DIY: Surfboard Amp For SMD On Any Proto Board

Some new buffers are only available in SMD packages. I suppose DIP sockets don't supply good enough thermal regulation for 1/2watt devices. Fortunately, there is a solution for the enthusiast who wants to build a good phone amp in a 1-of project using mostly general-purpose parts. Tiny surfboards adapt various SMD devices to the 100 spacing of standard, thru-hole protoboards.

The photo below is poor quality but it shows the LME49610 buffers on the left side. They are driven by a 4562 dual-amp which I happened to have. I built this contraption after reading the application notes. The significant different is the pots, instead of fixed resistors, in the feedback loop.  I am not an engineer, this is a hacker experiment which happens to work well. That is, it matches the performance of the Fubar3 amp.

The schematic is short on details. However, the 49610 datasheet provides good advice about bypass caps.

Note that there is plug for a separate power supply. There are also interconnects for the input RCA jacks the the earphone jack.  It is easier to build with a modular process and easier to fix the inevitable problems.

Casework is always a challenge. I could not find flush-mount jacks at the local Sayal so my trusty glue gun came to the rescue. A headphone jack from an old tape deck, is glued to a plastic standoff which is glued to the front panel.

All the common "grounds" converge to the volume pot which is the only panel item not insulated from the case.

The amp board seems relatively easy, or at least there is not much to think about. My next post will have details about headphone protection and bugs found.

Surfboard amp.. part 2.

Since the 49610 can drive over 200ma, I think it is important to limit the current in a way that does not hamper impedance matching or general performance.  So the linear power supply employs 2 LM317 regulators. The first is in a constant-current configuration arbitrarily set at 80ma, the other is an adjustable regulator which I set at 10v. Well, it has to be set to something

The regulator caps are 1000uf,25v.  I first used 16v but discovered that 12vac wallwarts output 14vac which becomes 18vdc on the caps. Don't forget the small caps for transient response.

The negative supply (not shown) looks like the above but uses LM337. I have not figured why this section powers up slowly relative to the positive.  The amp sees a big difference in the supply rails and tends to latch. This loads the supply which then limits the current thus maintaining the latch. Nothing is harmed since the 1st LM337 reduces the voltage to the second. And the headphone jack has only 100mv. I clear this condition by unplugging the phone or turning the volume to 0.

I did not see this problem when everything was sealed and connected for an overnight test. Perhaps the input signal makes a difference.  We will see.

Quote:

Originally Posted by fubar3 

Surfboard amp.. part 2.

 

Since the 49610 can drive over 200ma, I think it is important to limit the current in a way that does not hamper impedance matching or general performance.  So the linear power supply employs 2 LM317 regulators. The first is in a constant-current configuration arbitrarily set at 80ma, the other is an adjustable regulator which I set at 10v. Well, it has to be set to something

I have not figured why this section powers up slowly relative to the positive.  The amp sees a big difference in the supply rails and tends to latch.

 

Well, there were too many components packed onto one proto board.  One limiting regulator was defective probably as a result of poking around. So a simpler design was needed.

I found that the buffered amp works fine with +-5vdc from U1 and U2.  The power input side could be mistaken for a virtual ground. But no, it is a rail splitter which feeds +-8.6vdc across filter caps C1 and C2.  I am using a moderate operating voltage plus the resistance of the rail splitter to protect the phones in the event of an unexpected offset.

I applied a variable DC voltage while monitoring the output across MDR-V6 phones which have 45ohm resistance.  The worst case voltage was 300mv which the V6 seemed to tolerate ok.

Options: The components to the left of node1 and node2 can be removed. Then D2 and D3 can be connected together to one side of the AC input. The other side of the AC can be connected to the common node. Now you have a conventional half-wave input  which will give 18vdc across C1 and C2. If the output remains at +-5vdc, the regulators will get slightly warm but remain safe.  The 2k pots can be adjusted for +-12vdc. Caution: An amplifier fault will damage your phones.

If you go with the low power rail split, U1 and U2 can probably be replaced with fixed regulators.