1567 Copy

[Andre]

Hey Everyone,

After some time I've finally wrapped up one channel of a tube preamp based on the 1567a schematic.  Here are some pics and notes about the build...

So... What is it?
It will be a 2 channel tube preamp, with each channel closely resembling a complete 1567a mixer circuit, except that only 1 mic input is included for each of the parallel circuits.  The other mic input channels have been replaced by a resistor which creates a voltage divider similar to the parallel resistance of the mixing resistors when all other channels are turned all the way down.

About the transformers:
-The input and output transformers are Cinemags.  The input is wired 150R:50k and the output is 10k:600R.  The output differs from the original Altec which is 15k on the primary.
-The power transformer is a custom Edcor # XPWR018 - Please feel free to use this design without having to pay a customization fee.  This tranny was designed specifically for tube mic preamps.  It has 120V primary and secondaries at 230-0-230 (50mA), 0-40 (100mA) for phantom, and 6.3-0-6.3 (2.5A) for use with 12.6 V heaters.

Power supplies:
There are 4 power supplies:
-A regulated phantom power supply provides 44VDC.  It's breadboarded and mounted on the right side of the chassis.
-The main B+ power supply is next from the right and built on turret boards and resistors hanging off of the multi-can caps.  This starts with 230-0-230 VAC through a full wave rectifier.  It's filtered and then split down parallel paths to provide separate power for each of the channels.  It's output is about 10V less than what's listed on the schematic.
- The filament power supply is a made from the Slow Start Heater circuit sold for the Poorman's 660 build.  This is fed from a 12.6 VAC going through a bridge rectifier.  After much trial, errors and more errors, I found that I had to run the heaters at 12V, which means the 6CG7 filaments are in series.  As expected the voltage drops across the 6CG7s are not perfectly even, but this shouldn't cause an issue.

D.I.
Direct injection is done with a switched phone jack that cuts out the mic input transformer and feeds the instrument straight to the tube.  There is no 600ohm line input.

Sound.
I haven't been able to put it through much more than a couple of quick scratch tracks, but I'd describe it as glassy and effortlessly full.

Noise and hum:
There is a -84 dbfs (using Reaper's analyzer) hum that appears at 60 Hz and odd multiples of 60.  The hum moves to even multiples if I remove the second tube, which makes me believe the hum is due to power supply ripple.  If it bugs me enough I might trace it out and increase B+ filtering.  The unit doesn't have any issues with buzz or noise.

Gain:
Too much.  The original unit spec sheet indicates 92 dB gain.  I'm thinking about changing the second gain pot to a resistor and pot in series so that I can use more of the pots rotation.

That's enough for now.  Hope you enjoy it.

- Andre

 

[user 10827]

Hi Andre,

fantastic to see different stuff being built, and 92dB gain! woah, that is a lot, that's almost a weapon...

I'm not familiar with what a 1567 is, was it a Collins or some thing like that?

Best
Pete

 

[Jim50hertz]

Altec?  Looks verrrry cool, nicely built too!

 

[Andre]

The original 1567a is a 4 channel mixer made by Altec in the 60's.  It's become pretty popular to use as a standalone mic preamp.  Unfortunately I don't have an original to see how the copy compares.

 

[emrr]

Nice work, looks great. 

2 channel tube preamp, with each channel closely resembling a complete 1567a mixer circuit, except that only 1 mic input is included for each of the parallel circuits.  The other mic input channels have been replaced by a resistor which creates a voltage divider similar to the parallel resistance of the mixing resistors when all other channels are turned all the way down.

The original unit spec sheet indicates 92 dB gain.  I'm thinking about changing the second gain pot to a resistor and pot in series so that I can use more of the pots rotation.

Did you try losing all remnant of the original mixing bus resistances?  I would if you haven't; lose the unnecessary losses, and see if you get better noise levels and sound.  Noise already seems great for something with so much gain. 

Try losing R17 through R21, and lose either C2/3/4/5 or C6. 
C6 value more appropriate for R22 value, but I might consider losing C6 and lowering R22 value appropriately, so long as it doesn't throw off the EQ network.  I see potential advantages to that approach. 

You might consider altering the feedback network for less overall gain. 

There are other options I see, but they would necessitate a different panel.    :-\

 

[Andre]

Nice work, looks great. 

Did you try losing all remnant of the original mixing bus resistances?   I would if you haven't; lose the unnecessary losses, and see if you get better noise levels and sound.  Noise already seems great for something with so much gain. 

Try losing R17 through R21, and lose either C2/3/4/5 or C6. 
C6 value more appropriate for R22 value, but I might consider losing C6 and lowering R22 value appropriately, so long as it doesn't throw off the EQ network.   I see potential advantages to that approach. 

You might consider altering the feedback network for less overall gain. 

There are other options I see, but they would necessitate a different panel.    :-\

Thanks for the compliment.  I did cut out R17, R19, R20 and R21, along with C3, C4 and C5.  However, I added a 100k resistor to ground from the junction of R18 and C6.  This was chosen to approximate the voltage division that any one channel would see through the mixing network if all the gain pots were at zero.

It did cross my mind that I could eliminate C6, but in the end I figured it was safer to keep with the original design.  (O.K. so you can call me a prude.)

I'll agree that the noise level is quite acceptable.  Any changes at this point would only be done to reduce the gain.

Please share your ideas for altering the feedback network (decrease R32 and modify C16 maybe??).  I'm also curious about your other options, though you're right to guess that I'd rather not modify the front panel.

Thanks,
- Andre

 

[emrr]

You don't need R18 (the main one to eliminate), the extra load resistance, or C2/C6 both. 

My ideas about altering anything always involve a pile of parts and a soldering iron with test gear connected.  Every change will require looking at result and any interactions, and I haven't tried any of these. 

I'd consider an option to bypass V1 and connect the input transformer directly to V3, for lower gain usage. 

R29 is most of the feedback resistance.  C14 and C16 would need to change, with C14/R32 being most likely to need tweaking since it is designed to shape treble response, and probably specifically with the original transformers.  If you shorted R32 you would only really see the change in the top end from C14.  You could probably get away with sticking a 2M rheostat in as part of R29, dual 1M pot wired in series, reverse log if you can find one, probably linear if you can't.  Better, figure out the step size and put it on a 12 position rotary switch. 

R22/R31 grid resistances are very high, to maximize gain.  I would be tempted to lower them, and increase coupling caps appropriately.