• 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.

Pulsating Glow

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Pulsateing Glow

So, I just finished Guitar amp 3 of 4 today which was a fairly fun build, (Copying from a built amp is so much faster)
Fired it up, Seemed like instant success until I dimmed the lights in the room.

When I cord or puck a string, The orange glow from the tube changes (Gets Brighter). Not to be confused with the filament which stays constant, The tube clearly gets brighter. I also notice a bit of blue glow internally near the base of both output tubes. This also changes with any note or cord.
I had not either noticed this in neither of the other amps or it does not happen.

Its not red plateing but, I have not run it for a prolonged period to be certain.

Now, I have adjustable fixed bias and it has a fair amount of range at the pot, My question is more or less negative voltage?

I just tried to get a voltage reading at pin 3 of the output tube at the socket, The meter was set at 600V DC range and something inside the meter blew ( literally flashed or arced inside).
I have used this meter without issues prior to this incident. Needless to say, it no longer works.
I tried a pair of EL34's but it appears they may even be worse, I thought I even saw the tube arc near the top by this mica. The amp " Sounds Great, but clearly its not correct.

Any ideas?

Most of the voltages on the sketch are old references, I used this bias arrangement.

basic output schematic

On my way out for another meter :whazzat:

Trout
 
sounds like the screen dissipation is going high. Thats the only grid that will typically glow on signal peaks. The blue glow thing is totally normal and not of concern. Inserting a larger screen stopper resistor and/or changing the screen supply may be necessary. It may last a while like it is, but you never know... if its glowing its getting too hot.
 
Darn, no place open here on a sunday night to get a new meter.

I suspect somethings possibly mis-wired or more likely a componet off spec. I guess I will need to wait and get accurate voltages off the key points to see where it is.

My perf-board prototype works perfect and I am fairly certain I followed the design accurately, But, Its hard to re-check wiring under the eyelet board once installed. Removal is virtually disassembly of most of the amp.

I found it Really really odd that the meter blew!
Normally, I test from the ground to pin 3 of the output tube and get normal readings around 400V on the exact settings I used for this amp.
DC 600V scale

On this amp it however took out the meter big time, It litterly flashed inside visible thru the pastic case!! Very Scarey!!
Trout
 
The peak voltage on the plate is the DC voltage (400 volts) PLUS the peak AC signal voltage. The total voltage on signal peaks (under normal conditions) can go to 800 volts. If the amp is mis-loaded (8 ohm load on the 4 ohm tap, or wrong transformer) the voltage can go to several thousand volts. This is bad for output transformers, tubes (arcs) and obviously meters. As stated earlier the pulsating orange glow is likely a screen grid. This can usually be seen through the holes in the plate. The tube will not like this and will eventually fry, often catastrophically.

Check the wiring around the output transformer, tubes, and speaker jack. I have seen 6L6's erupt in fire because I put two 8 ohm speakers in series on the 4 ohm tap. The amps owner really liked the sound it made right before it blew up. We settled on the 8 ohm tap, and 6550's, they would take it.

I have given up on expensive meters, get about 5 of the $5 ones from Harbor Freight. That way you don't get too upset when they blow up. I have only blown one. Ohming out the B+ with the amp on will do it!
 
First off, the bias voltage for KT77 tubes should not be low enough to blow up a meter.

A typical tube of that type has an adjustable supply that goes no lower than -50V, which is a lot of room. This shouln't cause a problem with a meter rated for 600V.

The DC screen voltage may be too high, see the datasheet for ratings at different screen voltages.
 
The peak voltage on the plate is the DC voltage (400 volts) PLUS the peak AC signal voltage. The total voltage on signal peaks (under normal conditions) can go to 800 volts. If the amp is mis-loaded (8 ohm load on the 4 ohm tap, or wrong transformer) the voltage can go to several thousand volts. This is bad for output transformers, tubes (arcs) and obviously meters. As stated earlier the pulsating orange glow is likely a screen grid. This can usually be seen through the holes in the plate. The tube will not like this and will eventually fry, often catastrophically


I am 100% certain on the speaker load and correct Tap selected,
It Has 2,4,& 8 ohm selectable output.

I agee the screen grid is the source of the glow, I have seen this before. This one is no where near as bad as a pair of EL84's I melted in my hifi amp.
But it is bad enough that I need to fix it.


The JJ KT77 's have no plate holes to peek thru :( The glow changes are however visible via the mica spacers.



This transformer is rated for a 50W guitar amp so its fairly beefy and heavy. Basically its a multi-tap bassman transformer.
Specifications
Audio Watts (RMS): 50
Primary Impedance (Ohms): 4,200 CT
Secondary Impedance (Ohms): 2/4/8
Mounting centers: 2" x 3.5" vertical
Weight: 4.7 lbs

I re-checked everything ( Visible ), I have a perfect working unit sitting side by side with this one to compare to.
I'll post side by side pictures, I have re-designed the board, moved a few items, But the basic layout is the same.

Just for piece of mind, I hooked up the original unit, Cranked her up, No Issues. I bought all the matching parts from the same vendor.
The original perf-board version has passed the 3 hour gig test 6 times!
I just dont get it,
But I WILL!!! lol

I have another turret board already populated with all parts, I may replace it entirely.

I have given up on expensive meters, get about 5 of the $5 ones from Harbor Freight. That way you don't get too upset when they blow up. I have only blown one. Ohming out the B+ with the amp on will do it!

This was a $20.00 semi el-cheapo. I am going to grab a few at the local harbor Freight tomorrow, The pretty red ones! lol

When I touched the meter probe to pin 3 the amp produced a rather loud thud at the speaker, This time, I will check all the other voltages before trying that again!!

I am betting that its a bad part, or I missed a connection on the board underside somehow.

Trout
 
Duo said:
First off, the bias voltage for KT77 tubes should not be low enough to blow up a meter.

A typical tube of that type has an adjustable supply that goes no lower than -50V, which is a lot of room. This shouln't cause a problem with a meter rated for 600V.

The DC screen voltage may be too high, see the datasheet for ratings at different screen voltages.


My adjustment range is basically -37 to -48V.
It was not that bias supply that blew the meter, It was the plate voltage measurement at Pin3 of the KT77's I have previously measure my other amp at 405V on pin 3, 600V scale on meter.
Trout
 
Okay, yeah, b+ can achieve some really high voltages especially if you're running a high output level.

Trout's mention of possible oscillation is a wise suggestion in my opinion.

I had an amplifier with russian 6146 radio tubes oscillate and actually cause sparking at the screen grid.

With the addition of suitable screen bypass caps right on the sockets, the problem was gone.
 
poobah said:
Gene,

I didn't see the PSU on your schem. But I think a bad/incorrect (shorted?) inductor could send your B+ through the roof.

Just a thought...

Did the amp sound right?

:)

EDIT: Yeah George.... hold up on the "O" word - :dead:


The Amp sounds great, Totally normal in fact. Very loud, I did not use a choke in the PS, The schematic posted was just more or less the basic output section and bias supply.

Oh please do not use the "O" word lol.
I will know more tomorrow after I get a couple new meters, Right now its more or less guessing.

1 question though, How can I get the plate voltage and not blow another meter?
My notes show my other amp has a bias of -40.1V and plate of 397V. Good ballpark numbers.

Trout
 
The cheapie Harbor Freight meters are good to 600 volts. If you want to test a higher voltage use a voltage divider. I use two 100K resistors in series (yeah there is a small error) with the meter across the bottom one. Double the meter readings. This will protect the meter in case there is "something" causing a high "RF" voltage on the plate. If the readings are grossly wrong, suspect "something". The plate voltage should be slightly less than the supply voltage with no signal input.

It is possible for "something" to occur and the amp to function normally. There are other possibilities, but if the meter was on the voltage function, "something" caused it to draw enough current to fry. The impedance of the meter is high at DC but not at "RF".

On the other hand, maybe the meter's time was up, and it just blew up.
 
OK,

I got a few of those harbor freight multimeters. They seem to work well in fact!..

I ran around the not mounted extra turret board and checked all the resistors with all three meters, Them do indeed seems to match each other very well.

Now,

I'll start at the PS, No output tubes or preamp tubes, Just the rectifier(5U4).

Next I will put in the rest of the tubes, and re-check the voltages at the PS again.

It is possible that older meter was ready to go, But I find it odd that it did it when it did.

Here are a couple pictures,



The Older Working Unit

The New Not Working Unit

Basically, there are only a few layout differences.
#1- I moved the ground buss up onto the board requiring 2 less chassis holes.

#2- I moved the bias pot up onto the board requiring 1 less hole and making it less accessible to non-qualified fingers.

#3 Used a much sturdier board material, The original was 1/16th perfboard, The new unit has 1/8th inch garolite.

#4 Used all 20awg wire for hook-ups rather than the older 22awg I had originally used.

Other than those items, Everything is basically the same.
I cant help but to think I botched the under board wiring.
I ran around it with a meter and of course some readings interact but I do show connections where they are supposed to be
( I Think)

I'll post up the full schematic later, Right now, I want to see if I can actually get so voltage measurements.

Trout
 
Here is the working units Schematic

After a couple of hours of self debate, I decided clearly something's not right on the board, SO,
I am in the process of a transplant using a spare new board.
Interestingly, I have the board removed and all the wires are indeed correct on the old boards underside.

The only thing I am not certain about are the actual components. Possibly a bad cap, Eh, who knows.
Maybe the schematic is fundamentally wrong.

Should be complete in a few hours.
Results soon.
Trout
 
tubelab.com said:
Nothing fundamentally wrong with the schematic. I would add a small resistor in series with the screen grid of each output tube mounted right at the tube socket. I use 100 ohm 1 watt. It helps prevent "something" from happening.

OK, I think I understand this, I have seen this resistor setup commonly on Marshall clone amps, But there was just 1 shared 100/3W like in this draft.

Giving a peek at my layout, A single resistor would drop right in between the board & the socket very easily.

How about moving the CT of the OPT to 1st cap in the PS? I have seen both my original and the red edit in this draft.

Rough Draft

I am on break LOL, I have the new board 1/2 installed and momma made me take her to Dinner.

Trout
 
OK, Transplant complete!!

After a complete inspection of the underside of the eyelet board I feel a little better, It was wired correctly:D

Now, The new board was made with turrets rather than eyelets and much easier to install and replace parts if needed.

The new EL Cheapo Harbor freight meters work great.

I pulled voltages off The PS & the KT77's also added them to
This Schematic

In summary, I have not yet added the screen resistors since I wanted to not make to many changes.

Error #1

I re-checked several items including the speaker, 15.2 ohms on the meter means it must be a 16ohm, not an 8ohm. The opt only has 2-4-8 ohm taps.

My new 100W Jensen Neo speakers will not be here until friday, My old celestions are history due to prolonged overexposure to excess wattage! LOL

I am still getting the glow at the screen grid but, volume has to be set fairly loud to occure, At idle with no instrument plugged in, there of course is no issue.

Voltages seems odd.

Pin #3 432V
Pin #4 434V
Pin #5 35V

Center point on bias pot reads 43V

V3 Voltages seem odd?

Pin #1 216V
Pin #2 0V
Pin #3 1.8V
Pin #6 276V
Pin #7 4.49V
Pin #8 64V

So, Thats where I am this minute, Mild pulsating glow at mid-upper volume settings with mis-match output impeadance.

Trout
 
Voltages on V3 look pretty good... you can't really measure the voltage at the grid of Stage 2. It is a grid leak circuit and the meter loads it down and gives a false reading...

Output stage all looks pretty normal too. Just to be picky... you should express the grid and bias voltages as negative.

:)
 
poobah said:
Voltages on V3 look pretty good... you can't really measure the voltage at the grid of Stage 2. It is a grid leak circuit and the meter loads it down and gives a false reading...

Output stage all looks pretty normal too. Just to be picky... you should express the grid and bias voltages as negative.

:)


OPPPPPPs. I forgot the negative symbol lol

I am a bit puzzled about what would be considered the correct voltage at those bias points, I do not know the correct " Ideal " Setting.

Here is the exact spec sheet for the tubes I am using
KT77 Specs PDF

I know there has to be a formula to determine this at least to a ball park setting.

Now, The pulsating glow from the tube, Could that be as a result of the impedance mis-match? That under normally matched speakers I am already on the edge and this is just pushing it over the top?

Just for peace of mind, perhaps I should add the 100/3w resistor to the screen just as a precaution?

This is basically my first ventures into adjustable fixed bias so please forgive my lack of experience.
I only went this direction to grab the bit of extra power
An externally hosted image should be here but it was not working when we last tested it.


Trout
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.