Bottlehead Forum
Bottlehead Kits => Crack-a-two-a => Topic started by: Sp33ls on May 05, 2018, 10:09:00 AM
-
So, I've not had the best of luck recently with my kits :-[
When I went to paint the bell end of my Crackatwoa, I noticed shortly afterwards that a hum in the right channel would appear. After checking ground connections and tightening the nuts, it'd be intermittent.
After removing the paint around the top of the bell end holes, I checked with a DMM on the bottom side of the bell end, and it says I have a solid connection to ground for the bell end.
I passed all of the resistance checks.
But, more recently I noticed that I can reproduce the hum by pushing on the transformer, and now the right channel signal is either nonexistent or when it is present, the low-end seems to be quite muddy.
Also, the type of tube seemed to have an impact. My 6SN7's didn't have a right-channel, but my 12AU7's did, but some were muddier than others.
When doing the voltage check, now one side of the high-current boards doesn't have the LEDs light up, and one half of the low-current boar doesn't either. I also noticed that one of the LED's under the driver also doesn't light up now.
I've checked solder joins and loose connections, and have re-soldered suspicious joints, but nothing.
Any advice? I'm pretty bummed because I just picked up a nice pair of ZMF Atticus, and they sounded phenomenal for a little bit, but then I began to notice these issues.
-
Posting the voltages that don't align with what's in the manual would be really helpful.
An LED on the 9 pin socket that isn't illuminating and seeing something around 150V one one of the "O" terminals on the small PC board are common symptoms of one of a couple of issues.
1. Do both halves of your 12AU7 light up? If not, the solder joint between pins 4 and 5 on the 12AU7 socket isn't working.
2. If both halves of your 12AU7 light up, what is the actual plate voltage on the offending side? (OA on the C4S board above the 9 pin socket)
I would very, very strongly recommend just sticking to 12AU7s until the amp is working properly. Using an adapter and a bunch of different tubes and attempting to draw any conclusions is just bringing in extra variables that aren't at all helpful.
-
For sure, the only reason I threw a 6SN7 in there was in order to see if there was something funky with my 12AU7's. I was just surprised to see a difference in behavior.
It appears both side of the 12AU7 light up. One thing I noticed was that actually at first, both of the LEDs over the driver lit up, but then that one in the picture went dark again after a little bit (didn't notice it fade, just looked back after measuring it was dark).
I've swapped the 6005's but that didn't change anything. Unfortunately, I only have this one 6080 power tube on hand at the moment... you don't think it could be the power tube giving out, do you?
Here are the voltages:
Terminal Voltage (DC)
Low Current C4S
IA 150V (30)
OA 60-90V (27)
KregA 3-6V (.05)
bRegA 150V (30)
IB 150V (149.5)
OB 60-90V (70.4)
KregB 3-6V (5.8 )
bRegB 150V (149.5)
High Current C4S (offending channel)
IA 190V (212)
OA 150V (31.5)
bA 0V (0)
IB 0V (0)
OB 90-110V (26.5)
bB 150V (31.9)
High Current C4S (working channel)
IA 190V (202.3)
OA 150V (149.5)
bA 0V (0)
IB 0V (0)
OB 90-110V (93.8 )
bB 150V (149.5)
Headphone Jack
Tip 0V (0)
Ring 0V (0)
Thanks for the quick reply. I'd love to get this back to 100% this weekend if possible ;D
-
Yeah, so now that I see the voltages, your problem is unlikely to be related to the 12AU7.
Pull the 6AQ5s and let me know how KregA and OA on the high current C4S board change.
-
Pulled the 6AQ5s & this time both LEDs on the driver tube remained lit. Also, both of the high current boards had two LEDs lit this time (whereas previously one had all 4 and the other had none).
One thing that was interesting was that OA & IA seemed to gradually rise from about 50 to 70VDC in the time period of me measuring the system.
Terminal Voltage (DC)
Low Current C4S
IA 150V (66)
OA 60-90V (57.5)
KregA 3-6V (0)
bRegA 150V (66)
IB 150V (215)
OB 60-90V (70)
KregB 3-6V (1.4 )
bRegB 150V (215)
High Current C4S (offending channel)
IA 190V (220)
OA 150V (69)
bA 0V (0)
IB 0V (0)
OB 90-110V (60)
bB 150V (70)
High Current C4S (working channel)
IA 190V (217)
OA 150V (215)
bA 0V (0)
IB 0V (0)
OB 90-110V (119 )
bB 150V (215)
I placed an order for "Code Matched Pair NOS NIB 6005/6AQ5W GE 5 STAR Ruggedized, Precision 640 tested" on eBay just in case one of the tubes is bad. I did notice that there was a blue glow to one of them that I hadn't noticed before. Fingers crossed the NOS pair is decent.
-
It's very, very unlikely that a tube is causing this problem.
Now pull the 6080 and remeasure IA on the low current C4S on the offending board (or OA on the high current C4S). Also check Kreg again if you can.
-PB
-
With only 12AU7 plugged in, both of the driver LEDs light up, and all 4 of the low current LEDs have a soft glow. None of the high current LEDs light up.
IA (offending): 226VDC
KregA: 0.004
KregB: 1.45
With 12AU7 and 6005's in, the LEDs are the same as above, but now with two LEDs lighting up on the working channel's high current boards.
IA (offending): 74VDC
KregA: 0.206
KregB: 1.96
-
With only 12AU7 plugged in, both of the driver LEDs light up, and all 4 of the low current LEDs have a soft glow. None of the high current LEDs light up.
IA (offending): 226VDC
KregA: 0.004
KregB: 1.45
This is somewhat informative. With no 6AQ5s installed, we will have about 220V across the 147K and 2.49K resistors that help the regulator operate. There is some voltage on the B side Kreg, but nothing on the A side. This could happen if the TL431 is installed backwards, shorted out, or miswired. It's also possible that a TL431 an PN2907 got mixed up, but that usually creates a different looking set of problems.
To continue debugging this, we will do some tests on the amp with no 6AQ5s and no 6080.
What voltages do you get on the "X" side of the 2.2uF capacitor on each side of the center C4S board?
I would also be sure to triple check that the wire which leaves KregA and heads over to the 6AQ5 is properly connected at the 6AQ5.
-
Measuring the voltage across the 2.2uf cap, right? So, just placing the leads on both sides of each:
A side: 226VDC
B side: 224VDC
The amp seems to have been working fine for the past year (minus that one cold solder I had on the step attenuator), so wouldn't I have noticed a bad configuration on the TL431 before now? I wouldn't be surprised if I bumped it on accident while handling and caused another issue, though.
-
No, one probe to ground and the other probe on the "X" terminal.
-
Oops, sorry.
A side: 225VDC (closest to the corner) & 0.008VDC (closest to center)
B side: 225VDC (closest to the corner) & 2.25VDC (closest to center)
So, that's interesting (to me anyways lol).
I tried re-soldering the capacitor just to see if it was a bad connection, but there wasn't any difference.
Any thoughts on where to check given those readings?
-
The 0V on the A side is troubling. Disconnect the white wire leaving Kreg on that side, then remeasure this voltage.
-
I disconnected the white wire at KregA, and measured the X closest to the center of the board. No difference in the the voltages of 'X' on the A side.
-
Remove the TL431, then recheck.
-
Removed the 431 on the A side and rechecked the voltages at X on the A side. This time there's a difference.
225VDC and 3.75VDC.
Coincidentally, when I test the 431 with my Fluke's diode mode, I see 0.008VDC like I was seeing before at X.
-
I'd go ahead and replace it, but consider ordering some extras. These don't fail on their own, so I'd be very, very carefully inspecting the little resistors on that 6AQ5 socket to be sure they are present and well soldered, as a missing one or one with a loose joint would allow for the voltage provided to the TL431 to go too high, and then the 431 would fail.
-PB
-
https://www.mouser.com/ProductDetail/926-LM431BCZ-NOPB
^ This one works just fine, right?
Yeah, that's interesting... I wonder what did it. I'll buy several just in case I have a troubleshooting session ahead of me :o
Hopefully it'll be as simple as replacing it and being back in audio nirvana.
Very much appreciate all the help, PB. I'm traveling this week, but I'll report back next weekend on my findings with the new LM431s.
-
I believe we buy the ACZ version of the 431.
-
Yeah, I looked at the datasheet and didn't see a noticeable difference. Do you know what differences there might be?
-
I think there might be some slight noise differences.
-
Okay, I could see that. The only real difference I saw between the two was that the "initial accuracy" was 2% for ACZ and 1% for BCZ.
I sent Mouser an email to see if they could swap it to ACZ just to be certain there won't be any issues. I also picked up a few extra transistors and diodes just to have as backup.
Does the LM431 protect the capacitors from hitting their voltage limits?
Thanks again!
-
No, the LM431 regulates the high voltage to 150V.
-
So, I swapped out the LM431 and plugged everything (including all the tubes) back in to see.
Still having the issue. But, I've noticed something interesting.
When I first turn on the amp (cold start), the voltages actually appear to be closer to normal (150V & like 2.5V), but then after several seconds, they both just take a sudden dip. What could cause behavior like that... a capacitor?
-
Pull the 6AQ5s and look at your Kreg voltages again, as well as the "X" voltage.
-
With or without the white lead from the 6AQ5 plugged in to Kreg?
-
The white lead can be connected.
-
KregA: 0.4
KregB: 1.38
60 VDC & 1 VDC on A side.
214 & 2.18VDC on B side.
-
How about OA on the high current board on the bad side?
-
It was around 50VDC... oscillating a bit upwards.
-
With the 6AQ5 pulled out? (And the 6080 should be pulled out, but 12AU7 in) This is a bit of a tough one to debug since you said the amp was previously working. Low voltage on OA on the C4S board tends to point to a miswire or short downstream of that board.
-
Oh, sorry, that was with the 6080 still plugged in.
Only 12AU7:
X: 226 & 1.9
KregA: 1.1
KregB:1.44
IA: 227
OA: 226
-
Now try the 12AU7 and 6AQ5s with no 6080.
-
12AU7 and 6AQ5s with no 6080.
X: 150& 2.5 (163 & 2.5 on B side)
KregA: 7.8
KregB:1.95
High curent:
IA: 215
OA: 150
-
Would you be able to swap the 6AQ5s from side to side to see if the low KregB voltage follows a tube?
Seeing 7.8V at KregA makes me very, very suspicious that you have the R1 resistors swapped on the high current C4S board.
-
Swapping the 6AQ5s from side to side does not change much.
One interesting thing tho, is that the reading on KregA starts at around 9.2 and then gradually decreases in voltage at a rate of about .01VDC every few seconds. That gradual shift in voltage seems to be a common behavior I'm experiencing in a few spots. I'm just not sure what would cause a shift in voltage over time like that... ???
Since the left channel appeared to be working fine, would that rule out the power supply as being the culprit? Just wondering if part of the power supply is separately responsible for each channel.
-
A loose resistor on the 6AQ5 socket could do that.
-
Hmm.. none of them seem to be loose.
So, with the 6080 inserted, the two LEDs on the B side of offending high current board light up (next to the capacitor). The other two on the A side don't light up at all.
Without the 6080, the B side doesn't light up (this might be normal, though?).
-
The LEDs lighting up or not lighting up aren't really informative in and of themselves, but your voltages are.
The difference in Kreg voltages when no 6080 is installed does suggest an issue with the build. This could be R1 resistors that are switched accidentally, or a miswire. I would be paying very, very close attention to the 6AQ5 sockets.
-
The R1 resistors look to be fine. I'm guessing I probably would've noticed issues earlier, tho, since my kit was working fine for about a year, right?
This issue really only appeared after removing the bell end for the transformer to paint it. Not right after, tho. There was a small period of time where everything was fine. I did touch up some joints during this time tho... Mostly around the power supply.
I'm curious if you have any spots I can measure with the power supply and high current board in order to help trace down where the issue might like. I don't see any issues with it when scanning for bad joints, and I don't recall unwiring anything.
-
I've suggested the spots several times in this thread. When you took the amp apart to paint the transformer, you may have jostled a connection that wasn't solid, and hence these issues cropped up.
-
Okay, I greatly appreciate all of your help. I'm going to remove the high current board and maybe swap out some of the components on there just in case. And, it'll allow me to more easily take a closer look at the 6AQ5.
-
In my experience, blindly swapping out components on a board will just create new problems that may make debugging even more difficult.
You can measure the transistors with your meter set to beep for continuity. If it beeps between two legs, then you have a shorted transistor (rare). A shorted transistor will usually lead to much higher than anticipated voltages though.
-
Makes sense I'll go that route first, then.
I'll report back tomorrow. Thanks, PB.
-
The old 431 seems to be shorted. But, the new one I replaced it with still seems to be fine. I've checked the other transistors and none of them have the continuity going off.
When I touch the continuity meter to terminals 1-6 and 9&15 on the power transformer, I'm getting a beep with about 5 Ohms on the ones in 1-6 and 0 Ohms with measuring terminals 9&15.
I'm assuming that this is an issue? Or is that normal...?
-
That's normal.
-
So, after replacing all of the transistors on the offending channel, I pulled the LEDs on the A side of the high current board and noticed that they no longer light up when I use my multimeter in diode mode. I pulled the ones of the B side, and they still light up.
It appears I blew the LEDs on the A side of the high current board.
And it appears that in the circuit, they are required to work properly in order for the transistors to function properly, correct?
-
Yes, but if they blow, they generally explode into little pieces and are missing off the board. I haven't ever had one fail that was previously working without another related major failure.
-
Interesting... I wonder why those two were behaving differently than the other LEDs -- I pulled them from the circuit to test.
Well, either way I feel I might as well replace them now. Once I get some new ones in, I'll post back.
-
So, after replacing the LEDs that weren't lighting up, 3 out the 4 now light up on the offending board, previously none were lighting up.
The voltages seemed to drop closer to normal, but the 'X' closest to the center on the offending channel still drops after just a little bit to next to nothing. I noticed that it also shows continuity to ground, whereas the 'X' on the B side does not. Is this normal?
-
Nope, not normal.
-
It appears it was shorted due to the LM431 shorting out again. After replacing that, I noticed that the 'X' on both sides closest to the center are about 2.5VDC.
Terminal Voltage (DC)
Low Current C4S
IA 150V (182)
OA 60-90V (72)
KregA 3-6V (1.93)
bRegA 150V (182)
IB 150V (149)
OB 60-90V (65)
KregB 3-6V (5 )
bRegB 150V (150)
High Current C4S (offending channel)
IA 190V (185)
OA 150V (182)
bA 0V (0)
IB 0V (0)
OB 90-110V (105)
bB 150V (182)
High Current C4S (working channel)
IA 190V (196)
OA 150V (150)
bA 0V (0)
IB 0V (0)
OB 90-110V (91)
bB 150V (150)
I noticed that the heatsink of the MJE5731AG wasn't really warming up at all.
-
Have you tried resoldering the joints on your high current C4S boards? I still believe you have a bad solder joint on the 6AQ5 socket or the high current C4S board associated with that side.
-
Oh, man. It looks like I finally might have got it! :o
I ended up replacing the 2907A and an LED (that wasn't lighting up) on the high current board, as well as resoldering the red connections to the 6AQ5 socket.
After I use it for a bit, I'll report back. But, so far it sounds amazing.
Thanks, PB! Really appreciate all of your patience with me here. I definitely became a lot more intimate with how my kit works through this experience.
I'm curious... with new components, do you notice a break-in period for things like new caps & transistors? I know it's getting into more of the voodoo, but curious if in your experience you've noticed differences between recently built kits vs broke-in kits?
-
I'm glad things worked out. It can be a bit of a challenge to find those pesky loose connections.
-
capacitors have a fairly clear break-in effect. It varies widely depending on the cap's place in the circuit and the particular cap (brand, dielectric, and rating). Audio transformers have, for me, the second most significant predictable break-in effect. In most cases, both of these can get worse before the start to get better (around 10-20 hours) with close to full performance around the 100 hour point - but teflon caps seem to take a few hundred hours.
Tubes have some break-in effects, but since many are either used or old stock this can vary all over the place, so it's hard to generalize.
Also loudspeakers have some big effects. In this case, it's probably the mechanical breakdown/softening of the cone and suspension. Incidentally, this same effect occurs in string instruments such as guitars and the violin family and is well known in that context.
I can't say that I've heard anything I can reliably attribute to other components.
-
Thanks for the response Paul Joppa. I swapped out 10,000uF caps on the high current board with their "upgraded" Nichicon FG counterparts.
I also swapped out the caps on the low current board. I decided to try a Kemet paper film cap in the 0.1uF location just for fun. Not expecting to hear a noticeable difference the such a small capacitance, just thought it'd be fun :)
-
The 10,000uf caps on the big board are well out of the signal path.
-
Okay, so yeah, likely no discernable difference with those either. I mostly just bought replacements just in case I somehow damaged those with whatever connection issue occurred initially.
So far so good, though! There was a moment where I heard a faint, higher-pitched noise (mostly in the right channel again), show up, disappear for a minute, and show up again for several seconds. But, I haven't heard it since. Thought it might've been my phone, but it was on the opposite corner of my desk. I live in downtown Chicago, so I wouldn't be surprised if there were some interference, or noise on my AC circuit.