0
\$\begingroup\$

I have the following DAC output stage in a design I'm working on, which takes the differential outputs of a DAC to create an unbalanced audio signal, 5.7Vpp.

For the R and C I would like to use 0402 chip capacitors and resistors. How do I figure out the required watts for the resistors and if the standard 1/16 watt power rating will be adequate?

The differential inputs to the stage are 2.8Vpp signals centered around 5V.

I'm not an everyday user of opamps, but have used TINA-TI by texas instruments to simulate the stage to make sure it works (i.e. that it has the current gain and frequency response).

The circuit is powered by +14/-14V rails and I did put an amp meter in the simulation software on the rails, which gave me 4mA both for each of the rails, but that doesn't tell me everything?

I also have a second (similar) version of the below circuit in which I want to use the opa1622 headphone amplifier opamp to drive headphones, and there this question will be even more important since that opamp is capable of outputting a lot more power than the ne5532.

dac output stage

\$\endgroup\$
2
  • \$\begingroup\$ Hi b20000, welcome to eesx. I edited out the thank you and signature part from your question, this isn't something we normally include on this board due to the Q&A format, which is different from a forum. As per your question, since you already have the circuit in the simulator, add a load to your circuit and run a transient simulation with 1kHz sine wave coming in, maximum Vpp, I bet you can plot the power in each resistor. Factor in the tolerance and see if you fit 1/16W. \$\endgroup\$ – Vladimir Cravero Nov 20 '16 at 12:04
  • \$\begingroup\$ @VladimirCravero how do you typically plot power in the resistors? In the software I'm using (freely available, it's called TINA-TI) I can get a table of DC results which show currents and nodal voltages, but I can't seem to find an option to plot power in the resistors. \$\endgroup\$ – b20000 Nov 21 '16 at 10:26
2
\$\begingroup\$

Well there is clearly less then 6V in play anywhere, so that is ~2mA into a 3K ohm load, giving 12mW, so 0402 is probably fine from that perspective. There is however something to be said for physically larger parts as they tend to have lower excess noise.

I note you are specifying 0.1% parts, for resistors, but that PC92 & 93 are not so tightly specified (Good luck with that!)? This will hurt CMRR at high frequency.

Also R110 should be 100 ohms or so, to ensure stability when driving long cables.

In the headphone case R110 should be 20 ohms or so, but may dissipate some power if the cans are shorted, or the jack is only partly inserted, calculate this assuming a short to ground and full scale output, a 1206 or so is usually sufficient.

I note the lack of an output coupling cap, usually a good idea, 100uF or so with a 10k bleeder across the socket, particularly for headphones, DC offset matters (But for cans the coupling cap should be much larger to support low impedance units). If you think a phantom power accident is possible then make this cap a 63V part and wire it positive terminal to the socket, with the ESD diode where PC148 is, that way the diode is behind the build out resistor which will limit the surge current.

\$\endgroup\$
4
  • \$\begingroup\$ thanks for your reply, would it make a big difference to use 0603 instead of 0402? Regarding the specs of PC92/93/96, the datasheet of the DAC mentions the use of film caps, so I could use these (although they seem really expensive) - is NP0/C0G also an option or are film caps really better here? You mention "long cables" but how long are you thinking? This is in a studio application setting, so I'm thinking the typical user won't use cables longer than 10 meters (about 30 feet) and if they do then they probably go into a mixer first close to their setup? \$\endgroup\$ – b20000 Nov 21 '16 at 10:20
  • \$\begingroup\$ Regarding the headphone case, are your suggestions still applicable when using the opa1622 headphone opamp driver? It seems, when looking at the datasheet, that it takes care of the output requirements for driving headphones? \$\endgroup\$ – b20000 Nov 21 '16 at 10:24
  • \$\begingroup\$ NP0/C0G are pretty much blameless as caps, and are now available in usefully high values, I would much rather use them then film, especially if doing a SMT board (Film is a pain for thermal profile). In a studio situation I would always include the build out resistors, and I would at least make the line balanced if not differential, single ended audio gets sworn about. 0603 is better then 0402 in general, but if you have a space constraint I wouldn't worry about it unless you measure a problem. A few tens of ohms is a good idea with a headphone driver, if only to improve ESD resilience. \$\endgroup\$ – Dan Mills Nov 21 '16 at 12:52
  • \$\begingroup\$ thanks so much for your comments - if I wanted to make a balanced out variation of the above output stage what would be the best way to do it? \$\endgroup\$ – b20000 Nov 21 '16 at 17:39
1
\$\begingroup\$

Tiny chip resistors have very fast thermal response: 1mm size are approximately 11 milliSeconds. The 0102 resistors are thus even faster (scales by 1/size^2).

Problem is: if you use tiny resistors to SET THE GAIN, the low frequency response of the resistors is excellent because the R has adequate time to fully change value during bass music content. This response ----- as gain-settling resistors -------- causes AM-modulation of all the musical content, even the high notes.

Walt Jung warned about this. So use physically-large resistors to set gain.

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.