In addition to the above, I failed to recognize that resistor and capacitor as a Zobel network. It does cut frequencies, but that may be just to offsets the rise in high frequency response due to the inductance of the speaker. (Which will be all the more pronounced when that amp is switched to its mixed-mode feedback guitar mode.)  Also, this arrangement helps prevent high frequency oscillation from developing, like if there are long speaker cables that are capacitive.

I think I'm satisfied with that answer.

Pick the fruit from the low-hanging branches first, then proceed to the more difficult fruit.

Though I think that this problem is probably not caused by the mechanical connection in an effects loop jack, or between a jack plug, or the wiper contact in a potentiometer, it's not difficult to rule these out.

If you believe that a gain or volume potentiometer is causing this, you may be able to restore it by cleaning. Blow aerosol contact cleaner into the potentiometer and give it a few good turns back and forth. You may have to repeat a few times. If that doesn't help, find a potentiometer with the same physical dimensions (diameter, shaft length) and electrical resistance, and replace it.

Bad jacks can be replaced also (and I second the idea that if you have a jack-switched FX loop that you are not using, it may be worth it to just bridge it with solder connections so you have one less tone-sucking mechanical connection in your signal path).

If you rule out jacks and potmeters as being the cause, the next thing that might be it is a sliding connector inside the amp (if it has any). If there are any such connectors between the amp's circuit boards then (with the power off), gently unplug them and plug them back in. Be sure not to exert any pressure or tension on circuit boards. They are fragile, and sometimes the connections are hard to unplug or plug in. You may have to squeeze a little tab that opens a latch to allow the connectors to separate.  Sometimes unplugging and re-plugging a connector can restore a poor contact. (If doing that fixes your amp, I would consider rebuilding or replacing that connector.

If it's none of the above, the next thing we suspect is a failed solder connection on a circuit board, or possibly on some off-board component like a potentiometer.

I would proceed by using some non-conductive tool (wooden stick) to gently prod or tap the components (resistors, capacitors, transistors, hookup wires) with the power on. If any component causes popping, crackling or basically any noise when gently poked, it almost certainly has a bad connection. You may be able to repair it simply by melting the solder and allowing it to cool.

How can poor connections cause a fluctuation between two volume levels? Pretty darn easily.

For instance, between one amplifier stage and the next there may be a voltage divider (pair of resistors in series, going between the output of the previous stage and ground. The actual output to the next stage comes from the connection between these resistors). This voltage divider takes a fraction of the signal voltage to the next stage.  If the voltage divider's ground connection is disconnected, then the signal path still exists, but has a different level.  If the ground connection is flaky, then there may be a fluctuation between two levels as the voltage divider cuts in and out.

In the amp there may be fixed voltage dividers made with resistor components. But controls like gains are also voltage dividers. A bad ground connection to a potentiometer can thus do the same thing. Voltage dividers can be in feedback paths to; there, a flaky ground will cause level fluctuations by varying the gain.

Speaking of gain and feedback, in an emitter-follower amplification stage, a flaky connection on the capacitor that bypasses the emitter resistor will cause gain fluctuations. When the capacitor is in, negative AC feedback is shunted to the ground and the stage has a high gain. When the capacitor cuts out, AC signal flows through the resistor, inducing negative feedback, and reducing gain.

You don't have to understand any of that to find the component with the loose connection.

Please read the website's safety topics before doing any poking in live equipment. Don't touch any exposed conductors with an electrically conductive probe. If you do any repair, turn off the amp first, and wait a few minutes for all of the big power filter caps to discharge. You can still get a shock from the capacitors even if the device is unplugged from the power grid.

Experiment settles it: too much overall resistance in the feedback path is very harmful.

Firstly, the diminished current in it causes the feedback path to pick up hum (bad S/N ratio).

Secondly, frequency response is impacted quite severely: highs are rolled off.

Third, although there doesn't appear to be a noticeable rise in gain, the amp's biasing appears to be destabilized. It normally runs cold, but with this change, its output transistors warm the heat sink. I was watching for that and shut it off immediately.

Quote from: teemuk on July 19, 2011, 03:30:52 PM

QuoteIs this really just a low pass filter or is something else going on?

Yes there is. That "something" is......








....a drawing error in the schematic. 





Guess you didn't see that coming, right? Refer to channel 2 and corresponding resistor R64. The right value for the resistor is 10-ohms.

I was comparing the channels, but didn't pick up on that difference. Thanks.

I think I now understand the circuit.

The 10 ohm resistor makes sense. It ensures that the impedance of the shunt is never less than 10 ohms, regardless of frequency.  It doesn't make much of a difference to the treble cut at 10 Khz. The capacitive reactances at 10 kHz, 1 kHz and 100 Hz of the 0.22 uF cap, respectively, are -72.3, -723 and -7230. They all swamp the resistance value.  At 10 Khz, if the resistor were not there, there would only be a 0.8% difference in impedance: smaller than the tolerance on the cap.  We need the 10 ohm series to make sure that at high frequencies far above 10Khz, we are not short-circuiting the amp's output.

I now see that since the amplifier's load is small, even at 10 Khz, the high frequency cut will not be all that aggressive. My intuition for capacitor values is "biased" toward high impedance, small signal circuits. Not used to thinking about tone controls in power sections. A 0.22 uF parallel cap in a small signal, high impedance situation would roll of treble above 1Khz very badly, but so much here.

Quote from: phatt on July 19, 2011, 04:46:28 AM
There is no freaky super teckno trick going on with the *I* Drive trick.

Hi Phil, wrong thread? 

In the schematic of this amp (easily Google-able), at the very end of the output chain, the output is shunted to ground through a 0.33 ohm resistor R59 in series with a 0.22 uF capacitor.

Is this really just a low pass filter or is something else going on? That cap value seems quite aggressive. Wouldn't this kill the sparkle from the cleans?

Quote from: J M Fahey on July 19, 2011, 12:49:28 AM
Please don't multiply feedback values by 100, it does not work that way.

This is a good discussion topic. Can you explain why, for sampling the voltage to be fed back, a more resistive voltage divider can't used which has the same ratio between the resistors? Broadly speaking about amplifiers in general, what are the issues to watch out for with doing that. I'd still like to try it. If it doesn't smoke, it's good!  :duh I should be able to tell if the amp becomes obviously more gainy. If it oscillates, I will hopefully be able to kill it before it kills the speakers.

(One obvious issue might be that the feedback input may have a lower impedance than expected, which could makes the more resistive voltage divider less tall, so to speak. I.e. the same issue as with the potentiometer being too resistive is just relocated somewhere else, but now causing a real problem: less negative feedback, destabilizing the amp.)

Quote
Although the power amp is *some* kind of "big Op Amp", it's not a TL071 or similar Fet input one by a long way.

Yes, no FET input there. You cannot count on megohms of input impedance.

It's too troublesome to get a 1K pot which is dual-ganged, and I don't want separate controls for the left and right channel.

I think what I will try is to increase the impedance in the voltage divider into which the pot is mixing current feedback.

Now it's 100K:1K, but what if I just change it to 1M:100K, and make the coupling resistor 200K. Basically multiply everything by 100 except for the pot. So, change one resistor on the power amp board, two in my circuit, and I should be able to keep the 50K pot.

The one little uncertainty in this idea is the impedance of the transistor circuit where the feedback ends up, but I suspect it's decently high.

Quote from: teemuk on July 17, 2011, 02:01:19 PM
I agree, use smaller potentiometer resistance. Additionally, instead of using the typical volume control -style circuit to control the amount of CFB, you could just try wiring the pot as a rheostat in series with the 2K resistor. Most of those caps look redundant too, more than that, at the extremes of the potentiometer's dial they likely introduce some wacky RC filters that could hinder the intented operation. I'm quite sure you don't need at least the one that couples the 50K potentiometer to ground.

I made a note of that ground cap being useless in an earlier posting. The actual circuit I built does not have that capacitor, but does retain the coupler.

I did see the rheostat-like current feedback control in that high-level diagram in your book ,  but I wanted to be able to turn the current feedback completely to zero, so the Alesis can be returned to something similar to its original personality (modulo the diddled voltage feedback amount).

It would be fun to add presence and resonance to the circuit board, what the heck. I would probably want a Baxandall style tone control that is flat when the pots are centered (again, preserving the personality).

Thanks for the comments! And the book.

Quote from: J M Fahey on July 17, 2011, 08:47:38 AM
It will work, sort of, but your 50K damping pot is *way* too high, in practice it will work like a switch in both ends (max/min damping) .

Yes! I had exactly the same thought when I woke up this morning. At 50% position, we have 25K of the pot on top, but the bottom 25K is in parallel with 3K. Completely wacky.

I just used this because I had it on hand (bad reason, I know).

Gee, I'm as dim-witted as that guy in the other thread who wants to but a big ohm volume pot at the power end of his amp.

Thanks.

By the way, wow, how on earth (no pun intended) were you able to make a useful comment without a well-defined ground? 

Quote from: phatt on July 17, 2011, 12:06:16 AM
The best I can make out of your posted schematic is you have simply added a big fat resistor across the speaker which will just over heat the output devices,,to say the least.
You have no ground point that is clearly marked.
As such it's very hard to make a valued comment.
Phil.

Sorry, it's not a clear schematic, because it's basically just an excerpt of the whole system covering just the device being built and spliced in.

The terminals on the left represent the replacement of a 39K resistor, which is connected directly to point between the output devices: i.e. takes the output voltage.  This is the top resistor of the amp's voltage-feedback-sampling divider; so our "feedback return" terminal connects to the 1K bottom resistor. This has no DC ground, but an AC ground via an electrolytic cap. You can see that I've replaced the 39K resistor with a 100K one.

On the right side, the terminals are not speaker terminals, but the speaker return terminal and a ground. Normally the speaker returns to the ground. Here, we are spliced in between the speaker and ground. So the high wattage, small valued resistor is in series with the speaker: it is sensing the current and converting it to a small voltage, which is further sampled by the potmeter. This is the basis of the current feedback.  The output devices are not being destroyed. The GND terminal is a real ground to the power supply; I just didn't use the ground symbol anywhere.

Quote from: Kaz Kylheku on July 16, 2011, 08:17:11 PM

Now I will be playing with the pot to see if I can hear speaker damping being affected

It is. Quite terrific!

Firstly, the amp is more alive even in the maximum damping position. I expected this: I changed the voltage divider to feed back less voltage.

When you crank the damping control, there is noticeable drop in gain.

The speaker resonance on various notes is obvious.

My speakers resonate particularly well around C and D. (Think A-string rooted power chords at 3rd and 5th fret). The C and D roots are almost indistinguishable due to the speaker going ape-*s!!t*, and the obvious "oomph" is there long after you mute the note.

Way different beast from the stock unit.

Now I have to drill the hole to mount the damping control.  Also, put some feet on the new circuit board and fasten it in place.  Oh, and do the other channel.   That damping pot is double-ganged for a reason! Heh ...

For time being I will keep the right channel stock to do some A/B comparisons.

Quote from: Kaz Kylheku on July 16, 2011, 08:17:11 PM

Now I will be playing with the pot to see if I can hear speaker damping being affected

It is. Quite terrific!

Firstly, the amp is more alive even in the maximum damping position. I expected this: I changed the voltage divider to feed back less voltage.

When you crank the damping control, there is noticeable drop in gain.

The speaker resonance on various notes is obvious.

My speakers resonate particularly well around C and D. (Think A-string rooted power chords at 3rd and 5th fret). The C and D roots are almost indistinguishable due to the speaker going ape-*s!!t*, and the obvious "oomph" is there long after you mute the note.

Way different beast from the stock unit.

Now I have to drill the hole to mount the damping control.  Also, put some feet on the new circuit board and fasten it in place.  Oh, and do the other channel.   That damping pot is double-ganged for a reason! Heh ...

For time being I will keep the right channel stock to do some A/B comparisons.

So, everything worked, from design, to building, to working the first time on power-up with no debugging. :tu:

Passed smoke test!

Furthermore, hooked it up to pre-amp, guitar and speaker cab. @#$%-ing works!

Now I will be playing with the pot to see if I can hear speaker damping being affected.

Anyway, I haven't played guitar in like three days, damn!

I told myself: no guitar until this is finished.

Well, here is a pic of the kit, ready to go into the amp for a smoke test.

In the chassis, the matching connector is already spliced between the left speaker return and ground.

Cross your fingers! xP

No negative feedback from anyone!

Here is what the partially constructed project looks like. The hookup wires go to a connector (not shown). The 5 pin connector on the board, left to right, is FB-OUT, FB-IN, POT-GND, POT-WIPER, POT-HOT.

I reversed the order of the coupling cap and resistor, because it was convenient for assembly. An electrolytic 10 uF 50V is used.

I dropped the pointless capacitor in series with the 50K pot.

Apologies for the large dimensioned schematic image in my earlier post, by the way.