Danelectro Amp repair, low bass Ch2

[dmp]

I have a 1965 Danelectro DM-25 that has less bass on ch2 than ch1. I had all the controls set the same and when switching channels it's very noticeable that ch1 has more low end and sounded better (reverb and vibrato off in ch2).  I haven't played it in a few years, but I don't think it had this issue before.  The bass & treble pots on ch2 do seem to work, but turning up the bass on Ch2 has much less effect than turning it up on Ch1 (if that makes sense).
I was looking at it for a couple hours yesterday and could use some ideas on what might be the cause. At this point I haven't disconnected anything, since it's very clean inside. Nothing looks bad. Ch1 and ch2 are nearly the same except for the reverb send rtn and the vibrator optocoupler connection.

- I measured pots and caps in-circuit with a multimeter and the pots measure close in value between the channels and some of the caps measure near correct (like the .02 blocking caps). Obviously 'in circuit' measurements don't isolate a component, but I was looking for obvious differences between the channels (some of the caps weren't giving correct measurements)
- pulled the reverb tube, no difference- swapped the first 12AX7, no difference
- the optocoupler measures 430k with the amp off and the vibrato works when turned on. I haven't tried disconnecting this, but that's probably the first thing I'd try if I disconnect anything.
- With the amp off, I injected a 70 Hz, 100mV sine signal at the plate of the 12AX7 for ch1 versus ch2 and measured after the 330k mixer resistors on a scope, after the volume controls. Vol and treble all the way up, bass at noon. Ch1: 26.4 mV, Ch2: 24.8 mV.   Nearly the same level for both! Confusing. The guitar I've tested with is a telecaster with passive pickups.

Any ideas what's wrong or suggestions to diagnose this?

[dmp]

Channel 2

[dmp]

I disconnected the vibrato connection to ch2, and it now sounds healthy - the same as ch1 when switched back and forth.

[tubeswell]

Did you try a preamp tube swap?
Or it could be a leaky coupling cap in that channel? This could also weaken the signal. (Test caps for DC-leakage)
Or it could be a disconnected cathode bypass cap. Use your R-meter to check the connections for DC-continuity.

[dmp]

I did swap the preamp tube - no change.
But I think I found it last night by disconnecting the vibrato LDR.  With a resistance of 430k when the vibrato is off, since it is connected within the tone stack, it is forming a HPF.
I was using this with a big external speaker (Showman 15") which made it noticeable. The 12" in the danelectro cab 'hides' it
A quick sim in circuit lab shows the effect on the bass for 430k vs 100M (i.e. disconnected)
I'd guess the designers didn't mind that the vibrato channel had less low end.
Maybe the vibrato could attach after the volume pot

[dmp]

I'll probably leave it connected as is  - could use one of the front switches as a vibrato defeat since the ground switch is currently not used.

[PRR]

> I found it last night by disconnecting the vibrato LDR.  With a resistance of 430k when the vibrato is off

It really should go past several Megs after a few minutes. Why is it so leaky?

Has the light-seal around the roach been damaged? (Turn the room lights OFF- does that make a difference?)

Is the 12AX7 conducting? I think the cathode bias should turn it off; pull the 12AX7, does that make a difference?

Is that 12AX7 buffer's grid at ZERO? Or has it leaked positive?

Lastly (I hope): LDRs are "hot" chemical mixtures and may degrade over decades.

[dmp]

I'll check the vibrato 12ax7 and doublecheck the LDR resistance. The light seal looks fine.
Good ideas.

[dmp]

With all the lights off, no change. Ch2 lacks bass.
I pulled the vibrato 12AX7, and ch2 is fixed - sounds the same as ch1. So it seems it isn't the LDR, but the tube is conducting all the time a little bit, creating some light on the LDR
With the footswitch in (to turn off the vibrato), ch2 lacks bass.  So with grid at 0v the tube is conducting and ch2 lacks bass. 
The tube cathode measures ~2.14v. The 27k measures about 31k (in circuit). The 2.2M measures higher, about 10M (in circuit - questionable measurement). Without the footswitch, the grid still measures 0V when the vibrato strength is at 0.

[tubeswell]

Try replacing the roach

[dmp]

If the 2.2M is actually much higher resistance, like the 10M I measured, then the bias on the tube cathode is lower, which makes the tube run hotter - conduct more current.
Because the 27k and 2.2M form a voltage divider from the 220V supply. With 2.2M, the Vc=2.67 and with 10M, Vc=0.6Then the voltage increases with however much current the tube is conducting, I*27k
I think I'll replace the 2.2M?

[sluckey]

I think I'll replace the 2.2M?

If no joy after replacing the 2.2M, there's a simple mod that will absolutely shut off lamp driver triode. Disconnect the footswitch jack from the Strength pot. Now disconnect the 27K cathode resistor from ground and connect the end of the 27K to the footswitch jack.

[pdf64]

Yes, well spotted, that 2M2 resistor should be pulling the cathode voltage up, adding a bit of fixed bias to the cathode bias, so as to turn off anode current at idle.
With decades of constant DC, CC resistor values tend to drift up, often significantly so.

[dmp]

Schematic for some reason specifies a 1 W 2.2M resistor. But with 220 v I don't see how it could see more than .022 W (V^2/R)Am I missing something?

Now disconnect the 27K cathode resistor from ground and connect the end of the 27K to the footswitch jack.

Wouldn't that put 220v to the footswitch if the 12ax7 was dead or pulled?

[sluckey]

Half watt resistor is more than sufficient. A one watt resistor is twice as more sufficient.   

Wouldn't that put 220v to the footswitch if the 12ax7 was dead or pulled?

No. The only time there will be 220V on the footswitch is when the footswitch is open. So what? You ain't scared of .1mA are you? 

[dmp]

You're right - worst case it would be very small current, and unlikely to happen at all.
edit: Just looked up dangerous current levels - seems 10mA becomes painful,  100mA+ can be lethal So 0.1mA would be a tingle I guess

[sluckey]

I think you're on the right track with replacing the 2.2M resistor. This probably keeps a little trickle current flowing through that incandescent bulb roach. The bulb is likely to last longer than hard switching the tube on and off. Bulbs usually fail at turn on. And if the bulb in that roach fails, you will be unlikely to find a replacement. I ran into that with a failed incandescent roach in my Sunn amp. My solution was to use a VTL5C1 in a modified circuit.

[pdf64]

I think you're on the right track with replacing the 2.2M resistor. This probably keeps a little trickle current flowing through that incandescent bulb roach. …

I don’t see it like that. 
The 2M2 will increase the magnitude of the bias voltage between grid and cathode. The idea I think being to bias that stage into cut off.
Thus the bulb will only ever receive pulses of current, when the positive going wave from the modulation oscillator pulls the instantaneous grid voltage up, and allows anode current to flow, momentarily, for the duration of the positive part of the modulation wave.

[dmp]

These are sweeps of the Danelectro tone stack since it isn't in the duncan tone stack sim. Sounds nice!

[pdf64]

Wow, how’ve you got those charts?

[dmp]

Online simulator called CircuitLab
This is a Frequency domain simulation (20Hz to 20kHz).
I varied the position of the pots as 3 points per decade (log scale) for each line

[PRR]

Schematic for some reason specifies a 1 W 2.2M resistor. But with 220 v I don't see how it could see more than .022 W (V^2/R)....

Resistors have voltage limits, above a Meg this usually hits before dissipation. Especially in Carbon-Composition. Higher-Watt resistors are bigger and may support more voltage.

However 220V should be OK for about any half-Watt?

However if you are staring at a dead resistor it is usually wiser to get a bigger resistor (or two! series) than to do a lot of spec reading or torture tests.

{edit- punctuation}

[dmp]

I replaced the 2.2M and the problem is gone! Thanks!