Switched Attenuator help for an RCA BC-2B

[emrr]

[quote author="NewYorkDave"]I just thought of something: if someone was willing to provide me with a set of input and output transformers, I would be happy to build up a BC2B and work out the whole gain control issue, and post complete details for everyone to use. I'm just not willing to spend my own money for the transformers, especially the 15K:600 gapped output. I just bought a new car and I have payments to make. :wink:[/quote]

Dave,

I've got all this gain / feedback info worked out from experiments with a real unit. I'll try and remember to post it once I've got a second to breathe. Bug me if three months go by without update.

 

[Kid Squid]

I'll Remind you :wink:

Steve :thumb:

 

[emrr]

Okay, here's what I found on the real BC-2B with 1K sine wave tones.  I followed this with a listen to program material at all settings, and the response seemed to stay remarkably flat.  I didn't do sweeps to judge any response changes.  later revisit in red


feedback R / feedback amount

open / none
2.12M / 3.5 db
1.33M / 5.1 db
880K / 6.9 db
470K / 10.2 db  (40 db stock)
290K / 13.3 db
180K / 16.5 db
120K / 19.5 db good down to here with stock NFB cap
80K       / 22.5 db
50K       / 25.9 db good down to here with NFB cap changed to 0.47mfd
30K       / 29.3 db
20K       / 32.3 db

I don't think I'd personally trust stepping the gain down to those last two steps via feedback, but it seemed okay.  In comparison, the BA-11A has 27 db of feedback, so it's not terribly out of range.  I'd probably personally go +10 to -15 with the feedback, using a stepped switch so I know where the stock sound setting lives. 

I doubt you'll overload a stock BC-2B input transformer; open question for anything else.  You won't (sanely) overload that LS-10 if you use it. 
The K & K audio kit appears to use both an input pad and a stepped pot on the input transformer secondary.  I'd think you could safely go with 6-10 db of stepped loss on the input secondary; probably depends agin on the input trans quality.

The K&K mention of original BC-2B module price is now seriously outdated, BTW.

For any other antique pre you want to try, simply disconnect the feedback and take gain and resistance measurements with a variable resistance sustituted for the fixed feedback resistance.  It's really that easy.  Be prepared to sub larger caps in the feedback path and/or audio path for continued flat low frequency response at higher feedback quantities. 

There are later (late 50's-60's) tube preamps that have really high amounts of feedback to help make up for the use of smaller, cheaper, lower quality iron, and it can be a real eye opener to pull the feedback out and hear them go to sh*t.  Many of the early 50's units were still using iron that sounded good with or without feedback as part of their corrective massage, and you can make more changes with them. 

As far as input and output attenuators, Dave has left a lot of good advice already.  Standard 20 db pad on the front and any version of T, H, L, or pot on the out will work.  Many people get away with a 1K pot for output, though I like to splurge for proper T attenuators personally.  Most L's will automatically give you a 6db loss unless you also wire in a bypass switch.

Final note; I measured the step down through a stock BC-2B output transformer at -18.6db, which suggests a higher primary than 15k.  My gut tells me that RCA used 25K-30K for most preamp outputs (as did Western Electric), but I know I shouldn't necessarily trust my gut on this one.  :grin:  Some RCA tube preamps have output trans step down as high as -22 db.  The 15K will work just fine, though.


LATER ADDENDUM:

[quote author="NewYorkDave"]
someone on Usenet who owns an original BC-2B measured his preamp transformers as 30/150:33.7K input and 12K:600 output.[/quote]


I had an NOS/NIB input xfrmr once, and the RCA box stated 30 ohm is 1:46 ratio, 150 ohm is 1:20.5 ratio. That would appear to be a 63K secondary.

12K:600 for the output would be only a 13 db ideal loss. Most 15K:600 outputs I've checked measure close to a perfect 14 db loss. I measured 18.6 db loss through the output (BA-11A in operation with 1K tone) with a 620 ohm load on the 600 ohm secondary. Not accounting for losses, that would appear to be roughly 42K:600. Considering 1.1db of winding loss, that would translate into 33K:600, and a 2 db winding loss would be 27.5K:600.  Looks closer to my gut feeling expressed earlier.


More:

http://www.groupdiy.com/index.php?topic=21504

 

[Wonderlandaudio]

I thought those RCA consoles used balanced T-bridge attenuattors at the input, like Daven pots or similar. am I wrong?
I´m not really into that stuff so I might be telling bullshit...

 

[emrr]

The BC-2B uses Daven 150/300 impedance L's after the preamps and remote line input repeat coil.

You rarely see balanced pots in vintage radio mixers, or anywere else, for that matter. Way too pricey, and unnecessary technically in a localized enviroment. You will usually see a repeat coil after a T or L for balancing before you'll see the expense of an H (balanced T) attenuator. Almost every old tube limiter uses 600 ohm L's on input and output, and have tech notes sugesting use of repeat coils for additional isolation and earth balancing when necessary. You almost always see Daven (or similar) ladders after the preamps in old consoles, which feed the mixing buss. Line inputs usually come in through repeat coils that then see L's on the way to the buss. The Collins 212A/B are one of the few console series that even uses T's everywhere. Virtually everything else used the less expensive L's, which also have a noise advantage in one regard.

From Altec attenuator catalog:

"Contact noise in an L decreases in proportion to increased attenuation while it remains the same in a T circuit regardless of attenuation."

Altec also claims "another reason for the popularity of L circuits is that they provide 6db of isolation between the preamp and combining network."

 

[Wonderlandaudio]

Thanks man, this has been a real masterclass. In fact, for a T bridge balanced attenuator you need 4 poles and doesnt work well for less than 20db att´s if I remember well.... or was the opposite. I just remember from several years ago when I experimented with Langevin AM16´s. Going after the input tranny was dangerous so I tried that. But I wanted to play with feedback as well, which meant 6 poles, and the output... ended up with the 20db pad and output control, but added some feedback control.

Thanks for the info. I have a lot of A-10´s that I´m thinking into using them in tube pres...

 

[Wonderlandaudio]

The thing is I also have a bunch of these:




So, I should use them at the output, right?

 

[emrr]

Yes, those will only serve you well on the output. I have similar Clarostat wire wound units on my Collins 356A's. Remember you'll need a true 600 ohm load after the T to get proper scaling with it, since a 600/600 T is meant to see 600 in both directions for designated operation. 680 ohm is a good choice, with a disconnect switch like Dave's H pad design.

IMO you should sell the A-10's and get something better sounding. I think they are way overvalued due to the LA-2A link. There are better sounding choices for less $$. Even LS-10's are not terribly out of range price-wise, if you want a definite vintage upgrade and can afford the space requirements.

EDIT: oh, I see they are available for sale in the black market. Buy'em up, LA-2A people! :grin:

 

[NewYorkDave]

Terminology check: "L" usually refers to an L-pad, whereas it seems you're talking about ladders, which are not the same.

 

[emrr]

Correction noted; I'm using an incorrect shorthand that I see in many places. No doubt common from Daven's use of L and T in their part # prefixes.

RDH describes ladder attenuators that are both Pi section and L section; the Pi section is the standard version in audio. The L is more common in test equipment.

There is also a less commonly seen ladder type that varies the input impedance more considerably, but has a minimum loss of 3.5 db rather than the usually seen 6 db.