SMD redesign or back to through hole?

[Copper Dog]

Doing some research and found your forum. Like the input I see so here goes;

Designed an audio system I call a Dual Transceiver Interface or DTI which mixes the audio output of two FRS/GMRS transceivers. It enables a user to monitor both radios and select either to transmit on from a single speaker mic.

The prototype works great however I made the huge mistake of employing 4.7 uF ceramic SMD caps for coupling on the inputs. (On the breadboard I used polarized electrolytic's. It's what I had on hand). I've just added another stage to up the output for better sound when using the speaker as opposed to the usual plug in ear piece unfortunately the audio quality degraded considerably. I've cleaned this up by going to 47 uF/25 V coupling caps on the inputs and between stages and 220 uF on the output. Again on the bread board this is done with aluminum electrolytic capacitors.

Here's the concern - I can put SMD electrolytic caps in the design but things start getting big again. What SMD (0805 to 1210 range) caps do you guys use that are good for audio that would recreate or improve the bread board sound (terrific bass) I'm getting? Can someone recommend a good material? Most film styles don't seem to come in this range.

 

[TheSeekerr]

Do you understand the basics of the high-pass filter you're creating with the coupling caps you're using?

If not, would you mind posting a schematic so we can see?

 

[jitter]

Some SMD components actually take up more space on a pcb than their through hole counterparts.
Radial electrolytic caps are a good example of that, and I'm guessing it's what you're thinking of using. The trough hole part has the leads tucked under the housing whereas the SMD part has them sticking out at the sides. As a result, the footprint of the latter might need more space than the former.
When space is limited, a careful selection of components must be considered.

 

[Copper Dog]

Thanks for your interest. I often combine smd and leaded components in my designs and am very familiar with the various footprints. My usual stuff involves logic or motor control circuits on double sided boards. This is my first foray into audio. This particular circuit is a simple mixer and amp built around 741 and 386 op amps. The only thing fancy about it are the digital pots used for volume controls.

The circuit works great however I used ceramic smd caps because I had lots. I've refined the sound and boosted the gain to what I want but it's on a bread board with leaded alum electrolytic capacitors in the design. What I'm looking for is advice on selecting good quality smds for the pcb. Whats considered the standard in audio? Thin film? Polyester? This isn't a studio quality system. It's an audio mixer and transmitter selector that will be used harsh conditions. It's gonna take a pounding actually.

 

[rsavas]

I have used 10uF/50V/X7R in a 1210 package and they sound okay to me!!

 

[grommeteer]

If you use ceramic capacitors please Be aware that some materials tend to have lower capacity when charged to a voltage close to their rating. So they have a lower value than expected with a high DC voltage bias across them and distortion if there is large signal across them. Second problem may show in filters. X7R is better than other materials.

 

[rsavas]

Thanks for the info on MLC cap's, missing info on a data sheet probably. I use them for decoupling, for which they are excellent and for coupling, well you have read the papers about them, not the best there is, but it really depends on the system as a whole. Use them for signal coupling in a Class "D" amp, not a problem for me, very low bias voltage. For filters, probably not the best, due to accuracy, NPO(COG) another story.

Rick

 

[Copper Dog]

Thanks very much for the input. I'll take a look for X7R caps. I have plans to build myself an audio system this winter so I'll be checking back here for sure. It's great getting nice sounds out of a new circuit instead of clicks, beeps, flashing lights...or less lol.

I noticed some interesting changes to the sound while experimenting with this little circuit. Looking forward to trying my hand at his stuff.

Great forum.

 

[qusp]

if you can still get them, Wima PPS SMD film caps

 

[jitter]

This is my first foray into audio. This particular circuit is a simple mixer and amp built around 741 and 386 op amps.

For decent sount quality and low price, I'd use NE5532 instead of LM741.

 

[Ian Finch]

Yes, 741 are just awful for audio and were so, from day 1, more than 40 years ago. If you want only single op amps, use the 5534 version of the 5532 wisely suggested by Jitter or, with high impedance circuits, even JFET input type TL071. These opamps are very old too but yet to be beaten on bangs per buck.

 

[Copper Dog]

The 743/386 combo is working well for this particular circuit. The sound really is fine for the application. It's strapped to a tactical vest and the guy's wearing them only care that it be intelligible, loud enough and keep working. For the next audio system I will definitely use better chips.

I've spent a few hours here looking at various audio forums and articles regarding caps and the story is...conflicting. Thanks for your input it's all great info.

 

[Copper Dog]

Decided to go with the 4.7 and 47 uf Panasonic SMD electrolytic caps I have on hand for coupling. This probably isn't what a lot of you pros would do but the bread boarded circuit delivers some impressive sound with electrolytic caps in it. I'll upload this circuit from Eagle to get some opinions soon as I figure out how to do it here.

 

[Boscoe]

Put a 220nF C0G in parallel with the larger value SMD coupleing caps. See what it sounds like then even try 100pF ECHU as well, then see if you can hear the difference?

 

[richie00boy]

Those caps will be fine. Bit of a strange thing worrying about their impact on sound quality then saying you are happy with the 741 op-amps.

Although 741 are pretty rubbish as far as audio goes, they may end up being better for EMC than the others that have been suggested here. And for your application their audio quality will not be a problem.

 

[Copper Dog]

Those caps will be fine. Bit of a strange thing worrying about their impact on sound quality then saying you are happy with the 741 op-amps.

Although 741 are pretty rubbish as far as audio goes, they may end up being better for EMC than the others that have been suggested here. And for your application their audio quality will not be a problem.

Lol, you have a good point. I'm a neophyte here and it shows. This site is all about quality audio with good looks while this device is going to be strapped to a tactical vest and bashed in the dirt. Doesn't mean it shouldn't have good stuff in it I know but I'm working with what I have. The next version I'll upgrade. I'll include a picture of the prototype soon as I get a chance.

 

[counter culture]

I wouldn't worry too much about the 741, it really is marginal in a hi-fi application and doesn't have that good an input impedance (2M ohm?) however it is fine here. The NE5532, while it has very good fidelity, has an intrinsic input impedance of only ~300k IIRC. In a communications application ceramic caps in the signal path are quite acceptable. Values up to 10uF are available on a 1206 footprint. The bandwidth of interest is the telephone bandwidth, 300-3000 Hz, more bass than this may not actually contribute to intelligibility. The 3k upper frequency means that the slew rate limitation of the 741 is pretty irrelevant. It's unnecessary to pursue very high values of capacitance, you can simply raise the input impedance of the amplifier by increasing the resistances in the gain setting network. Don't let them get too high though, hiss may become pronounced. Make sure you have some shunt capacitance at the input to keep RF out of the audio circuits. 6dB down at the stated frequencies is often considered adequate in military radios.

 

[Copper Dog]

Excellent input Counter Culture. Are you suggesting a capacitor to ground in the mixer circuit after the coupling cap and prior to the 741 input? The circuit doesn't like closely held cellphones I know that.
The power saver mode of some radios causes a spike and annoying repeating click which I've managed to filter with a 27uH inductor however it can still be heard at near max volume. The system will normally ve used with an ear piece so the click is not an operational impediment however I still know its there and shouldn't be.

This smd board is ready to etch however I'll be posting this schematic for forum critique in a couple days before proceeding. I cringe at the beating I'll probably take over what to this forum is a simplistic audio circuit. It's not fancy but it has to work be and resistant to interference. In my experience simple+elegant = reliable. I look forward to the criticism

 

[counter culture]

Are you suggesting a capacitor to ground in the mixer circuit after the coupling cap and prior to the 741 input? The circuit doesn't like closely held cellphones I know that.

Yes. I would expect that regardless of whether you are using an inverting or a non-inverting configuration, you will have, at some point in the input network, a resistor from signal to ground, to define the input impedance and possibly to set the opamp output voltage.

If the opamp supply is single rail, you will have a resistive divider to bias the output voltage to half the supply voltage. You could be feeding the signal in at this point with a gain-setting network connected to the inverting input or you could be feeding the signal into an inverting configuration, but whichever the exact configuration, dual rail, single rail, inverting or non inverting you should have a resistor to ground between the signal and ground at some point which will define the input impedance independent of the opamp impedance, which can be expected to be considerably higher.

If you don't have such a resistor you should put one in. Its value should be in the tens or at most hundreds of kilohms (the opamp being ~2M). If it's a resistive divider the 2 resistors must be treated as a parallel combination, because the PSU has a low impedance, and the power rail is a ground from the point of AC.

You should shunt (parallel) this resistance with a capacitor whose capacitive reactance (Xc) is equal to the resistance at the lowest frequency that you want to exclude from the amplifier, or, to put it another way, the highest frequency you want to enter the amplifier. This means that half the (un)wanted power is shunted to ground, but if you pick an audio frequency, or one in the tens or even hundreds of kHz, then those frequencies that most readily propagate (the rest of the spectrum) will be attenuated to an even greater extent. Most hi-fi amps will be rolled off ~200kHz (so that the effect is minimal @ 20kHz) but no higher, and frequently a good bit lower. Many audio amps are intrinsically limited. Try and get the cap (and res) close to the opamp pin for the most predictable effect.

 

[Copper Dog]

Only had a few hours between business trips and couldnt get a schematic to upload. I'll try again tomorrow however in the meantime I'll just describe this simple circuit.

The Ua 741 is used as a single supply, inverting mixer. Supply is a 9V battery via a resistive divider. Each radio's audio output is fed through a 4.7uF cap into a DS1809 digital pot (10K) for individual volume control, another 1K resister and through a common 4.7uF cap and 1K resistor into the Op Amp. The gain is set at 20 and the output from this goes through a 10K audio taper into the LM386. The output from that is via a 47uf cap to an external speaker. De-coupling is done at each pin, the battery and a 5V zener rail for the digital pots.

Using counter cultures suggestion I calculate a 3uF cap and (I think) a 100R resistor in parallel to ground, prior to the 741 input, would limit the input frequency to just over 3 kHz or the top of the audio spectrum.

Am I out to lunch on this counter culture?