Topping L30 Headphone Amplifier Review

[snowboy]

I've heard a lot of people in different countries report that fedex, dhl etc deliver goods and then later invoice for this or that service. It seems the answer is to ignore them, or reply "your contract is with the supplier who paid you to deliver the goods, not the customer" and the problem goes away. Governments set various taxes and duties; it's not for a business to demand this or that "handling charge" or whatever. Fleece some other sucker.

In my country they charge the fees before delivering the goods.

 

[NgtFlyer]

These are not simple transformers but switch mode DC power supplies that are literally densely packed with electronic components. These PS can easily fail and often have to convert a LOT of power in a relative small dimension.

Transformers can fail though. Maybe @JohnYang1997 knows if these transformers have thermal fuses. These type of transformers (actually the fuse inside) fail more often than transformers not having them.

Exploding is a big word though when primary windings suddenly short it may give a nasty 'flash' and the coating in transformer wires smells terribly when something goes wrong.

For a transformer to get UL certification, it has to have a thermal fuse. This reduces the change of catastrophic failure (smoke, melt, fire) but it could still happen.

 

[EHC]

Most probably the transformer is destroyed consequence of a larger current (as normal) forced by the amplifier. Meaning, the modified L30 are as (un)reliable as the unmodified ones.

As long the OPA1612 noninverting input(s) are conected to ground ONLY through the potentiometer cursor(s), this things will keep happening.

 

[solderdude]

Most probably the transformer is destroyed consequence of a larger current (as normal) forced by the amplifier. Meaning, the modified L30 are as (un)reliable as the unmodified ones.

It doesn't look like it. The transformer is rated for 15W (1A @ 15V).
This is not drawn, it will be much lower. More likely just a cheap transformer with a short in the primary winding. This has nothing to do with the L30. In order to draw 15W in that dimension the L30 and transformer would get very very hot.

As long the OPA1612 noninverting input(s) are conected to ground ONLY through the potentiometer cursor(s), this things will keep happening.

No... why would it ? You mean you cannot use the L30 with the volume control turned all the way down otherwise it will draw large currents (without getting hot) able to blow up transformers ?

 

[EHC]

No... why would it ? You mean you cannot use the L30 with the volume control turned all the way down otherwise it will draw large currents (without getting hot) able to blow up transformers ?

OPA noninverting input must be always connected to ground (trough a resistor) for stability purposes. But potentiometers are unreliable components, and sometimes the cursor, even for microseconds, can disconect itself from the resistive path.

In this amplifier configuration OPA's noninverting input in the air could spell disaster. See the schematic here.

 

[solderdude]

Except the potmeter wiper is connected to the + input of the opamp.
Personally, I would have put a 10k resistor from + input to ground to prevent the input from floating. a resistor costs next to nothing and would agree that the + input needs to be connected to a reference and not only a wiper.
The schematic is drawn incorrectly b.t.w. It could never work this way but you knew that already (or should have known)
I don't believe this 'destroys' transformers.

 

[EHC]

Except the potmeter wiper is connected to the + input of the opamp.
Personally, I would have put a 10k resistor from + input to ground to prevent the input from floating. a resistor costs next to nothing and would agree that the + input needs to be connected to a reference and not only a wiper.
The schematic is drawn incorrectly b.t.w. It could never work this way but you knew that already (or should have known)
I don't believe this 'destroys' transformers.

Yes, the solution is an extra resistor. Regarding the transformer, a destroyed IC can draw (a lot) more current.
I really appreciate the elegance simplicity of the design, but you cannot bet on a potentiometer reliability.

 

[solderdude]

There are voltage regulators in there that limit the current drawn so there are no large spikes on the power supply.
When the wiper would not make contact and the opamp would go really fast to one of the supply rails that would make really, really loud sounds.
No one complaints about this so is not an issue.

A loose wiper (when really loose) would cause a DC on the output and the DC protection would kick in. There won't be a sudden high current destroying the transformer. It is abundantly clear this cannot destroy transformers.

When a pre-2012 L30 has failed (with smoke and getting really hot) and in this condition was left switched on for a long time period then yes the transformer could have been overheated but not fail. Such a transformer could have suffered latent damage and give up the ghost.

The reported failure clearly shows it was not a thermal fuse but looks like a short in windings somewhere.

 

[companyja]

Hey, benign question but does the L30 always click about a second after you switch it to HP out? The sound comes from the unit, not the headphones attached to it

 

[solderdude]

That's a relay.
They are supposed to click. You just sjould not hear anything from the speakers/headphone.

 

[EHC]

There are voltage regulators in there that limit the current drawn so there are no large spikes on the power supply.
When the wiper would not make contact and the opamp would go really fast to one of the supply rails that would make really, really loud sounds.
No one complaints about this so is not an issue.

A loose wiper (when really loose) would cause a DC on the output and the DC protection would kick in. There won't be a sudden high current destroying the transformer. It is abundantly clear this cannot destroy transformers.

When a pre-2012 L30 has failed (with smoke and getting really hot) and in this condition was left switched on for a long time period then yes the transformer could have been overheated but not fail. Such a transformer could have suffered latent damage and give up the ghost.

The reported failure clearly shows it was not a thermal fuse but looks like a short in windings somewhere.

If OPA's noninverting input is in the air even for a microsecond, a current exceeding 10mA (maximal admissible cf datasheet) can develop from TPA output trough low resistance resistors in the feedback loop into the inverting OPA input. Chances are, both OPA and TPA will destroy themselves.

 

[solderdude]

The schematics are incorrect so what you say can happen can't happen.
The - is always connected to ground, the amp wouldn't work otherwise.
Again... the schematic is incorrect what concerns the feedback path.

 

[Raoul Duke]

Just got a 21xxxx replacement to my 2007xxx, free of charge, and not needing to send the old one (that I will destroy). I must say I'm impressed by the way they handle this.
However can someone please confirm that the issues are resolved by the new version without needing to change the AC adapter? There was no adapter shipped, so I am still using the one that was sent with the 2007xxx model. If I understand correctly there is also a problem with the adapters?

 

[GGroch]

If I understand correctly there is also a problem with the adapters?

There is no problem with the adapters. The adapters have not changed.

 

[Raoul Duke]

There is no problem with the adapters. The adapters have not changed.

Thanks!
I must say I never had a problem with the L30, but then it very rare I give any electrical discharge (unlike my wife who probably could light our lamps some days by touching them). Anyway I found the information on that issue in this forum, so thanks for the heads up!
The new L30 seems to be a lot less hot. I used to put an old CPU heat exchanger on the old one, because it was running at around 40°C.
Listening to some Hespérion XXI classic music in DSD with my Ultrasone 2900 : pure joy!

 

[Veri]

However can someone please confirm that the issues are resolved by the new version without needing to change the AC adapter? There was no adapter shipped, so I am still using the one that was sent with the 2007xxx model. If I understand correctly there is also a problem with the adapters?

Everyone's using their new unit with the old PSU. That's the way to do it.

 

[NgtFlyer]

Hey, benign question but does the L30 always click about a second after you switch it to HP out? The sound comes from the unit, not the headphones attached to it

Yes, that is the protect relay. It's working as it should.

 

[NgtFlyer]

Everyone's using their new unit with the old PSU. That's the way to do it.

The AC transformer (the unit uses 15V AC) has not been changed. No need to replace it when its design has not changed - it wasn't necessary.

 

[companyja]

Decided to open my new 201262xxx L30 on a whim to see what the fix looks like even though I heavily suspected it's there, this is what it looks like:

That looks very clean and unlike some of the other photos, I imagine this might be how they come out of the factory so maybe my board was made like this already, which would be cool. Maybe some day when I'm brave enough I'll open the front as well to give you a photo of the revised 2011+ L30 circuit and to try to see if there's anything I can do about that horrendous volume pot is there a possibility of developing larger channel imbalance by removing and putting the pot back on?

 

[sandymc]

If OPA's noninverting input is in the air even for a microsecond, a current exceeding 10mA (maximal admissible cf datasheet) can develop from TPA output trough low resistance resistors in the feedback loop into the inverting OPA input. Chances are, both OPA and TPA will destroy themselves.

No. For an OPA1612 to suffer that kind of damage requires a voltage on the input pins 500mV+ greater than the supply voltage. Section 7.3.2 of the TI data sheet (https://www.ti.com/product/OPA1612) address this specifically. Voltages like that can't come from the output, because that's limited to the supply voltage, unless the schematic is very wrong. The point about ESD is that it can easily be greater than the supply voltage. Which is how the mango hits the fan.

Best practice would be to have two resistors - one grounding, and another in series with the input to limit fault current to less than 10mA.