Magnavox 8802 tube bias questions

[Kidmoe]

So just for giggles I decided to take the voltmeter to my Magnavox 8802 amp that I have lying around after reading about the bias levels that Magnavox shipped them out the door with. I was curious how mine would measure and this is what I found on the 6V6:

286v plate
285v screen
17.8v cathode
180 ohm resistor

So if my math is right I figure that each tube had just over 23ma of plate current running through it (when I removed the stock 180 ohm resistor I found that it actually measured 199 ohms in my unit which was causing the tubes to bias even cooler than that) with only about 6w of dissipation. So in the interest of getting this amp's blood pumping a little I replaced the 180 ohm resistor with a 120 ohm after which I got these numbers:

268v plate
267v screen
15.6v cathode
120 ohm resistor
6.57v heaters

Again, excluding faulty math, I get about 31ma per tube and right about 8w plate dissipation after changing the resistor. My question is should I stop there or can I squeeze just a little more out of this amp? I have a 100 ohm resistor that I could try as a next step. I know that the 6CA4 rectifier is only rated to handle 150ma of current so getting to about 35ma on each of the 4 tubes wouldn't leave much for the rest of the components. Also, should I care that the screen voltage is pretty much equal to the plate voltage although the schematic shows the screens riding 10v lower than the plates?

 

[jazbo8]

My question is should I stop there or can I squeeze just a little more out of this amp? I have a 100 ohm resistor that I could try as a next step. <snip>
Also, should I care that the screen voltage is pretty much equal to the plate voltage although the schematic shows the screens riding 10v lower than the plates?

By sqeezing a little more out, do you mean more output power? If so, you are heading the wrong direction with the cathode resistor. The screen voltage does look high, you may want to check R2 (470R) to see if it is still good.

 

[Kidmoe]

Thanks Jaz, In a few older threads that I came across it was advised that the bias should be raised to between 30ma and 35ma per tube from the stock 25ma. What benefit would this provide?

 

[battradio]

Little or none , that is with in tolerance of a 5% cathode resistor and the tolerance of a cheep meter .

 

[jazbo8]

+1, 5% makes no difference. But I do wonder why the screen voltage is not lower than the plate voltage as shown by the schematic.

 

[dcgillespie]

By sqeezing a little more out, do you mean more output power? If so, you are heading the wrong direction with the cathode resistor.

If I may, in this case, the OP is in fact headed in the right direction for producing more power. Besides the fact of the power loss from the output stage being well over biased to begin with, there is also the basic fact that the cathode resistor is an element of the basic 4 element series circuit consisting of:

1. The power supply,

2. The output tubes,

3. The cathode resistor, and

4. The OPT primary winding.

These elements are all ultimately working together to produce current flow in the OPT primary winding. Anywhere the resistance of the first three elements can be reduced means that more peak current can flow in the fourth element. Therefore, reducing the value of the cathode bias resistor (to a point of course) will increase the power output -- and the quality of the power output as well, in this case.

Other issues may certainly crop up, as (for example) when the output stages move towards becoming properly biased, the little 6CA4 will be screaming bloody murder -- properly biased, both output stages together will typically be drawing the full rated current flow of that tube. The power transformer also may not like the new arrangement, unless you're tapping into (now) unused tuner power that this unit might have originally provided.

If the power transformer is up for the challenge (due to originally powering external devices no longer used), and an extra heater winding is available that powered those devices (tuners usually had their own heater winding), then along with properly biasing the output stages, it might also be a good idea (if possible) to put that extra heater winding to use by powering a second 6CA4 rectifier to eliminate that weakness from the design.

With this unit apparently being well over biased, and that being the finding of my 860(4) project as well, I wonder if this wasn't a common design tactic that Magnavox used to maximized tube life......

Dave

 

[jazbo8]

With this unit apparently being well under biased, and that being the finding of my 860(4) project as well, I wonder if this wasn't a common design tactic that Magnavox used to maximized tube life......

Dave, I don't think this one is under-biased, in fact, it is biased pretty high for class AB operation, i.e., it's more A than B compared to the other makes. Do you know the OPT's spec? I just used 8k for my calculations, so may be like you said, the OP was moving in the right direction by lowering the cathode resistor.

 

[analog addict]

Heh you could always take the car racing philosophy and push it until you see a problem or something breaks. Red plating the power tubes/rectifier arcing would be the possible endpoints, but I've seen 6CA4s take a fair amount of abuse as well...:scratch2:

But doing it methodically and measuring your results at every step is the way to really learn what the design is capable of....:thmbsp:

 

[wa2ise]

. My question is should I stop there or can I squeeze just a little more out of this amp?

You'll never notice the difference. It takes 1dB to make a just noticeable difference. And that's about 25% more power.

 

[dcgillespie]

I haven't had time to run any actual data, but from my experience, 23 ma of plate current (per tube) for cathode biased push-pull 6V6 class tubes into an 8000 ohm load represents a pretty OVER biased condition as I originally stated in the opening paragraph (relative to the required grid bias voltage), and therefore very limited Class A operation -- less than a watt of power developed by napkin calculations before class B operation commences.

In my closing paragraph, I misspoke when I inadvertently wrote "under" biased in that sentence. I'll go back and correct it when I get a chance.

Dave

 

[jazbo8]

Da Tubes Again...

I haven't had time to run any actual data, but from my experience, 23 ma of plate current (per tube) for cathode biased push-pull 6V6 class tubes into an 8000 ohm load represents a pretty OVER biased condition...

Here we go again, it's déjà vu... the darn tubes in the unit are throwing off the results again. You would be correct to assume that the 6V6's are over-biased from the 23mA figure that Kidmoe showed above, it does look pretty low.

Since the voltage readings above showed that they were pretty far from the ones shown in Magnavox's schematic, so the actual Pda in the unit is a bit higher, however, if the tubes were running closer to what the datasheet say, then they'd be running at ~70% max Pdiss and should not be pushed higher.

As to why the voltages are so different, I don't know, may be it's the higher mains voltage combined with the tubes?:scratch2:

 

[Kidmoe]

Jaz, My mains voltage is about 121.5v so the readings are off compared to the schematic because of that with the highish filament voltage of 6.57v being a giveaway.

Also I can't verify this info but a poster in this thread states that the opt are 6.6K 4ohm according to the Sams Photofact he has. Would be nice to know how much current capability the power transformer is rated for. With the original tubes biased around 25ma each and also powering the tuner in the console I'm guessing around 140ma unless the pt was overspeced, which I doubt.

http://www.audiokarma.org/forums/showthread.php?t=247818

 

[jazbo8]

Jaz, My mains voltage is about 121.5v so the readings are off compared to the schematic because of that with the highish filament voltage of 6.57v being a giveaway.

Also I can't verify this info but a poster in this thread states that the opt are 6.6K 4ohm according to the Sams Photofact he has. Would be nice to know how much current capability the power transformer is rated for. With the original tubes biased around 25ma each and also powering the tuner in the console I'm guessing around 140ma unless the pt was overspecked, which I doubt.

http://www.audiokarma.org/forums/showthread.php?t=247818

The plot thickens... I missed the filament voltage ealier, it's not that big an increase though, in any case, even with the higher emission, the plate current is still lower than what it should be. Based on my calculation, with 286v plate, 285v screen and 17.8v cathode, the plate current should be ~31mA not 23mA. Thanks for the information on the OPT, so are you running 4 Ohm speaker with it?

 

[Kidmoe]

17.8/180 then divide by the number of tubes (4) then knock off about 5% or so to account for some screen current and my math gets right around 23ma unless my method is incorrect. So then 268.2(.023) gets about 6.1w dissipation.

 

[Tom Bavis]

"Over" and "under" biased - could be confusing since bias voltage gets confused with idle current, and of course they change in opposite directions, so higher bias voltage gives lower idle current. "Hot" and "cold" bias may be less ambiguous terms.

 

[jazbo8]

17.8/180 then divide by the number of tubes (4) then knock off about 5% or so to account for some screen current and my math gets right around 23ma unless my method is incorrect. So then 268.2(.023) gets about 6.1w dissipation.

Sorry I was not clear, your math is correct for your set of tubes, but they do not match up with the datasheet, i.e., they are running a bit on the low side - but still within the acceptable tolerance.

 

[Kidmoe]

"Over" and "under" biased - could be confusing since bias voltage gets confused with idle current, and of course they change in opposite directions, so higher bias voltage gives lower idle current. "Hot" and "cold" bias may be less ambiguous terms.

Very true. I always struggle for the correct way to describe bias and whether it is under/over. Tom, do you have a Sams for the 8802? I'm curious what the official current handling of the power transformer is.

 

[chazix]

Also, should I care that the screen voltage is pretty much equal to the plate voltage although the schematic shows the screens riding 10v lower than the plates?

Eh, it's only eight years or so later - can I re-visit this aspect?

I don't know whether one should care, but I think the reason is that the screen supply was also meant to serve a separate tuner section. If there's no tuner section connected, there's no appreciable voltage drop in the screen supply (compared to the supply that serves the output transformers).

So, back to should I care? It seems like it would be simple to insert a Zener diode in series between the two supplies to make the voltage difference similar to original. Should I?

Thanks!

 

[chazix]

Never mind. After reading the interwebospheres, I conclude that there's no reason that the screen voltage needs to be lower than the plate (within their respective limits, of course).

 

[gadget73]

yeah I wouldn't worry too much about it. Can bump the resistor a bit if you want, or add a drain resistor at the tail end to simulate the draw of the tuner if you really wanted to but I don't think it would make any practical difference.

amps that tie the screens to the same B+ point as the output transformers typically have the plates 5-10v lower than the screen just from drop in the output transformer.