View Full Version : Residual hum in Altec 1569A amp- any suggestions?


GordonW
01-30-2013, 06:10 AM
Have a couple of Altec 1569A amps here. One of them was dead when it arrived, and required a complete recap.

After doing that, and replacing a blown screen resistor caused by previous tube misplacement before the amp got here (putting a 5U4 into a EL34 socket is NEVER a good idea), it had what I considered a bit more background hum than I would like.

One thing I noticed, that the bias supply had only ONE cap for filtering. Diode, bias adjustment pot, then a cap, then the resistors feeding the output grids.

I wanted to change the bias pot arrangement, to at least have one bias pot per SIDE (so that each side of the output transformer would have a separate adjustment, so I could equal out total current into the two transformer sides), so in the process of changing from one pot to two pots, I also added an additional RC filtering stage before the pot (diode, resistor, cap, pots, caps, then the output tube grid supply resistors). I wound up using, IIRC, a 220uf 100v cap and about a 1K resistor, between the diode and the 220uf cap- then that fed the pots).

This helped considerably- especially when I used the pots to zero out the idle current differential between the sides of the output transformer. But, despite brand new caps (stock values, pretty much), there's STILL a bit of low-frequency (60Hz) residual hum, at the speaker. It's lower level than the other (stock, working, unrestored) amp, but it's still more than what I'd want...

Anyone done any further work on one of these 1569As, to try and snuff out that last little bit of hum? I've already tried switching out the input tubes (6CG7) to no avail... and grounding the input has absolutely no effect whatsoever...

Regards,
Gordon.

nerdorama
01-30-2013, 11:29 AM
Hi Gordon,

What's the signal ground system like? Star arrangement or bus? I've had decent results on other amps by reworking the signal grounds. Often the original layout is wired for convenience of wiring rather than best noise.

BTW, thanks for all the posts you've made over the years documenting your projects. Lots of fun and interesting stuff.

Best,
John

Gary Kaufman
01-30-2013, 03:30 PM
Gordon -

I rebuilt a pair of 1569A's a few months ago - one had the original 40/40uf@500v caps, the other had cobbled together axial caps under the chassis.

The 40/40uf@500v caps seemed to test fine for leakage and value so I left them in place. I rebuilt the other unit with some 50/50uf@500v clamp mount caps.

The rebuilt unit ended up having much lower hum levels, and I ultimately rebuilt both with the clamp mount caps.

Both are quiet now - with almost inaudible hum even with Altec A5's.

Are you using the 15095's for input?

Did you try clipping on some additional capacitance after the rectifiers?

You could un-ground the center tap of the 6.3v filament transformer and add a hum potentiometer.

Are the output tubes well balanced?

- Gary

GordonW
01-30-2013, 04:18 PM
Gary, I put in separate bias pots for the two sides of the output transformer, so I could balance the idle current, even closer than just well-matched EL34s could do by themselves.

There are no input transformers on these- they're running unbalanced input (an adapter to RCA plugs connected to the screw terminals).

I've got a 50/50 and a 32/32 cap in this one- one section of the 50 and one of the 32 in parallel for the first 80uf section, then the other 50uf for the diff amp, and the other 32uf section for the front end B+. That keeps the stock 80uf value for the rectifiers.

I think I will try some additional capacitance, for the front end. Maybe clip in a 22uf or 33uf in parallel with the 32uf...

Regards,
Gordon.

Ths364
01-30-2013, 04:29 PM
I have an Altec 353a: after replacing the big mains caps with new It still had a low level hum. Replacing the coupling caps made the hum disappear.

Gary Kaufman
01-30-2013, 04:38 PM
I missed the separate bias pots. I've started doing that more lately. Finding well matched quads of anything can be a challenge. I could easily start a rant about "factory matched quads" or anything sold as matched on Ebay - but lately the only thing that appears matched is the number of pins.

The pair of 5U4GB's certainly doesn't mind 100uf.

Nice amps, and fun to work on. Altec obviously had field servicing in mind with the open layout and rack mounting.

dcgillespie
01-30-2013, 05:01 PM
Gordon -- If the hum is 60 Hz in nature, then nothing you do to the HV supply will affect it at all, since any thing there is 120 Hz. Ground loops also tend to be 120 Hz in nature as well, so I would not suspect that.

60 Hz hum is typically either AC line/ground oriented (via the typical AC bypass cap), poor lead dress, or heater circuit grounding/lead dress oriented.

Have you got a schematic you can post?

Dave

jaymanaa
01-30-2013, 05:36 PM
Gordon -- If the hum is 60 Hz in nature, then nothing you do to the HV supply will affect it at all, since any thing there is 120 Hz. Ground loops also tend to be 120 Hz in nature as well, so I would not suspect that.

60 Hz hum is typically either AC line/ground oriented (via the typical AC bypass cap), poor lead dress, or heater circuit grounding/lead dress oriented.

Have you got a schematic you can post?

Dave

This, and to be sure, measure the AC that's riding on your DC.

zandru
01-30-2013, 08:25 PM
http://ampslab.com/SCHEMATICS/Altec1569A.jpg

dcgillespie
01-30-2013, 11:41 PM
Zandru -- Thanks for the schematic!

Gordon -- A couple of things really stand out after seeing the schematic:

1. The output tube screens have NO additional filtering beyond that of the plates: Since there is no cap after the 250 ohm screen resistor, the ripple of the plate supply is being directly injected into the screens of the output tubes. To see how much of a problem this is, pull the phase inverter and driver tubes and see how much hum remains. If the hum remains, consider placing an appropriate cap after the 250 ohm screen resistor, and then isolating each screen from that point with its own 100 ohm screen stability resistor to protect against any tendency towards arcing.

The driver stage receives its B+ from no more than a single section CRC filter from the rectifier which is inadequate by most standards. Consider either:

A. Replacing the driver stage dropping resistor with two resistors of 1/2 the value of the original dropping resistor and adding a new cap at their junction, or

B. Sourcing the B+ for the existing dropping resistor AFTER the 250 ohm screen resistor IF a new cap is added at that point, or

C. Do both.

With the original design as published, it's no wonder the unit is extremely sensitive to the DC balance of the output stage with regards to minimizing hum. I'm sure that improving the filtering in the bias supply helped as well. Altec was clearly counting on a TON of cancellation going on in the output transformer, because the filtering provided will certainly not produce low hum levels with any kind of imbalance there.

I hope this helps. Let us know what you find!

Dave

kirk57
01-31-2013, 12:24 AM
I'll be watching this one...I have a pair of these and have replaced all the PS and coupling caps but still get pretty much identical amounts of hum from both amps unless I adjust the bias far below the recommended value (forget what that value is as it's been a while since I've looked at them)

Does the 'no hum at very low bias' give any hint as to the source of the hum?

dcgillespie
01-31-2013, 07:32 AM
Kirk -- Very much so! Adjusting the bias current to very low values has a huge effect (negatively) on the effective gain of the output stage.

For any hum originating at the input or picked up due to poor lead dress, such an adjustment would have little effect because the NFB loop would react to largely negate the loss of gain in the output stage. Therefore, the basic sensitivity of the unit remains unchanged, and with it, unchanged with regards to hum pickup in the input stages as well.

However, while the NFB loop will maintain the overall sensitivity of the unit, the output stage itself will still become much less sensitive to the ripple supplied to it by the power supply at lower quiescent current levels. That the hum in these units is so dependent on the quiescent bias and DC balance setting in the output stage then is a very strong indicator that basic power supply filtering is at issue here.

Dave

GordonW
01-31-2013, 04:57 PM
Dave, I think I'm with you on this one. I fired it up without the second 6CG7, and there was still residual 60Hz hum.

If the guy who owns this one wants it "silent", adding a screen cap and separate screen resistors is going to be the first thing I suggest...

Though, you'd think, that since the hum into the screens would be the same four all four tubes, that the PP nature of the outputs would cause cancellation. Of course, if all the screens aren't drawing the exact same idle current- it could be just like the plates, when they have unbalanced current...

Regards,
Gordon.

dcgillespie
01-31-2013, 06:07 PM
Yeah, I think this amp was intended for distribution service, where the kinds of hum levels for high fidelity service just aren't required. To bring this amp into that realm, the B+ filtering is going to need a good going over.

Dave

Sam Cogley
01-31-2013, 06:23 PM
Sounds like a good choke would help.

kirk57
01-31-2013, 06:32 PM
Yeah, I think this amp was intended for distribution service, where the kinds of hum levels for high fidelity service just aren't required. To bring this amp into that realm, the B+ filtering is going to need a good going over.

Dave

Well there certainly is lots of room under the hood for bigger PS caps.

Any recommendations?

dcgillespie
01-31-2013, 06:56 PM
Kirk -- Bigger caps would not be very effective in this case. What is needed is more stages of filtering, as I was suggesting earlier.

Installing a 50 uF cap after the screen resistor, and then installing individual 100 ohm screen stability resistors (carbon film, .5w each) at each screen terminal would go a long way to helping the issue. Additionally, sourcing the B+ dropping resistor for the driver stage at the output of the 250 ohm screen resistor -- rather than it being supplied directly off the rectifier tube as in the original design -- would put some icing on the cake.

Dave

zandru
01-31-2013, 07:04 PM
I'm in favor of subtraction rather than additional components.

Sam Cogley
02-01-2013, 01:11 AM
I'm in favor of subtraction rather than additional components.

Not necessarily the best plan if the PS is already under-designed.