Bottlehead Forum
Bottlehead Kits => Mainline => Topic started by: carlsor on September 27, 2018, 10:02:00 AM
-
I spent the last week assembling my new mainline headphone amp. Most thorough instructions I have ever seen.
Now on page 73 and getting the right voltages on the A side, but all high voltages on the C side.
I swapped 6C45 tubes - no change. I tried a different 12AU7 tube - no change.
No smoke or blown fuse.
The first time I powered up - the LEDs on B side of the C-C4S board went on for 2-3 seconds and then went dark from then on. The LEDs for the A side of the C-C4S board never went on. All 4 LEDs light up on the A-C4S board. All the LEDs still pass the diode test on my meter.
Readings: +275V = 286V, +6.3 = +6.28V
Breg: A=220, C=285
Kreg: A=10.95, C=1.24
OB: A=149, C=284
LM431A base: A=2.478V, C=2.123V
It looks to me like a defective LM431, but some other problem may have caused the Vreg failure. I have spare LM431A Vregs so I can replace if needed. I double checked all the point-to-point wiring to the 3 boards and all was fine.
Suggestions please?
Ross
-
When you say you double checked the wiring, was this a visual verification of the connections, or an actual continuity check? I only ask this because I was fooled by a wire that looked good but was fractured below the insulation on a somewhat similar repair I did last week. The connecting wires to the board in question had been prepped with a stripper that nicked the copper and created a stress riser that eventually broke just below the insulation. I replaced them all after the first one broke when I gave it a test wiggle.
-
Will do a continuity check on all wires to boards.
I was hoping for some guidance on what part of the circuit or components to focus on.
If all checks good - should I replace the LM431?
Thanks....Ross
-
Maybe reflow the ends of the black jumper that runs diagonally under the board from bB to bA? It kinda sounds more like a floating ground connection or some other missed connection is keeping the tube from conducting than a bad chip.
-
There are a few common things that I've run into when facing problems like this:
1. Are the LM431 chips where they are supposed to be? Are they facing the right direction?
2. Are the solder joints on the MJE5731A transistors super thorough? (Especially on the middle leg?) One flaky solder joint on the A side MJE5731A on the offending channel would produce all of these problems.
3. If you pull the 12AU7, the LM431 is magically removed from the circuit (on both channels). Both Breg voltages that should be ~220 will rise, but my guess is that the OB on the offending side will remain about where it is.
-PB
-
I stepped away from this project for a couple days to clear my head and take a fresh look at the problem.
- The LM431 Vregs are oriented correctly with the outlines on the PCB.
- I check and resoldered all the the component and wire connections including the MJE5731A transistors.
I removed and reversed the two C4S boards between the two sides because only 4 wires were involved with each side.
I fully expected the problem to move to the other side, but it stayed on the same side!!?? so the C4S boards are both OK.
Time to focus on the Bias board. Tomorrow I will change out the LM431 on the bad side unless I get other directions.
Thanks for all the suggestions.
-
You can also post a photo of the top and bottom of the center PC board. It is also possible to get problems like this when the Kreg connections or regulated B+ connections are accidentally reversed.
-
Problem finally solved.
-I had neglected to connect the filament to B4 also to B5. Fixing this made the 12AU7 provide 220v to both 6C45P tubes.
-There was still too little current going to the right 6C45 which was traced to a faulty 2N2222A transistor which measured 1.434V at the base, whereas the working channel base measured 0.674V. I replaced it with a 2SC2458L-GR from my parts stash which had the closest HFE and base voltage characteristics to the 2N2222A. Now the voltage to the coupling cap is only 10V above the working channel. I was able to adjust both channels to 145V. I need to buy a metal can 2N2222A which has a much higher heat dissipation and is the official part.
Thanks for all the help.
-
Did first listening test today using balanced cable to my Focal Clear headphones (55 ohms 105db effic).
Volume set at -27db with Low Impedance setting which is plenty loud. Sounds great with all types of Jazz and classical music.
Sound is dynamic, musical, full range, excellent instrument separation, plenty of bass.
Should get even better after break-in and parts upgrades. Everything stock now.
I noticed that the 12AU7 heaters got very bright for a couple of seconds during powerup and then backed off. I didn't see anything like this with the 6C45 tubes.
-
That heater flash on a 12AU7 generally indicates that its a Mullard tube.
-
Carlsor since this is resolved. Are you also Carlsor on diyaudio? If so. How does Mainline pair with the dddac?
-
Yes, I am carlsor on DiyAudio.
Right now I am listening to my fully tweaked buffered single board DDDAC as the source to my new Mainline.
My DDDAC sounds as analog as my turntable setup. I could listen to this combo all day - no fatigue whatsoever.
Best headphone listening experience I have ever had by far!
-
This super analog type of sound is what got me interested in the DDDAC project.
Since your on Bottlehead, any experience with a tube output stage (Broskie)? That's what i had in mind for a single board DDDAC.
How does the Mainline 25K variable input impedance handle output impedance of the DDDAC (don't know what sort of output you're using)?
Edit: @Mods, let us know if it's not to be discusses on this forum.
-
How is running the DDDAC through a Mainline not putting a tube output stage on the DAC?
-
Pardon me, still learning on the subject and don't have the technical knowledge to judge what would work best with Mainline. I thought the output from the DDDAC would be to little to use with Mainline.
Edit: And i suspect some DC passage prevention is needed?
As far as i can tell you can connect the DDDAC to another component in a few ways.
1. Direct coupling the DAC balanced output (i read this would give weak bass, little output from DDDAC if i'm correct?, maybe more parallel DAC boards would solve this problem? and Mainline works with SE input?).
2. Capacitor coupling (this is the standard way DDDAC is sold as a diy kit).
3. Output transformer (supposedly the best way to get the output level needed).
4. Ross' JFET buffer stage and then output transformers.
5. Tube buffer. This is what is suggested for making the balanced DDDAC play with a SE amplifier like Mainline. (Broskie Cathode Follower was the one i was eyeing).
I'm sure Carlsor can answer this better. I learned a lot from lurking over his posts on the diyaudio forum.
On the DDDAC website there's lots of information on what de DAC board puts out.
-
- The differential outputs of my single board DDDAC are terminated to 133 ohm resistors which provide signals about 2.7V above ground. My buffer uses BF862 JFETS which present virtually no load to the PCM1794 DAC outputs but more forcefully drive the differential signals through Cinemag 15/15B transformers. The single ended Cinemag outputs are terminated with a 10K LDR resistor and 150pf WIMA which gave the best SQ. Top quality power supplies for the digital and analog sides of the NOS DDDAC have a lot to do with the analog sound. The 2SK208O JFET on pin 20 is a must in place of the 6K resistor.
- Presently, my DDDAC is connected to a "Truth" buffered before and after LDR volume controller which also has outputs to my speaker amps and is my source selector. The Truth is so transparent and uncolored - it is easier to just leave it in the system. Connecting the Cinemag outputs directly to the mainline would require an output resistor that would end up with a combined load of about 10K at the normal listening volume of the mainline.
- A buffer tube unbalancer should sound great with a single board DDDAC, but I have no experience with this.
- Using an output transformer without a buffer would require a stack of at least 4 DDDAC boards. Some persons have gone 12 high!!??
- No one reports being totally happy with capacitor coupled outputs.
- Direct coupling won't work with the Mainline because of the 2.7V offset.
-
Update,
I received some metal can 2N2222A transistors and replaced the defective one. All voltages now within reason and output to coupling cap adjusted to 145V. I tested the Mainline with all stock parts which sounded better than expected.
I replaced each 10uf Dayton coupling cap with two 3.9up Mundorf silver/oil bypassed with a 0.1uf TFTF V-Cap for a total of 7.9uf on each side. I doubt that I can hear the difference between a low frequency cutoff of -3db at 5Hz versus 4Hz. I also read that trying caps larger than 10Hz made no audible difference. Besides, these were caps I had lying around from earlier projects and they filled all the allowable space. No shortage of bass!
I also replaced the 270 ohm resistor in the PS with a 270 ohm 5H choke because it's part of my religion to do that with power supplies for tube amps. Didn't do a before and after test, but for $15 it can't do any harm.
After 30-40 hours of break-in with music - wow! Very musical, detailed, dynamic, tight and flowing. Best instrument separation and soundstage placement I have ever heard from headphones. No temptation to screw anymore with internal parts. There are better headphone cables, power cords, and interconnects to try.
Ross
-
Just an FYI on the choke thing: The hybrid shunt regulator functions like a 30,000H choke isolating the power supply from the signal carrying components. In a circuit without this radical decoupling, using a choke will lower power supply noise quite a bit more than a resistor with the same DC resistance, but these benefits are diminished when using the regulator.
-
Ross, about the capacitors you put in. Maybe just using one mundorf yields better results.
I haven't ever heard a coupling capacitor bypass that I thought did anything positive.
-
After the break-in period is complete I will remove the V-Cap Teflon bypass caps and listen for any change in the sound.
I don't see the problem with putting two identical coupling caps in parallel in order to reduce the bass cutoff frequency. There should be no odd interaction between them. I would like to know if there is any theory or experience that indicated a problem with this.
-
I notice that the high voltage goes to the three tubes at the same time the filaments start heating up.
My other tube projects used tube rectifiers or a time delay-on relay to allow the heaters to warm up before the tubes were slammed with high voltage. Supposedly this delay extends tube life.
Any downside to adding a delay to my Mainline?
-
Putting two identical caps in parallel isn't really bypassing. Bypassing refers to putting a very small capacitor across a much larger one.
A time delay-on relay could potentially be added to the Mainline, but you'd probably also need to add a power transformer to power it. I haven't yet been able to exhaust a pair of 6C45 tubes, so it's quite difficult to make any real statement about their life expectancy and how that might change with various modifications.
You can populate the A side of each high current C4S board with a cap and a resistor to more slowly apply B+. We haven't tested this, and I'm not 100% sure the 10,000uF caps will fit all that well if larger parafeed caps are installed, but it's an inexpensive experiment that doesn't involve drilling holes in the chassis or figuring out if your time delay relay is going to work with a voltage doubler.
-
Thanks Paul.
I recommend 12-rocketman in ebay as a source of 6C45P tubes. Price is $10.90ea with reasonable shipping from Russia.
Tubes sounded great too.
carlsor