Stereo valve amp not working.

[maninashed]

Don, please forget that -34v bias voltage, its irrelevant. You need to measure the CURRENT through the valves, by measuring the voltage ACROSS the cathode resistor we can work out the current then we can tell how the valve is operating.
I agree with the comment made by Magnetic there does seem to be some heating of the PCB but its hard to tell how bad this is from the photos we have. It needs close inspection. Never run a valve amp without speakers connected, even with no input. The reason for this is because you will risk damage to the output transformer/valves. If the amp becomes unstable and oscillates at high frequency you will be unaware of the problem until it's too late.
Bill

[Don_Zalmrol]

Well I looked trough the manual for the bias configuration for my amplifier, but they don't mention it anywhere (see the dropbox link).

I've also looked up the measuring range of my fluke tester it's an T5-600 and it's sensitivity is in 1V steps…

So this is probably why I only see 0V on P8.

I will check tomorrow the voltages over and the current trough the 10R resistors, I've found them on the PCB printout.
How do I measure the current trough the valves? By just connecting my meter between the 10R resistor and the circuit?

Manual:
P1 http://dl.dropbox.com/u/10146488/EL3...e/Part%201.pdf
P2 http://dl.dropbox.com/u/10146488/EL3...e/Part%202.pdf
P3 http://dl.dropbox.com/u/10146488/EL3...e/Part%203.pdf

Also the PCB looks a bit burned around the E34L sockets, but it's a thick PCB plate which can withstand quite some heat.
But I think I will re-solder them for more strength and resilience against the heat, just to be sure.

[Station X]

As you say your meter is not suitable for measuring the voltage across the 10 ohm resistors and that's probably why you're measuring 0V.

You meter should be suitable for checking the 10 ohm resistors. These are clearly visible on the second page of your first pdf document. Set your meter to the ohms range and connect one lead to each side of the resistor. With any luck you'll get a reading close to 10 ohms. If not the resistors will have been deliberately shorted out and you'l need to trace the PCB tracks to find out where.

It is normal to adjust the bias voltage so as to give a specified anode current. As vallves vary in their characteristics it's likely that the bias voltage will be different for each valve. I suggest you obtain a better meter and then measure the voltage on pin 8 of all the output valves. Each 0.01V you measure equates to 1mA of anode current.

[Diabolical Artificer]

Hi Don,by measuring the voltage across the cathode resistors you can then calculate the current using Ohms Law,look it up on the net. It is a simple equation that allows you to "see" how voltage- (V) current- ( I ) and resistance/load- (R) all interact/effect each other. So if for example you measure 100v across a 10ohm resistor,you DIVIDE the VOLTAGE by the value of the RESISTOR; which in this example is 100 (V) divided by 10 ohms (R) which equals 10. That means you have 10 Amps of current (I) flowing through the resistor (R). Its the 100volts (v) thats shoveing all those electrons (I) round the circuit/load (R)

This is putting it simply,and sorry if you all ready know this. If you want to truly understand what is happening in your amplifier your going to have to use your brain. Dont worry though,we,re all not geniusus and I for one struggle with the maths. ( I,m not sure my example is correct for instance,I had to do the equation several times to make sure I didnt embarass myself!) Good luck,Andy

[Don_Zalmrol]

I've bought a new meter for hobby usage which is precise to 3 decimals after 0.
The Fluke meter is a very good meter, but is one for fieldwork and semi-industrial use, so it's not precise…

Tomorrow I will measure the amplifier and repost the new values of the pins and resistors voltage & current trough the valves.

@glowinganode

I took this picture in a big hurry, so it is blurry. I had to shut the amp off before damage could occur. This situation is a tube arc waiting to happen. Plate current was over 300mA and climbing. This would have killed a wimpy power supply and can fry an output transformer. A loud hissing and crackling sound emanated from the speaker just as I hit the master kill switch. If you see a purple glow like this inside (between the elements) a vacuum tube it means that the tube has an abnormal gas inside (usually an air leak). TURN THE AMP OFF IMMEDIATELY! A glow like this is normal in a gas tube like voltage regulator tubes (0A2, VR150 etc). A BLUE glow that appears ON THE GLASS of a vacuum tube is a normal condition.

All of my valves have a light purple gaze inside the filaments. But valve #2 is emitting the most purple gaze, now with the temporarily bias adjustments this seems to be lesser than before.

@magnetic

Thx, I still now the law of Ohm :p
But it never hurts to refresh it. And i'm also not a mathwiz.

I will post the results of the measurements tomorrow with the calculations of the current trough the 10R resistors.

[Don_Zalmrol]

Merry x-mas to all of you.
And again thanking you all for helping me so well!

I'm measuring the voltages on all the valves, because the meter is so sensitive I've included a small range with the lowest & highest measurement of 1mV

#1 valve:
P8: 230~231 mVDC

#2 valve:
P8: 252~253 mVDC

#3 valve:
P8: 221~222 mVDC

#4 valve:
P8: 263~264 mVDC

I've measured the Ohms over the resistors and the meter show them around 15 Ohms
The voltages above the cathode resistors (10R/1W):

#1 resistor: 236~237 mVDC
#2 resistor: 256~257 mVDC
#3 resistor: 219~220 mVDC
#4 resistor: 282~283 mVDC

So the current trough the resistors on the lowest measured voltage is:
#1 resistor 1: 0.0236 A or 23,6 mA
#2 resistor 2: 0.0256 A or 25,6 mA
#3 resistor 3: 0.0219 A or 21,9 mA
#4 resistor 4: 0.0282 A or 28,2 mA

And the current trough the resistors on the highest measured voltage is:
#1 resistor 1: 0.0237 A or 23,7 mA
#2 resistor 2: 0.0257 A or 25,7 mA
#3 resistor 3: 0.0220 A or 22,0 mA
#4 resistor 4: 0.0283 A or 28,3 mA

I've also looked a bit closer at the PCB and the "burn patterns" that you see is from my soldering irons tin I've used for soldering the valve sockets.

Do I need to re-solder the sockets or not?

And the PCB is around 5mm thick, so this one I think could withstand stand temperatures from up to 225°C

Do you need other measurements from me?

[Station X]

I'm puzzled as to why the voltage readings taken (presumably) between pin 8 and chassis differ from those taken (presumably) by measuring directly across the resistors. These are effectively the same measurement and the extra resistance introduced by the PCB tracks should be minimal. If the resistors are 15 ohms rather than 10 ohms, the combined anode and screen currents will be lower that those you calculated. For example 236mV across a 15 ohm resistor gives a current of 15.73mA.

Can you clarify how you made the measurements please.

[Don_Zalmrol]

Sorry, I calculated them on 10 Ohms instead of 15 Ohms, my bad

So the current trough the resistors on the lowest measured voltage is:
#1 resistor 1: 0.01573 A or 15,73 mA
#2 resistor 2: 0.01707 A or 17,07 mA
#3 resistor 3: 0.01460 A or 14,60 mA
#4 resistor 4: 0.01880 A or 18,80 mA

The P8 measurements where done by putting the measuring pin on pin 8 and the common on the GND of the amplifier.

The resistor measurements where done by putting the measuring pin on one side of the resistor and the common on the other side of the resistor (in parallel, so to speak).

[Station X]

OK.

Amplifier turned off. Meter set to ohms range Negative black lead of meter to chassis.

What is the resistance measured to the following points. Probe them with the red lead.

1. End of 10 ohm resistor connected to chassis.
2. End of resistor connected to pin 8.
3. Pin 8.

What is the colour code on the 10 ohm?? resistors?

[Don_Zalmrol]

I'm going to re-test them with the new (more) precise meter because I'm starting to doubt the accuracy of the fluke T5-600 meter…

1) 10,8 Ohms
2) 10,2 Ohms
3) 10,4 Ohms (I've measured between P8 and the chassis)

The colour code is Brown Black Black Gold Red (10 Ohms, 2% tolerance, 50ppm/K)

So my previous calculations on 10 Ohms where "correct" with just a deviation of +- 0,2 Ohms)

The voltages above the cathode resistors (10R/1W):
#1 resistor: 236~237 mVDC
#2 resistor: 256~257 mVDC
#3 resistor: 219~220 mVDC
#4 resistor: 282~283 mVDC

So the current trough the resistors on the lowest measured voltage is:
#1 resistor 1: 0.0236 A or 23,6 mA
#2 resistor 2: 0.0256 A or 25,6 mA
#3 resistor 3: 0.0219 A or 21,9 mA
#4 resistor 4: 0.0282 A or 28,2 mA

And the current trough the resistors on the highest measured voltage is:
#1 resistor 1: 0.0237 A or 23,7 mA
#2 resistor 2: 0.0257 A or 25,7 mA
#3 resistor 3: 0.0220 A or 22,0 mA
#4 resistor 4: 0.0283 A or 28,3 mA

[Don_Zalmrol]

Quiete a bit accuracy error on the fluke T5…

It shows again 15 Ohms over the 10R/1W.
And with the new tester (a Testo 382 meter) only 10,2 Ohms over the 10R/1W resistor,
which is in the acceptable range of this resistor (10 Ohms, 2% tolerance, 50ppm/K).

[Station X]

OK. We've established by colour code and measurement that the cathode resistors are 10 ohms. If your meter is accurate on the mV range, then the combined anode and screen current of each valve is about 25mA.

25mA seems rather low to me and I wouldn't expect it to to cause a valve's anode to glow red. We'd better wait for the amplifier experts to come on line and see what they have to say.

[Herald1360]

As a matter of interest, what does the new meter show on ohms when the probes are connected together (shorted)? 0.2 ohms sounds likely to be the resistance of the test leads, in which case the resistors are even closer to 10 ohms.

[glowinganode]

First and foremost, happy Christmas one and all. I hope you're all as full as I am!
Graham, yes you're right those standing currents do seem very low, it's a good sign they're all similar values though.
Whilst they are now a lot more accurate, they may be out of context if aprogressive or intermittant fault has not had chance to develop.
Being a stereo amp, we have all made the assumption it is an audio hifi class AB amplifier.
The low standing current and fixed bias arrangement has me wondering whether the circuit is actually based on (two) class B guitar amplifiers.
Why add the complication of fixed bias, unless maximum power output (eg. guitar amp / P.A. equipment) is the ultimate goal?
Fixed bias also has the disadvantage that thermal runaway of a gassy, high slope valve can occur, as it doesn't have the negative feedback effect of a cathode resistor (autobias).
Radio hams that run linear amplifiers will most likely be aware of the effects of this, severe damage to power supply components and transformers.
I have witnessed this personally, and thankfully I got away with just a couple of rectifier diodes, it happened so quickly.
Since then, I tend to err on the safe side, and replace (or at least test) any valves I suspect are gassy.
However with only 25mA standing current, and quite modest HT of just over 400V, the combined dissipation is around 10 Watts, around 40% of that permissible.
It could be though, that the fault occurs after a good thrashing for a couple of hours, Don you need to answer my earlier question of when or how quickly this fault occurs.
Also you say you run head phones from this amp? Are these attenuated / high impedance? What is the relative postion of the volume control when you run headphones? Have you matched the output impedance appropriately?
I'm glad you've got a better meter to use, you seem to be really getting into this! It's as much about observation as measurements.
It would be handy to leave a meter permanently connected to the HT supply, to monitor the total HT current. Run the amp for a couple fo hours, whilst observing the current trend, note the standing current during the silence between tracks.
Good tidings,
Rob.

[jjl]

Yes, those standing currents, as Graham and Rob stated are much more in line with class B biasing than class AB1.

I wonder whether, running the amp with no load has caused sufficient high voltage spikes to produce breakdown and perhaps resultant tracking in the valves, valve bases, PCB or output transformers. I hope not, but this would be much more likely to happen with the widely varying conduction of the output valves under class B conditions. If such breakdown has occurred in a class B amp it may not be evident when the amp has no input signal.

Some very detailed notes on this sort of behaviour in guitar amps here

http://sound.westhost.com/valves/analysis.html#s6

John

[Don_Zalmrol]

Originally Posted by Herald1360

As a matter of interest, what does the new meter show on ohms when the probes are connected together (shorted)? 0.2 ohms sounds likely to be the resistance of the test leads, in which case the resistors are even closer to 10 ohms.

I starting to think the same, when I put them together I get many results switching, but 0,2 Ohms pop out most of the other results.

Originally Posted by glowinganode

However with only 25mA standing current, and quite modest HT of just over 400V, the combined dissipation is around 10 Watts, around 40% of that permissible.
It could be though, that the fault occurs after a good thrashing for a couple of hours, Don you need to answer my earlier question of when or how quickly this fault occurs.
Also you say you run head phones from this amp? Are these attenuated / high impedance? What is the relative postion of the volume control when you run headphones? Have you matched the output impedance appropriately?
I'm glad you've got a better meter to use, you seem to be really getting into this! It's as much about observation as measurements.
It would be handy to leave a meter permanently connected to the HT supply, to monitor the total HT current. Run the amp for a couple fo hours, whilst observing the current trend, note the standing current during the silence between tracks.
Good tidings,
Rob.

Sorry, the fault occurs when I place valve #2 in valve #2 after maybe a minute (so practically immediately).

When valve #2 is in valve socket #1 or an other. The overheating doesn't occur, perhaps it will like you say it will happen after some hours.

When I put another valve in valve socket #2 I notice the anode is glowing brighter, this might be a result from the bias configuration…

So I'm starting to think this valve is malfunctioning and is gassy. I will try to set my amp up for testing some time, but I need to buy some measuring clamps for ease…

Thx, I really want to learn and fix my amp by myself so I know what to do or build another one in the future.

Originally Posted by jjl

Yes, those standing currents, as Graham and Rob stated are much more in line with class B biasing than class AB1.

I wonder whether, running the amp with no load has caused sufficient high voltage spikes to produce breakdown and perhaps resultant tracking in the valves, valve bases, PCB or output transformers. I hope not, but this would be much more likely to happen with the widely varying conduction of the output valves under class B conditions. If such breakdown has occurred in a class B amp it may not be evident when the amp has no input signal.

Some very detailed notes on this sort of behaviour in guitar amps here

http://sound.westhost.com/valves/analysis.html#s6

John

My amp is a HiFi stereo amplifier for home usage and not guitar or PA usage.
Although it might be based on the same design and uses AB specs.

I will read your link for more information.



I'm also starting to think my valves just might be "used up".
I heard something interesting today from my mother, that she accidentally powered my amp on when cleaning my room and noticed this later on when she came back to fill my closet and powered it off…

So my amp was powered on with no active input signal. The next morning the "overheating" of valve #2 happens.



The headset I'm using with this set up is a Sennheiser HD558 the impedance of this headset is 50 Ohms. And the "ballast" resistors inside the amplifier are 2x 100 kOhms 5% tolerance (Brown Black Yellow Gold) resistors. The volume is at maximum and I control the volume locally with a small inline volume-control switch from Sennheiser.

Note this output is not original on my amplifier, I later added this thanks to Piet from Jukebox-Revival.
He first advised 470 Ohms resistors for my headset (this was a Synq headset with 32 Ohms).

Later on I replaced the resistors from 470 Ohms to 100 kOhms, because the output volume was to high even when the volume was low.
Note I use my computer as media center and the volume is transmitted trough HDMI to my TV, from my TV to the amplifier by shielded analog cables.

http://www.sennheiser.com/sennheiser...dphones_504631

[jjl]

Don

The measurements you have posted so far are quite puzzling given the obviously distressed valve you are seeing. The cathode currents are very low.

The only scenarios I can think of that make sense are :-

1. The amp is biased for class B conditions.

2. The amp is biased for class AB1 conditions (most normal for hi-fi) and all of the E34L valves have low emission i.e. they are worn out; with healthy valves I would expect to see at least 60mA cathode current for each.

3. You have somehow made an error in your measurements.

Scenario 2 more or less rules itself out; low emission valves will dissipate less power i.e. will run cooler.

How do you connect your headphones to the amp?

No input signal is not a problem, no output load when there is any input signal is a potential problem.

John

[Station X]

The headset I'm using with this set up is a Sennheiser HD558 the impedance of this headset is 50 Ohms. And the "ballast" resistors inside the amplifier are 2x 100 kOhms 5% tolerance (Brown Black Yellow Gold) resistors. The volume is at maximum and I control the volume locally with a small inline volume-control switch from Sennheiser.

Note this output is not original on my amplifier, I later added this thanks to Piet from Jukebox-Revival.
He first advised 470 Ohms resistors for my headset (this was a Synq headset with 32 Ohms).

Later on I replaced the resistors from 470 Ohms to 100 kOhms, because the output volume was to high even when the volume was low.

When I use my amplifier, I always listen with it to the television or my computer. And I play the sound trough a headset which is connected to the amplifier with a jack-input with decoupling (or how you say it in English) function to turn the speakers off and activate the headset.

Where are these ballast resistors connected? I don't see how they can soak up the surplus power without a serious mismatch.

I would suggest that you connect some speakers and then adjust the bias pots so that all the valves draw 25mA. Make a note of the bias voltage on pin 5 of each valve at this point. Then observe the valves for overheating. If a valve starts to overheat measure the current again and see if it's increased significantly. If it has then recheck the bias voltage to see if that's changed.

[murphyv310]

Originally Posted by glowinganode

Hi Don, see here re. purple glow.
Also here, scroll down to "A very gassy 211"
Rob.

Hi Rob.
Indeed on the links you have posted the large purple glow is caused by a leak of gas. This was seen quite often in line and frame output valves in the heyday of Valve TV's. It was often seen in CRT's that were nearly or completely down to air.
The OP's valve though as far as I can see is depicting the blue glow that can be normal with some makes of valve or operating conditions.
I would really like to see the output on a scope from the amp and also do some sniffing with the scope probe too.

I totally agree with Graham, the amps output is severely mismatched and with the load the amp sees is very likely to flash over damaging bases, valves and the output transformers.

[glowinganode]

Hi Trevor,
I agree with you that this glow can be normal for some makes of valve or operating conditions. Indeed I recall seeing an EL38 in a high voltage stabilised power supply that I built many years ago and it worked perfectly fine. If all four were identical, I wouldn't be concerned.
But one of the valves is obviously in distress, and I think it's important Don knows what could happen if not sorted.
Unfortunately we can't look at the output on a scope, much as I'd love to too.
All we can do is offer educated suggestions, and a reasoned scientific approach (without a circuit diagram!)
I think we're slowly getting to the root of the problem, and I suspect one limb of one of the output transformers has a short circuit turn(s), caused by the virtual open circuit load of the headphones.
As a result, the associate valve is working into a much lower impedance causing excessive anode dissipation.
Don, could we have some current readings for the four valves whilst playing music or preferably a test tone, whilst connected to speakers.
I would recommend you do not use this amplifier to run your headphones, as you risk causing expensive damage with such a mis-matched load.
Rob.