The BA-3 as preamp build guide

Anyone try using a 48 Switching wall wart?

I am looking to build one of these BA-3 as preamp. I have already built the B1 and Korg B1 buffers with wall warts and they work great. I can't hear any noise anyway. I was thinking this Mean Well 48V 800mA . It is rated at 80mVp-p ripple and noise.

GS06U-11P1J MEAN WELL USA Inc. | Power Supplies - External/Internal (Off-Board) | DigiKey

I was thinking modifying the power supply for the LX mini active crossover for the higher voltages.

http://www.firstwatt.com/pdf/art_lxmini%20crossover.pdf

I searched this thread but haven't found any mention of this. If I can't hear any difference I'd rather skip the whole linear power supply thing. If it is noticeable I'd go ahead and do the regular linear supply.

Any comments will be appreciated.
 
BA-3 is split supply +/-, not single rail like Korg b1.

So unless you get a bipolar (split supply) wall wart or manufacture a divider it will not work.

Also, since the supply noise/quality is dominated by the Korg B1's distortion it is not that impacted by supply quality. Supply quality may be much more apparent in a lower distortion circuit. Korg B1 also has a filter after SMPS and before the circuit that BA-3 does not have.

Lastly, current demand of BA-3 preamp is MUCH higher than Korg B1. More than double.

It's possible to use a amps or wall wart with Ba-3...but it's more effort than with the Korg and will probably show the shortcomings of SMPS.
 
BA3 as an Integrated Amp with MoFo

Posted this is the MoFo thread as well. I thought it made sense to post here too.

After a short hiatus I jumped back on my MoFo build. Got it up and running today along with the BA3 Preamp. The combination sounds great so far. Playing on my 0.53x mini-karlsonators the BA3 and MoFo go plenty loud. The combo is also dead silent with no turn on thump. This is turning out to be a great integrated project. Still have to get the phono stage soldered up but that’s another project for another day. In the meantime I’m going to get some more listening time on the MoFo. Thank you Michael for a fantastic amplifier and thanks again to 6L6 for the BA3.
 

Attachments

  • CE8E43E4-10BD-4A71-807A-B71E05EFD7E2.jpg
    CE8E43E4-10BD-4A71-807A-B71E05EFD7E2.jpg
    1,006.1 KB · Views: 544
  • A65950BA-2C51-498D-B54B-5F254814C3BB.jpg
    A65950BA-2C51-498D-B54B-5F254814C3BB.jpg
    982.8 KB · Views: 533
  • 3D1AECD6-5972-40B9-90FE-B15671734755.jpg
    3D1AECD6-5972-40B9-90FE-B15671734755.jpg
    891.3 KB · Views: 520
Posted this is the MoFo thread as well. I thought it made sense to post here too.

After a short hiatus I jumped back on my MoFo build. Got it up and running today along with the BA3 Preamp. The combination sounds great so far. Playing on my 0.53x mini-karlsonators the BA3 and MoFo go plenty loud. The combo is also dead silent with no turn on thump. This is turning out to be a great integrated project. Still have to get the phono stage soldered up but that’s another project for another day. In the meantime I’m going to get some more listening time on the MoFo. Thank you Michael for a fantastic amplifier and thanks again to 6L6 for the BA3.




Cool, fill us in a little bit on the CPU? heat-sink thingamabob. There has been a thought bouncing around of a build in a pc case using, perhaps, water cooling or some such thing, it's just random neurons at this point, but it's there.
 
Aha... the select-3. When you talked about -2, I though you were wanting the AC switch.

How about this— comes with an attenuator, and input switching for 3 sources. I’ve found the fine/coarse setup in John’s attenuators to be wonderfully useable. Much better then you’d think, they aren’t awkward at all. FrontPanel


Just an FYI:



Been over a week and the AC switch and Front panel have not even shipped out yet. I did get the power supply kit in about 6 days, but nothing else. Oh ya, the kit was missing two caps as well. I just wanted to put this out there... The email said one would ship in 2-3 days and the other within 24hrs, but that was on the 6 and 7 days ago. If you need something in good time, you might want to look elsewhere.
 
Last edited:
Lowest possible voltage supply. Sharing PSU from Mezmerize B1 Buffer.

Hi - I'm building BA-3 as a separate/standalone preamp for F5T/F6 (that are currently being under construction). The preamp will contain BA3 and Mezmerize B1 Buffer as input selector + volume control. The B1 has high quality on-board power supply.... This led me to possibility of cost optimisation: increase B1 voltage (blue leds?) to the level that will not kill it and still be good enough for BA3.

Is such sweet spot possible to find? :scratch2:
 

6L6

Moderator
Joined 2010
Paid Member
(Zenglish translation --- the B1 can't take the voltage that the BA3 needs, and you'd need to add cascode transistors to the B1 to protect them from the higher voltage.)

BA-3 doesn't need the B1 in front for proper operation. How many inputs do you have to switch?
 
Hello people! I have the preamp up and running and I think I know what's going on, but wanted to get your input here before I make a change. First, I have noticed what some others have with the twitchy pots and some movement of the dc offset, but considering the output caps, I think we are okay with that. IN fact, as I come up on the bias, I am seeing less drift, but then I hit a wall... the wall is, I think in the power supply. To be more precise, my choice of the RC value in the supply.



Supply is the Glassware Dual/Bipolar BA-3 latest version from the DIY store Parts from the BOD here and the C11 are Clarity CSA 10uF


(Glassware) The documentation states if the load is light, a 10R can be used, The RC would then be 4700uF and .01uF and 10ohms. (I used 10R) I didn't know, at the time, that this amp was to be biased to 1v! Wow!



When I bias the circuit up to around .6 volts the voltage drops back to zero or close to it until I move the pot back and then it comes back up, but trying to go higher induces, what I think is an over current in the supply.



Proposed action is to decrease the value of R to 3.3r or 1r which is what John suggests for higher current load. I could just parallel a 3.3R with it to bring it down from 10R to 2.48R and see what that gets me, or do you suggest just going back and removing the 10R and replacing it with a 3.3 or 1R? I still don't understand all the math. Perhaps, one of you can share your thoughts....



thanks


jt
 
Last edited:
Confirmed as the supply not being able to supply enough current. I temp soldered the 3.3R on top of the 10R for 2.48R and the supply did not bonk when the bias was turned up to 1v. This is the first channel she's stable a about 0-15mv dc. That damn pot is twitchy as hell and the dc moves around a bit, but with the top on and about an hour it settled pretty well at about 13mv offset at 1.012 bias. I don't think I will try and get the offset any better as it was a bit like a monkey F'n a football trying to get it there. :)


Anyway, for those who build this amp, don't bother chasing your tail until you get a cover on it and let it cook for a couple hours. I found that once you get the bias up there and allow everything to heat up good and stabilize, that the offset will settle down.


Before I figured this out, I would set it and watch it go from close to zero to 100mv or more and from the depending on the pot, it would go + to - or visa.



My advice is to take 6s advice and run one side to .750 and keep and eye on it only making corrections if the dc gets out of control. After a couple hours with the lid on get in there and get it close to zero. If it stays + or - 20mv dc leave it be. As I was typing this it was about 20mv and has fallen back to about 6mv. I will let it run the rest of the day keeping an eye on it and then proceed to the second channel.



I have read about a few things people have done with resistors and different pots to make the adjustment a little easier, but I'm not going to go down that road. I do think that this design is very sensitive to fluctuations in temperature, so it does give me some pause to give it to my friend, who will have it in his man cave/garage. I will take the meters over when it starts getting warm out and see what I see.



This was my 3rd build 2nd Pass and I thought I had something wrong because of the drifting DC, fear not new builders, she is sitting at 1.2mV as I close here. :)
 

Attachments

  • 20191104_073924.jpg
    20191104_073924.jpg
    147.3 KB · Views: 448
Last edited:
(Zenglish translation --- the B1 can't take the voltage that the BA3 needs, and you'd need to add cascode transistors to the B1 to protect them from the higher voltage.)

BA-3 doesn't need the B1 in front for proper operation. How many inputs do you have to switch?

The min. is 4. I know that B1 in from of BA3 is redundant, but is appeared to me as an elegant solution for vol and input selector. Are there any negative side-effects by combining the two?
 

6L6

Moderator
Joined 2010
Paid Member
I do think that this design is very sensitive to fluctuations in temperature,

And so does Nelson, which is why it's got the big output capacitor. :yes:

:) :) :)

When setting bias it really helps to short the inputs and have some kind of load on the outputs - grab some 2k or 5k or something in that general ballpark and attach from output top ground so the circuit has something to look at.
 
And so does Nelson, which is why it's got the big output capacitor. :yes:

:) :) :)

When setting bias it really helps to short the inputs and have some kind of load on the outputs - grab some 2k or 5k or something in that general ballpark and attach from output top ground so the circuit has something to look at.


Thanks 6, I think I read that on page 5274. :santa2: I will try it when I bias the second channel tomorrow. :D
 
Well, she's up and running... have it, for testing, plugged into the TU-8600. The combo it pretty darn sweet. Running it into my Sen HD800s for now, it's after 2am here, so the speaker test will have to wait for a more reasonable hour. Next step to take it over to my friend, for the F5T I built him.



Bonehead Alert: What I thought was the power supply running out of gas, was not. Hate to admit it, but the new meter I was using to keep track of the offset has an autoscale... Yup, it was switching form mV to Volts at about 600mV which I still don't understand, but that's where it makes the switch to volts. So.... Ya, a bonehead operator error. I have since pulled the 3.3R out of the circuit and went back to 10R and she works fine.



Note: Anyone using the Glassware front panel and switch, I would recommend you get the fine adjust resistors. I have the course and I think it's a bit... well... course. :p
 
Last edited:
I just finished building a Mezmerize dcb1. The board has a dual polarity power supply and input switching. Just leave out the buffer FETs and up grade the power supply to handle the BA3 .I see it as an elegant solution to input switching and power supply.

That's a really good idea. I'll save matched FETs for other usages.

This moves me back to question: what is min reasonable BA3 supply voltage?
With blue/white leds I'm able to get 16-17 V... is it enough?
 
That's a really good idea. I'll save matched FETs for other usages.

This moves me back to question: what is min reasonable BA3 supply voltage?
With blue/white leds I'm able to get 16-17 V... is it enough?

It will work. But I think over 20 or 22 would be better.

Also, you need to change the current consumption of the shunt...BA-3 consumes way more current than B1.
 
I am currently attempting to bias my BA3 and having some issues. Steps I took:

- Powered up with variac (no smoke)
- Confirmed +-25V at the rails
- Put one Voltmeter across R11 & another Voltmeter across R10
-Put a third voltmeter at Node of R12 and C3 Ground
-I turn one Trimmer at a time one turn. P1 then P2, P1 then P2, until i start to see a reading on my voltage meters.
- If I continue doing this method, I am unable to get to 1V and they max out at around .400mv

Am I supposed to be trimming down the offset as I go? And which trimmer would use to lower offset?

I currently do not have inputs or outputs hooked up to the Ba3 and using FQP3 Mosfets with 1K trimmers.

Any help would be appreciated..