6L6 amp capacitor question.....

[Kamakiri]

I don't have a schematic for this thing, nor can I find one, but I'm stuck on something and hope someone can help.

This is on an RCA MI-12202-B amplifier, tube complement is 6L6 x 2, 6N7 x 2, 6J7, 5Z3.

I'm recapping the amp, but there are caps on the 6L6s that I just can't read, they look like they were typed on with a typewriter before the cap was rolled, and the ink is almost totally worn off.

The caps are identical, and go across pins 3 and 6 of both of the 6L6s. Put one on my capacitor tester to see if I could get a reading on it, but it's reading 200 pF. Could that be right?

[maxhifi]

Pin 6 is a tie point, not connected inside the tube. These are probably Across the primary of the output transformer, can you confirm what else goes to pin 6? I have a pair of RCA mi series amplifiers too, they're awesome beasts

[Kamakiri]

Pin 6 just has a 120 ohm resistor that goes back to pin 1, nothing else attached.

[maxhifi]

Pin 1 is ground on a metal 6L6, but on a glass one it is sometimes used as a tie point just like pin 6. Where is pin one connected to? If ground what you have is just a little network to stop the amplifier from oscillating. The 200pF rating sounds about right, it is not too critical, but what is critical is the voltage rating of this capacitor - it's got to be quite high, I would probably use a 1.5kV capacitor here, maybe higher, definitely not a little 400V capacitor.

[Kamakiri]

These are actually metal 6L6s in this thing. I can't even read the voltage rating, but it's about the size of an average wax cap, which really threw me for a loop if it is indeed a 200 pF. If it's 200 pF @ 1.5 kV, that makes a lot of sense.

The other end of pin 1 goes right to chassis ground.......

[maxhifi]

http://www.wowhififever.com/RCA_mi9377/pic_3.gif

The above link shows an RCA amplifier with the same network yours has, with different component values. Only difference is in the above URL the network is connected to the centre tap of the output transformer (B+), rather than to ground. From the perspective of an AC signal, these points are at the same potential.

Consider that the peak to peak AC voltage on the primary of the output transformer is what will determine the voltage rating of the capacitor, NOT the value of the B+ supply. That's why I would go for some nice big 1.5KV capacitors or even 2kV just to err on the side of safety.

RCA MI series amps are beautiful equipment, I wish all electronics was built to that standard!

[bob91343]

The peak to peak voltage applied isn't the right number. That voltage never exists at one time. It's the peak that matters, as compared to the trough experienced by the other terminal. And vice versa. It is subtle but the capacitor rating has to be above the maximum instantaneous voltage applied across its terminals, which has to be at one instant where it's maximum.

[maxhifi]

Ok this sounds about right - haven't thought about it for a while. The number is also affected by B+ being added to the peak voltage, since in the amplifier we are talking about the RC network is connected to ground, right?

So say for a push-pull output stage which from the RCA tube manual puts out 26.5W, with a 6600 ohm load on the output transformer, and 360V of B+, the following applies. From ohm's law the RMS voltage across 6600 ohms to get 26.5W is 418.21V, this is from plate to plate. From plate to centre tap, this will be 209.1 RMS, or 295.7 peak volts. Add this to the 360V B+, and the sum is 655.67V. So in this case, a 1kV capacitor should do the trick. I suppose that if the amplifier does not have a load, or has a load of a higher impedance than rated, the maximum output voltage can also rise a little bit, so the capacitor should have a decent safety margin. I haven't designed any tube equipment in a while, feel free to correct me if I am wrong!

So I will assume that the RCA amp in question probably contains voltages similar to above, add a safety margin and a contingency for the fact that the voltages and power level may be higher than in my example above, and I am quite sure a 1kV or 1.5kV capacitor is perfect for the job.

-Max

[Kamakiri]

In the above marked diagram, the cap that I believe I'm referring to is C3/C4? Showing a reading of .0033 uF on that schematic.

I took a reading off the second cap, taken out of circuit, and this is what I've got.

[dtvmcdonald]

Why would a high class amp use a phase inverter circuit that
ruins the even harmonic distortion canceling properties of push-pull?

In this circuit there is no excuse that extra gain is needed.
A standard circuit with equal plate and cathode load resistors
would do fine and has low distortion.
Was it some sort of patent scam? But RCA was king
of the patent hill.

[bob91343]

I didn't check the arithmetic but the capacitor voltage calculation is correct in principle.

[maxhifi]

Second harmonic distortion still cancels in the output stage... In the voltage amp there's not all that much distortion - RCA used that circuit over and over again and to be honest it sounds alright... it's a scaled up version of the circuit in most radios which have a push-pull output stage. The amplifier lacks global negative feedback, and the amplifier's overall distortion will be dominated by output stage distortion anyway, like a guitar amplifier. I have a pair of mi-12188a amps and have always liked them.

Keep in mind we're looking at 40s public address amplifiers here not a marantz 9 - this is from the era when records were 78RPM, tape recorders were from the future, and radio was mostly AM - speakers were full of all sorts of resonances and breakup modes and largely designed by trial and error in a pre-computer age. I have no idea how much THD and IMD a 1940s paper cone woofer with a paper surround makes below 100Hz, but I bet it's way more than the phase inverter in this amplifier! In the context of its application, these old amplifiers were over-qualified for the job, and it's amazing how well they hold up to this day in the context of much better speakers and program material.

[maxhifi]

Originally Posted by Kamakiri

In the above marked diagram, the cap that I believe I'm referring to is C3/C4? Showing a reading of .0033 uF on that schematic.

I took a reading off the second cap, taken out of circuit, and this is what I've got.

Yes, C3 and C4, however in the circuit I sent you a copy of, they are connected to the centre tap of the transformer rather than to ground. This means they don't have to deal with B+ in addition to the signal voltage, so can have a lower voltage rating versus your amplifier, but the function is exactly the same.

This is an RC network, and exists to roll off the high frequencies (above the audible band) to prevent the amplifier from oscillating under certain circumstances, for example really long speaker wires. You could calculate the roll-off frequency using the source impedance of the output stage, and the RC values, similar to designing a speaker crossover network, however I really don't think the actual value is too critical - it's just got to be enough to stop the amplifier from oscillating. I also think it's unlikely you will install the amplifier in an auditorium with really long speaker wires as per original design, so chances are you could get away with deleting this network entirely and not even notice it's gone.

[Kamakiri]

After a LOt of searching, I FINALLY managed to find the actual schematic for the amp. Seems that those caps I was trying to figure out were way off. The Riders reads them as .005 @500v

[maxhifi]

Looks like it's a low powered 6L6 amp - so the voltage rating is nothing extreme. I really like the construction of those MI amplifiers, I saw the photos of yours in the phono forum and it all makes sense now. The only cheap feature is the wafer tube sockets, but at least they used real sockets on the 6L6s.