Bad power tube, filter cap, resistor?

[joulupukki]

Now looking those numbers there is way too high voltage loss on other half of OT primary. Perhaps tube (anode) socket is bad? Or solder? Or tube pin has corrosion?

Right. I would think that except that when I swap the two power tubes to the other sockets the bad measurements follow the red-plating tube (I just checked).

Wouldn't this indicate that the bad tube is causing the issue?

 

[2L man]

Right. I would think that except that when I swap the two power tubes to the other sockets the bad measurements follow the red-plating tube (I just checked).

Wouldn't this indicate that the bad tube is causing the issue?

Possibly? Its plate current is 80mA and plate power dissipation over 23W! Typically that kind of misbalance can happen when coubling capacitor leak but that should have stay in same socket.

 

[joulupukki]

Yeah, I wish I had another pair of tubes to double-check it with. I called around locally and nobody has any so I'll probably just order another pair tomorrow. This evening I'll take the power tubes out and take a few more readings with B+ turned on just to make sure I'm not seeing anything odd. The pins on the tube that's red-plating don't look corroded at all, fwiw.

 

[King Fan]

Well, it's definitely not looking correct.

13.2 V across the 130R cathode resistor (seems normal?)....

I may have goofed, but here's a simple way to calculate cathode bias. It may not work as well for EL84s, and maybe it's not great for poorly matched tubes; dunno. But it's simple.

V4 = 122% MPD?

V5 = 116% MPD?

As I say, I may have goofed, and this method may not be perfect, but I think maybe you, and definitely several other folks, have suggested your bias is too hot. That's what the 18W.com thread was hinting at too I think. I'll await thoughts from others, including @Ten Over , who requested these data.

BTW in these measurements was your redplating tube V4 or V5?

 

[joulupukki]

@King Fan Interesting. I tried that method of calculating but then when I read Rob's "Measure Bias With the Output Transformer Resistance Method" section on his page (which he says is his preferred method of measuring bias) that's the way I've been doing it since. That's how I originally came up with the 100% and 110% plate dissipation values before the V5 output tube went bad after I dimed the amp and played it for a while (cab was in a separate room than me).

BTW in these measurements was your redplating tube V4 or V5?

In these latest measurements, the redplating tube is V5. If I move that tube to V4, then the voltage drop measurements go weird on the V4 circuit. I'd be shocked if the redplating tube isn't damaged/bad.

Basically it sounds like I should:

- Get a new pair of EL84 tubes
- Raise the value of the cathode resistor from 130 to something higher (not sure what value to use yet).

 

[corliss1]

Try the new tubes first. Make sure the amp is stable. THEN move on to worrying about everything else

 

[joulupukki]

Try the new tubes first. Make sure the amp is stable. THEN move on to worrying about everything else

Agreed, though if I'm already having to order tubes (and pay for shipping), I may as well toss in a few values of 5W resistors for the cathode resistor, right? I mean, just in case I want to bias it a little cooler. I was thinking maybe getting a 150, 200, 250, 300 ohm 5W resistor with the order.

 

[dan40]

Try a pair of JJ el84's in there. My Friedman amp runs the el84's at 17 watts of dissipation and JJ's are the only el84's that will take the abuse.

 

[joulupukki]

Try a pair of JJ el84's in there. My Friedman amp runs the el84's at 17 watts of dissipation and JJ's are the only el84's that will take the abuse.

17 watts! JJ EL84 tubes are what came with this kit and I found out from Mojotone today that they were burned-in and matched. So, I dunno what to think. Maybe somehow I damaged the one in some other way. I'll re-order a pair of JJs from a different vendor.

 

[corliss1]

Sure, resistors are cheap

 

[joulupukki]

With the power tubes removed and the amp all powered up these are the readings I get:

V4
Plate: 391 V (pin 7)
Screen: 377.5 V (pin 9)

V5
Plate: 391 V
Screen: 377.5 V

 

[2L man]

Basically it sounds like I should:

- Get a new pair of EL84 tubes
- Raise the value of the cathode resistor from 130 to something higher (not sure what value to use yet).

And install 1ohm resistors to both power tube cathodes! Then there is no need to measure voltage losses over OT primary when biasing. Measuring voltage over combined cathode resistor won't tell much in case tubes are different or amp has other problems.

 

[joulupukki]

And install 1ohm resistors to both power tube cathodes! Then there is no need to measure voltage losses over OT primary when biasing. Measuring voltage over combined cathode resistor won't tell much in case tubes are different or amp has other problems.

Can you point me to an example of this? I’m not entirely sure what this would look like in practice.

Edit: Nevermind. I read more about this here: https://www.aikenamps.com/the-last-word-on-biasing

 

[joulupukki]

If I did put 1 ohm resistors in there, would you use a terminal strip and do it something like this (take out existing yellow wires from the cathodes and move them to the lines drawn in red)?

 

[King Fan]

Good discussion. General questions for my smart friends:

1. Is adding 1Ω 1% resistors *to a cathode bias amp* more accurate than simply using the (actual, measured) cathode resistor ohms? I love 1Ω resistors in fixed bias, but I haven't thought about it here.

2. Is the OT resistance method of measuring bias more accurate than the cathode resistor voltage drop method? In general, or in certain cases? If so, why or when?

 

[peteb]

Good discussion. General questions for my smart friends:

1. Is adding 1Ω 1% resistors *to a cathode bias amp* more accurate than simply using the (actual, measured) cathode resistor ohms? I love 1Ω resistors in fixed bias, but I haven't thought about it here.

2. Is the OT resistance method of measuring bias more accurate than the cathode resistor voltage drop method? In general, or in certain cases? If so, why or when?

the cathode resistor is fine, but like 2L man pointed out, this amp has one combined cathode resistor, making it impossible to know current draw of each tube.

what is most accurate, in my opinion, is to measure cathode current and measure the plate current and compare.

if the plate current is 90-95% of the cathode current, it’s good.

also, measuring the screen current is an additional check.

it is good to use multiple sources.

 

[Ten Over]

4.02 V looks problematic.

0.49 / 100 is 5 mA

4.02 / 100 is 40 mA

no bueno

He is using 470R screen resistors.
V4 has 1.1mA of screen current.
V5 has 8.6mA of screen current.
Screen current is around 11% of plate current.
Both screen currents are consistent with the plate currents.

 

[Ten Over]

If I did put 1 ohm resistors in there, would you use a terminal strip and do it something like this (take out existing yellow wires from the cathodes and move them to the lines drawn in red)?

View attachment 999399

Install a 1 ohm 1% resistor between Pin 3 and Pin 8 on V4 and another one on V5. Connect Pin 8 on V4 to Pin 8 on V5. Connect Pin 8 on V5 to the 130 ohm cathode resistor.

 

[2L man]

Good discussion. General questions for my smart friends:

1. Is adding 1Ω 1% resistors *to a cathode bias amp* more accurate than simply using the (actual, measured) cathode resistor ohms? I love 1Ω resistors in fixed bias, but I haven't thought about it here.

2. Is the OT resistance method of measuring bias more accurate than the cathode resistor voltage drop method? In general, or in certain cases? If so, why or when?

1. Measuring voltages over 1% 1ohm resistors produce very repeatibility reading using current multimeters but multimeter accuracy has significant effect because we hobbyist do not calibrate them. On Pentodes the Cathode current consists Screen and Plate current as well which is fine when Bias is idle current and Screen current is only few % of Cathode current.

2. OT Primary half resistances also produce very accurate plate current voltage readings, when measured from B+1 to Plates. If B+1 is first measured against ground and then Plates against ground and deducted result is poor because there comes two voltage readings which have also less accurate decimals!

How long the OT has warmed has some effect to copper coil resistance and perhaps primary halfs warm up different rate. So resistances should be measured after certain use and then Bias voltages after about similar use.

Measuring voltages from B+1 to Plates and switching Mains Off the OT Primary resistances can be measured after few seconds when tubes have quit conducting. Naturally with caution because there might be some B+1 voltage in Electrolyts but there is no voltage over primary coils anymore so resistance measuring is safe for Multimeter.

I have read that inserting probe to power tube anode can make amp oscillate but it has never happen to me.

 

[joulupukki]

Install a 1 ohm 1% resistor between Pin 3 and Pin 8 on V4 and another one on V5. Connect Pin 8 on V4 to Pin 8 on V5. Connect Pin 8 on V5 to the 130 ohm cathode resistor.

Ah, that makes more sense. Essentially using the unused pin 8 of the EL84. I read elsewhere that some people use a 10 ohm resistor instead of a 1 ohm resistor, you just have to divide the reading by 10. I placed my order of extra stuff last night and didn't get any 1 ohm resistors added in time. I do have a couple of 1/4 watt 10 ohm resistors but ... hmm.