various RCA tube preamps - progress and experiments

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

emrr

Well-known member
Joined
Apr 12, 2006
Messages
8,525
Location
NC, USA
5005010052_4b1791d24f_b.jpg


5004401597_5cb4a2ba7c_b.jpg


5005002360_6c8b86e755_b.jpg


5004997758_87ee3fdb19_b.jpg


5005004748_8d54f0954e_b.jpg


5004988492_b437b164b4_b.jpg


5004381829_66dbf3c9e5_b.jpg


5004379625_b6bd551eb9_b.jpg


I may add prettier rear labels another time.  I have the cosmetic rack face plates, but there are functional issues with installing them, and then having to pull them off constantly.  I wish RCA had placed access holes in the faces for use with a full six modules, rather than the four or five hole pattern they chose. 

I machined the connector bars out of 1/8" aluminum L bar, rather than hard wiring to a patch bay at a distance.  I move pieces around too often to rely on today's judgement call about wire length.   

I am currently providing B+ and filament to the non-powered modules with a Heathkit tube regulated bench supply, and am sorting out what will be more permanent. 

In many cases the input and sometimes the output transformers had a center tap tied to ground, which can be problematic in modern interfacing.  In every case, I have disconnected these center taps and allowed the inputs and outputs to float.  It is common for people to try and use one of these preamps today with a condenser mic and a phantom power modification, and many fail to realize that a center tap will shunt the 48VDC to ground through the transformer primary. 


Below are response plots on the BA series rack, after watching real time input and output impedance loading effects, and making some decisions.    No resistances replaced, all caps except 1) oil bathtubs, and 2) the pair of mica caps in the BA-3.   

My source impedance is roughly 150 ohms here.  This is pink noise with a long settling time, 0.7Hz resolution.  Straight wire measurement draws a straight line, for comparison.

Input loading varies per unit.  The BA-1 top peak grows larger with a 20 dB U pad in front of it, as is commonly needed with microphones.  The 3 and the 11 are affected much less. 

Output load resistance makes a lot of difference here.  The values I've picked make the most sense in my system, where they will sometimes see a 10K load, sometimes a true 600 load. 

The BA-1 especially is perfectly flat with a 600 ohm load, this shows a 750 ohm load.  This is the only case I've encountered of a preamp showing bottom resonance with so little a difference in loading.    This is pentode/pentode 1620 with DC in the output transformer. 

The BA-11 is loaded at 1k8, it's top peak is still there a little with 600. This is pentode/triode 1620 with DC in the output transformer, and I once measured about 27 dB of negative feedback in the loop.  The second BA-11 matched in measurement.

The BA-3 differs in that a 600 output load takes a lot of treble away. This shows a 1K load, with some additional high frequency boost as described in the manual. A 10K output load with original boost value gave similar response as shown, but I didn't feel great about such a light load on a 6L6 amp.  The BA-3 is specified as a line amp, many people these days use it as a low power monitor amp, and  hi-fi guys will bypass the input transformer for Hi-Z input.  In this case the unit will be used for microphone input, so I disconnected the first of the two ganged pots, now input transformer going to open grid.  This is pentode/pentode/beam tetrode 1620/1620/1622 with DC in the output transformer, cathode biased 1622.    As secondary evidence, I have another of the input transformer, and it shows a similar sharp treble drop when tested by itself.  It is possible the 1622 plate zobel has too much capacitance, or the resistance has dropped.  These stability networks tend to be burned up and way out of spec on a lot of broadcast power amps. 

The steepness of top and bottom roll-off in the transformers is interesting, and seems quite on purpose.  Many other preamps and line amps of the era have much more gradual slopes at the edges.  The BA-1 and 11 both have additional small caps in the feedback paths to kill more top end; I have not yet experimented with pulling them out.  I also expect to play with the amount of NFB in the 1 and 11, to see where the acceptable limits dwell. 

Click this link for the full size plots

4999274490_3ec894dc60_z.jpg






Next is the 80 series rack.  This is an 85-B1 preamp, a pair of 85-X isolation amps, and an 87-A.  I left all three of these unloaded, so they see as high as 10K on their outputs.  I found no obvious frequency changes in any of these amps with varying output loads.  Distortion will of course change.  Again, these amps will see different output loads depending on the signal path needs of the day, sometimes being 600, so I don't want to double load any amp if unnecessary. 

The 85-B1 is a single 1620 triode amp.  Identical to the preamps found in the 76 series consoles.  There are two plots here, one with direct 150 ohm line source, another with a 20 dB U pad.  I suspect my source has fairly weak drive at 150 ohms, and may be the cause of the greater top roll-off.  The load may be re-equalizing the source.  I suspect a 50 ohm source with a 100-150 ohm build out resistance would give a different plot.  The pad exhibits the typical top end boost that an input shunt resistance sometimes causes.  With the BA-1, 85-B1, and 87-A, the shunt in the pad was expanded experimentally up to 1K and the top boost remained, so I see no ability to trim it to perfection.  All original parts in this one; no electrolytics involved. 

The 85-X is a single 1620 pentode amp with plate to grid feedback, feedback variable from about 11 dB to 18 dB.  It's input is tapped at 20K and 600 ohms, and was intended primarily for unity gain isolation at medium line levels.  At the 20K tap, gain ranges from slightly above to slightly below unity.  At 600 ohms you get roughly 12-19 dB gain.  This plot is the 600 ohm input at maximum NFB.  There's a bit of bumpiness which I found later to be radio interference from a poor ground.  I apparently didn't save the 20K plot, but recall that it had a top rise similar to the 85-B1, and more bottom roll-off.    The second 85-X matched in measurement mostly, with a little less bottom end.  All original parts in these; no electrolytics involved. 

The 87-A is two triode 1620's with interstage volume pot.  There are two plots here, one with direct 150 ohm line source, another with a 20 dB U pad.  Same comments as for the 85-B1 and BA-1.  As a side note, these are found in the field with one side of the input transformer grounded, for use as turntable amps.  When connected this way, you get a fairly large rise in the high end.  New caps in this one, excepting the 330 pfd mica high boost cap in the interstage connection. 


Click this link for the full size plots

5003782507_ed28610c6d_z.jpg

 
Excellent work!  Now that's one big happy family  ;D


For those ruing the endless barrage of clone posts lately there's a small textbook of info in this post that goes beyond racking - probably about 10 separate threads worth.  If you take a look at the original schematics for each of these circuits a number of good questions will pile up - some are covered well here.  Plenty of interesting discussion to go forward on.
 
so many tubes and iron - makes you want to build something...

Thanks for posting
 
Nice!!!

This is the same equipment I had to play with when I started in the broadcast business almost 50 years ago.
Thanks for the schematics as well.
A lot of RCA stuff has been lost over the years and very little information exists on the 'net.'
I still have several pieces of RCA gear including a console and a BA6A.
 
I've been recollecting RCA info and stuff for years.  I need to publish something in linear form when I can.  There's a fellow in Australia who's compiled an amazing amount of info from catalogs and other lit.    These are what I used to consider the old stuff, until I learned about the really old stuff, so I've been chasing that for awhile too.  I have a pair of their first AC powered monitor amps from the mid-1930's, PPP 45's and a motor assisted volume pot for remote control(!). 
 
very purty. Love the baklite knobs.. Do you have any RCA lit, drawings, etc for any of the speaker line? most notable the bass enclosures mi-9462 aka ubangi as wellas the lc9a's?
 
There are hi-fi sites that follow the speaker stuff.  I really haven't followed it, other than casual observation.    Worth looking here for starters:

http://www.oswaldsmillaudio.com/archive.html

There's a hi-fi oriented forum there too, normal warnings/etc, but some people with a solid technical grasp, and some who either retired from RCA or worked in theater install for years. 
 
thanks for the link although I have already been around there... no one has any cabinet drawings and such just the standard promo lit.  Allen sides was  rumored to have ripped off the cabinet designs for said models but then again a hole in a wood box with what looks like a bass horn could be anybody I reckon. thanks though... keep up the beautiful restornation work
 
The site moderator there mentions comparing notes with Allen Sides, and goes into a bit of detail about it.  It's here:

http://oswaldsmillaudio.com/forum2/index.php?topic=384.0
 
Effects of feedback variation in the BA-11A.    Also included is a plot of V1 bypassed, with input transformer feeding V2, with no NFB loop present.  NFB resistor value can be seen in the plot names, items with checked 'V' are included.  

Full size plots here.

5074483798_45da75f7b3_z.jpg


The two stock NFB resistor values I find in print are 100K and 150K, which measure as roughly 24-27 dB NFB.  

It appears a value around 500K gives the flattest response on the top end, once the 22 pfd NFB cap is removed.   I can't find the 22 pfd cap doing anything with a NFB value equal to, or greater than the stock value.   With much less NFB, it kills the highs, being in parallel with the NFB resistor.  

The response without NFB is shocking; the amp really falls apart with less than 12 dB of NFB.  Here NFB addresses the use of a pentode front end with these surroundings.  The top end looks fine in the single triode plot (white).  
Digging in RDH3, on page 5, regarding pentode amplifiers:
1) plate R affects upper freq limit
4) miller effect of subsequent stage has a more pronounced effect than with a triode source
500K plate resistance given as having upper freq limit of 5KHz, generally speaking
 
Not much useful to report here, and I didn't store many plots on this one.  This one is a real house of cards.  

White and blue are the standard NFB R of 150K, with blue showing the removal of the 33 pfd NFB cap.  

Yellow is response with NFB removed, a little over 22 dB total at 1kHz.   Treble response still good due to dual pentode connection?

With the 33 pfd cap present, you can't go any higher than 200K before highs begin to attenuate.  If you remove the cap, you can go to about 400K (7.5 dB less NFB) before bass begins to suffer.  

Greater than stock NFB immediately gives a boosted bass and treble resonance, so there's not really anything useful in this direction.

It should be noted that these transformers perform fine in the BA-2 triode amp, with no NFB loop at all.

full size plots here.

5074483864_1e31d68226_z.jpg

 
Whoa - that's a big difference.  Could you pass audio signal cleanly with no FB?

Any idea how quickly THD rose with higher FB resistor?  I couldn't get my bench version to even ~ 185K before distortion was unbearable and attenuating input signal only helped to a point.  With stock FB resistor it sounded excellent.  I never bothered to go lower with it.

What's your take on the decisions to use 2 pentodes to get 40db vs 2 triodes to get 50db? - BA1 vs BA2
 
No listening tests yet, just straight measurements. 

Can't address THD, no way to measure it and get this much done in a single session.  At this point I'm chasing useful freq variation points as benchmark before going further.  Looks like there's no point with the BA-1, while the BA-11 has potential range.  Long before I had any way to make detailed freq measurements, I put a 2M rheostat in a BA-11, and it sounded fine, though with obvious sonic change at the extremities.  Important to note that I don't mind control ranges going too far, with the acknowledgement that I will generally drive somewhere in the middle of the lane rather than in the ditches. 

BA-1:  1st RCA broadcast preamp module using loop NFB.  They wanted more NFB than 2 triodes would allow and still hit the industry standard 40 dB gain point.  Can't address why they went with a pentode output stage also; they never did it again.  Given the specificity of the feedback tuning I am seeing, I can guess why they went back to a triode output stage in the BA-11. 

BA-2:  they needed an amp with variable gain control for turntable preamplification, so they stuck with the classic open loop triode plan, and an interstage volume control.  Note that typically preamps did not have any sort of gain control, and the BA-2 is a special case. 

Previous to the BA-1, RCA ran feedback loops in power amps, the OP-6 remote amp, and some other higher gain voltage and program amps.  The BA-3 program is the first of the broadcast types to incorporate feedback loops. 
 

Latest posts

Back
Top