SparkFun Forums

[lordsteve]

I'll add that I am powering my project from a Texas Instruments TPS61070 3.3V boost converter, not from batteries direct or from a wall wart.

[riden]

After following this thread for a bit, I decided to take one of these connectors and make some measurements. Referencing the data sheet that the OP provided, I can confirm the following information:

1) Ground (sleeve) is pin 1

2) Tip is pin 5

3) Ring is pin 2

These are exactly as described in the data sheet. However, the following is also true:

When no plug is inserted, pins 2 and 3 are shorted together and pins 4 and 5 are shorted together. That information is not shown in the schematic.

So if you connect pins 3 and 4 to ground, the amplifier input could be shunted to ground when a pin is not connected. If you don't want this to occur, leave pins 3 and 4 floating and connect pins 2 and 5 to your amplifier.

[lordsteve]

Riden: Thanks for your post. I do not connect pins 3 and 4 to ground on my board. Could you also do a continuity test like I did? I checked continuity between pins 2 and 5. When no plug was inserted, the resistance was infinite (humor me). When I inserted the plug into the jack, there was a point when pin 2 was shorted to pin 5 -- the resistance dropped to 1 (one) ohm -- then went back to infinity after insertion was complete.

[riden]

I checked continuity between pins 2 and 5. When no plug was inserted, the resistance was infinite (humor me). When I inserted the plug into the jack, there was a point when pin 2 was shorted to pin 5 -- the resistance dropped to 1 (one) ohm -- then went back to infinity after insertion was complete.

That is totally consistent with what esklar81 and JonChandler said earlier. I can get this to happen with my connector as well. When you insert the plug there will be a time when the tip of the plug will short pins 2 and 5 (the tip first touches pin 2. As the plug insertion continues it will also touch pin 5. Finally when the plug is entirely inserted the tip will not be touching pin 2).

If shorting the inputs together is fritzing your amplifier, you may have to turn off of the power while you insert the plug.

Hope this clears things up a bit.

[esklar81]

The jack was not wired to anything and there was no plug inserted. The resistance I deemed a short was 1 (one) ohm. When I say left is shorted to right, I mean that the left audio signal is shorted to the right audio signal. This happens momentarily when a plug is inserted. When the short occurs...brown out. When I manually short the left audio channel to the right audio channel, same thing...brown out.

I'm still failing to understand your hardware. If "the jack was not wired to anything", then how can "the short occurs...brown out"??? What's being shorted or browned out?

I'm assuming this is a common problem and must happen frequently. There has to be some way to handle it.

While it's a "common" occurrence, I don't think it's all that common a "problem". In typical audio applications, people tolerate not having good audio in either the speakers or the headset while they're inserting the plug. As long as the signal quality isn't impaired by the short for more than the fraction of a second it takes to insert the plug, I doubt most people notice this or would care about it if they did.

That said, if you're doing something that won't tolerate the transient, then perhaps you should change to a different type of connector. There are lots of multi-pin connectors from which to choose, starting with the RJ-xx connectors that are currently used for telephone and ethernet cabling.

Eric

[lordsteve]

As far as my hardware, when I tested a jack for continuity, it was a spare, not-connected jack.

The problem in my project is not audio quality, but rather that when inserting the headphones plug, the shorting of the left and right channel inputs causes a brown out condition.

Thank you.

[riden]

The problem in my project is not audio quality, but rather that when inserting the headphones plug, the shorting of the left and right channel inputs causes a brown out condition.

I have a few questions:

1) If you have the power off, plug in the headphones, and turn the power on, does it work as expected?

2) Could you define the "brown out" you are seeing? Is this is a temporary condition or does the amplifier stay in this state requiring a power down of the circuit?

3) If you aren't driving the amplifier with any signal (e.g., music) and you plug in the headphones do you get the same result?

I agree with your earlier post that it is very common with most 3.5mm jacks to have the tip and ring briefly connected together when the plug is inserted. I would think that a vendor of an amplifier chip designed to be used with these types of jacks would address this situation so we're probably missing something.

[lordsteve]

Hi, Riden.

1) If the power is cycled with the headphone plug fully inserted, the device works properly and I can hear the audio. If I then pull out the plug, the device "browns out".

2) By "browns out" I mean that the device shuts off: the microcontroller stops running, the screen shuts off, etc. It is a temporary condition; the device powers back up when the plug is not shorting left audio to right audio. (If there is a better term than "brown out", let me know.)

3) I had not thought to try that, but will do so and post back.

[riden]

I think that we can conclude that the problem is caused when both outputs of the device are momentarily connected to each other, the ring is momentarily connected to the shield (shorting one channel to ground), or a combination of both conditions. The MAX4410 has short circuit protection. My guess is that your power line is dropping too low for too long for the microcontroller (i.e., your brownout situation). I would try providing some power line isolation and filtering that will maintain power long enough for the short circuit protection in the MAX4420 to take over.

Try something like the following. I think it may solve your problem:

Put a 47 uf or 100 uf capacitor across the microcontroller/LCD VDD (power) pins. Power the MAX4410 from the battery or power supply. Then connect either a low value of resistor (10-15 ohms? I'm pulling that value out of the air) -OR- a Schottky diode (because of the low forward voltage drop) -- anode to the battery/supply and cathode to the microcontroller power. The idea is when the output is momentarily shorted to ground, the capacitor will be able to keep the microcontroller running until the MAX4410 shutdown circuit limits the current or the short is removed.

BTW, the reason I asked my third question was to determine if the problem still occurs when both outputs are at a zero level. Probably not real important now seeing that we have a better picture of what is happening.

[lordsteve]

What does the resistor do in that configuration? Does it regulate current out of the capacitor when it discharges in a "brown out" condition?

[riden]

Yes, it slows the discharge of the capacitor until the protection kicks in. You might have to play around with the value. It might be easier to use a diode if your circuit can tolerate the forward voltage drop of the diode.

[lordsteve]

One thing I noticed in the MAX4410 datasheet is a sentence about the soldering profile. It says that hand soldering is not allowed. I suppose it is possible that hand soldering, especially my very n00b soldering, could do something to the short circuit protection of the MAX4410. The datasheet says very little about the short circuit protection anyways. I guess I will have to use another headphone amplifier IC, because I want this project to be productized. Any suggestions are welcome. Thanks all.

[lordsteve]

Could I add a small surface-mount isolation transformer to each channel to isolate the headphones from the chip? What parameters of the transformers do I need to look for?

[riden]

One thing I noticed in the MAX4410 datasheet is a sentence about the soldering profile. It says that hand soldering is not allowed. I suppose it is possible that hand soldering, especially my very n00b soldering, could do something to the short circuit protection of the MAX4410.

Hand soldering is not "allowed"/recommended for the UCSP version (the package with the bumps). The SSOP version is rated for 10 seconds at 300 degrees C, so you're fine. I doubt your soldering would damage the short circuit protection while the rest of the chip works fine.

[riden]

Could I add a small surface-mount isolation transformer to each channel to isolate the headphones from the chip? What parameters of the transformers do I need to look for?

As the chip is designed to directly drive a 32 ohm load, I have no idea if adding transformers will help or hurt you. I would think the later.

It seems from the past discussion that it is the microcontroller and/or LCD that is getting fritzed by the temporary shorting of the output terminals. As the MAX4410 has short circuit protection, I'd focus on protecting the power to the microcontroller and LCD. Try some of the strategies I suggested earlier.