The Objective2 (O2) Headphone Amp DIY Project

[RocketScientist]

Originally Posted by dewasiuk

I finished building the rest of the amp today and all the measurements seemed correct, however the regulators seemed pretty warm, almost hot to the touch. How warm should they get? I checked over the pcb and all the solder joints seem fine so what else could be happening?

What AC wall transformer are you using? With 12 VAC the regulators should be warm but not hot. With 16+ VAC they get hotter, especially if the batteries are also installed and charging or the amp is playing loud into difficult headphones.

The actual case temp is around 35C - 55C (95F - 131F) depending on which transformer, if the batteries are installed, how hard the amp is working, etc. The 7812 on the inside runs a bit hotter than the 7912 on the outside as it doesn't have the aluminum case (or as much free air if the board is not enclosed) to radiate heat to.

Given all the talk about the regulators lately, it's worth sharing the actual calculations for anyone who's curious with the amp playing music at a typical "loud" listening level with my DT770-Pro 80 headphones. The regulators are rated for continuous operation with a junction temp of 125C:

12VAC = 39C
14VAC = 47C
16VAC = 62C
18VAC = 71C
20VAC = 79C

From the above you can see why I advised Jokener to aim for 14 VAC for the custom transformer for the European group buy. It's high enough to meet all the needs of the amp without getting the regulators very warm at all. But anything up to 20 VAC is still fine.

Here are the calculations for anyone who's curious. The average amplifier power consumption is under 40 mA, the batteries need around 10 mA, and the quiescent current of the regulators is (worst case) 5 mA, so that's 55 mA total current per regulator.

With the WAU12-200 12 VAC transformer the unregulated voltage is 16.1 VDC so Vdrop = 16.1 - 12 = 4.1V. 4.1 * 0.055 = 0.23 watts total dissipation. The worst case theta junction to ambient is 60 C/W. So 0.23 * 60 = 13.8C temp rise over ambient (some datasheets list a value of 40 which would mean considerably lower temps). With a 25C ambient, that gives a junction temp of 39C (102F) and a case temp well below that (under 35C).

With the WAU16-400 16 VAC transformer, the unregulated voltage is 23.1 VDC so Vdrop = 11.1V. That gives 0.61 watts total dissipation and a rise of 37C above ambient or 62C junction temp with a case temp around 50C (122F).

With a 18 VAC transformer Vdrop is 14V giving .77 watts and 46C rise or 71C junction temp.

With a WAU20-200 20 VAC transformer, Vdrop is 16.5V for 0.9 watts and 54C rise for a 79C junction temp. This is still well under the 125C continuous rating and the 150C auto shutdown temp. But it is hot to the touch.

If you work the amp really hard, it draws about 60 ma for both channels, which increases the total current to 0.075 amps. Even with a worst case 20 VAC transformer, you get 1.24 watts and a rise of 74C above ambient or around 100C junction temp. If the ambient is really warm you might get close to the 125C rating but you'd have to really work at it.

[RocketScientist]

I was posting the above while dewasiuk and MrSlim posted their comments. With batteries installed, the voltage across C2 and C3 drops slightly so my 23.1 vs dewasiuk's 23.8 is about right.

With the amp idling, and no batteries installed, the current should be around 30 mA total (25 mA for the amp 5 ma for the regs). So Vdrop = 23.8 - 12 = 11.8 volts * 0.03 amps = 0.35 watts per regulator. That's temp rise of 21C over ambient or a junction temp around 46C. The case temp should be comfortably under 40C (104F). That should not be too hot to keep your finger on. If you can find any 9 volt batteries, even cheap disposable ones, you can measure the current as in the documentation. Or you can lift up (desolder) one end of D1 or D5 and insert your DMM in series using the current jack and range to measure the current on AC power. Do that without headphones connected just in case the power management circuit isn't working right and you end up with only one rail working.

[nezbleu]

Originally Posted by RocketScientist

Or you can lift up (desolder) one end of D1 or D5 and insert your DMM in series using the current jack and range to measure the current on AC power.

Good point; if some people are having issues that would point to excess current draw, the obvious next step would be to measure the current draw! Do that with no opamps installed. If the number is even slightly high you have a short somewhere. If not, install opamps one at a time, measure again, until you find a jump in current draw.

This really isn't rocket science , if a regulator is overheating then, as a DIYer, your job is to find the fault. Start by checking voltages, if that doesn't pan out start measuring actual current.

[nezbleu]

I just wanted to add another data point to the discussion. I already posted here to say that I built two O2's, and they worked just fine and sounded superior to other HPA's I have heard.

However, I hadn't used them much, and I was powering one at a time via some batteries that I didn't have a lot of confidence in (they don't even have a mAH spec printed on them). Before I read RS's comments about high-mAH-rated batteries usually having high internal impedance, I ordered 4 from eBay rated at 450mAH. They arrived a couple of days ago; I checked their voltage and it was 9.05V, so I assumed they had arrived fully charged (more-or-less) and I inserted them into both amps and turned them on. One was left on but just idling (no input, no load), the other had input from an mp3 player on repeat and was driving a pair of 32-ohm iem's at slightly uncomfortably loud levels. Of course, I had neglected to check the battery level of the mp3 player, and it died long before the amp. The upshot is that about 16 hours later I checked them and both LEDs were glowing but neither amp was producing output. I did not observe any hysteresis effects, but perhaps if I had checked sooner...

Today I recharged one set of batteries in a charger; when they were done the measured voltage was 9.6V. During the same time the other set was in an O2 plugged into AC; after about the same time period the measured battery voltage was around 9.1V, so I assume the second set charging in the O2 was not fully charged.

Tomorrow, after the one set has run down I will plug the amp into the AC (I am using a 500mA 16V xformer), play music at unhealthy levels, and use the temperature probe of my DMM to measure temps of voltage regs, mosfets, opamps, etc.

To summarize: 2 amps, no observed low-batt hysteresis effects, no unusual (or even remarkable) observed temperatures, outstanding performance. If only the front panels would arrive (I understand they are on their way)!

[nezbleu]

Just to be complete, I will add the following: All the parts I used were either the preferred part specified in the BOM or one listed as an acceptable substitute in the BOM, with the following exceptions:

R1,R2: I used ERG-2SJ221A Panasonic Electronic Components Metal Oxide Resistors
They fit, they are rated at 2W instead of 1W, and they at least specify a temperature coefficient (although not impressive).

Electrolytics: I used Panasonic FR's, which have very low ESR and generally excellent specs, and I used 35V parts even for C8 and C9 (based on a general belief that caps rated for a higher voltage do a better job even at lower voltage -- like most capacitor-related opinions this is based on hearsay and is essentially religious in nature).

Oh, yeah: Mouser were out of the right-sized 2.7M resistors for R25, so I used 2.8M. I gather that should have made the hysteresis issue worse, but as I said I have not yet observed any problems of that nature.

Like I said, everything else is bog-standard.

[robodoc]

Originally Posted by agdr

Agreed! That is why the voltage measurements. If you get +/-12Vdc with the switch off, then your power supply is OK, at least with no load. And when you turn the switch on (all the chips still out, of course) if you get +/-12Vdc then things get interesting, since the power LED and a series resistor are right across the power supply rails after the power switch.

Okay the power supply seems to be okay, without U2 and the other amps I measure +11.91 Vdc and -11.96 Vdc with switch "off".
Switch "on" changes the measures to +11.75Vdc and -11.8 Vdc.
Installing U2 changes nothing with this measurements. Seems okay to me.
I will do some more measurements around the LED later and report again. Thank you so far.

[robodoc]

Originally Posted by RocketScientist

So if the LED isn't coming on, on AC power, the only possibilities are the LED itself, R6, a bad solder connection/trace, a short or something drawing way too much current (the regulators would get hot), or something upstream of the power switch which is easy enough to check if you have a DMM. Pins 3 and 6 of the switch should be close to 24 volts with the power switch on. If not, are the regulators getting hot? If so, there's a short or bad component somewhere in one or both rails to ground or shorting the rails to each other.

I have 23.6 Vdc there. Nothing is getting warm.

Originally Posted by RocketScientist

If there's 22+ volts across pins 3 and 6 of the power switch, check the voltage across R6, it should be about 21-22 volts. If it's the same as you measured across the power switch, the LED is either shorted, or there's a solder bridge. If the voltage is much less than 21 volts across R6, the LED has either failed open, or isn't soldered fully, or the traces to it are damaged.

Okay we're probably on the right way to find a solution ... I have 23.2 Vdc across R6 and 0.4 Vdc on the LED. Should be less on R6 and more on the LED terminals, right?

Originally Posted by RocketScientist

You can check the voltage directly at the LED terminals. It should be around 2 volts. If it's much higher, the LED is bad (or in backwards). If it's much lower but not zero the LED is bad. If it's very close to zero the LED is bad or there's a solder bridge.

The LED is bad, I finally soldered it out and replaced by one from another set I have here. Fine to see the new LED is doing the job!
There measuement are now 21.8 Vdc across R6 and 1.8 Vdc across the LED.
I will do the initial tests now.

I still have serious concerns what happened to make the opamp pop. Do you have any hints what I shall prove again before connecting headphones? I can't see any obvious faults on soldering or component disorder.

Originally Posted by RocketScientist

Are you using battery or AC power for your testing? One other person installed the battery terminals backwards which will also keep the LED from coming on and can damage the op amps but probably not make them explode (the otherwise high reverse current is limited by the 220 ohm resistors R1 and R2). So, regardless, make sure the "+" terminals of the batteries are the left (input side) of the board.

Testing just with AC power. I don't even have the battery connectors soldered yet.

Thanks a lot for your help!!

[cheapskateaudio]

Originally Posted by RocketScientist

With the WAU16-400 16 VAC transformer, the unregulated voltage is 23.1 VDC so Vdrop = 11.1V. That gives 0.61 watts total dissipation and a rise of 37C above ambient or 62C junction temp with a case temp around 50C (122F).

Thank you, how did you get that case temperature?

I have .615 watts dissipation based on my measured numbers. My unregulated voltage is over 24 volts.

The amp draws 35ma, and the regs have 15ma at the outputs.

Uncased, I measured the regs at 77C after an hour or so.

[qusp]

all of the math you need has already been posted at the top of this page (or the previous page if this was pushed over)

[Turbon]

Originally Posted by cheapskateaudio

Thank you, how did you get that case temperature?

I have .615 watts dissipation based on my measured numbers. My unregulated voltage is over 24 volts.

The amp draws 35ma, and the regs have 15ma at the outputs.

Uncased, I measured the regs at 77C after an hour or so.

It thought 20VAC was the max...

Brgds