Too much plate voltage - Ampeg B15 Portaflex

The schematic calls for around 450v on the plates of the 6L6's in this guy, I measuring close to 500v on the plates under load. The amp is basically eating expensive NOS tubes for dinner.

Ampeg calls for a 5U4/GZ34 rectifier which I'm sure contributes to the problem. I've not been able to find any information on suitable subs for the GZ34 which will allow lower voltages that will seem to work without damaging the amp.

I'm thinking of putting a 10W resistor after the rectifier to knock down the voltages to a reasonable level or replacing the GZ34 with something else.

Anyone have a suggestion for substitutions in this circuit?

Schematic is here.

http://www.kbapps.com/audio/schematics/tubeamps/ampeg/images/b15n.gif

Mark

Have you checked how much voltage there is at the AC wall outlet?
The schematic calls for 117V... Should you have say, 130, that could explain your problem.
You'd probably check if you have more than 6.3V on the heaters, which would also need to be addressed...

A resistor (or a choke) after the rectifier would be a good solution, but it'd better handle a bit more than 10 watts...

Axel

I've got about 117V coming out of the wall right now.

That gives me heaters @ 6.3V and a plate voltage just under 500v.

If I drop the wall voltage down to 107V, that gives me the proper plate voltage of 450 but
my heaters drop down to around 5.4V.

Not good.

There's a Mullard GZ34 in there, which I love the sound of. I wonder if other types of GZ34 would get me closer.

Mark

could always drop the tube rectifier in fave of a bridge diode .

117V is right, and 6.3V confirms that the problem is not there.

Another tube? GZ32 comes to mind, unfortunately it needs a bit more filament current, and we are not sure that the original PT can deliver...

Anyway, the 5AR4 has a voltage drop of 17V at 225mA (Pmax). You would need 40V more to achieve your goal.
Even if you could find such a tube, I'm afraid it wouldn't be able to deliver enough current.

@Pucho:
A diode bridge wouldn't have the 17V voltage drop, therefore it would imply a power resistor...

Axel

pucho812 said:

could always drop the tube rectifier in fave of a bridge diode .

Click to expand...

The GZ34 probably drops more voltage than the 0.6v a diode would drop.  Therefore you might stand a good chance of the HT being greater than it already is.

40V worth of zener diodes between the PT center tap and ground. Ask Ohm what the wattage of said diodes should be.

Changing rectifiers won't lose that amount of voltage.

You could use a choke - it should drop the supply about 50V.  That is true when it is used in a choke input configuration meaning it comes first - before the cap.  When used after the first big cap I'm not sure if you get the same amount of supply drop.

http://www.aikenamps.com/Chokes.html

Cheaper option is buy some hi wattage cement resistors and lose the excess that way.  Several lower values require a lower wattage per resistor than a single big one of course.

If you use full wave diode rectifier you will get a higher output than with the tube.  Constant for a diode is ~1.4 X winding, 1.3 for GZ-34, 1.2 for 5U4, 1.14  for 5Y3.

If you convert to bridge diode with that same transformer you will get the entire secondary X 1.4 at half the current. I don't think you want to do that.

6L6GC will stand 500V all night for years. (Yes, it says 450V but it is the same plate-stuff as TV tubes rated higher.... it was a marketing trick to make you pick a higher-price tube for hi-volt operation.)

What is your idle current and dissipation?

PRR said:

6L6GC will stand 500V all night for years. (Yes, it says 450V but it is the same plate-stuff as TV tubes rated higher.... it was a marketing trick to make you pick a higher-price tube for hi-volt operation.)

What is your idle current and dissipation?

Click to expand...

I've been having an ongoing problem with this guy.

Basically one power tube will go into Chernoble mode with the plate glowing cherry red and shut down the amp. I was able to monitor the plate voltage when it happened the last time and basically the voltage drops down to almost zero and the amp shuts down. I'm assuming that the tube is pulling a huge amount of current in this case. Last time it happened there was a considerable amount of smoke rising up from around the tube bases, but no char marks on any resistor or the board.

In the past I've been able to recover by shutting the amp down for a few minutes and bring it back up. Plate resistors have been replaced along with almost all the elco's, aside from the 40/40/40 main filter cap.

I will check the idle current today.

Mark

Have you checked the valve holder and valve base to make sure it hasn't developed a burnt area between pins 2 and 3 (heater and anode)?
The last 3 amps I dealt with had this problem. Once the bakelite chars, it becomes very conductive, relatively speaking.

Bias voltage, and resulting current draw per tube are of primary concern here.  It sounds a lot like incorrect bias condition.  I have repaired several Fenders that lost their bias supplies, and they act in a similar manner.

It does sound like you may have a tube socket issue, as an option. 

I am not sure if it would work for your application, but on the Drip 175 that I am building the 5Y3GT3 is recommended in place of the original GZ34, in order to lower the voltage a bit.

I am not sure if the GY3GT3 can provide adequate current in the B15 though, unless you want more of that voltage sag thing at higher volumes...

--Peter

The bias voltages appear to be correct, I'm getting just over -48v, schematic calls for -50v.

I've also checked the tube socket for signs of damage.

There's a large divet on the top of the socket and some charing on Pin 5. So it appears that it may be a socket related issue. I'll have to pop out the circuit board to get at the bottom of the socket.

Mark

yep, sounding like socket. 

I have seen cases where bias voltage had to be different to get expected current and voltage readings with a particular tube set.  I've spent some time with a power tube matching rig that handled 100 tubes at a time, and there are always at least 10% of a 100 batch that fall well outside of the average.

And as per a previous thread, set the bias by measuring the anode current, rather than measuring grid volts.

I'm a bit surprised nobody recommended a 5U4 in there.  It will easily drop 40 volts out of that circuit over the GZ34.  Of course, I hope the client knows he will lose some wattage with any B+ reduction.

But then it just could be a socket...+1 one on anode current biasing.  You may want to consider a couple 470 ohm screen resistors, too.

> one power tube will go into Chernoble mode

Why? Just one tube? Always the same socket?

Something goes wrong, but isn't wrong while you are watching meters.

> eating expensive NOS tubes for dinner.

Perhaps time to get some good new Russian "6L6GC" for smoke-testing? They are not bad tubes. Hoffman has EH 6L6 for $44/pair. I'd rather smoke a few $22 bottles until the problem is found, than keep smoking-up the dwinding supply of 1970s 6L6GC.

The old 6L6GC sure were over-volted in commercial products; 500V is over-spec but thousands of gitar amps did it and they lasted fine.

If you are very concerned, 6550 is nearly a drop-in, much higher ratings, similar sound. Check that socket Pin 1 lug is not used as a tie-point: pin 1 is N/C on most 6L6 but is the metal base on 6550. If pin 1 were a tie-point for G2 resistors (but this plan shows none), that would be lethal. If tie for G1, then stray metal objects would short-out G1, no-bias, quick smoke. Heater power is higher, probably not enough to notice. Feel the PT for added heat after 20 and 60 minutes on idle.

Grid coupling caps C13 C14 are suspects. If they occasionally short, grid bias changes from -50V to +250V, or rather to maybe +5V held-down by 6L6 positive-grid current. With 450V on G2, 6L6 will pass about a half an Ampere, start to dissipate 250 Watts, stuff will sag and smoke.

Change the socket. At 500V you should be using ceramic or some extra-good phenolic; factories tended to cheap-out. They may have got away with it for a 30-day warranty, but long-term the insulation breaks-down and 30 years later the cheapness catches up with us.

Oscillation is possible. Has the OT plate lead layout been changed at ALL?

If both sockets are involved, the bias supply is a good bet. There is more strain on that 100K 2W R36 than a napkin calculation suggests. It is "OK", except 30 years on carbon-composition working over half its rating is far too much: the clay/carbon mix cracks. It may be "touching" one moment and OPEN the next. When open, bias goes to zero, tube current goes high. I would consider two 47K 2W metal-film to get a 4W total rating in a less fragile material.

DO put 10 ohms 0.5W in each 6L6 cathode. This is a handy place to check current, reduces gain enough to turn barely-marginal oscillation into almost-marginal oscillation, and in disaster may smoke-out faster (and cheaper!) than the tubes or OT or PT.

These are great ideas Paul, I will give them a go.

Good tube sockets and replacement bias supply resistor are on the top of the list.

I love the sound of this amp, but with it being unreliable it's just not usable in a session.

Mark

It sounds like too much plate voltage may not be the problem here, but if you want to lower it some try a 10k at R38 and/or R37. 165v on the pres and 235v on the P.I. gives you some wiggle room.