This is a teardown follow up to my review of Schiit Lyr Hybrid Tube Headphone Amplifier. The unit is on kind loan from a forum member.
Taking apart the Schiit Lyr, like their heavier units is a bit of a challenge as it uses an odd metal box inside another metal box. After taking out the outer box, you are greeted with the internal exoskeleton:
I could take this enclosure apart but that would require unscrewing the output transistors but since this unit is on loan, I chose to not do that.
Typical of most headphone amplifiers, the architecture is rather simple. To the left, covered by the box are a couple of transformers connected to AC mains (protected by a fuse). The output is rectified by a bridge rectifier and then filtered (cleaned) by the black electrolytic capacitors that are visible. They are made by Panasonic (very good brand) and rated at 105 degree C (much prefered to cheaper 85 degree C).
The first stage of the amplifier is a dual-triode tube configuration. Unlike simple tube amplifiers however, this is a hybrid unit meaning the rest of the amplification and output to headphone jack occurs through a pair of MOSFET transistors. I did not try to reverse engineer the circuit to see if that is a class-A operation or class AB. Either way, the transistor stage provides the very low output impedance required without using of transformers, etc. Documentation says no feedback is used which would explain the rather high distortions I measured in my review.
You can get a better view of the output MOSFET transistors from this angle which are made by the respected company, International Rectifier ("IRF"):
As you can kind of see, the metal portion of the transformer is "crusty" for the lack of a better term. Not sure if this is due to rust, use of second-rate transistors, or just the way they look. I am used to seeing them without this oxidation/coloration.
The case is used for heat sinking so best to make sure there is good airflow underneath the box in addition to above (needed for the hot running tubes anyway).
BTW, the dual enclosure design means that the drilled holes in the outer case are fake. They are blocked by the internal case so do not provide any extra cooling that way.
All of these heat generating components mean there is convection/air movement which will suck in dirt. And dirt is what we had in this unit:
This is a good reason to avoid hot running electronics if you can. Solid state version of this device could be built completely sealed, avoiding this problem.
There is an IC opamp from respected company, ST Micro but I did not try to figure out what it does.
I was impressed to see a lot of 1% rated resistors in there. Not very impressed (but I guess mandatory in tube amps at this price point) is a variable resistor per channel:
That is going to drift over time and change based on temperature of the unit which partly explains my variable results depending on how long the unit was warmed up.
A rather nice volume control was used in the unit but for some reason the label was torn up:
Not so nice is what I noted in my review: the volume control knob had come loose. This is caused by a round shaft and a set screw that is attach it to the knob:
If you are going to use a knob with a set screw, the knob needs to have a flat side on the shaft. Or some way of inserting into that slot that is visible. Without it, the set screw will lose its grip constantly forcing you keep tightening it.
A bigger miss is the way the rather large and heavy transformers are secured (or not in this case):
As I have noted, the transformers have four round holes in the corners meant to fasten them to the chassis. Whoever designed the PC board made accomodation for this by putting those holes in there. But no screw or any mechanical fastener was used in either transformer. Instead, the mounting relies entirely on the soldered terminals.
To be fair, they are using rather thick PC board and I could not flex the PC board by hand. Still, it is just not good design practice to have such heavy components not have good, solid, mechanical fasteners to the chassis. They put the screw holes on them for that reason. Seems like someone down the line decided that saving 20 cents on screws and a minute to put them on was a worthwhile cost optimization.
FYI I saw the same practice in Schiit Yggdrasil DAC so this is not an exception but rather than the rule.
More concerning and a failure waiting to happen is the placement of these two electrolytic capacitors right next to the hot running tubes:
The heat will dry them up and cause them to fail over time. They have been put on their sides away from tubes or some other scheme of distancing them from the hot tubes/tube sockets right next to them. Users should keep an eye on these capacitors and change them at the first sign of bulging, something leaking out of them, or as a routine measure every so often.
Conclusions
When taking apart electronics, I always look for design/manufacturing features that either put a smile on my face or a frown. In this case, we have some of each. Precision resistors, brand name parts, beefy chassis are good to see. Alas those are far outweighed by other flaws which would have been very easy to avoid such as screwing in the transformers, distancing capacitors from heat sources, etc.
This is certainly NOT a product that I would be proud to own or be associated with as an engineer.
Taking apart the Schiit Lyr, like their heavier units is a bit of a challenge as it uses an odd metal box inside another metal box. After taking out the outer box, you are greeted with the internal exoskeleton:
I could take this enclosure apart but that would require unscrewing the output transistors but since this unit is on loan, I chose to not do that.
Typical of most headphone amplifiers, the architecture is rather simple. To the left, covered by the box are a couple of transformers connected to AC mains (protected by a fuse). The output is rectified by a bridge rectifier and then filtered (cleaned) by the black electrolytic capacitors that are visible. They are made by Panasonic (very good brand) and rated at 105 degree C (much prefered to cheaper 85 degree C).
The first stage of the amplifier is a dual-triode tube configuration. Unlike simple tube amplifiers however, this is a hybrid unit meaning the rest of the amplification and output to headphone jack occurs through a pair of MOSFET transistors. I did not try to reverse engineer the circuit to see if that is a class-A operation or class AB. Either way, the transistor stage provides the very low output impedance required without using of transformers, etc. Documentation says no feedback is used which would explain the rather high distortions I measured in my review.
You can get a better view of the output MOSFET transistors from this angle which are made by the respected company, International Rectifier ("IRF"):
As you can kind of see, the metal portion of the transformer is "crusty" for the lack of a better term. Not sure if this is due to rust, use of second-rate transistors, or just the way they look. I am used to seeing them without this oxidation/coloration.
The case is used for heat sinking so best to make sure there is good airflow underneath the box in addition to above (needed for the hot running tubes anyway).
BTW, the dual enclosure design means that the drilled holes in the outer case are fake. They are blocked by the internal case so do not provide any extra cooling that way.
All of these heat generating components mean there is convection/air movement which will suck in dirt. And dirt is what we had in this unit:
This is a good reason to avoid hot running electronics if you can. Solid state version of this device could be built completely sealed, avoiding this problem.
There is an IC opamp from respected company, ST Micro but I did not try to figure out what it does.
I was impressed to see a lot of 1% rated resistors in there. Not very impressed (but I guess mandatory in tube amps at this price point) is a variable resistor per channel:
That is going to drift over time and change based on temperature of the unit which partly explains my variable results depending on how long the unit was warmed up.
A rather nice volume control was used in the unit but for some reason the label was torn up:
Not so nice is what I noted in my review: the volume control knob had come loose. This is caused by a round shaft and a set screw that is attach it to the knob:
If you are going to use a knob with a set screw, the knob needs to have a flat side on the shaft. Or some way of inserting into that slot that is visible. Without it, the set screw will lose its grip constantly forcing you keep tightening it.
A bigger miss is the way the rather large and heavy transformers are secured (or not in this case):
As I have noted, the transformers have four round holes in the corners meant to fasten them to the chassis. Whoever designed the PC board made accomodation for this by putting those holes in there. But no screw or any mechanical fastener was used in either transformer. Instead, the mounting relies entirely on the soldered terminals.
To be fair, they are using rather thick PC board and I could not flex the PC board by hand. Still, it is just not good design practice to have such heavy components not have good, solid, mechanical fasteners to the chassis. They put the screw holes on them for that reason. Seems like someone down the line decided that saving 20 cents on screws and a minute to put them on was a worthwhile cost optimization.
FYI I saw the same practice in Schiit Yggdrasil DAC so this is not an exception but rather than the rule.
More concerning and a failure waiting to happen is the placement of these two electrolytic capacitors right next to the hot running tubes:
The heat will dry them up and cause them to fail over time. They have been put on their sides away from tubes or some other scheme of distancing them from the hot tubes/tube sockets right next to them. Users should keep an eye on these capacitors and change them at the first sign of bulging, something leaking out of them, or as a routine measure every so often.
Conclusions
When taking apart electronics, I always look for design/manufacturing features that either put a smile on my face or a frown. In this case, we have some of each. Precision resistors, brand name parts, beefy chassis are good to see. Alas those are far outweighed by other flaws which would have been very easy to avoid such as screwing in the transformers, distancing capacitors from heat sources, etc.
This is certainly NOT a product that I would be proud to own or be associated with as an engineer.