ACMP-81 hum sounds

[antichef]

OK, I got bored - I took off the lid of my ACMP-81 pres and messed with the hum issue. I recorded my messing with one of them - here's the mp3 file along with some observations.

*TURN DOWN YOUR MONITORS BEFORE LISTENTING* -- I have everything boosted up, and then for some reason, the mp3 export "normalized" it way up again.

The EQ stage is engaged the whole time. There are two circuit boards that have inductors on them, and these, along with the toroidal power transformer, seem to be involved. Each of those boards (like the others) has a gain pot that's adjustable on the front.

I found that if I turn each of those pots down almost (but not quite!) all the way, the hum reduces down to a much lower level -- that's the first thing you'll hear -- the hum decreases first when I turn one pot down, then the other. Next, I put some twisty pressure on the toroidal transformer and the hum goes away completely. Then I cut to another take, where both pots are back at 12 o'clock, and I do the same thing with the toroidal transformer, and the big hum goes away completely.

Wish that was the end of the story -- I was able to back off from the unit and have it be quiet, but it was real touchy - when I moved it, the hum came back, and then as I messed with the toroidal, eventually it refused to go away at all, and now spinning/moving/unbolting/removing/flipping/hanging-outside etc. the toroidal has no effect on the hum at all. The effect of reduction when I turn those two pots down is still there, though.

Wrapping the toroidal transformer in mu-metal peel & stick has no effect. Same with the inductors. (careful with that stuff, btw, I cut the crap out my fingers with its sharp edges and didn't even notice until I saw the blood).

On one of my units, there's a brass screw missing from one of the inductors -- maybe someone was looking at it before me. Doesn't seem to make any difference on sound or hum, though.

Anyway, I have some replacement transistors on the way, and under normal operating conditions, the hum is really low for me anyway.

 

[antichef]

BTW, as I mentioned in the GB thread, I replaced all the BC547 transistors with BC441 transistors, and most of the BC557 transistors with BC461 transistors, and now the EQ does not add any noise. It still does pop when I turn it on, but after that, it's the same as being off in terms of noise.

The EQ board that is leftmost still has a bunch of BC457 transistors, I believe, and they're in a different configuration from the other EQ boards (where you see the two together) -- I haven't swapped these out, but probably will this weekend.

 

[Slowjett]

So you got a clean pre now?

How much did the trannys cost you?

 

[thajeremy]

we need video of someone fixin up one of each model.....

 

[antichef]

They're fairly clean - could still be better. The main thing is that it doesn't make a difference whether the EQ is on or off - that is, there's no EQ hum. And there's no hum at all on one, but there is some hiss when it's turned all the way up. The other has the hiss, too, and just a little hum that I haven't tried to get rid of yet. Both are quieter than my ACMP-73s. Neither is as quiet as my Black-Lion-Audio modified MOTU Traveler (but I *think* they're quieter than it was before it was modded).

Oh, and so far, it would be a pretty boring video -- 3 pins * 2 transistors * 4 boards * 2 preamps = 48 solder joints

For each EQ board, you have to remove the big (but not the little!) knob, unplug a ribbon cable or two, remove a "bridge plug", and unscrew 3 to 4 screws holding the board to the chassis - then the board comes right out, and you need to use a solder sucker to desolder the two transistors (which are in the same two positions on each board), then put the new ones into the holes (facing the opposite way of the old ones, since the pinout is reverse) and solder'em in and clip the leads that are sticking way out of the bottom of the board. Then put everything back in place. I did one board at a time and tested each one, because that's a good idea for someone at my neophyte skill level. Two times I thought I had bricked the unit because 1) I forgot to put back a bridge plug and 2) I put in a new transistor backward. oops. But all was happy when I figured out what I had done.

Half the transistors were 71 cents each and half were $1.37 each, so, um less than $20 I guess.

 

[antichef]

OK here's my instructional picture:

I should have said that the big knob actually has two screws (the other one is under the visible portion of the knob in the picture), and sometimes you have to loosen both - you probably want to tighten both when you put it back on.

Some variations are: one or two boards have an additional ribbon cable or two, but it should be obvious, and one board has a big brass post (for the lid screw) instead of a regular screw in one of the positions (it's visible in the picture to the right of the subject board, just under the inductor with the comment about my sloppy mu metal fix-attempt), but you just unscrew that, too.

after you do the stuff in the picture, you can pull the board out, then replace those two transistors with the new ones. The picture is post-mod. The little dots on the tops of the new transistors are there to show the orientation of the pins.

I repeated this for each of the 4 eq boards that have those transistors in Q4/Q5. The first EQ board has a different configuration, and a bunch of BC557 transistors (like 7), but no BC547. I left those alone for now, because the hum on my EQs is gone.

I have ACMP-73s, and they don't use the same transistors.

Another edit: Hopefully my posts don't give anyone the impression that I know what I'm doing -- the transistor swap was prescribed by zmix on another board, and anything else I've done has been suggested by someone else who knew better than I did. Zero original thought here. I just thought it would be convenient and maybe encouraging to see this stuff.

 

[antichef]

Oh, yeah - and on the issue of the part number -- zmix said that you *do* want the "6" on the end:

Quote:
( Q4 must be a BC461 - 6 and Q5 must be BC441 - 6). The input board and HP / LP boards are fine as they are.."

Is the number 6 after the part number the quantity we'll need?
Thanks for sorting this issue out!

No. The suffix "6" distinguishes the part. The BC441 and BC461 without the "6" can only handle 25% of the current that the "- 6" part can.

You will need 4 each of the BC461-6 and 4 each of the BC441-6 for every TnC-81 that you have.,

My transistors do *not* have the 6 on the end, and so they're not to his spec -- so far, it doesn't seem to be an issue, but if I notice any issues I'll post.

edit: trying to research this -- the data sheet for the -6 version has different variables maybe from the data sheet for the no-number version, but the -6 version says that Ic(cont) is 2 amps, and the no-number version says that Icm (collector peak current) is 2 amps. Maybe the -6 version is saying that it can handle 2 amps continuously, and the no-number version is saying it can only handle it for brief peaks? Clearly, I'm over my head, but the difference isn't jumping out at me.

 

[dgatwood]

Be sure to loosen both screws on the big knob. If you only loosen one, the knob is hard to remove. Possible, but not very good for the pot.

 

[antichef]

Be sure to loosen both screws on the big knob. If you only loosen one, the knob is hard to remove. Possible, but not very good for the pot.

squirrel speaks truth. About half of my big knobs pretty much fell off after loosening the one screw, though -- poor QA working in my favor for once -- Be sure to tighten both when you put them back, in any event.

 

[antichef]

Here's an update on my mu-metal adventures -- wrapping the inductors seems to have no effect at all. Wrapping the toroid does have an effect -- I just couldn't notice it before, because of large amount of hum coming from the EQ boards. Now on both my 81s and both 73s, I have the toroid wrapped and rotated until it gets quiet.

 

[vicenzajay]

Antichef...

Thanks for this thread, by the way. You're giving me hope that I'll be able to do some of this when I get home.

You're actually becoming quite proficient at it as well, it seems!

Jay

 

[antichef]

vicenzajay - hope indeed -- these things are great after a little tweaking!

I made this post in the GB thread, but I'll repost here to keep it accessible -- this isn't hum related, but it's pop related.

I ghetto-ized instructions posted by 0dbfs over at prodigy pro:

by moving that R43 lead from where it is over to to the R44 lead, position 7 on the gain switch becomes an "off" position, but the nasty pop is gone -- I even measured that 7R8 resistor that cooked on Jason's unit, and it stayed at a steady 2 volts throughout the entire gain range! I did this on both of my 81s and both 73s, too (I think the resistor number may be different on the 73, but it's in the same place - 4 over from the left).

0dbfs recommended replacing R43 with a 1.5k resistor and putting it in that position, but I decided to try it with just moving the existing 680k one because I'm lazy and ignorant, and I didn't even have to pull the board to do it (just pried up the lead with a little screwdriver and the touch of the soldering iron, then soldered it to the other lead).

Losing position 7 is kind of a bummer, but I heard that the original N*ves were like that too (and I guess we know why).

0dbfs didn't puke all over what I did, so I'm going with it - he did mention that he strategically used a piece of shrink wrap to reduce the risk of undesired connectivity between the moved lead and the old pad -- good idea.

As usual, if you see any unexplained conflict between my instructions and just about anyone else's -- go with theirs, because they're probably right!!

 

[dgatwood]

Moving my comments on this pre over here because they get lost in the GB thread.

At this point, I've been fairly successful at fixing the preamp hum without changing any transistors just by fixing a fundamental screw-up in the board layout. The grounds are woefully insufficient.

The one disadvantage to using the huge caps I used is that the circuit doesn't stabilize until those caps charge up fully, which takes about fifteen or twenty seconds, during which time it is massively noisy. In hindsight, this mod should probably be done without raising so many of the cap values to be that high.

At the moment, there's just the tiniest bit of hum left, but I haven't fully star grounded it yet, so that last bit might go away when I do.

Problems still remaining:

• Massive noise until the huge caps charge up.
• A very tiny bit of buzz.
• A scratchy sound near both ends of several pots. Probably some critical interaction between an unfortunate DC bias and some resistor in parallel with that pot.
• A strange, repetitive popping sound after adjusting the pots. It suddenly starts popping (quietly) at a fast rate, then gets less and less frequent and gets quieter. I have no clue what could be causing this one.
• A lot of hiss when the EQ is engaged.

Photo time. Here is the underside of a board from the ACMP-73 (which does not hum). Notice that almost the entire underside of the board is a ground plane, with connections not only to the jumper connectors, but also to metal rings around the screws.



Exhibit B: a similar board on the 81 (which hum).



Notice that there are no ground traces anywhere on the board except where absolutely required. None of the screw holes are connected to the ground plane. This is fairly typical of the 81's boards.

It doesn't take a genius to take one look at this and say "Wow, what a difference in layout engineering quality...."

Now, for my mods. Here is a full photo of the modded unit:



Notice all those extra red wires and the giant metal tanks. Those are the mods. In order from left to right: 2200uF, 1000uF, 2200uF, 1000uF.

Each of the minus sides is connected additionally to a heavy wire. The other ends are attached to a terminal spade. These are then attached to nearby ground screws, and at the moment, some of them are also connected to the power supply ground via the screw at the upper right corner of the power supply board.



Also, the second board from the left has smaller filter caps. I left those as-is, but grounded their minus sides as well.

 

[acegunn]

dgatwood, how are you testing for the 81 transistor oscillation? I didn't think I had the transistor oscillation on my 81's, just hum, but it turns out it didn't show up on a certain setting:

On my 81's, if I have gain and output all the way up, I only hear hum, which is only slightly above the noise floor. If I turn the output down a bit, or the gain know down one position, the transistor oscillation becomes very apparent/loud.

 

[jdbakker]

Running extra ground wires to "nearby ground screws" as well as the PSU ground screw (rather than having a star grounding scheme) sounds like a recipe for getting lots and lots of ground loops. Some of those loops will add so the hum, others will cancel. Then again, just putting in a solid star grounding scheme may well hurt more than it helps, as the signal connections between the boards is unbalanced and ground referenced.

JDB.
[no ACMPre here yet, so no way to do anything other than being an armchair EE]

 

[rodabod]

A strange, repetitive popping sound after adjusting the pots. It suddenly starts popping (quietly) at a fast rate, then gets less and less frequent and gets quieter. I have no clue what could be causing this one.

That sounds like motorboating to me. Maybe you will still need to swap those transistors.

 

[dgatwood]

Running extra ground wires to "nearby ground screws" as well as the PSU ground screw (rather than having a star grounding scheme) sounds like a recipe for getting lots and lots of ground loops. Some of those loops will add so the hum, others will cancel. Then again, just putting in a solid star grounding scheme may well hurt more than it helps, as the signal connections between the boards is unbalanced and ground referenced.

That term really bugs me. There's no such thing as a "ground loop". That's a total misnomer.

So-called "ground loops" (I won't use the term without quotation marks) are caused when two devices should be grounded, but one of them isn't. This causes all the electrical noise in the device with the poor ground to flow through the shield of an audio cable towards the device with the good ground so that it can flow to ground. Because that shield is wrapped around an audio cable, hum is induced on the line and enters into the audio.

Within the context of this device, the best thing that could be done to avoid ground hum is to star ground each of the boards and slice the grounds on every jumper board, but if you star ground it with heavy enough wire, slicing the existing ground traces (which do parallel the signal traces) shouldn't make things much better at all.

 

[dgatwood]

That sounds like motorboating to me. Maybe you will still need to swap those transistors.

I wouldn't describe it as a motorboating sound, so. It's a very crisp high frequency tick, tick, tick, tick, tick. It's very odd.

I suspect that the transistors are getting so hot that this is some sort of thermal noise. Two transistors per board are scalding to the touch.

Oh, yeah. I forgot one more mod I did to this unit. I looked at the lead length on the transistors and concluded they were way, way too long, so I pulled them through the board and snipped them off. I realize this isn't in a frequency range where that tiny difference is likely to cause problems, but I figured it couldn't hurt, and it got the transistors out of the way of the giant caps.

 

[dgatwood]

A couple more notes (mostly for my own benefit):

As best I can tell, the high frequency whooshing issues are mostly board 5, but all the boards are doing it to some degree.

I've also noticed that adding capacitance to 5C7 by touching the end of it with my finger causes the hum to jump into the stratosphere. This occurs to a much lesser degree on all the boards, but board 5 is really bad in this regard.

Note to self: the two spots on the 10k pot that cause problems are about 23 ohms and 330 ohms. This is probably significant.

One more note: even with all the pots disabled, there's still some high frequency trashiness. This absolves the main feedback paths on all of those boards because the feedback path is gone with that connector unplugged.

Board 6 is introducing 60 Hz hum still. Interesting because I always assumed most of it was coming in from boards 4 and 5. If I split the lines and run the signal jumper from the leftmost board to the rightmost, I still get the subtle 60 Hz hum when the EQ section is engaged, so the LF board is partially at issue.

I've also done bypasses to move the right end of the circuit progressively to the left (connecting each board's output to the return. From this, I have determined that boards 4 and 5 are both contributing to the high frequency issues, and that board 3 does the whoosh thing only when fed into board 4. If fed straight into the return, it does not behave in this way. From this, I am led to suspect that the problem here is that the weird oscillation is an interaction of multiple boards rather than a single-board issue on any single board.

With boards 3, 5, and 6 skipped, board 4 hums in the middle of its pot instead of on both ends, so yeah, there's definitely inter-board complexity to this issue.

With boards 5 and 6 skipped but board 3 still in the signal path, board 4 hums at the extremes and not in the center. This tells me that the scratchy pot/rushing/squealing issue in board 4 is partially caused by board 3 in some way, as it only occurs if board 4's input is driven by board 3 and not when it is driven by board 2 directly. Interesting. Or perhaps board 4 is causing problems for board 3, which still only has tiny filter caps.

With boards 3, 4, and 6 skipped, board 5 hums only with the pot at the very bottom in a very sudden fashion.

With boards 2, 3, 4, and 5 skipped, board 6 hums only slightly, and is linear in the level of the pot. This remains true with board 2 engaged. This probably largely exonerates board 2.

The most disturbing thing is the massive hum if I have only boards 2 and 4 engaged. The fact that board 4 hums that badly in isolation but not when between boards 3 and 5 is curious. The fact that board 5 hums only with the pot at the very bottom when run in isolation is equally curious.

 

[ausrock]

I've been watching these GB pre "fix" threads with genuine admiration for the guys that are slowly finding solutions to the problems, but parallel to that admiration is "dismay" that the people behind the GB are virtually nowhere to be seen, the warranty that "they" claimed existed, seems to be being conveniently ignored while others sort out the problems.

Regardless of where the problems originated, the minute a brand name is applied, the responsibility to the end user normally shifts from the original manufacturer to the owner(s) of that brand name...............ie: the onus falls not on GB participants and/or the Chinese manufacturers but squarely on TnC.

Like a lot of other o/seas guys that have gear ordered, I stand a snowflake's chance in hell of receiving any TnC support if I encounter "issues"