Converting signal supply +24 designs to bipolar PSU

[Potato Cakes]

Hello, Everyone,

Lately I have been trying to figure out how to redraw various BA circuits to work with bipolar power supplies. Since my grasp of the what's and why's when reading a schematic it tenuous, I have been making comparisons to the 2520 op amp. My initial thought is that the 0V connection on a BA card could be swapped with a -VDC source, but things like an additional reservoir caps would have to be added but then possibly the need for a bias trimmer could be done away with. The thing that conceives a bit of uncertainty is what will still need a ground connection. Here is a list of items I've made so far that will need an earth reference:

XLR Pin 1
Input transformer secondary (-) polarity
Output transformer primary (-) polarity
VDC reservoir caps
Input Gain switch/pot resistor reference
Input zobel network (if needed/desired)

Some 2520 schematics use a ground reference for a cap between the input transistor and others have trim connection. But versions like the SL-2520 do away with the ground/common connection all together, which makes me believe that the same could be applied for the BA card minus needing reserve power caps.

It also seems that a +/-20 - 24V power rails can be handled easily with modern components as BC550Cs and MPSA18s can handle 45V and DB139/140's can do 80V or if a TO-204-3 package is needed newer replacements for the 2N3055 are rated anywhere from 80-140V. The voltage ratings for caps would naturally need to be raised accordingly.

Another thing that is beyond my understanding is how (or if) to set the bias current as I have seen on at least one version of 2520 schematic redrawn that does not have a ground common connection.

I know this can be done and I also know it's not as simple as subbing out the 0V connection on a BA card with a -VDC, but from just making observations I don't feel that there is too much to alter this circuit to work with a bipolar PSU. Those few things, however are beyond my ability to understand. I would love to hear some thoughts if anyone has knowledge on the subject and would like to share.

Thanks!

Paul

 

[PRR]

Some lack of basics here.

To amplify audio, ALL you really need is one power supply. You bias the major audio points (especially the output) at about half the supply, and swing both ways from there. To get audio out without leaking DC, you use coupling capacitors.

Meanwhile in another universe, people were computing problems where the answers could be plus or minus, and might stay the same for indefinite time. Operational Analog Computing. To do this you need plus and minus supplies around common (zero, ground).

Many bad things happened in the 1960s. One of them was audio-guys getting hold of modular transistor op-amps from analog computers. Yes, a fast computer opamp "will" pass audio. And since it can be set so In and Out idle at zero DC, "we can avoid all those coupling caps!!" Well, it didn't work out that way. We NEVER want "audio down to DC". Our speakers can't make DC pressure in a room, and they strain trying to reproduce sub-sonic garbage.

But it got worse. '709 and '741 chips fell below $10, below $1. And op-amp thinking audio-guys made very good audio amps in computer DC-coupled style, 990 and 2520 etc. And op-amp use can be quite mindless. Soon all of audio was infested with DC-coupled amps, but necessarily coupled with caps to drop <20Hz.

There's one great place to use DC-amp thinking in audio. A *stereo* loudspeaker amp. With single supply you need one power cap and two speaker cap, all big. With +/- supply you need two power caps but no speaker caps. 2 is less than 3. And at speaker current, the older caps were clearly flawed. You could deliver a wopping 10% more power OCL (hey, looks big in the ad) and better damping at 20Hz. OK, when a transistor went short it took-out the speaker too. But 2<3 yada yada.

While basic amplifier design runs the same for one or two supplies, nearly all the details differ. Trying to hack a single-supply amp onto a dual-supply may be like putting two more wheels on the bicycle. Yes, we do have more cars than motorcycles; but it's been a long time since "cycle cars". (OK, my 2002 Honda sedan gearbox shows motocycle roots, but it is the last of that line.)

 

[Potato Cakes]

A lack of basics is basically all I have when trying to understand audio circuit fundamentals. I do appreciate the historic notes as I like to know why things are/were done to combat marketing hype and untrue things that are said by others I work with in the audio field.

What had spurned this line of thinking was I am working on builds where single source and bipolar PSU requirements exist in the same unit. I had seen that newer designs like the RND Shelford series runs on +/-24V which boasts benefits of more headroom and other claims attributed to modern electronic designs. While I am not trying to recreate/copy what they are doing, it did get me to think about what/if there would be any benefit to adapting a single source +24V design to a bipolar scheme and how that could be possible, which then led me to start digging around and comparing schematics to see if any answers would be forthcoming by finding similarities.

The other question I now have is would there any audio benefit to simply increasing the voltage used by the same single source circuit? Modern parts have an overall higher voltage rating and use the same physical space (sometimes less) than the same component values used when classic designs were new, so then something like a BA could run on +48V. I don't know if this a solution in search of a problem, but it would for me (with sub basic circuit knowledge) make accommodating all the different power requirements a little easier.

Thanks!

Paul

 

[Fuzz Face]

A lack of basics is basically all I have when trying to understand audio circuit fundamentals. I do appreciate the historic notes as I like to know why things are/were done to combat marketing hype and untrue things that are said by others I work with in the audio field.

What had spurned this line of thinking was I am working on builds where single source and bipolar PSU requirements exist in the same unit. I had seen that newer designs like the RND Shelford series runs on +/-24V which boasts benefits of more headroom and other claims attributed to modern electronic designs. While I am not trying to recreate/copy what they are doing, it did get me to think about what/if there would be any benefit to adapting a single source +24V design to a bipolar scheme and how that could be possible, which then led me to start digging around and comparing schematics to see if any answers would be forthcoming by finding similarities.

The other question I now have is would there any audio benefit to simply increasing the voltage used by the same single source circuit? Modern parts have an overall higher voltage rating and use the same physical space (sometimes less) than the same component values used when classic designs were new, so then something like a BA could run on +48V. I don't know if this a solution in search of a problem, but it would for me (with sub basic circuit knowledge) make accommodating all the different power requirements a little easier.

Thanks!

Paul

My understanding is that the op amp output cannot exceed the supply voltage swing, so increasing the supply voltage simply increases headroom by allowing a greater max output swing

I am certainly not one of our more knowledgeable members though, so I look forward to being corrected/learning more about this subject

 

[PRR]

> allowing a greater max output swing

How much swing do you need to smoke the input of the next box in the chain?

 

[Fuzz Face]

> allowing a greater max output swing

How much swing do you need to smoke the input of the next box in the chain?

I’m not saying the output should reach the max potential swing in practice, but isn’t this the reason for increased headroom?

 

[dogears]

ok. Double your voltage you get 6 more dB. Worth it?

 

[Potato Cakes]

I've finally managed to get back to this figuring out this silly idea of mine. To answer the last question, it's not necessarily about achieving more headroom (although that's not a bad thing), but it's about having being about to use bipolar power supplies with a tried and true audio circuit which will simplify integrating other circuits like those using ICs to make what can be integrated in a build much easier.

I took a BA283NV circuit I had and rigged it to a +/-15V SMPS that I am trying out for a prototype build and hooked it up to the circuit as is and powered up with no complication and passed VAC as expected. I even connected a trimmer for adjusting the circuit gain (top half of schematic, T - R10 [trimmer] - V [ground] as shown on the original 1073 schematic) and that functioned properly even though in this test circuit the V connection point is now -VDC and not ground. Also of note, there is no VDC measured between the U (audio in) and V (now -VDC).

The board to board connections for the BA283 style amplifier I am working with will still need shielded cable for noise purposes and now the shield will need to directly to chassis but not to the PCBs since I have to take in consideration the other circuits I plan on including which will be using ICs and/or DOA. The input transformer's secondary - polarity connection normally ties to ground with the V connection in the original schematic, but as I have stated the V connection is now -VDC. Should I still tie the input transformer secondary (-)audio connection to V as prescribed? My initial tests with just my signal generator's XLR output suggest that I do, and that the fact I am not getting any VDC between U and V with my bipolar PSU scheme suggests it would be safe to do so. However when making connections to other circuits that are using a +/-VDC plus ground, it would normally use the ground as the reference for the audio signal. So my other question is if I tie input transformer secondary (-) audio connection to ground, can I get away with just connecting the audio at the U connection on the BA283? If not, would there be any benefit to buffer the input transformer (-)secondary with an electrolytic cap so that the buffered side can go to the V connection and the non buffered side would go to ground and provide the audio reference for the other IC/DOA circuits?

Thanks!

Paul

 

[Potato Cakes]

I did some listening tests tonight and it is dead quiet and handles full audio bandwidth program material with no problem at all. I wired up a 10468 parallel on both primary and secondary windings (+6) and even with about 28dB of gain added in the circuit it still reproduced accurately whatever audio I passed through it. The input transformer secondary was connected to V which as stated earlier is now -VDC.

I had to wire the output to a DI which seemed to provide the correct impedance to the headphone amp I was using. I had initially tried an LO1166 but it seemed to load down the circuit way too much.

Now that I know my theory works, I'm going to apply it to my prototype and see if all of my dreams will come true...

Thanks!

Paul