Bottlehead Crack Headphone Amplifier Kit Review

[MZKM]

Thought about running out to the store and showing off my clean teeth but couldn't figure out how to do that with a mask.

What my sign/foreign language teacher coworkers use:

 

[solderdude]

I get that tube amp by design makes it difficult if not impossible to have a low impedance output, but I have seen mentioned in the reviews and the comments that it don't have output transformers, but in real life, do we know ANY any headphone amps with output transformers?

Actually a lot of tube headphone amplifiers use output transformers. Depending on where NFB comes from these can be quite low impedance as well.
This is one of the few OTL tube amps (OTL = Output Transformer Less)

This would make it complicated since transformers don't work by bridging impedance, no notion of damping factor. They transfer power by impedance matching. For speaker, it can be simple, you need a tap at 4 ohm, and one at 8. But wondering how it could be of any use in a headphone amp, with load impedance all over the place. unless the amp is designed specifically for a particular set. or have multiple taps AND multiple headphone jacks, which would be a bit odd.

Some tube headphone amps have taps for 16-50, 50-100 and 100-600 for instance. The output resistance of those taps is usually far below that of the lowest headphone impedance so the amp acts as a voltage source but is power limited.

 

[YSC]

so it seems the matching synergy thing in amp and headphones does work, I remember this amp was said to be the to go amp for HD650

 

[milosz]

MANY "audiophile" products are not linear and are used in certain combinations to provide a "certain sound" that the owner-listener enjoys. I think this is what is behind the SET craze, and partly what motivates tube amp folk.

I have heard some setups with amplifiers that measure really badly, with speakers that are nowhere near what a well-designed speaker should be, and yet with certain recordings of certain music the system produces what sounds very much like a live performance. This seems to work with certain recordings of small combo acoustic jazz, blues, vocal, singer-songwriter and sometimes even small classical ensembles like string quartets or horn choirs. These systems are NOT accurately reproducing the SOUND of the recording yet they somehow are able to produce the MUSIC in an exemplary way. But typically such a system only works well with certain recordings and certain musical genres. And who's to say that's it's "wrong" to build a system around your favorite Louis Armstrong or Lightnin' Hopkins recordings? I "get" it but I don't listen that way so I don't have such a system.

Here with the Crack amp, there is a bit of this, there's a particular "nice flavor" with the 600 ohm 'phones, and there was the fun of building the thing.

But it's kind of goofy, you know? It's almost- but not quite - engineering.

RE: The hum in this amp, it would not be hard to put the filaments on a regulated supply.

RE: channel imbalance - since the output is capacitor coupled, and the capacitor used has a 20% tolerance, there can be as much as 40% difference in reactance through the capacitors at 1 kHz - the caps are 100 uF @ 20%, so worst case one cap could be 80 uF and still be in tolerance, and the other could be 120 uF and still be in tolerance.... capacitive reactance at 1 kHz for 80 uF is 2.0 ohms, and for 120 uF it is 1.3 ohms.... that alone could cause the imbalance. Add to this the likelihood that both sections of both of the dual-triode tubes in the thing have slightly different gains... well, there you go. Careful hand matching of parts could probably tighten this up. But then again my EARS have channel-balance problems. My right ear is often not working as well as my left, depending on the condition of my sinuses / eustachian tubes.

My source for headphone listening- my computer / flacc files - has a balance control. So that is really handy.

 

[solderdude]

so it seems the matching synergy thing in amp and headphones does work, I remember this amp was said to be the to go amp for HD650

It's not matching/synergy thing though understandable that people say this.
It is a matter of using the device within its intended purpose.
This amp thus 'synergizes' with all high impedance headphones not only HD650 nor is it designed with only the HD600/650 in mind.
It is one of the most sold high imdedance headphones though.
You just should not use it if you have low impedance headphones or don't want the sound to be modified.
You shouldn't use a 600 Ohm headphone directly from a phone either.

 

[milosz]

Here you go:
View attachment 80850

Amirm, that is the basic Crack amp schematic, and does not show the optional Speedball circuits. These are constant-current cascode plate loads for the 12AU7 and constant-current cascode cathode loading for the 6080. From the looks of the innards of the amp tested here, it does have these improvements, which make the circuits more linear. Without this option, it would be even worse.

 

[Tks]

1dB boost is that serious? Call me skeptical.

 

[KxDx]

It needs silver wiring inside... and rainbow foil stickers, and maybe some chakra stones or quantum dots for sure.
And a DACT, audiophile grade capacitors, a tube rectifier, special audiophile resistors and .....

Fuses. Quantum fuses.

 

[Francis Vaughan]

This would make it complicated since transformers don't work by bridging impedance, no notion of damping factor. They transfer power by impedance matching. For speaker, it can be simple, you need a tap at 4 ohm, and one at 8. But wondering how it could be of any use in a headphone amp, with load impedance all over the place. unless the amp is designed specifically for a particular set. or have multiple taps AND multiple headphone jacks, which would be a bit odd.

This is a common notion, but is actually wrong.
There are so many really common myths to address here.
(And I'm really not picking on you, these are so insanely common that they have become almost a received truth.)

1. Transformers absolutely work to "bridge" impedance. The turns ratio of a transformer defines the ratio of impedances. (Actually the square of the turns ratio.) For a tube amp you have hi-z on the tube side, and low-z on the load side. This is the reason you have a transformer.
2. Tube amplifiers absolutely can have feedback, and in particular, global negative feedback. This is very common. There are some interesting devils in the details, but adding negative feedback is easy, and will act to lower the output impedance.
3. Power is not transferred by "impedance matching". Indeed, it isn't clear what sort of mechanism this would be. However what I suspect this idea tries to address is the idea that power transfer is maximised when the source and load impedances match. This is probably the root cause of the final myth.
4. The output impedance of a tube amp via its transformer is not matched to the speaker impedance. Really. This surprises a lot of people. They get told that in order to maximise power transfer the tap on the output transformer is chosen so that the output impedance matches the load impedance. This isn't the case.* The taps on the output transformer are chosen so that the reflected impedance of the load as seen by the output tube provides the correct load on the plate, allowing the designer to select the desired load line for that tube, typically one that maximises the linearity of the tube. The marking on the back of an amp means - connect your X Ohm speaker to the X Ohm terminal so that the tube sees the desired load and its transfer function is the one the designer wanted.

* Yes this is a condition for maximum power transfer, but it isn't the design goal, and the impedances of a tube amp do not meet this condition unless by accident. Loudspeakers are designed to be fed by voltage sources.

 

[Severian]

I have one of these (non-Speedball) and I love it. It was very enjoyable to build, it looks beautiful, and it sounds great with the HD650, at least for certain types of music (e.g., rock) where the noise floor isn't as noticeable. I don't think it works well for, say, classical music. The biggest issue is that I get a lot of crackling, which I suspect is one of the tubes, and it's extremely susceptible to PC ground loop noise. I typically use it with a Bluetooth DAC instead.

 

[solderdude]

RE: channel imbalance - since the output is capacitor coupled, and the capacitor used has a 20% tolerance, there can be as much as 40% difference in reactance through the capacitors at 1 kHz - the caps are 100 uF @ 20%, so worst case one cap could be 80 uF and still be in tolerance, and the other could be 120 uF and still be in tolerance.... capacitive reactance at 1 kHz for 80 uF is 2.0 ohms, and for 120 uF it is 1.3 ohms.... that alone could cause the imbalance.

Nope it couldn't. The difference between the channels would be 0.013dB with a 300 Ohm load.
The difference in series resistance between 188 and 188.7 is negligible (0.35%)

 

[dfuller]

I should mention, I bet that the channel imbalance is due to poorly matched sections within the tube. Unfortunately, to get good matching between sections you need a curve tracer and a bunch of tubes, which isn't cheap. Some resellers do offer that as an option though.

 

[PeteL]

This is a common notion, but is actually wrong.
There are so many really common myths to address here.
(And I'm really not picking on you, these are so insanely common that they have become almost a received truth.)

1. Transformers absolutely work to "bridge" impedance. The turns ratio of a transformer defines the ratio of impedances. (Actually the square of the turns ratio.) For a tube amp you have hi-z on the tube side, and low-z on the load side. This is the reason you have a transformer.
2. Tube amplifiers absolutely can have feedback, and in particular, global negative feedback. This is very common. There are some interesting devils in the details, but adding negative feedback is easy, and will act to lower the output impedance.
3. Power is not transferred by "impedance matching". Indeed, it isn't clear what sort of mechanism this would be. However what I suspect this idea tries to address is the idea that power transfer is maximised when the source and load impedances match. This is probably the root cause of the final myth.
4. The output impedance of a tube amp via its transformer is not matched to the speaker impedance. Really. This surprises a lot of people. They get told that in order to maximise power transfer the tap on the output transformer is chosen so that the output impedance matches the load impedance. This isn't the case.* The taps on the output transformer are chosen so that the reflected impedance of the load as seen by the output tube provides the correct load on the plate, allowing the designer to select the desired load line for that tube, typically one that maximises the linearity of the tube. The marking on the back of an amp means - connect your X Ohm speaker to the X Ohm terminal so that the tube sees the desired load and its transfer function is the one the designer wanted.

* Yes this is a condition for maximum power transfer, but it isn't the design goal, and the impedances of a tube amp do not meet this condition unless by accident. Loudspeakers are designed to be fed by voltage sources.

Thanks for your answer. Yes my vocabulary and full understanding of the specifics is erroneous. but I thought that the statement: ''The issue here is that without an output transformer, we have a high output impedance of 188 ohm (compared to close to zero for most solid state designs) so a ton of power is lost in the device itself" seams amiguous and suggest that having the lowest possible, with the help of a transformer would mean less power loss. Yes, maximal power transfer is not the ultimal goal, but achieving a high damping factor is not the goal neither. I agree (and understand better) that the goal is to optimise the linearity of the tube.

 

[anmpr1]

A) I like the idea of building one's own gear.

B) I don't like the naming conventions. I guess the drug references are supposed to be clever, hip and trendy?

C) I recently read an old Letter to Ed (Stereophile circa 1990) from Stanley Lipshitz about the 'tube v transistor' thing, which seems pertinent in this context. His point was why spend a lot of money for what is essentially a tube based tone control disguised as an preamp/amp when it's a lot cheaper to buy a SS amp with a tone control?

D) This little device needs a tone control to get rid of the bass hump, and a balance control.

 

[solderdude]

Thanks for your answer. Yes my vocabulary and full understanding of the specifics is erroneous. but I thought that the statement: ''The issue here is that without an output transformer, we have a high output impedance of 188 ohm (compared to close to zero for most solid state designs) so a ton of power is lost in the device itself" seams amiguous and suggest that having the lowest possible, with the help of a transformer would mean less power loss.

The 'ton of power' also is a bit over the top. That is if you find 0.14W a ton of power when 0.2W is drawn.
Considering the HD650 proabably will be used with an average of 1mW (playing comfortably loud) the amp itself is 'loosing' 0.7mW.
Transformer losses + output resistance of the driver stage will be less than the 0.14W is probably what Amir meant.

 

[Severian]

A) I like the idea of building one's own gear.

B) I don't like the naming conventions. I guess the drug references are supposed to be clever, hip and trendy?

C) I recently read an old Letter to Ed (Stereophile circa 1990) from Stanley Lipshitz about the 'tube v transistor' thing, which seems pertinent in this context. His point was why spend a lot of money for what is essentially a tube based tone control disguised as an preamp/amp when it's a lot cheaper to buy a SS amp with a tone control?

D) This little device needs a tone control to get rid of the bass hump, and a balance control.

I haven't noticed any channel imbalance in mine except for at the very, very bottom of the volume range - lower that I would ever listen - which I understand to be mainly due to the volume pot. Not to say it isn't there, but it's not audible to me.

The slight bass boost is part of the appeal and is quite benign with the HD650 given that I would give it a slight boost with EQ when listening via a solid state amp like the JDS Labs Atom anyway. Beyerdynamic headphones don't appeal to me, so I have no plans to ever use my Crack with another pair of headphones.

Personally I would never pay this price for a commercial OTL amp given the inherent limitations but as a DIY project and a secondary special purpose amp, it's just plain fun. I'd never try to convince someone to buy the kit on the basis of high fidelity sound.

 

[solderdude]

I haven't noticed any channel imbalance in mine except for at the very, very bottom of the volume range - lower that I would ever listen - which I understand to be mainly due to the volume pot. Not to say it isn't there, but it's not audible to me.

You happen to have a tube with about equal gain in both tube halves and probably equal within 0.5dB.
Imbalance caused by a potmeter usually is only in the bottom part of the volpot range.

 

[MadMan]

Whenever someone says that 2nd order harmonic distortion isn't a fidelity problem, I sense the anguished screams of Alex Voishvillo in the distance...
View attachment 80851
(the excerpt here comes from a truly lovely, albeit aging, pair of AES papers on nonlinearity and multitone testing)

Edit: In that spirit, @amirm do you reckon we could see a loaded multitone?

Yes, it would be interesting to see a multitone of this amplifier. Multitone is usually quite boring for the high-fidelity stuff, but for an objectively low-fidelity unit like this, seeing the IMD from 'pleasant harmonics' would be nice.

 

[KaiserSoze]

For over a decade (or two) the Bottlehead Crack was the de-facto partner of the HD600/HD650.

Interesting statement, but I can't help but wonder what it means.

 

[KaiserSoze]

Whenever someone says that 2nd order harmonic distortion isn't a fidelity problem, I sense the anguished screams of Alex Voishvillo in the distance...
View attachment 80851
(the excerpt here comes from a truly lovely, albeit aging, pair of AES papers on nonlinearity and multitone testing)

Edit: In that spirit, @amirm do you reckon we could see a loaded multitone?

To take a low-distortion headphone like the HD600/HD650 and pass its input signal through a cheap little distortion-inducing tube amp like this exemplifies the extreme lengths to which some audiophiles are willing to go to prove that they have gone completely off the deep end.