Saturday, 26 February 2011
6080 Stage - CCS (sink) -50V at grid @ 50mA, plate voltage around 120V, bringing B+ to 170V.
5687 Stage - CCS (source) -3.2V at grid @ 20mA, plate voltage around 120V, with swing up to 170V at grid voltage of -5.5V.
Might use K and K Audio's Cascode CCS Kits instead of the Bottlehead Speedball, being cheaper. Going to use a tube rectified choke input filter with 3 stages of LC or RC filtering. With these changes, you really can't call it the Bottlehead Crack anymore, and it becomes another OTL Headphone amp.
B+ is 166.3 V at the 6080 tube socket pin.
AC Hum at the headphone jack; Right is 0.005 Vac, Left is 0.013 Vac
85 mA through the first PSU Filter resistor.
After doing this, I thought I give the HD650 a crack before I send them away, and lo and behold, the left channel works again!
However, in the two sessions with the DT880 and the HD650, I found that the right channel is significantly softer then the left channel. I think some is amiss here. Also measuring the voltage across the the large 2W resistor in the driver CCS, there was a big different between the left and right channels, something is not quite right in the amp, that is for sure. Maybe this also has something to do with the imbalance in Vac across the output.
On another note, it might be the HD650 connectors on the left side may be loose, and hence it worked again since I unplugged the cable ready for shipping, and then re-plugged it in after. Again, something is amiss.
Friday, 25 February 2011
Thursday, 24 February 2011
Wednesday, 23 February 2011
Tuesday, 22 February 2011
My basic design premise is to stick to the basic design, with a potential substitute for the driver tube. At the moment, the choices are: 5687 family, 12SN7/12SX7 (I have a stash), 6n6p and 12BH7 in the dual triodes. But more experimentation to try to use a single triode or a single triode strapped pentode. This could definetly open up more options, but this will have to be made soon, as heaters and bias points are a consideration.
The PSU will be a choke input in common mode, followed by a LC stage and then a seperate RC stage for each channel. But again, I might decide later on to do a LC instead for the last stage, to provide that inductive seperation between channels. The transformer should be a 240VCT to produce around 170V at the B+.
So far, I have committed to parts for the resistor side only and had the various PSU caps already on hand. Going to try the Charcroft metal foil for the shunting resistors at the output and using Kiwame 5W (in parallel) for the bleeder and RC filters. Also got some Amtrans carbon for the grid stoppers if I use a higher Gm tube. The PSU caps will be Obbligato oils, and tenatively using 70uF for the common stages and 30uF on the individual channel stages.
More details to follow.
Monday, 21 February 2011
After the failure of my HD650, I have ordered a pair of Beyerdynamic DT880 in the 600 Ohm version as a replacement. I still will be getting my HD650 fixed, but is an excuse to make an earlier purchase on the DT880. It will be fun to compare the two through the Crack amp. Hopefully they will arrive by the weekend, so that I have some time to try them out. Probably end up burning them in for 10 or so hours first before I give them a good listen.
During the research for different headphones, I decided that I will try to get a pair of Grado SR60i, they seem fun and cheap and should be a better match for the computer outputs direct and iPad. Another few headphones I'd like to try eventually are the AKG K702, the Denons and the Audio Technika. But, they all seem to be lower impedance headphones, and Id prefer to stick with the higher impedance ones so they can all be listened through the crack.
The wait begins.
Saturday, 19 February 2011
After checking through my wiring again on the Crack, I found one old solder joint, which was fixed. However, there was still no sound on the left channel, after some more checking, I decided to use a different am, and to my surprise, the left was still not working! Some more checking and swapping of the headphone connectors, confirmed my fears, the left headphone speakers is not working!
Well, problem found and solved. But the problem changes to getting them fixed. But for the mean time, I will be without any cans until they are fixed. Perhaps this will be a good time to purchase my next pair of cans... Not good.
Sunday, 13 February 2011
Just started this evening session, and to my horror, only the right channel is working. The left channel was completely silent, no hum, nothing. I immediate thought it might be the tubes, so I replaced both tubes with new ones, and same problem. I then proceeded to carefully check the wiring. I know that by substituting the larger Aeon caps, the soldering and mechanical support for those large caps was not optimal. Trying to gently tap the left output caps did get the sound to come back sporadically. Something strange I also noticed that when I turned the power off, the left channel would come back!
Well, all of this points to a bad solder joint in the left capacitors somewhere. I think I will need to re-solder and recheck that all the parts of still good. But after a modest bump to the sides of the chassis, the left channel was coaxed back into action. But not the ideal situation. Guess this is another joy of DIY, some measure of risk to tame the rewards that it brings.
Friday, 11 February 2011
Without being too critical, the sound is as good as I remember it. It does sound like the whole band is in my head, with excellent presentation of the music. Highs are crisp, bass articulate and vocals so wonderful! I must say, that Higgin's voice through the headphones are very addictive, and does put you in that place.... Especially with one of my favourite tracks, 'Sound of White' from her first album.
The only issue with the crack that is troubling is the hum. I can hear it, its not too loud, but it is there. Especially when you take one headphone off, and only have the other to your ear.
Moving onto something quite different with The Muse, and their latest album, Resistance, and the first track, 'Uprising', the bass line is so powerful, even through the headphones. Ballamy's dark and penetrating vocals to this track really hits you, along with the strong drums, the slam really do hit you and make a great contrast to the rest of the track. The guitar rift/solos have a real sense of presence and definition. Definitely nothing to complain about the setup.
Anyway, time to stop writing, and enjoy the rest of the session!
For this project, I was contemplating using a permalloy output transformer, and Tamura makes them in the 3k, 3k5 and 5k, which are ideal for this project. More details to come later.
Looking at a basic design for the amp, I started with the F2a operating points. See the curves below. The loadline is at 7k5, operating point at 375V on the plate, 70mA bias current and grid bias at -16V. The approximate power is around 3W. This requires a driver stage that can swing 32 V peak to peak. Not a bad set of parameters.
Now, I wouldn't read too much into this, as the formulas are only approximations and the curves are only approximations and the determination of the various points are only approximations. But, it does give some confirmation that this is a good starting point.
One might be thinking that it is a pretty expensive way to get 3Wrms of power with the plate disspation at 26W. Perhaps a different load line is required, this might warrant a 5k or 3.5k loadline to get more juice out of this tube. More to follow.
Thursday, 10 February 2011
Tuesday, 8 February 2011
Saturday, 5 February 2011
Using the CCS to feed the VR Tube ensures that during start up, the in-rush current is limited, and provides a stable operating point for the VR. a further refinement would be use some form of LED or CCS for the tube biasing, to ensure that the operating points remain stable. as suggested on the epc.cc site, tying the ground of the VR to the top of the cathode, may achieve is, and then the resistor below the cathode, simply sets e same current as the CCS. This will eliminate the cathode bypass resistor, and give the bias point of the tube a stable bias.
Friday, 4 February 2011
With the headphone amps, my first build was a PCB (original version) of Peter Millett's Hybrid 12AE6A amp. This was a great project. Easy to build, fun to use and good sound. It was my default amp for a few years. Using pretty much the standard parts, it has proved to be a reliable and enjoyable amp.
The second venture was with a full point to point kit, in Maple Tree Audio's Ear+. This was the version with the line stage. The build was done with the Blackgate, Tantalum and Noble pot upgrades. This was really a step up for me in terms of building and testing. Being a full point to point kit, it proved to me that good solid experience is required to ensure a smooth build. Overall, I thought the kit was well designed, laid out and provided good to very good sound. My main issues with this amp are:
1. There was hum. (Bad soldering and wiring??)
2. Left channel was always buzzy and intermittent. (Bad tube socket??)
3. Sound could be considered a little flat or laid back.
Whilst items 1 and 2 can most likely be attributed to bad construction on my side, I could not find a way forward with item 3. Hence, my quest continued in search of a better headphone amp. With this design, I have always mooted about doing a full re-build, with the courtesy of a few years of experience. This would include taking out the line stage, full rewire and layout, replacing the cathode coupling capacitors and general parts tweak and upgrade. Also would try to get rid of the 12AX7 as the input tube and replace with something more interesting. This I will leave for later, as there are more exciting projects in my mind.
After a few years, I decided that the next upgrade would be in the MiniMax Millett. I chose the PCB version and it basically replace the OPAMP buffer on the original with a discrete diamond buffer, and it replaces the current sources in the original with more robust IC versions. This time around, I also tried to build an enclosure of timber sides and metal top panel. The net result was a great sounding amp, that has been really fun to use. The build was fairly straight-forward, with the only parts tweaking being the coupling capacitors. I added a Russian K40 and Vishay MKP in parallel with the Nichicon Muse electrolytic.
During this period, life caught up, and babies were now in the picture. Headphone hifi seemed to be the most logical method, as speakers would inevitably get damaged or disturb the family. Though loud singing when listening to headphones could still be a source of nuisance. By this time, I had built a transformer coupled (cathode) amp, two hybrids, and wanted to try an OTL. At the same time, Bottlehead released their Crack. $200 and a few months later, the kit was delivered.
First, I can say that the value for money, from the Bottlehead kit is far above the Maple Tree Audio. Though the latter did have better components, it was not worth THAT much more. The kit was built with the Speedball upgrade at the same time, to avoid having to build and test twice. The end result was a great amp. The sonics seem to combine the best the amps to date. From this point, I can see that there are a few items that can be upgrade and improved.
1. Volume control, replace with nice GoldPT or DACT attenuator.
2. PUS Caps, replace with Oils (means a new enclosure).
3. Input and output jacks replace with quality Neutrik ones.
4. Potentially adding a choke to the PSU.
5. New Tube sockets.
A lot to upgrade, but I believe that they will bring it amp from a 90/100 to a 98/100. But right now, I am just enjoying the amp for all its worth.
Where to from here? Well, there are two upgrade/rebuild paths for my existing amp. But I have always wanted to build a standard SE Transformer coupled headphone amp. And to that extent, I have taken a few designs from the net, and made a few changes, and made my own. For me, I have coined it my final headphone amp, the one amp to rule them all! Perhaps not. But, the basic design premise is similar to the designs by Andrea Ciuffoli, Lynn Olson, epc.cc and a few others out there. Nothing unique.
The design is a D3a feeding into a Amorphous core Lundahl transformer. The PSU is glow tube VR150 regulated and the PSU itself is a choke loaded in common mode, followed by one shared LC stage and one channel specific LC stage. More details and schematics to follow. But this should be an interesting build.
Having looked back on all of this, I think after completing the D3a headphone, I will go back and rebuild the EAR+ and the Crack and then have on hand 5 quality tube based headphone amps. This will I think give me a good overall impression of tube headphone amps, and probably serve as a foundation for any future headphone amps. Finally, considering an upgrade to my HD650, and the two choices so far are the Beyers T1 and the HD800. But that, will have to wait for some time.
Anyway, will be posting more info and build notes on the D3a amp, and I will hope to write sensible and proper reviews on the other amps later this year.