Another RCA BA-2 Build

[dmp]

Building a RCA BA-2 preamp - without a level pot or the EQ RC, it has 54 dB of gain and sounds great.
But I have a little 60 Hz hum. I'm using a Antec Power Transformer.
http://www.antekinc.com/pdf/AS-1T230.pdf

Without center tapped windings on the transformer, I'm not sure the best way to hook it up. Should I add resistors to the heater windings to ground, or elevated from ground at some point?

 

[dmp]

Another pic

 

[dmp]

And the back

 

[emrr]

Cool, you scored some real iron.  Not easy to do. 

Why no level pot?  That's probably the #1 thing that makes people want a BA-2 over any other RCA. 

Use two 100 ohm resistors to ground from each side of the filament winding for starters.  You could use DC bias on that artificial center tap you are creating, plenty of descriptions around here. 

What concerns me most is the proximity of power transformer to output transformer.  Have you rotated power for lowest hum, and then also output?  Maybe even input in this case.  If I were starting over, I'd put the audio iron as far away as possible from the power iron.  If still an issue after adding artificial center tap, and trying iron in different orientations, consider adding another metal shield internally. 

 

[dmp]

I haven't decided on the panel controls. I definitely will have a stepped switch for level but not sure what else.
At this point I wanted to test the layout for problems - like noise / hum / etc...
Thanks for the tip, I thought I would need a heater connection to ground so I'll try the resistors.
Hopefully I won't have to put the PT in an external psu but it will be easy to test for interactions by moving it. This is the first preamp I've tried an internal psu.

 

[letterbeacon]

I found the first stage to be really susceptible to hum, and it was picking up a lot of hum that was in the PC that was sitting a few spaces below it in the rack.

It's a great sounds pre though!

 

[buildafriend]

Thanks for posting.

Are those the original input and output transformers?

If yes: How did you find them?

 

[dmp]

Are those the original input and output transformers?
If yes: How did you find them?

Yes, they are originals that I bought on ebay - 2 inputs and 1 output. About $150 each I think, so they were not cheap. But not a whole lot more than a good transformer from Jensen, etc.

 

[dustbro]

Were the original transformers from UTC?
if so:
http://www.magnetika.com/

 

[emrr]

Not UTC that anyone can prove. Unbranded. 

 

[buildafriend]

Are those the original input and output transformers?
If yes: How did you find them?

Yes, they are originals that I bought on ebay - 2 inputs and 1 output. About $150 each I think, so they were not cheap. But not a whole lot more than a good transformer from Jensen, etc.
With my day job I have much less time for this hobby than I would like so I don't skimp on parts, but try to make things that are really special.
If anyone has a really good need for an original input let me know.

I would consider buying your extra original input iron.. how much?

 

[dmp]

I added the 100 ohm resistors to ground for the heaters and the hum dropped significantly.
Measured there is still a bit of 60 Hz above the noise floor, but barely. I swapped tubes and the noise floor didn't change (but I only have 4 1620s total to try).
Moving the PT out of the case made no difference.

 

[dmp]

And the frequency response (white line is the BA-2, green line is just the soundcard as reference)

 

[emrr]

Sounds like your layout works.

It could be endlessly debated, but I'll state, assuming 1620's on hand are quieter than 6J7's on hand, that there's no reason to waste a 1620 in the V2 position.  I use them in V1 position, and V1 noise will always swamp any contribution from a good 6J7, and at half the price and 6x less the scarcity.  If we want to keep 1620's remotely affordable and available, it's good practice.  This is true here in this amp lacking loop negative feedback, and vastly more true in any design with loop feedback. 

I'm betting the low end bump is a tube self noise or interference artifact, at any rate what's shown on the extreme bottom end is not representative of this circuit or these transformers. 

Your top rolls off faster that what I've seen.  May be your in/out source/load conditions.  You should try the EQ network and see what it does for same conditions.  Also look at what response does with a typical 20 dB U pad.

 

[dmp]

I tried some measurements to see where the top end roll off was happening, and it seemed to be present at the input Tr secondary. The 60 Hz hum (and harmonics) was also present there, but I think that may be due to the ground layout, not transformer pickup. High end rolloff in the transformer would be stray capacitance to ground, yes? Or could a terminating resistance flatten it out?
My test setup was driving the input directly from the soundcard. Output Tr is terminated with 620.
I did not ground the primary of the input transformer as shown in the schematic. Could this be an issue?

Here is the measurement with the 330pf||33k filter before the 100k pot (white line). This did not sound good to my ears listening to music through the amp. Brought out unnatural high freq content. But it wasn't a blind listening test so I may be full of it.

 

[emrr]

It doesn't look like appreciable change in curve shape, only extension.  Interesting observation. 

Try a U pad in front, that should give a treble boost much greater than what you see from the EQ. 

No reason to ground the input CT, and don't follow the schem with the pri grounded as unbalanced. 

A terminating R won't help the high loss in terms of extension.  C to ground, yes.  Try reversing the input secondary and see if the plot looks better or worse; changes the C to ground relationship usually. 

Look at all my posted RCA plots for more examples if you haven't. 

 

[dmp]

Thanks Doug.
I'll try a pad and a phase flip on the secondary and look at the response. I'm not worried though - it sounds great as it is.
I guess it makes sense if a 1st order filter is causing a rolloff (low pass) that countering it with a 1st order filter (high pass) can give an offset in the total response. In this case the HPF (330pf||33k) has a 3dB point of 14kHz (i.e. boost +3dB at 14kHz, 6dB boost/octave). Below the HPF, the addition of the 33k gives a drop of 2.5 dB due to the voltage divider i.e. 20log(100k/133k). Seeing about 0.5 dB (3-2.5) boost at 14 KHz. Maybe that makes sense? I'd have to try a plot with the equations to be sure.
The LPF coming out of the input transformer seems to have a different (higher?) 3dB point.
Adjusting the RC to match the 3dB point of the inherent LPF might get rid of the bump in the 2kHz to 8 kHz range apparent in the white line.

One thing I'm confused about is I'm measuring about 54dB of gain. I thought these were max 40 dB gain? I connected the output to term 3-6 instead of 4-5, since it is labeled as 500 ohm. This would give a little more output, but not that much. Tube dependent since there is no neg feedback?

 

[emrr]

Supposed to be 50 dB 250 to 250, so 54 is right. 

filter calcs sound right from memory. 

I can't see a bump in the 2K-8K range, if it's there it must be 0.1 dB?  Any modern transformer will impart as much, easily. 

 

[dmp]

I can't see a bump in the 2K-8K range, if it's there it must be 0.1 dB?

You are correct - nevermind.

Also look at what response does with a typical 20 dB U pad

Is 'typical' the Jensen 619 / 169 / 619?
I've used 1200/220/1200 for most 20 dB pads in preamps.
Higher values will affect transformer response more? I'd assume so since the output impedance of my soundcard should be low. Not sure I get it though.

 

[emrr]

I found any reasonable shunt value causes a resonant bump, and you'll have to use an input pad with some sources, so it's two different sounds, any way you slice it.