• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

Resistor used for measuring bias keeps failing -- what can I do?

I have an old Anthem Amp 1, which is a pretty standard push-pull EL34 power amp, suing 4 EL34 tubes. It works great, but over the 25 years that I've had it, there's one part that fails every ~5 years or so.

In my amp, bias is set by measuring the voltage drop across a 10 ohm 1W resistor connected in the cathode circuit of each output tube. It's the one with a blue circle around it in the upper right corner of the attached schematic. There's one for each tube. I use the ground terminal of the speaker output and a test point on the circuit board to measure this. There's a potentiometer that is used to adjust the voltage drop. My target is a 0.4 V drop across this resistor.

Anthem Amp 1 schematic annotated.png


What happens is that every 5 years or so, this resistor fails. It starts by noise showing up in the left channel, and then the resistor eventually goes. It's always the left channel, and It just happened again. I've had the amp serviced over the years, but this is the only issue I've had outside of the normal replacement of tubes every so often. I'm not sure why this keeps happening. Here's a picture of the actual resistor.

Bad resistor circuit board.jpeg


At this point I'm comfortable enough with the situation that I can replace the resistor myself. What I'd like to know is whether there is anything I can do to make this more robust. My thought was to use a 2W or higher rated resistor instead of a 1W, thinking that a resistor rated for higher current would be more robust. I don't know if this is a good idea or not.

Can I use a 10 ohm 2W resistor instead of a 10 ohm 1W resistor in this situation? Or should I use the same value resistor? Any input would be appreciated. Thanks!

Also, is this an issue with this sort of bias adjustment method, and people just live with it? That's the other possibility I thought of.
 
AX tech editor
Joined 2002
Paid Member
To me this resistor looks too small for a 1W dissipation.
However, at 0.4V across the dissipation is only a fraction anyway.
Is it possible that over the years the cathode current drifts, maybe because of the trimpot drifting or developing an erratic wiper contact, which causes the cathode current to greatly increase thus burning the resistor in a (succesful!) attempt to save itself?

Jan
 
To me this resistor looks too small for a 1W dissipation.

Well, I trust the tech that has worked on my amp before. Maybe I shouldn't? :D Looking at pictures on Mouser and similar websites shows 10 ohm 1W resistors that look the same.

However, at 0.4V across the dissipation is only a fraction anyway.
Is it possible that over the years the cathode current drifts, maybe because of the trimpot drifting or developing an erratic wiper contact, which causes the cathode current to greatly increase thus burning the resistor in a (succesful!) attempt to save itself?
Maybe, but I check the bias every year, and it's pretty stable. The most it's moved is from 0.4V down to 0.35V, which should result in less current, right? I would expect that if something like this was going to happen, the bias would go the other way.

Your thought about this being sort of like a fail-safe in case the cathode current increases is an interesting thought. Sort of how you don't put a 20A fuse in a 15A circuit. Would you stick with a 1W resistor, then?
 
Member
Joined 2004
Paid Member
Those bias measuring resistors, under normal operation, dissipates only 16 milliwatts. A 1000 milliwatt (1 watt) resistor should be sufficient. You are also right that as power tubes age they conduct less current under the same bias conditions.

If the failure is with R53 only (and not also R54, etc) and it had happened repeatedly, then you should look for a problem elsewhere. I would definitely not use a higher wattage resistor because the ”real problem” may be prevented by R53 when acting like a fuse.
 
  • Like
Reactions: 3 users
Here are my thoughts. I have never had a bias resistor get that warm, but have had constant current source transistor resistors and such get pretty warm doing their job.

First, what is the current through the resistor? Disconnect one end and measure the current based on the full range of the bias setting and then the actual biasing specification. When operating at each end it is quick to not red plate the output tubes, you are looking for a possible operating range of current through the resistor and check the current draw at dead cold start up. In the Dynaco ST 70 the bias circuit has to catch up to the B+. The ST 70 uses stock 15.6 ohm resistors. Second , the way to check for current is to measure the voltage drop across the resistor doing the same test and calculate the current. P=I x E will give you the power in watts. Third, do all the bias resistors fail in the same manner, overheating? What is the burned resistors ohms? Meaning does the resistor stay in spec as it is discolored? Please check the duty rating for 25 deg C on the that type of resistor for 1 watt and then 1.5 watts and then 2 watts. There is a possibility that the resistor is heating up near the top end of its continuous duty range. All resistors I have encountered are a minimum of 2 times the specified watts and i one case 3 times. i.e. the circuit the pulls 1 watt would get a 2 or 3 watt resistor. Also, I prefer the ceramic wire wound resistors or something similar in a larger package that can dissipate heat better. Lastly, a disclaimer here I am not a pro electronics repair person, just a hobbyist. I just know that after rebuilding a Dynaco ST 70 was quite a lesson in such things. I found on the PAS 3 units the wattages specified were not the wattages originally installed at the factory and a host of other peculiar things. I hope this can help in some way. Please let us know watt you find.

One last thing check the screen voltages to make sure that the cathode and screen are not pulling excess current. And any leaking .1mfd capacitor c11,12,13 and 14 the is even slightly leaky will cause excessive current to flow.
 
Last edited:
If the problem only happens on one channel, and looks rather violent based on the condition of the resistor, it is probably some kind of flashover/breakdown event. The culprit could be an insulation failure, cap failure, accumulation of static charges at an unexpected place, etc.
Keep the same type of resistor until you have located the real origin of the problem, otherwise the damages could be more extensive
 
  • Like
Reactions: 1 users
Intermittent bias pot wiper (as mentioned in the post above)?

C12 voltage rating?
Versus . . .
Maximum B+ voltage at Power-up (before the output tubes warm up and load the B+, and before the 12AU7 warms up, so all of B+ goes across C12, and the Bias voltage is at the other end of C12).

Example to illustrate a possibility:
+450V B+ at power up;
-25V bias
475V across C12
C12 rating of 400V.
Some caps are self healing, but before they do, Wham!
 
  • Like
Reactions: 1 user
To me this resistor looks too small for a 1W dissipation.
However, at 0.4V across the dissipation is only a fraction anyway.
Is it possible that over the years the cathode current drifts, maybe because of the trimpot drifting or developing an erratic wiper contact, which causes the cathode current to greatly increase thus burning the resistor in a (succesful!) attempt to save itself?

Jan
Agree. It looks pretty puny to me for that power, although I have seen some small resistors with seemingly high wattage ratings that claim to be able to operate at high temperatures which would raise the wattage rating. I think Dale made them. Those in your amp don't look like that, though. Go with the higher wattage rating after you verify that the tube with the burned resistor is not drawing excessive current.
 
Are you measuring that 0.4V over the resistor? With your meter connected to each side of the resistor?

I agree with what has been written above:
  • 0.4V over a 10 Ohm resistor equals 0.016W (P = voltage*voltage/resistance)
  • this looks like an overvoltage, not an overcurrent/overpower. In case of too much power, your resistor would have changed color all over, not just a hole in it.
 
As was said, a blast hole on an otherwise clean looking resistor is the sign of a quick overload. Either voltage or current. A charred all over is the sign of long term overload and resultant heat. it might be a faulty wiper in the bias pot that intermittently opens or a bad coupling cap that also intermittently shorts putting the bias way too high. I would just change to pot, cap and resistor and be done with it. It’s a lot cheaper than a meltdown!
 
Agree. It looks pretty puny to me for that power, although I have seen some small resistors with seemingly high wattage ratings that claim to be able to operate at high temperatures which would raise the wattage rating. I think Dale made them. Those in your amp don't look like that, though. Go with the higher wattage rating after you verify that the tube with the burned resistor is not drawing excessive current.
Looking at illustrations on the Mouser website, there seem to be plenty of 10 ohm 1 W resistors that are close to the size of the ones in my amp. This one looks nearly identical. YAGEO 10 ohm 1 W resistor on Mouser.com
 
An arcing tube could cause this problem but it would most likely not be a five year interval between resistor failures! You might also see an associated flash from the tube. Look at the coupling cap for the cause. Many types are so called self healing and burn out the bad spot and work fine until the next bad spot. That bad spot is a short that will send voltage straight through to the output tube. That will disrupt the bias and then bang the resistor pops! Measure the voltage on both sides of the cap to ground. Compare it to the same reading on the other coupling caps. If it has healed it will test normal. Change it for safety sake.
 
I've had the same problem on self built amps and ended up using Welwyn wirewound resistors in the end. Some resistors don't fail short or open but go high resistance, not what you want at all. A WW resistor fails short or open.

Is there a possibility of high frequency oscillation? That's also something to bear in mind as a possible cause. Lastly a few reversed biased diodes from anode to ground might be beneficial to prevent HV peaks.

Andy.
 
This cathode resistor blowing is an old problem which I debugged on someone elses prehistoric Velleman exhibiting the same issues. In short, film resistors cannot sustain power transient impulses, i.e start pushing fast rise-signals through the power stage and these small film resistors will eventually blow. Cure:- as above DA post and others mention; use wirewound types.
Thinking about this; on output stage analysis, the power stage in any tube amp has quite a job to do when it comes to dealing with the inductive and capacitive leakage parasitics of the output tranny. The output tube(s) have to faithfully discharge transients that with a poor design could even be resonant, which requires peak currents that is reflected on to the cathode resistor, if fitted. That is a reason using cathode bias, the bypass capacitor must be a good quality type with a low ESR.

rJ
 
benchbaron,
(and others)

The amplifier Uses Fixed Adjustable Bias.
It does Not use Self Bias . . . so there is no bypass capacitor for the output tube cathodes.

A brief medium current transient does Not have enough Integrated heating power to burn out a 10 Ohm 1/2 Watt resistor.
But a very large current transient may have enough Integrated heating power to burn out the resistor.

EL34 data sheet plate curves show plate currents of 280mA (plate at 250V, screen at 250V, control grid at 0V).
The EL34 maximum long time integrated current is 150mA.
Think of conditions that might put out 280mA briefly.

If the bias pot wiper opens up temporarily, the control grid will be at 0V. Ouch!

Your EL34 screens are at more than 400V. That is a lot more than 250V, and the plate curves are off the top of the chart (curves are created with 250V on the screen).

It looks to me as if your EL34 tubes are Sideways (your picture).
A hot screen might "fall" into the control grid. Ouch!

Find the cause of the problem.
Otherwise, using a higher wattage 10 Ohm resistor will not burn out . . .
But instead the output transformer, power transformer, or output tube might burn out.