Problem: Weber 5F6A with very low output!

[Diablo]

Test resistance between the power side (+) of the power caps and ground. The resistance should be infinite, or in your case, the value of the bleeder caps you installed.

[jibe]

The bias circuit is bugging me. I only get -39, I haven't checked what I have before the voltage divider but I don't get why I'm off on bias voltage, I can't see anything wrong in the circuit. I have a 39k in series with a 50k pot and I get -36 to -42 when I go from min to max on the pot, something is not right here.



//Jibe

[Diablo]

The bias circuit is bugging me. I only get -39, I haven't checked what I have before the voltage divider but I don't get why I'm off on bias voltage, I can't see anything wrong in the circuit. I have a 39k in series with a 50k pot and I get -36 to -42 when I go from min to max on the pot, something is not right here.
//Jibe

Check the voltage before the voltage divider - from C23 to ground. See if you have -63 volts there. Are you using a W022798 PT from Weber? For the primary leads on the PT, did you connect the blue lead (125V) or the brown lead (120V) to the AC power switch? It should be the brown lead.

[tubeswell]

The bias circuit is bugging me. I only get -39, I haven't checked what I have before the voltage divider but I don't get why I'm off on bias voltage, I can't see anything wrong in the circuit. I have a 39k in series with a 50k pot and I get -36 to -42 when I go from min to max on the pot, something is not right here.



//Jibe

The two resistors in the bias circuit (the 15k and 56k in the stock circuit) form a voltage divider. There is a calculator here:

http://ourworld.compuserve.com/homep..._Bowden/r2.htm

(If you measure the voltage drop across the 15k resistor and the voltage drop across the 56k resistor (in the stock amp), you can work out the current going through the voltage divider and see how the voltage at the point between the two resistors varies in proportion to the value of the resistors you are using).

You have taken the 56k and turned it into a 39k in series with a 50k wired as variable resistor, but lets assume for the moment that we are talking about 'one resistor' when we talk about the 'ground leg' here.

The higher resistance you make the ground leg, the more the voltage will be at the point between the two resistors, up to a point (because there are only so many volts (lets say -69VDC for arguments sake, if the bias tap puts out 50VAC) in front of the whole divider), so ultimately you can only push the voltage at the dividing point almost as high as the originating voltage.

If you are getting -36VDC when the variable resistor pot is set at zero and if you had -69VDC to start with, that 15k resistor is dropping 30V, and the 39k fixed resistor is dropping the remaining 39V. As you increase the rotation on the 50k pot to max, the 15k resistor is dropping 27V, and the 89k pot and resistor is dropping the remaining 42V. If you want the negative voltage at the divider to go higher (i.e. more negative), you could either increase the value of the fixed resistor in series with the 50k pot, or increase the value of the pot in series with the 39k resistor, or decrease the value of the first (15k) resistor. But ultimately you will only go as high as -69VDC.

When you have -69V on the grid, the tubes will be running quite cold and the plate voltage will be low. So if you want things to warm up a bit, you have to lower the bias voltage (which means making it less negative, which means making the 39k resistor you have a the tail of everything smaller) - try 33k or 27k there.

[erndawg]

Is something strange here or am I chasing ghosts?

If you have accurately represented the facts, then yes, you do have a problem. You don't lose 40 VDC and drop your bias for no reason.

The speaker cab I'm using is brand new, the speakers are not broken in. Maybe that could explain something.

New speakers may be harsh, brittle, or cold, but their efficiency will rarely change a significant amount. What drivers are you using?

I'm not sure why my B+ dropped after fixing my short and installing 220k bleeders.

Are your 220k resistors red, red, yellow? That is, are you certain they are 220k?

Turning my 50k bias pot to the max doesn't give me more than 442v on the plate.

What is the bias voltage when the pot is maxed out and the plate has risen to 442 VDC?

The bias circuit is bugging me. I only get -39

Is that with the bias pot maxed? That sounds contradictory to your quote above...


Answer the questions I posted here, then we can start troubleshooting with the following:

Measure AC input to the rectifier: Place the Power switch to On, keep the Standby Off, then place each meter lead on pins 4 and 6 and measure AC. Is the voltage the same as what you have measured in the past?
I'm trying here to rule out a PT failure/degradation/shorted turns. Let's not forget that you were severely stressing your amp with the original short and very hot bias, and may have damaged one or both transformers.

See if your 6.3 VAC and 5 VAC have dropped at all from previously measured values.

Have you substituted the rectifier with a known good one?

Let's start with this. Then, if needed, we'll isolate each stage and also measure the actual power out. The bias currents you measured above only indicate the quiescent (static) current draw. We need to objectively know what your amp is really putting out, irrespective of subjective speaker performance.

[erndawg]

When you have -69V on the grid, the tubes will be running quite cold and the plate voltage will be low.

"...and the plate voltage will be low". Not so. The plate voltage will rise to its highest possible value, constrained by the maximum value of the power supply voltage without a load. So, all else being equal, a less negative bias voltage (hot bias) will give you a higher plate current (hot tube) and a resultant lower plate voltage.

[jibe]

I'm using the 022798EU variant of PT, the 240 primary since I'm from Sweden.
I'm pretty sure I measured the bleeders before I installed them but I'll double check.
I'll also check the rest of the circuit so I don't switched resistors by mistake.
I built my own cab out of 2 16 ohms Celestion Blue.
From min to max on the pot I go from bias voltage -36 to -42 wich seems low for a 39k in series with a 50k pot, I'm not sure yet what I had to begin with.
If I go more negative the B+ will rise to 442 at max, but cathode current will go up above 35mA as well.

I remember checking the Secondary PT output it looked ok, around 3xx AC something from ground.
I'll do another round of full measurements, maybe something have changed. I'll also check what I have before the voltage divider.

Apart from the newley installed 1ohm cathode resistors, the 39k in series with the bias pot, the 220k bleeders and the bypass of the RCA jacks everything is as before.

I'll go back to work with this and report back.




//Jibe

[tubeswell]

"...and the plate voltage will be low". Not so. The plate voltage will rise to its highest possible value, constrained by the maximum value of the power supply voltage without a load. So, all else being equal, a less negative bias voltage (hot bias) will give you a higher plate current (hot tube) and a resultant lower plate voltage.

Actually erndawg, if you have -69V on the grid it will pull down the plate voltage more than if you had say -40V on the grid. The negative voltage on the grid counteracts the effect of the positive voltage on the plate. Try measuring it and see. Think about it, if you took a +100VDC potential and a -100VDC potential they would cancel each other out leaving you with zero potential.

[JoeM]

Actually erndawg, if you have -69V on the grid it will pull down the plate voltage more than if you had say -40V on the grid. The negative voltage on the grid counteracts the effect of the positive voltage on the plate. Try measuring it and see. Think about it, if you took a +100VDC potential and a -100VDC potential they would cancel each other out leaving you with zero potential.

??? The tube's biased colder with -69 volts on the grid. B+ should rise with the lower load. With -40v the tube will be conducting more, pulling the B+ down.

[tubeswell]

Okay so I did this with my 5G9 earlier to check this again.

With -23V on the grid, the tube current was 31mA and the plates were at 371VDC, with -27V on the grid, the tube current was 27mA (more negative bias) the plates were at 364VDC.

This surely must be because the increased permanent negative grid voltage is lowering the overall positive plate voltage, which in turn would reduce the current flowing from the cathode to the plate. (less voltage = less current for the same resistance)

Remember we are not talking about a pre-amp stage here. There is no large resistor loading the plate in an output stage. The output tube has a reactive load through the OT, which has very little DC resistance, so the quiescent plate voltage of an output tube is virtually at the same voltage as the B+. So to my way of thinking, the only other thing that can change the permanent idle voltage at the plate is another permanent negative voltage applied between the cathode and the plate. (And the bias voltage is drawn from the PT in the same way that the B+ voltage is - so they are both contraposed to one another so to speak) Anyway that's my take on it.

Having said all this I must qualify that I am a mere (tho' nonetheless passionate) hobbyist in this field, and so I have a personal philopsophy of never putting myself beyond the point of being able to be persuaded to an opposite point of view. Therefore I stand to be humbly corrected if somebody can kindly convince me where my understanding is flawed.

[black_labb]

i skimmed through the thread, but could the issue with the jacks possibly be that you connected the hot on the sheath part of the jack and the ground (connected to ground) on the tip, when the jack wasnt insulated from the chassis/ground? not saying the jacks are high quality, but they cant be that bad that they cause problems from the start... can they?

[jibe]

Nope, not the case, In my book they break when you put a solder iron to them, I don't know if I'm not causious enough but break it did.




//Jibe

[JoeM]

Okay so I did this with my 5G9 earlier to check this again.

With -23V on the grid, the tube current was 31mA and the plates were at 371VDC, with -27V on the grid, the tube current was 27mA (more negative bias) the plates were at 364VDC.

That's exactly opposite to any tube amp I've worked on. If you carry this to the extreme, at some point a high negative voltage will have the tube cutoff - it wont draw any current - the reduced load on the PS will let the B+ drift up. The amount of current drawn by the bias circuit is negligible.

[erndawg]

I'm using the 022798EU variant of PT, the 240 primary since I'm from Sweden.

Do you know what the spec is for that PT's High Voltage AC?


Does anyone have any experience with that European Weber PT? Specifically, what are the loaded and unloaded high voltage AC specs (the voltages at pins 4 and 6 of the rectifier)?

I'm pretty sure I measured the bleeders before I installed them but I'll double check.

Yes, just in case...

I built my own cab out of 2 16 ohms Celestion Blue.

Those are very efficient speakers and should sound great and *quite LOUD*.

From min to max on the pot I go from bias voltage -36 to -42 wich seems low for a 39k in series with a 50k pot, I'm not sure yet what I had to begin with.

Report back with the voltage on C23. Measure it with Power ON, and Standby OFF then turn the Standby ON. Wait ~ 60 seconds for the bias voltage to stabilize for each reading.

I remember checking the Secondary PT output it looked ok, around 3xx AC something from ground.

You need to know what that 3xx was before, compared to what it is now. You may have a PT failure. We need info from someone who is familiar with this European PT.

[erndawg]

Actually erndawg, if you have -69V on the grid it will pull down the plate voltage more than if you had say -40V on the grid.

Not true. With -69 VDC bias, you are biasing more into the tube's cutoff region - meaning it will draw less current - it will eventually *cutoff*. Consequently, the plate voltage will rise.
With -40 VDC bias, you are biasing more into the tube's conductive/saturation region - meaning it will draw more current - it will eventually *saturate*. Consequently, the plate voltage will drop.

To drive home the significance, remember this: If you had *no* bias voltage, that is, the bias supply died, a bias wire came loose, etc, you will very, very quickly burn out your power tubes because they are way into saturation and drawing a huge amount of current. That, of course, causes the plate voltage to drop.

Conversely, going more negative with your bias - say from 0 VDC to -40 VDC - will *limit* your plate current as you move towards cutoff. That causes the plate voltage to rise.

The negative voltage on the grid counteracts the effect of the positive voltage on the plate.

Again, not true. There is no "counteraction" taking place. Bias is merely a control voltage - nothing more, nothing less.

Think about it, if you took a +100VDC potential and a -100VDC potential they would cancel each other out leaving you with zero potential.

That is not how a tube works! It's got nothing to do with summing at a node.