Ampeg B15NF bias slowly raising

[agedhorse]

I didn’t realize that you had the 100uF revision. They revised the earlier one 100uF cap version in favor or the two cap bias circuit, it is quieter and has better performance.

Nor did I, the CRC filter is more effective as well as reducing ripple noise AND injected ground current noise.

If the voltages at G1 are stable, generally this indicates a fault with the tube itself though this assumes that the quiescent screen voltages are present and stable.

Have you tried changing the output tubes with a known good set? I would look at this before assuming there's a defective coupling cap.

Also, the bias current can rise with temperature somewhat and it's not uncommon for the current to rise over say 10-15 seconds but will asymptote at a reliable value based on G1 the bias voltage. It might be a good idea to verify that you do indeed have a real problem first.

 

[boroman]

The tubes were borrowed from the other B15N (whole set) which is acting normal, so it's not them. I changed them back to the ones I've had from the beginning.

I replaced 100uf cap two times (thought that new one might be bad...). Trying 30uf cap there makes the "right" voltage in about 10 seconds.

I could suspect also bad new filter cap, but the amp sounds great when it finally catches the right bias. No drifting then. Stable. Super thick sound. No humming.

What bothers be I'm getting around 200 volts on the PIN3 and on the main caps when not in playing mode... (thatks for the point). Why? I don't know...

 

[agedhorse]

Did you confirm that the bias reaches a terminal value and remains stable at that value?

If so, you are chasing ghosts and there is nothing wrong (including "bad caps")

This is exactly WHY we first understand the circuit, make measurements and compare to what is expected in a properly functioning circuit, then evaluate possible causes for the incorrect measurement values.

Why do you measure voltage in standby? It's entirely possible that you are measuring a voltage that is electrostaticly coupled to a very high impedance node. There are lots of parasitic circuit elements that can cause presumably bad data if you don't understand how this works.

 

[Wasnex]

Is the standby circuit the same on this amp as the others? As I mentioned in my last post, I believe the bias circuit should go dead when the amp is put in standy rather than charging up. Either the original bleeder resistor or the pot you installed for variable bias should drain the bias off when the amp is put in standby.

However if the amp was modded, so the standby is after the tube rectifier instead of on the center tap, then the bias supply will continue to work while the amp is in standy. Then I believe the bias supply may also charge up a little more for two reasons: 1. The amp is not pulling current through the tube rectifier so there is less sag in the AC. 2. The tubes are not conducting so there is no reverse current flowing through the grids to draw the bias down.

When the modded standby is engaged, the AC will sag, and it will take a bit of time for the bias to drift down to it's normal operating level.

With the original standby setup, the bias has to charge up when the standy is switched to operate, so the current ramps up as soon as the plate supply is charged.

If the amp stabilizes after 30 seconds and holds bias for several hours without going into thermal runaway, I have to wonder if there is a problem to be concerned about. It could just be due to differences in how the amp is wired.

 

[boroman]

That's a very good point @Wasnex / @agedhorse !
I didn't think about it, as I connected it like it was before. Can't tell at the moment if other B15 is wired the same in this matter (not in the studio)

But looking at the pictures, I drew it:

Is it something odd here?

 

[beans-on-toast]

The standby switch is normally inserted between the high voltage secondary center tap and ground. It isn’t the best approach.

 

[Wasnex]

The standby switch is normally inserted between the high voltage secondary center tap and ground. It isn’t the best approach.

So where is the high bias voltage coming from when the amp is in standyby?

 

[Wasnex]

The tubes were borrowed from the other B15N (whole set) which is acting normal, so it's not them. I changed them back to the ones I've had from the beginning.

I replaced 100uf cap two times (thought that new one might be bad...). Trying 30uf cap there makes the "right" voltage in about 10 seconds.

I could suspect also bad new filter cap, but the amp sounds great when it finally catches the right bias. No drifting then. Stable. Super thick sound. No humming.

What bothers be I'm getting around 200 volts on the PIN3 and on the main caps when not in playing mode... (thatks for the point). Why? I don't know...

This was going to be the next thing I suggested as a potential source of the behavior. You're basically changing the RC time constant that effects how fast the bias circuit bleeds down after the amp is switch out of standby. I would say this is normal. If you want the amp to behave like the other amp, it needs to same bias circuit components. It's not something I would worry about. Consider what you have to be a soft start circuit.

Is the 200V residual when you turn the power off? Or does this voltage build up from 0V?

 

[agedhorse]

When circuit elements float, voltage measurements with a high impedance DMM will not be accurate without solid references and impedances.

 

[Wasnex]

When circuit elements float, voltage measurements with a high impedance DMM will not be accurate without solid references and impedances.

I could see a phantom reading with the plate supply. I don't believe there is a path to ground except through the tubes, which are not conducting since continuity is lacking in the power supply, so the voltage could float up. I am guessing this is the electrostatic reference you mentioned. I don't claim to totally understand.

I would think the bias supply would be referenced to ground through either the bias pot or the 56K and 47K dropping resistors. Also the behavior of the amp is suggesting the elevated bias reading in standby is accurate. Otherwise the bias would need to charge up when the amp is set to operate, rather than needing to bleed down.

What am I missing?

 

[beans-on-toast]

So where is the high bias voltage coming from when the amp is in standyby?

HV Power supply is off when the ground is lifted. The heaters, 6.3VAC and 5VAC are energized. Closing the switch, standby off, can be hard on the switch. With the standby switch open, DC is off but the AC finds a ground reference for the bias to work. Impedance paths to ground are important.

 

[Wasnex]

HV Power supply is off when the ground is lifted. The heaters, 6.3VAC and 5VAC are energized. Closing the switch, standby off, can be hard on the switch. With the standby switch open, DC is off but the AC finds a ground reference for the bias to work. Impedance paths to ground are important.

This represents the power supply in standby since the center tap is not grounded.

I guess I'll just have to take it on faith that it works, because I see no path to ground through the transformer for the bias supply.

Here is a normal half wave rectifier. You need continuity between both sides of the secondary or the circuit does not work.

Our circuit is a tube equivalent of C & D in the image below.

If the ground is lifted off the center tap I would expect both outputs to go dead. Are you telling me the + supply goes dead and the - supply continues to work?

The only guess I have is the path to ground is through the capacitors. So perhaps this circuit would produce + and - voltages.

Is this on the right track?

 

[beans-on-toast]

Bias circuit are often difficult to properly model because they are more complex than they appear. In some cases, the values of capacitors and bleeder resistors are important because of their impedance in the return path. A good example is the oddball capacitor coupled bias circuit used by Ampeg in the V4B and other V series amps, some Marshalls used the same approach. If the components aren't just right, there isn't enough usable voltage. I found this out when I wanted to implement a dual adjustable voltage. All this because they didn't want to pay for a dedicated bias transformer tap.

 

[beans-on-toast]

More explanation of the capacitor coupled bias supply. See the last section. The coupling cap in this circuit takes a beating and is prone to failure.

The Valve Wizard

 

[agedhorse]

Any time you lift the ground connection at the center tap, you also remove the ground reference. Otherwise, the supply side of the bias circuit may float since that winding ham no reference.

 

[boroman]

This was going to be the next thing I suggested as a potential source of the behavior. You're basically changing the RC time constant that effects how fast the bias circuit bleeds down after the amp is switch out of standby. I would say this is normal. If you want the amp to behave like the other amp, it needs to same bias circuit components. It's not something I would worry about. Consider what you have to be a soft start circuit.

It has the same bias compoments like its identical brother...

Is the 200V residual when you turn the power off? Or does this voltage build up from 0V?

I have 200V on the caps just after I hit power switch.
I must check the other B15M (which I just finished and didn't want to open again...) but I need to check if I have those 200V on the plate just after powering it on. I don't have any on my V4B after doing that.

 

[beans-on-toast]

There is a ramp up time difference between a fully discharged power supply and one that has retained a charge such as when an amp is being used and is turned off, then on again. From a design perspective, how the bias comes on from a fully discharged capacitor state is most important. How the B+ comes on relative to the bias supply is also important for power tube life.

 

[boroman]

OK. I brought back 2nd identical one to measure it all again. I confirm they are both wired and recapped EXACT same way.

So the differences are: with the "good" one, after
- powering it on but
- NOT in playing mode

Power supply has like -4V acroos bias resistor, one side of diode, 100K resistor, four supply (+) caps and anode. It all stays in about minus 4 volts.

With the "bad" one: I have -75V on diode, bias cap and bias resistor, almost 240V on 100K resistor near diode, around 200V on the plates and 170-200V on (+) power supply caps. When I hit power on, that voltage is slowly raising unti it reaches 170-200V, so it's not an "instant" ramp up

Any thoughts?

EDIT: if I could blame on of the caps - one would be wrong but not all would carry voltages, right? If I could blame standby switch is either working or not, not possible to pass just about "half" voltage (200V), right? The same for transformer...

 

[Wasnex]

OK. I brought back 2nd identical one to measure it all again. I confirm they are both wired and recapped EXACT same way.

So the differences are: with the "good" one, after
- powering it on but
- NOT in playing mode

Power supply has like -4V acroos bias resistor, one side of diode, 100K resistor, four supply (+) caps and anode. It all stays in about minus 4 volts.

With the "bad" one: I have -75V on diode, bias cap and bias resistor, almost 240V on 100K resistor near diode, around 200V on the plates and 170-200V on (+) power supply caps. When I hit power on, that voltage is slowly raising unti it reaches 170-200V, so it's not an "instant" ramp up

Any thoughts?

EDIT: if I could blame on of the caps - one would be wrong but not all would carry voltages, right? If I could blame standby switch is either working or not, not possible to pass just about "half" voltage (200V), right? The same for transformer...

I haven't really figured out the return path for the electrons in the bias circuit when the amp is in standy, but I believe they have to go through one side of the tube rectifier. See my thought train in post #32. If I understand your notes, one amp is developing 200V on the plates in standby and one is not. So it's possible the bias circuit is exactly the same on both amps, but something in or related to the plate supply is different in one of the amps. I would pull the tube rectifier and see if the bias voltage still develops. My guess is it won't.

If the bias voltage does still develop, I would wonder if the PT is getting a partial short to ground somewhere. Bad standby switch, bad wiring, bad PT? Just spitballing. There has to be a return path for the electrons and you have to find it.

 

[boroman]

I haven't really figured out the return path for the electrons in the bias circuit when the amp is in standy, but I believe they have to go through one side of the tube rectifier. See my thought train in post #32. If I understand your notes, one amp is developing 200V on the plates in standby and one is not. So it's possible the bias circuit is exactly the same on both amps, but something in or related to the plate supply is different in one of the amps. I would pull the tube rectifier and see if the bias voltage still develops. My guess is it won't.

If the bias voltage does still develop, I would wonder if the PT is getting a partial short to ground somewhere. Bad standby switch, bad wiring, bad PT? Just spitballing. There has to be a return path for the electrons and you have to find it.

Yes, without recitfier tube its all OK... But I swapped two different 5AR4s and one 5U4GB and it's the same on them all. So not a tube fault by any means. Tomorrow I will check the socket and will put another standby switch... Ech.. thanks for all your effort guys!