EL34s and 6L6s working together in perfect harmony

[jmaf]

jmaf, the attenuator schematic is correct and it works about 2 years without issues.

The 25 ohms resistor in parallel, in the attenuator input works like a dummy load. The output transformer is setted to 16 ohms to feed a 16 ohms Vintage 30 loudspeaker.

Perhaps like you said the combination of the mismatched tubes with the attenuator it caused the issue, but why only after 8 hours of use?




MWJB, the amp in fact is not underbiased, but biased with the quiescent current at the standby switch of 100 mA as recommended at the Carvin VL100 schematic.

People says that the ideal is 70% of the plate dissipation which would result in a quiescent current of 185 mA for 4 EL34 tubes.

I thank you both for the answers. Soon at my vacation work I will open the amp to verify the voltages, current, diodes and power supply capacitors. After this I will change the tubes.

I think the eight hours was just a coincidence, it was time for the tube to fail.

Current measurements are only a rough guide without knowing the plate voltage at that current - to know for sure how much you're dissipating, you need to know both. 70% dissipation is the norm for guitar amps, but it is a hot setting nonetheless. I'm not saying that was the cause to your tube failure. I just think it failed because it failed. What is your +B on that amp at the time of failure?

I don't remember what it was about your attenuator that I found back then, but we can go over it again sometime if you want... I just saw your nickname and that you're from Brazil and remembered I had this conversation sometime before with "another" Ramsay. So I thought this must be him.

[jmaf]

JMaf - +/- 10% is a pretty tight margin, 3mA @ 30mA mean? Many vendors won't guarantee matching better than +/- 5mA. There are 6l6 & EL34 that will bias up within these perameters anyhow....even if they don't quite, keep one of each on either side of the primary, match current draw within reason on each side of the primary, take the hottest biased tube as the limiting factor...I can't help feeling I've said all this before..;-)

I bet you have

Yeah, 10% is tight. In 2005/2006 I used to buy tons of tubes and try to match them to 5%(quiescent current under the same conditions, cheap matcher that I built). When 5% failed, I went for 7%. When 7% failed I went for 10%....

Where did I get these references? Off my gut ..... I knew that a 20% mismatch caused loud hum, sounds bad. I knew 10% sounded ok. I went for 5% with the tubes I bought just to be "extra quality". My extra quality standard has always been to double things. When I need 300V I use 600 V caps...so I just transferred that over to tube matching. Not very scientific, I know...

Perhaps that's why my shop went broke too....but folks were happy with their tone hehe.

[MWJB]

"I bet you have"
Sorry Jmaf, that last sentence wasn't specifically aimed at you.

Ha ha, I usually get my tubes from Watford Valves, matched pretty close (+/- 1-2mA) unless you are after something very esoteric, then you obviously need to go with what's available within a limited supply. I did get into the vicious circle of having to buy lots of "matched" sets from another vendor, in a desperate bid to use up unmatched tubes from a few original sets from the same place (8mA was their cut off for matching)....the more pairs I bought, the more odd tubes I was left with ....grrrrr! ;-)

[Ramsay]

Neither of these perameters tells us what each tube is drawing independently of the others. Therefore we do not know the idle current dawn by each tube...the amp is not corrctly biased. 2 tubes could be drawing 5mA the other two 45mA. I wouldn't want to see more than 160mA drawn in total...more likely 130-140mA if all tubes were drawing equal current.

70% of 25 is 17.5W, 17.5W/450v = 39mA...17.5/500v = 35mA...so I don't understand where the figure of 185mA total comes from?

The calculation is (17.5W/450V = 39mA x 4 = 156 mA) + (10,45V/350 ohm = 29,8 mA) = 185,8 mA.
350 ohms is R57 which is a series dropping resistor and 10.45V is the voltage across this resistor.

So If the amp is operating only at 100 mA of quiescent current, I can say in a sense that it is underbiased, currect?

[Ramsay]

I think the eight hours was just a coincidence, it was time for the tube to fail.

Current measurements are only a rough guide without knowing the plate voltage at that current - to know for sure how much you're dissipating, you need to know both. 70% dissipation is the norm for guitar amps, but it is a hot setting nonetheless. I'm not saying that was the cause to your tube failure. I just think it failed because it failed. What is your +B on that amp at the time of failure?

I don't remember what it was about your attenuator that I found back then, but we can go over it again sometime if you want... I just saw your nickname and that you're from Brazil and remembered I had this conversation sometime before with "another" Ramsay. So I thought this must be him.

Yes, I am the same Ramsay of that topic, but I am thinking, perhaps this thing would not have occurred if I had not mixed the EL34s with the 6L6s.
So it's all my fault.

[jmaf]

"I bet you have"
Sorry Jmaf, that last sentence wasn't specifically aimed at you.

Ha ha, I usually get my tubes from Watford Valves, matched pretty close (+/- 1-2mA) unless you are after something very esoteric, then you obviously need to go with what's available within a limited supply. I did get into the vicious circle of having to buy lots of "matched" sets from another vendor, in a desperate bid to use up unmatched tubes from a few original sets from the same place (8mA was their cut off for matching)....the more pairs I bought, the more odd tubes I was left with ....grrrrr! ;-)

Can definitely relate to that....

Didn't know about Watford Valves, will look for them(Edit: Thanks for the tip, BTW! Writing this on a hurry, at work....). Being outside the USA and Europe(and outside of Russia, Czech Republic for that matter) I have limited access to stuff. I've bought tubes over eBay and son, but one time I received 6 burned Sovtek 6L6's on the mail....I tried to contact the seller, they disappeared.... I've also bought some of those Soviet leftovers that were being offered on eBay a while ago, the tubes are visually very sturdy, military types I guess, but I haven't found the time to try them yet. So...more tubes left over and lying around heheh...

[jmaf]

Yes, I am the same Ramsay of that topic, but I am thinking, perhaps this thing would not have occurred if I had not mixed the EL34s with the 6L6s.
So it's all my fault.

If I were to blame just ONE culprit, yes it'd be the mixing of tubes. BUT, like I told you, I don't think it was just one factor. It's hard to guess not looking at the amp, the tubes, etc.

You've been poking around your amps for a while now. Do you have a lamp current limiter on your bench already? If you do, then I'd replace the blown fuse and get it up and running again, but through the series limiter. Watch the 6L6's closely for arcing and so on. If they're redplating then it's a bias thing, but if they're arcing, popping and smelling like burned phenolic resin, the tubes are gone.

[MWJB]

"The calculation is (17.5W/450V = 39mA x 4 = 156 mA) + (10,45V/350 ohm = 29,8 mA) = 185,8 mA.
350 ohms is R57 which is a series dropping resistor and 10.45V is the voltage across this resistor.
So If the amp is operating only at 100 mA of quiescent current, I can say in a sense that it is underbiased, currect?"

Unless you have tight tolerance 1ohm resistors on the power tube cathodes, don't bother with "calculating" your idle current. Measure it. Parts like your dropping resistor might vary in tolerance so you are largely assuming it is 350ohms. You'll drive yourself mad trying to accurately allocate all those mAs to specific parts/stages.

If you know the power tubes are pulling 100mA, you only know that you are pulling 100mA, you don't know which tubes are under or over biased. Underbiased means drawing too much current, overbiased means not drawing enough current. You have 4 tubes which could be doing any/all of those things.

You don't bias an "amp", that's just a turn of phrase, you bias the tubes in the amp...each one specifically. I have seen manufacturer's recommended tubes fitted to an amp that simply failed to work, because the tubes drew zero idle current! So you need to make sure that the 100mA (though on the cool side if measured at 25mA per tube) isn't 2mA + 2mA + 48mA + 48mA.

[Ramsay]

Next week I will put the amp with four EL34 and will bias it with 156 mA of quiescent current.
It means 70% and I will try it with my power attenuator!!
What do you think about??

[soundguruman]

"So, I dimmed the ambient light and played guitar for about 20 minutes, always checking to see if the card was a red valve, which fortunately did not happen."

A technician to follow my own heart. You should just put two sets of bias controls and two bias rectifiers...
You just back it off slowly, till the red (just) goes away.

[Ramsay]

Finally I found the issue.
The outlet AC of my home marks 131V AC.
With this voltage the B+ of the amp marks 507 VDC. (the insulation voltage of the filter caps is 500VDC).

So that redplating 6L6 died cracked the glass and burned the fuse how you can see at link below:http://i223.photobucket.com/albums/d...n/DSC00761.jpg

So I put a voltage stabilizer and the B+ down from 507 to 470VDC.

So I made a few calcs, and readjusted the bias and the B+ is now at 450VDC, in other words, the standard voltage.

By the way, I won't try again put together 6L6 and EL34.
that's enough for me.

[Alan0354]

I biased my 6L6GC in lower bias voltage of slightly less than -40V, I want to run them more in Class A, work just fine for me. So if you put the bias at -40, I think everyone will be happy. You bias EL34 in -50V or more, you really run it in Class B or even step into Class C. Of cause, running 6L6 at -40 will shorten the life as you run more DC current being in Class A.

[Ramsay]

you're wrong.

Changing the voltage bias don't change the amp class to A, AB1, B or D.
Because it depends of the construction and project of the amp in itself.

[jmaf]

you're wrong.

Changing the voltage bias don't change the amp class to A, AB1, B or D.
Because it depends of the construction and project of the amp in itself.

He's not wrong. Bias is what sets the working class except, in your list, class D.

[Steve Conner]

This has been debated to death before. I like to make the distinction between design class and operating class. For instance, you could take an amp designed to operate in Class A, and adjust the bias so that it operated in AB. Results would not be as good as if the amp were designed to run in AB from the start.

Or, you could have a solid-state amp that was meant to be AB, but had such bad parasitic oscillations that it was actually Class-D.