SCA35 and AU-70

[mr2racer]

Hey All,

I found a Dynaco SCA35 chassis on eBay, cheap. I also found a Sansui AU-70 for parts including all the transformers. They both used EL84 outputs and Sansui transformers were made by Tango, so I thought, this will be easy.

When I took a closer look at the Sansui schematic I found a really different way of wiring the outputs and the OPT. First off it is not an ultralinear transformer. They introduce B+1 (413v) to the center tap of the OPT primary. This resuslts in 410 volts at the plate. Then they introduce B+2 at the middle grid of the EL84. This is reminiscent of what Eli told me was called "Full pentode mode."

Then, there doesn't seem to be a cathode resistor. The cathodes are connected cross coupled to the end taps of the opt transformer secondary. Between these end taps are a ground, a 4 ohm tap and a 16 ohm tap.

I have a few questions regarding this schematic, if anyone is familiar with it. The first question is do I need to follow it to achieve the 25 watt output and the sonic quality of the transformer? Second, Since the Sansui has fixed bias with pots for each EL84 is it possible a cathode resistor is unnecessary? By connecting the end taps of the OPT secondary to the cathodes is this a cross coupled local feedback circuit? And lastly, without a cathode resistor What do you measure to adjust the bias? Negative voltage from grid to ground?

[mr2racer]

I'm attaching a pdf file that includes the schematic.

[Eli Duttman]

Notice that the 7189 grid resistors are connected to C3 and C4, not ground. C is the notation used for negative grid supplies. The circuit uses fixed bias.

As for 25 W. from a PP pair of 7189s, "No way Jose!". Using the tubes at the abs. max. of 13.2 W. of plate dissipation each yields a total of 26.4 W. Theory tells us that the max. possible efficiency for pure Class "B" is Π/4. So, 20.7 W. is the theoretical limit. Class "AB" operation, along with other real world considerations, brings you down to at best 17 W. Inflated claims have, and continue to, plague us.

The O/P trafos may be rated for 25 W. That provides the magnetic headroom needed to cope with a GNFB low freq. error correction signal.

[HollowState]

do I need to follow it to achieve the 25 watt output and the sonic quality of the transformer?

No, not necessarly. The sonic quality of the transformer is determined more by it's design and build quality. And power out is a function of power in. However 25 watts from a pair of 7189's is unrealistic.

Since the Sansui has fixed bias with pots for each EL84 is it possible a cathode resistor is unnecessary?

Correct

By connecting the end taps of the OPT secondary to the cathodes is this a cross coupled local feedback circuit?

Yes it is. The cross coupling is a function of the winding direction (phasing) of the transformer. This is a practice used by more then one manufacturer to reduce distortion generated in the output stage itself.

What do you measure to adjust the bias? Negative voltage from grid to ground?

Measuring the negative grid voltage will give you the bias level. Remember bias is the voltage that determins tube idle current. If you measure, and then know, the resistance of the transformer secondary between the ends and 4 ohm ground, then measuring the positive voltage at the cathode will let you determine tube current with simple ohms law. I=E÷R

[mr2racer]

Thanks All,
Dyanco rates the sca35 at 17.5 watts per channel. I guess Dynaco is not inflating the spec as Sansui did. Sansui and Tango worked closely together in designing and building these transformers so I think I'll keep the cross coupled feedback. So in effect they're using the resistance between the end taps and ground as a cathode resistor, or at least I can use it to measure the idle current.

And they are using the end taps to make the local NFB applied even? If I were to move the ground to the end tap instead of the green wire, which we are assuming is a 4ohm tap, would it still function in the feedback loop? In that way I could recover the 4 ohm output? Also, If I were to insert a small resistor, say 1 ohm, between the cathodes and the secondary taps it would decrease the feedback slightly but give me an easy place to measure idle current wouldn't it?

But then there is still the question of applying the B+2 to the middle grid. Is that necessary? Why do they do that?
Kevin

[Wavebourn]

4 ohm output + another 4 ohm output means twice voltage, that means same power on 16 Ohm load. Yes, feedbacks applied are even.

They applied lower and better filtered voltage to screen grids to make output stage with such load resistance and anode voltage more linear.

[HollowState]

So in effect they're using the resistance between the end taps and ground as a cathode resistor

That's not the main objective. The objective is the signal feedback. The resistance is just there and one is stuck with it. But you can use it for measurment none the less.

And they are using the end taps to make the local NFB applied even?

Yes, it's balanced feedback.

If I were to move the ground to the end tap instead of the green wire, which we are assuming is a 4ohm tap, would it still function in the feedback loop?

NO! you would unbalance it . Leave it as is.

In that way I could recover the 4 ohm output?

You haven't lost it. It's still useable, only the opposite end is at ground.

If I were to insert a small resistor, say 1 ohm, between the cathodes and the secondary taps it would decrease the feedback slightly but give me an easy place to measure idle current wouldn't it?

Yes you could do that. The amount of decrease in feedback is miniscule with a 1 ohm resistor.

But then there is still the question of applying the B+2 to the middle grid. Is that necessary? Why do they do that?

It's very necessary. Without screen grid voltage the tube won't function.

[jakruby]

dynaco operated the sca35 in ultra linear mode using cathode bias with about 350V. The sansui uses fixed bias, operates the tubes in pentode mode and runs at 400V - all of these things will result in more power output so 25W is indeed possible

[mr2racer]

One last question? The 7199 is awfully expensive. I've seen, with minor modification, 6U8A, 6GH8A or 6AN8A can be substituted. Does anyone know which is the best to use? Or is it best to bite the bullet and buy 7199's. I noticed Jim Mcshane has a couple choices. for this type of tube is there a superior manufacturer? From what I've gathered Sylvania's are better than RCA's but beyond that I have no idea.

[Wavebourn]

Try 6F1P, it is greatly underpriced.