue, couldn't i just add a resistor serial with the capacitor to test (this seems a little safe and saner to me than trying the 1117 without a electrolytic...?).
Thanks again, Regards
- Dario
PS: I will probably try my other board later... maybe its also a very-very-slightly-defective-ak?!
i've cleaned the LMs a little, not that much. I've soldered with no-clean flux, i've measured the resistance of the LM pins to each other, everything is kinda unmeasurably high for my DMM.
Audio is coming out nicely, there is no hearable distortion on both headphones (directly on the dac) or speakers. I could start to measure ...
The caps after the 1117 are same as also FCs, 47uF.
I'm still not sure what to make of this ... If this i really a ESR issue, couldn't i just add a resistor serial with the capacitor to test (this seems a little safe and saner to me than trying the 1117 without a electrolytic...?).
Thanks again, Regards
- Dario
PS: I will probably try my other board later... maybe its also a very-very-slightly-defective-ak?!
Hey,
i've just tried the other board with same parts (same result, hot regs+ak) as well as some regular cheapo noname caps for the 7812, 7912 and 1117-5v (those are certainly NO low-esr) ... still the same ;(.
I'll make a few nice closeups later, but please don't laugh at the chaos i've produced...
Thanks again!
i've just tried the other board with same parts (same result, hot regs+ak) as well as some regular cheapo noname caps for the 7812, 7912 and 1117-5v (those are certainly NO low-esr) ... still the same ;(.
I'll make a few nice closeups later, but please don't laugh at the chaos i've produced...
Thanks again!
I have a question. Except for the opamp, does the 15.0.15 V line powers anything else? I'm curious if I can just remove it, because I am going to do so with the opamp and put a tube output.
I can check myself, but I have the ideas now, and the DAC is enclosed. If no one replies, I'll disassemble it tomorrow and check it out. Or I will search for the PCB image.
I can check myself, but I have the ideas now, and the DAC is enclosed. If no one replies, I'll disassemble it tomorrow and check it out. Or I will search for the PCB image.