HELP! low bias and red plates after Lar/Mar mod

[fretwizard]

Hi All

I've just installed a lar/mar master volume in my ceriatone 1959 and everything seemed to be fine, except I replaced the PI output caps with NOS mustards because the original TAD caps by Ceriatone were stuck in the turret (because of the original wires going all the way up the turret) and I broke the leads while trying to pull them out

I wanted to try the NOS mustards in that position anyway...

My dual 250k pot wasn't really balanced so I chose different resistor values. With a 4.7M and a 2.7M I managed to get both halves to 223k when the volume is flat out.

Once I've finished I wanted to be safe and make sure there was no problems so I powered the amp slowly with the variac. Everything seemed ok.

Then I turned the controls all the way down, except the PPIMV of course and started to check the bias voltage and....

IT"S VERY LOW

I measured these posistions:

PT bias tap: 103 Vac
27k resistor before diode: 70 Vac
after diode: -30 Vdc

Now it gets bad. At the joint between the second 10uF cap and the 15K resistor I get -14 Vdc when the bias pot is maxed. And the minimum I get is -12.5 Vdc.
After the 47K resistor I measure -6.9 Vdc.

If I turn on the amp the power tubes are red plating at any bias-pot setting and I get values like 300 Mv and the bias test points.
The plate voltage measure about 388 Vdc.
If I turn the master volume down I can see it's doing something but the plates are red all the time...

What could have happened? I remember I only had to disconnect the PT bias supply so I could lift the board and remove the old output wires. the continuity from the bias filtercaps and the ground seems fine...
Unless I've pulled some other wire in the process but still why would the voltage change so much after the caps if the ground is ok? I didn't even mess with that part of the circuit...

R

[fretwizard]

I should pull all the tubes out and re-measure the voltages before checking under the board again....

[fretwizard]

I've pulled out all the tubes.
The PT bias supply is 104Vac, and B+ seems fine

The measurements are pretty much the same as before.
27k resistor before diode: 70 Vac
after diode: -30 Vdc
Then down to 15 after the 10uF filter caps.

Now at the V1 pin 1 I'm measuring over 400V!!!! and goes higher as I get closer to the power tubes. On the PI plates I even have the multimeter beeping for overload...

I've disconnected the choke/HT/OTCT junction and also the choke/OTprimaries junction and lifted the board. I removed some solder blobs on the bottom of the turrets but none of them seemed to cause any shorts with the jumper wires. I've checked for shorts along the PI supply and the bias supply and I get nothing wrong...

Preamp grounds seem to be ok, so are the can caps grounds...

R

[toner]

It's normal to have higher voltages at the tube sockets with no tubes installed.

Can you post pictures of the bias section and the PPIMV?

[fretwizard]

[fretwizard]

I've just rewired everything back... with the preamp tubes in, I get plate voltages closer to normal and 500v on the power tube pin 3.
In the bias section of the circuit I measure the same voltages I listed earlier. Thought now I notice that on 1 side of the PPIMV I get 14vdc, while on the other side I get 8vdc ... same difference between V6/V7 and V4/V5 at the base of the grid stoppers.

The pot resistance still measures about 223k when maxed, on both halves. As I said before the pot was not balanced, it was 240k on one side and about 230 on the other, so I took a 2.7M and a 4.7M instead of 2.2M on both sides. The side with the lower voltage seems to be the one using the 2.7M resistor.

I know that my mustards are wired in opposite directions (that was a mistake) but this shouldn't affect the dc voltage right?

[fretwizard]

Maybe I also fried the bias diode with too much heat when I tried to reconnect the PT bias wire into the turret, maybe that's the reason why it outputs only -30 vdc...

As soon as I have time I'll try to remove the PPIMV and reconnect the 220K bias resistors and see if I have the same problems with the mustards...

[fretwizard]

Alright...

I've disconnected the input side of the PPIMV, including the bias wire and checked the voltages again.

It all seems fine, the diode is outputting -60 vdc , same voltage all the way to the turret where the 220k resistors used to join. So I didn't fry the diode...

Also pulled out the power tubes and tested the voltages at the mustard caps with B+ applied, I get DC only on the plate resistor side, so I guess the mustards are not leaky?

Now I guess the voltage will go down as soon as I reconnect the PPIMV bias... so I wonder what might be the problem with the potentiometer...

R

EDIT: As I expected, voltage drops as soon as I reconnect the bias wire from the pot... even after I disconnect the output shielded wire from the grid stopper resistors

[fretwizard]

I've disconnected all the wires from the pot and left the resistors only...

If I connect it to the bias supply with an alligator-clip the voltages are fine... I can measure -60 vdc on the center lug.
I must have screwed something while soldering the shielded wires...

[joey]

I've disconnected all the wires from the pot and left the resistors only...

If I connect it to the bias supply with an alligator-clip the voltages are fine... I can measure -60 vdc on the center lug.
I must have screwed something while soldering the shielded wires...

It's certainly possible you can short the shield and inner conductor together if you are not careful with the iron.

That said This is a little OT, but Honestly shielded anything is not required for this. I went back and read the thread that inspired it,but the real ideas behind the whole reason why you would ever consider it is to maximise cmrr in a true differential system, you are not using the shield as a shield for this, just as a second conductor. but surprise surprise guitar amps are far from balanced so cmrr is good but not great. Does it hurt? No! Although you can do the same thing by just twisting both the input and output feed, or keeping them all parallel with a zip tie, which is what your really trying to accomplish by twisting them in the first place. It also only takes three twists per inch to maximise the effect.

On another note do not tie the shield to the bias supply if you are using the shield conductor as a true shield, it is a very high impedance suppy in almost every guitar amp, and far far from being just as good as ground. If you are just using it as a conductor than fine.

Lastly you don't Need to worry about the difference of 10k between each section of the pot, it's silly! You are well with in tolerance of the pot, and quite frankly and more importantly it wont matter! your not disrupting anything, trying to fix this buys you zero. there can be more between both sides through out the sweep if it's audio taper.

[fretwizard]

In the end the problem was very stupid. The pot was shorting with the chassis, or some lugs shorted with the chassis when I tightened the nut on the panel. The former is more likely because I remember to have double checked shorts...

Now I'm using 2 isolating washers from an unused cliff jack to keep everything farther apart and I've solved the problem. The thing is, I found out just after I've rewired the pot from scratch, even using new shielded cable. I was testing the voltages each time I connected a wire to the pot and everything was fine, until I tightened the pot on the back panel...

At least it looks even tidier now but I could have saved a lot of time if I checked more closely...
All good now

Joey, this is an audio taper pot. The 'bias' side of the pot is now balanced with the resistors but obviously the 'volume' side is still 240k vs 230k... I wonder if this will contribute to the mild ghosting I already have when the amp is cranked.
I knw for sure that the ghost is gone once I lower the PPIMV to 6/5.

Where do you think I should connect the shield? Can I use the output jack ground next to the pot?

[joey]

In the end the problem was very stupid. The pot was shorting with the chassis, or some lugs shorted with the chassis when I tightened the nut on the panel. The former is more likely because I remember to have double checked shorts...

Now I'm using 2 isolating washers from an unused cliff jack to keep everything farther apart and I've solved the problem. The thing is, I found out just after I've rewired the pot from scratch, even using new shielded cable. I was testing the voltages each time I connected a wire to the pot and everything was fine, until I tightened the pot on the back panel...

At least it looks even tidier now but I could have saved a lot of time if I checked more closely...
All good now

Joey, this is an audio taper pot. The 'bias' side of the pot is now balanced with the resistors but obviously the 'volume' side is still 240k vs 230k... I wonder if this will contribute to the mild ghosting I already have when the amp is cranked.
I knw for sure that the ghost is gone once I lower the PPIMV to 6/5.

Where do you think I should connect the shield? Can I use the output jack ground next to the pot?

Well what I was trying to say was that any attempts at trying to balance such a small imbalance in effective grid load is pointless (it's not actually called bias feed that a made up term that stuck, it is still just a grid load), and doesn't do anything in practice. the only difference between 230k and 240k has in the circuit is the effective ac plate load. With 230k and 100k it's effectively 69.9k to ac, with 240k it's 70.4k, thats mice nuts, and most fixed resistors even have bigger variance than that. It doesn't do anything usefull to correct a slight error like that. no current is drawn there, so that's not a concern.

Look at your output valves when you hear the ghosting (is it ghosting or a whine?) , do they suddenly glow when you hit a big Chord?

If you are indeed using a shield as a shield it must go to a low z path to ground potential or else it's not effectively a shield and can make things worse. Never connect it to a power supply positive or negative (negative is really just the same as a positive just opposite in polarity) unless it is a low z supply and even then DONT, if ground potential is an option. just connect one end to the most convenient ground. You could go further and have a whole separate return system for you shields, but you would again be wasting your time in practice.

That said shielded cable is not needed for this if you are actually trying to use a shield, it's pointless even at low levels as there's decent enough cmrr. The only real benefit to shielded cable is to use the shield as the other conductor in order to maximize common mode ac coupling, (same principal using it to wire your heaters) and even so it's easier to use a twisted pair for this purpose.

[neikeel]

Edit - wrong thread

Glad it was sorted!

[fretwizard]

Look at your output valves when you hear the ghosting (is it ghosting or a whine?) , do they suddenly glow when you hit a big Chord?

Well they do glow blue when hitting a big chord, but I also play it at high volumes with an attenuator. Before the Lar/Mar, it used to ghost at volumes just higher than3/4, bright channel, especially with neck pickup.

I'm pretty sure is ghosting... I had a thread about that, where I started asked for some voltage problems, and after that, the ghosting:
http://forum.metroamp.com/viewtopic.php ... 3&start=15" onclick="window.open(this.href);return false;

clip here:
https://soundcloud.com/overtone85/sets/ ... host-notes" onclick="window.open(this.href);return false;

but this is OT I think... anyway I don't know if it's worth balancing the PI. I know that the plate resistors are 110k and 90k (large tolerances...), rather than 100/82. I read about that in another thread where Novosibir was mentioning that, and also something about screen voltages being very close to the plate voltages... http://forum.metroamp.com/viewtopic.php?f=3&t=22964" onclick="window.open(this.href);return false;

[joey]

I think you missed the point a little again, and maybe didn't quite understand what I was suggesting. The point is that there is NO NEED to use those resistors on the pot for purpose of trying to balance it. It does nothing for you in that regard. you can keep them there if you are worried about the pot failing, but also keep in mind in the case that would happen, large values like that will most deffinitely exceed maximum spec'd g1 load resistance for fixed bias, which can cause bias drifts even as far as redplating anyway due to gas bombardment