Wright Mono 8 Schematic

[jdrouin]

I thought I'd ask this question here because some people used to know George Wright, and some of his ideas contributed to early BH products. I've made a breadboard to try different amplifier circuits, starting with old standards like the Fi Primer and JE Labs 300Bs. I love Wright equipment, and have successfully breadboarded an amp that is the front end of his WPA 3.5 (2A3) but with a 300B output stage.

I'm curious to try the Mono 8, and while I haven't found a schematic anywhere (I know, he never wrote them down), I did attempt to reverse-engineer it from hi-res pics in a recent internet sale.

There are 3 components in the 6SN7 stage that I can't quite figure out the values of. Does anyone recognize them or happen to know what their values are likely to be?

I've attached pics from a recent internet sale, a drawing which duplicates and simplifies the layout, and my attempt at rendering a schematic.

What I don't know are the following three things in the 6SN7 stage (squared-in with red pencil in the drawings):

* Small red coupling cap from plate 1 to grid 2 (pins 5 —> 1)

* Second cathode (pin 3) resistor and little white bypass cap (peeking out from under the big red coupling cap)
(common values here are usually 23K-30K and 47uF 160V, but who knows?)

* Silvery-blue “F1" cap (I might see numbers 047 -- possibly 0.47uF?)
(it connects between the 270K and 75K resistors that go to the ground leg of the first [rightmost] 22uF/450V PS cap)

Many thanks for any insight you might provide, and sorry for posting so much info.

[Doc B.]

George used a combination of new parts and parts from a huge stash of surplus parts he got from an old surplus electronics place called Radar Electronics. The parts that are not marked could well be from the hoard. Like many smaller manufacturers he seldom built things the same way twice. Best bet is to use your estimate of the most commonly used values, as George's stuff was pretty traditional.

[Paul Birkeland]

The first coupling cap would be fine as a 0.047uF.

I drew a loadline for a 6SN7 assuming about 300V of B+ available and a 27K load. 4V of bias and 5mA of plate current looks reasonable.  That's a cathode resistor of 800 Ohms (820 is fine) and a 470uF/10V bypass cap should do the job.

If you build the amp as-drawn in your schematic, you'll smoke 2A3's and potentially your output transformer. 

[jdrouin]

Thanks, Caucasian Blackplate. I added those values into a SPICE model and got the bias and plate current you suggested. Wave form looks good too.

Do you think 0.47uF for the other unkown cap -- the one that connects between the 270K and 75K resistors -- is reasonable?

[Paul Birkeland]

Hmm, I didn't notice that the choke is in the ground leg (I had initially thought there was positive voltage being applied to the 2A3 grid, which isn't such a great idea).

Normally we would see the 270K resistor go from 300B grid to ground.   We also have that 1uF cap that's coupling some power supply noise to the output stage likely to cancel out some of the ripple in the high voltage supply. Your mystery cap could form an AC voltage divider with the 75K resistor to inject some more power supply noise onto the grid of the 300B.

For what it's worth, if you're building from scratch, I would leave out the 1uF cap, the 75K resistor, and the mystery cap with the 75K resistor.

[jdrouin]

Thanks. I hear what you're saying. Wright designs seem to have an unorthodoxy that goes against modern wisdom. I'll try it as-is on the breadboard once I understand what it's doing (which you've helped me to do -- so thank you), and can then make changes like you suggest to hear what happens.

The thing that's bugging me now, though, is that 750R cathode resistor on the 300B. The earlier Mono 8s had a 6SN7 voltage amp and driver and the MQ FS-030 OPT, which is rated at 70mA quiescent current. So unless I'm missing something, in order to use a 750R resistor and not exceed that 70mA quiescent current, you'd have to run the 300B at 275Vak (according to my LTSpice experiments, anyway), which would result in like 5W output, not the 7W (12W peak) that George claimed on the earlier models.

I must be missing something.

[Paul Birkeland]

George changed the designs on these amps frequently and was known to not really keep notes about what he was up to.

The 70mA rating on the FS-030 is a recommendation.  I wouldn't expect a performance meltdown running it at slightly higher currents.

Assuming that we can just make the B+ whatever we want but are constrained by a plate current limit of 70mA, and output transformer impedance of 3K, and a cathode resistor of 750 Ohms (this is not normally how we design a tube amp, lol), then we have  a bias voltage that's about 50V.  That would indeed lead one to conclude that you aren't going to get a lot of power out of the output stage! 

[jdrouin]

"this is not normally how we design a tube amp, lol"

I know, right?

Attached are two LTSpice screenshots that show the circuit with your originally-suggested res/cap values.

One shows the 300B running at 275Vak/70mA, and the voltages work out as follows:

6SN7 1: 153Vp, 6.4Vk
6SN7 2: 161.6Vp, 4.6Vk
300B: 328.5Vp, 53Vk

The other shows the 300B running at roughly 350Vak/91mA, with voltages as follows:

6SN7 1: 195Vp, 8.2Vk
6SN7 2: 204.4Vp, 5.9Vk
300B: 420.1Vp, 68.9Vk

I don't think either of these can be right.

Would an FS-030 be capable of handling 91mA?

Also, in LTSpice each of these can take a maximum signal voltage of just 0.25V before clipping. The Mono 8s were designed to work with his line preamps, just as the WPA 3.5 was. The WPA 3.5 requires an input signal of 3-4V for full output.

[Paul Birkeland]

At the old MQ website, the product info for the FS-030 said it could be gapped for up to 80mA.  My guess is that gapping for more than 80mA would reduce the primary inductance enough that Mikey didn't want to do it.

Your simulations show that the first stage has more bias voltage than the second stage.  Though you are able to clip the amp, I don't consider this ideal. You could go for 8V of bias and 4mA (2K cathode resistor) and see if things improve.

Your comment about the gain is spot on.  I remember reading a Stereophile review of a 300B amp some time ago where the review thought that the amp didn't have enough gain but totally ignored that the input sensitivity of the amp was 1.5V for full output (it was a power amp, intended to be driven by a preamp).

[Paul Joppa]

The original spec for the FS-030 was 60mA maximum. That is of course Mike's conservative rating. I believe this was chosen to match the 300B, that being the reference operating point in the WE data sheet - 300v p-k, -61v bias (cathode resistor ~1000 ohms, OPT 3000 ohms).

The RCA spec for the 2A3 is 250v p-k, 60mA, -45v bias (750 ohm cathode resistor), 2500 ohms load. George usually followed these specs. At one point, he used the DS-025 output transformer, matching the 2A3 spec of 2500 ohms.

All his SE 2A3 circuit variations I have seen had a cathode follower for the second stage, with the first stage plate direct coupled to the second stage grid. This is basically the same topology as the original Foreplay, which was designed by George. The coupling cap from the second stage cathode to the 2A3 grid was fairly large, 0.5 to 1.0uF as I recall. This allows the 2A3 to be driven into grid current (class A2), making possible a claim of 8 watts, as long as it sees short, infrequent transients, and you don't ask uncomfortable questions about distortion. It's still a 3.5 watt amp measured in the normal way. [I have myself used 275v/55mA/3000 ohms, which works pretty well.]

[Paul Birkeland]

Since you are starting from scratch, and have FS-030's, I'd not worry too much about what was actually in the Wright amp, but rather where you might want to be with the output stage.  Western Electric was thorough to a fault, and I'd certainly start by reviewing their operating points (this may be a little quicker than tweaking the Spice model).

[jdrouin]

Caucasian: Oh yes, I know that WE op point sheet well.  Also, your suggestion of 2K for the second 6SN7 cathode resistor is close. I read your comment on my phone, and then looked at one of the pics I'd attached and could clearly see the stripes are brown-black-red, or 1K. Unless it's grey-back-red, which would be 8K, but probably not. Funny, the colors didn't show up as clearly on my large computer monitor.

Paul: Yes, the WPA 3.5 as well as WLA 12 amps I've looked at are exactly as you described. Interesting thing about the WPA 3.5 is that it takes the signal from the 6SN7 2nd cathode, rather than the second plate as in the Mono 8, and capacitor-couples it to the 2A3 grid. That's what I did with the bastardized WPA 3.5/300B that's playing *very nicely* on the breadboard now. Btw it runs the WE-recommended 300B op point of 350Vak/50ma into a 5K load (I've got big 5K Electra-Print OPTs from an earlier project, not MQs).

Back to the Mono 8...

I suspect that one or both amps were tampered with, since they not only have some different component brands but different values as well. For instance, in one amp the 6SN7 cathode 1 resistor is apparently a 292 ohm carbon comp (red-white-red-black-black) -- unless that means 2-9-2-0-0 or 29.2K. In the other amp is a much larger wattage 3.9K 5% (orange-white-red-gold) carbon film.

But I hadn't thought about that second black stripe, so let's run with this happy accident...

Suspending our disbelief in my stripe-reading abilities for just a moment, let us speculate that 29.2K is a viable option in 6SN7 1 cathode. In the model, that drops the input sensitivity to about 2V before clipping (2.25V with a 39K resistor) -- much more realistic for use with a line preamp. If the same resistor in the other amp is not 3.9K but 39K (I don't think it is, but still), then there's still a difference between the two that calls into question the authenticity of the specimens before us. But let's try it anyway.

The below trials were done by running a 300B with 750ohm cathode resistor at 300Vak, which results in 77mA current. That's a little north of the FS-030 manufacturer specs but possibly within its realistic limits. That might just reach the 7W that George claimed for that particular version of the Mono 8.

With a 29.2K 6SN71 cathode resistor, we get these rounded results:

6SN71 -- 296Vp, 15Vk
6SN72 -- 190Vp, 5.8Vk

With a 39K:

6SN71 -- 308Vp, 16Vk
6SN72 -- 190Vp, 5.8Vk

And for the hell of it, with 8K on the second cathode resistor:

6SN71 -- 318Vp, 16.7Vk
6SN72 -- 312Vp, 14.5Vk

Edit: That last one can take up to 3V and a little more before clipping, which is closer to the WPA 3.5 input sensitivity.

All are within spec -- if a bit hot -- for the 6SN7 but it's odd that the first triode has higher voltage than the second, as I almost always see it the other way around.

Not sure how that would sound, but that's what the breadboard is for. Should I try it?

[Doc B.]

Funny, the colors didn't show up as clearly on my large computer monitor.

Welcome to the 21st century world of over-hyped, supersaturated smart phone displays. It's like Apple and Samsung want everyone to feel like they're on acid all the time.

[Paul Birkeland]

If you have a bread board, try a single 6AB4 with a constant current source running 2-3mA of plate current and biased at 2-3V. 

[jdrouin]

Thanks for the alternative tube suggestion. Something to consider when I start designing my own circuits down the road.

An owner of the Mono 8 with 6DN7 input/driver (ca. 2001) sent some good pics of the interior, and I was able to generate a schematic (attached). I'm more interested in the 6SN7 version, and more examples will be necessary to get a handle on that one.

Still stumped on that 750R cathode resistor for the 300B. Would require a low voltage op point (ca. 250Vak and only 4W output power) to stay around 60mA plate current, unless George had his FS-030s custom airgapped for the 70mA range.

Is there some possible alternative that I'm not thinking of here? I've been studying schematics for 2 years -- a lot less than you guys -- but have never come across a 300B SET that's designed like this.