Bottlehead Forum
Bottlehead Kits => Mainline => Topic started by: dubiousmike on November 02, 2013, 01:34:53 PM
-
[EDIT - for those reading this thread in connection with their own trouble shooting, my ultimate issue, as described at the end, turned out to be a bad joint on R3 on the reg board, the 2.49k resister that, combined with R4, determines the variable output of the LM431 (labeled Kreg, which sets the cathode voltage on the 12au7). If Kreg is too low, the regulated voltage going into OA/IB on your C4S board will be too low to light up your B-side LED's!]
Just completed the initial stab at my mainline, and apparently I took a misstep somewhere along the way. The two LED's on the B side of the C4S board above my A socket aren't lighting up.
All resistances and most voltages check out, except for the Breg and Kreg measurements associated with that board.
Bad Side Good Side
Breg 163 220
Kreg 1.8 10.8
My tubes are glowing, and I just finished double checking the orientations of all the components on the problem board. Are there any other measurements or steps I might take to help isolate the problem? Thanks in advance for your help!
-Mike
-
Check that the resistors in each R1 position on your faulty sides are the correct values. Swapping those can throw things way out of whack.
-
Thanks for the quick response. Just checked the resisters (R1) on the C4S boards, and they are as the should be. 37.4 ohm on the A side of the board and 49.9 on the B side.
Any other thoughts as to what is in series with the two LEDs that I might check, measure or resolder?
Here's a link to a photo of the problem board if that would help diagnose the issue: http://cdn.head-fi.org/a/a7/a7e8c330_20131102_160013.jpeg
-
What do you have at IA on the funky board?
How about OB/OA on the center PC board? (Does turning the trim pot affect these voltages?)
-
I'm not familiar with the circuit but looking at your photo there appears to be a difference in the # of wires connected to the bA and bB pads. Maybe something is not connected that should be?
Robert Lees
-
Oh, also, the TL431 and PN2907 look pretty similar, they are easy to swap by accident.
-PB
-
Thanks for the further suggestions. PB - I double checked and do have the tl431's and pn2907's in their proper places. Robert - The A and B sides of the C4S board are supposed to be wired asymmetrically, but I plan to go back over that section of the manual tomorrow with fresh eyes.
Is it possible to measure one or more of the pins of the 2907's or main transistors to see if I might have a dud (or I suppose I could have damaged one while installing the heat sinks)? Do you guys have the circuit in splice or some software that would allow you to tell me the target voltage at some other points on the C4S board?
-
Somewhere there is probably a sticky about checking transistors. Basically a transistor is a set of diode junctions. So the way to test one is to do a "diode check" (resistance) between each possible combination of leads. They should all read about, say, 1K ohm. A transistor is dead if any pair has a very low resistance reading, like around 50 ohms.
-
Can you remove the screws from the PCB and roll it over, then shoot another pic?
Also, did you swap 6C45's side-to-side? Do both 6C45's glow? The problems you describe could all be attributed to the 6C45 on your offending side not being able to draw any current. This could occur if you miss the step labelled in red when building the center PC board, not having the heater feed going to tube socket pins 4/5 connected properly (which would mean no glow), or having a miswire at the outer sockets themselves.
-
Thanks for the further help Doc and PB! First, let me address the below because I forgot to respond to it before:
What do you have at IA on the funky board?
How about OB/OA on the center PC board? (Does turning the trim pot affect these voltages?)
IA on both boards is 278v. OA appears to be where things get interesting on the problem C4S board (with the unlit LED's). This would seem to make sense, because if OA isn't delivering the right voltage to IB, then perhaps the LED's on the B side wouldn't light?
Breaking it down - on the bad board, OA reads 163 - whereas it is 220 on the good C4S board. This same voltage (163) is at terminal 22, on +reg of the B side of the regulator board, and on pin 6 of the B socket. Does this mean my real issue is on the regulator board?
In answer to your other questions - OB and OA on the regulator board both read around 2.0, and the trim pots do affect these voltages on both sides. Also, I double checked the red instruction (bending the trim pot pins), and it looks like I did that correctly. All tubes glow, and I have tried swapping them with no effect. I have also tried a different 12au7 that I know is a good tube.
For further comparison purposes between the C4S boards:
Bad Board Good Board
IA 278 278
OA 163 220
IB 163 220
OB 161 174
Thanks again for your ongoing help!
-
I would be inclined to check the A side transistors on that bad board.
-
You can also yank the 12AU7 and run the amp without it, though I doubt that will fix your issue.
I would focus on comparing the wiring between the 9 pin sockets. (There are some differences, as some pins are tied together internally).
I'd also reheat the joints on the jumper that goes from bA to bB on the offending C4S board.
-PB
-
Thanks again so much for the further help!
I did discover an apparent error: my problem C4S, A-side MJE5731A was installed in the inner set of three holes, instead of the outer (as depicted in the manual). Sadly, moving it did not fix the issue. In fact, as I only pieced together afterwards - the two sets of holes appear to be shorted to one another.
On the upside, when I went to measure this - I think I pretty much confirmed doc's theory that that particular transistor is the problem. The middle/collector pin on the A-side MJE5731A is evidently the first spot in the circuit where my undesirable 163v shows up. By contrast, the collector pin on my A-side MJE5731A on the good C4S board reflects the proper 220v.
Does it make sense for me to reach out to Josh about a replacement MJE5731A at this point?
PB - I forgot to try pulling the 12au7 when I had her fired up, but I did go back over all the wiring of all three sockets and the wiring from them to the various boards, and it all looks right. I also metered bA to bB and it has continuity and 0 resistance - so that shouldn't be the issue.
-
Yeah, a backwards MJE5731A will cause this problem.
We can send you a replacement if you like, they are not all that easy to obtain locally. I'd also recommend replacing the PN2907 that works in conjunction with it, just as a precaution.
-PB
-
Thanks PB - that would be great! I just sent an email to [email protected] confirming my request for those two transistors, my address info, and including a link to this thread.
Any chance you could steer me towards some reference on the best way to desolder a PN2907 if it turns out to be necessary? I assume picking up a solder sucker might be a prerequisite?
-
Any chance you could steer me towards some reference on the best way to desolder a PN2907 if it turns out to be necessary? I assume picking up a solder sucker might be a prerequisite?
Yes, get the kind that looks like a long tube, that is spring actuated.
To get the PN2907's out, cut the transistor out first, then suck the solder/remaining component lead out of each hole. I would just replace the PN2907 out of the gate.
-
Much appreciated. [Edit: just received an email confirming shipping on the replacement parts. BH's customer service (and products!) really are second to none!]
Also, regarding removal of the existing transistors, will a ratshack or cheap amazon solder sucker (plunger type) do the job, or is this one of those tools where its worth dropping an extra $20 for something like a "soldapullt"? I only had wick on hand when I moved the faulty MJE5731A, and that was quite a chore. I think I actually have one of these ( http://www.radioshack.com/product/index.jsp?productId=2062731 ) buried somewhere in my garage, but I couldn't find it the one night I finally had occasion to break it out heh.
-
Plunger solder suckers work well on PCBs. Not so well on tube pins and terminals. I have both but have gravitated to solder wick for all.
-
If you are trashing the transistor anyway, just cut its head off, then you can deal with the legs one at a time, then a little wick on the holes if needed...John
-
I think I actually have one of these ( http://www.radioshack.com/product/index.jsp?productId=2062731 ) buried somewhere in my garage, but I couldn't find it the one night I finally had occasion to break it out heh.
The solder suckers with the bulbs are reasonably annoying to use, I'd blow another $15-20 on a spring actuated sucker.
-PB
-
Well my spring actuated sucker arrived this weekend, and it worked really well on the MJE5731A pins. They came out nice and clean. Unfortunately, I had a much harder time getting suction on the smaller pins of the beheaded PN2907 and ended up damaging an adjacent LED and the 37.5 ohm resister (37R4F) in the process. Pro tip: this is harder than it appears on youtube and it is evidently still important to pay attention to what your iron is touching even when you are focused on the precise placement of the sucker nozzle with your other hand :-[
I at least managed to extract the destroyed LED and resister without causing further havoc and sent in a supplemental part request late last night. Will plan to provide a further update once the replacement transistors, LED and resister are in place.
-
When I use these devices I hold the solder sucker on one side of the PWB and then touch the other side with the soldering iron. Then when solder is molten/flowing pull the trigger. The solder will be gone. This assumes one cuts the part out as John & PB said.
Good luck with the debug.
Tim
-
Thanks Tim! Makes sense. I'll definitely give that method a go the next time around. For now, I'm thinking happy thoughts and keeping my fingers crossed that the replacement parts will do the trick when they arrive.
Seems pretty clear at this point that my A side MJE5731A was the source of the problematic 163v measurement. Hopefully, that's the only issue I've got to resolve.
I'm thinking that once I have a totally functional mainline, I would like to go through and compile a longer list of voltage measurements, at all of the labeled terminals on the boards, in hopes of helping others down the road with possible trouble-shooting. Along with fully labeling the schematics provided in the manual, I imagine it would be a great learning exercise for me. (The latter, of course, I would not publicly post in deference to bottlehead's IP.)
-
Looks like I've been barking up the wrong tree. My replacement C4S transistors, LED and resistor all arrived as of this evening, and the installation went smoothly. Unfortunately, when I powered her up, the two B side LED's were still dark, and the same problem ~160V measurement was present at OA, IB, 19U, +reg, etc.
After going back through the entire manual (again) and cross-checking my wiring, it dawned on me that I had never actually tried PB's suggestion early on of pulling the 12au7. Bingo. Suddenly all my LED's are up and running and voltages on both C4S boards mirror one another. I tried an alternate 12au7 (pulled from my crack), but that put me right back with 2 dark LED's.
I triple checked the wiring on the B socket, but can't find any errors. I'm also including photos below in case my eyes are playing tricks on me. Does this mean my most likely issue is something on the A side of my regulator board? From the schematic, it looks like the only other active component (aside from the tube) feeding the reg HV to the B side of the C4S board (what should be 225V) is the LM431? Would that be the most likely culprit? Any further suggestions or measurement points you guys can provide would be greatly appreciated. [I actually have a couple of spare 2n2222a's on hand, so I could try replacing that if you think it's worth a shot.]
Thanks again - and here are the pics of the reg board and b socket:
http://cdn.head-fi.org/1/18/18cd82ef_20131115_215233.jpeg
http://cdn.head-fi.org/4/4d/4d34750f_20131115_215323.jpeg
http://cdn.head-fi.org/e/e1/e17b8ab2_20131115_215359.jpeg
-
I'll have to look at our Mainline at the office when I get in on Monday.
I can see that your C4S board has the R1 resistors in the correct spot, which leaves plenty of current for the 12AU7. Having one side functional tends to eliminate miswires, so I'll have to do a side-by-side comparison.
-
Just solved it!!! Thanks so much to PB, Doc, Grainger, 2wo, TubeTim, vetmed for all the support along the way - and a special shout-out to the Queen for her incredibly gracious emails and encouragement in connection with fielding my replacement part requests.
The culprit (as explained in some detail below for those who are interested), turned out to be a simple bad joint on a resister which threw off the LM431's variable output, which meant inadequate plate voltage on the 12au7 on that side, which in turn meant inadequate voltage on +reg/OA/IB, which caused the LED's to stay dark. It is so immensely satisfying to have finally figured this out - made all the more so given the incredible sonics of the mainline. I'm listening to Pierre Fournier's recording of Bach's Cello Suites right now on my hd800's. I've listened to this album a hundred times on numerous systems, and I can't recall that it has ever sounded this good. The mainline's effortless transients, truly lifelike timbre and micro-detail as Fournier digs deep into the strings are absolutely amazing. Trebles are detailed and extended, but never strident, and the bass presentation is faultless - deep, impactful, and incredibly tight. I will always love my crack (which is a wonderfully enjoyable amp in its own right and will soon take up residence in my office) - but for anyone who has been sitting on the fence awaiting comparisons, be assured that the mainline absolutely takes it to the next level and is an absolute steal among TOTL amps.
I promise to post more impressions once I've had the time to do some real critical listening. For now, here's the concluding chapter in my trouble-shooting saga:
After discovering that my C4S boards fully lit up (and measured as mirror images of one another) if I removed the 12au7, I realized my issue had to be a regulator or b-socket problem. I started by measuring continuity on all the relevant b-socket pins (passed). Then, I spent a couple of interesting hours this afternoon fully labeling the schematics with tube pins nos., OA/IB/+reg, kreg, OB, Oa, Ba, etc...along with comparative voltage measurements at each of the foregoing.
In the course of annotating the schematics, I reviewed the datasheet for the LM431, and finally wrapped my head around the fact that the voltage on its output pin (Kreg) is a function of R3 and R4 on the regulator board. Lo and behold, touching up R3 (the 2.49k resister) was the cure to all my woes. In essence, if Kreg (i.e. the LM431 output - which is the 12au7 cathode voltage) is too low, the regulated HV going into OA/IB on your C4S will be too low to light up your B-side LED's! This was a heck of a learning experience. Frustrating at times, but all the more satisfying in the end - in fact, I may have literally been jumping for joy when I flipped the switch, saw all the LED's come up and then confirming that all my voltages were dead on. Thanks again to everyone who posted for your help getting there!
[Edit - to correct a material misstatement: Kreg is the LM431 output and 12au7 cathode (not plate) voltage.]
-
Once again I will just point out that we aren't being lazy when we suggest repeatedly that builders reflow all of their solder joints before deciding that there is a bad part involved. When a kit comes to us for repair and the fault isn't obvious that is the first thing that is tried.
-
Duly noted. But to be fair, I did reflow the whole C4S board early on - which we all thought was the source of the issue up until last night, after I installed the replacement transistors as recommended.
For future reference, the bottom of the reg board is also a little harder to access to reflow (because the leads specified in the manual aren't long enough to fully flip it over like you can with the c4s boards). I wasn't previously inclined to desolder and remove the reg board for a touch up job when the joints looked reasonably good (at least to my untrained eyes) and we all thought the problem was on a different board. In the end though, you are clearly right that pulling and reflowing all three boards would have turned out to be the most expedient cure by a long shot in this instance - albeit without the satisfaction of specifically identifying the issue.
In any event, very happy to be up and running now and greatly enjoying listening to the k2hd version of Oscar Peterson's We Get Requests.
-
Hi Mike
Nice job on the debug! That is great detective work and a good learning experience no doubt.
I feel your pain about the solder joints. I've soldered since '66 and double checked each joint on PWB with a loupe. I managed to find 1 or 2 missed or questionable. Also had missed some on tie points. Luckily, all (?) were found before power-up and it worked right out of the gate.
The wires going to PWBs are either short or long depending on perspective. If PWB never needs to be lifted from standoffs, they are to long. For debug it's a tad short but do-able. To me the hardest part was the reg board and soldering in the wire leads. I confused side A and B, and undid the wires on that part and had to redo. Luckily it all worked out.
Enjoy!
TubeTim
-
Thanks for the kind words TubeTim - I'm so happy to have figured it out and learned a ton in the process! Sounds like your practice of double-checking each joint with a loupe is extremely well advised. I'm going to pick one up and make sure to do that on future projects. My super cheap "helping hands" knock-off has a magnifying glass on it, but it is really next to useless. And I've got a partially assembled mini3 in the works where the spacing is so tight I know my naked eyes are not to be trusted.
Also a good point about the length of the various leads. Having an extra inch or so made the C4S boards pretty easy to access, but I suppose shorter really is better in the long run.
Hope you are enjoying your mainline as much as I am. I was up listening until close to 3am last night - just couldn't take my hd800's off. Transients and separation between voices are so good and clean that you can even hear the differences relative to other amps in super simple music like the Wailing Jenny's Parting Glass. I'm totally enamored thus far.
-
A magnifier is a huge help in checking solder joints. But it cannot show you if the joint is conductive. The only way to verify this is to do a continuity test, touching the the wires, terminals, leads, etc. on either side of the solder joint.