DCAudioDIY.com

Listening to the Topping E70V has revealed the need for a balanced connection between the DAC and preamp. Using the unbalanced outputs results in a less than satisfactory sound. Even using an XLR-RCA cable where half of the signal is thrown away and CMRR advantages of balanced interconnects are lost, makes a noticeable improvement.

So the need for a line stage with a balanced input is here. Looking at a few possibilities, the chosen one here is John Broskie's "Unbalancer". The urge to just buy one of his PCBs was strong, but unfortunately, they are out of stock. No reply from Mr. Broskie as to when they may again be available.

A couple of things were concerning about the overall design. Most notably, the input differential amp has an LM334 as the current sink, and it's tied to ground. The LM334 has a maximum current rating of 10mA. Less than the 7-8mA of each half of a 6SN7 type that would likely be used here. Never been comfortable with running a device at 100% of capacity and hoping for longevity. Exploring options and after speaking with Dave Berning, a simple bipolar transistor with pi network filtered zener reference has been chosen. Guy Riffle has been tirelessly simulating different scenarios, and we think that tying the sink to a negative 16-18 volt of so negative supply, using a 6.2 volt zener should work. Even in simulations, there was an improvement in noise performance adding the 1k/100uF filter between the zener and base of the 2N4400 (lots on the shelf).

The other issue with the original design was the unregulated, very basic power supply. Instead, we are using vacuum tube damper diode rectification, all film caps in the B+ supply, a 6Hy 150 ohm choke, a depletion mode MOSFET cascoded current source, then 0A2/0B2 gas tubes for a ~258 volt B+. Finally, two 150mH 2amp chokes followed by 10//0.1uF caps at the circuit loads. Again, the simulations look very good. (Using PSUDII).

Four inputs, reed relay switched, and two outputs, one effectively high pass around 100Hz with my crossover, and the other full range. Two inputs are balanced, two unbalanced.

A 4PDT-CO switch is being used as a phase reversal and mute switch.

Each channel has a separate gain control. They are 100k stereo stepped attenuators (from eBay). Very small and they use surface mount resistors. Should provide very close tracking +/-.

Using a simple MOSFET soft turn on setup with a 150k resistor and 100uF cap at the base of the MOSFET.

Heaters will be DC, but not regulated. Simple Pi network filtering. Negative 18 volt supply will be similar and also unregulated. The 5 volt relays will get power from the heater supply followed by an LM7805.

Using a chassis from an abandoned project: hardwood sides and 1/4" plastic top plate. So, yeah, true to form, it's not pretty!

Basic construction is done, creating circuit boards now: high voltage, low voltage, relay and amplifier.

It'll probably be a few more weeks until complete. Life does get in the way.

Anyway, not many discussions going on here, so figured this might be of some interest.

Stuart

BTW, there is an unfinished Izzy Wizzy phono stage here. It has transformer coupled output and no coupling caps. Seems like finishing that will be a great next project!

More pics

A great way to go from balanced to unbalanced or unbalanced to balanced is the use of two devices from THAT corporation. I have used them very successfully in the past and except for high voltage isolation, they have many advantages over transformers. I did a two part article on using them in July of 2009 published in audioXpress. I'm enclosing the PDF files on them and also a layout of the PCB is did. I have the ExpressPCB file for that if anyone wants it. That circuit will convert a stereo signal in either direction.

I'll read up on those Tom. The Unbalancer doesn't use any signal transformers. I'll post the schematic later. The Izzy Wizzy phono stage has a differential output stage. The plates are connected just like a push pull amp to an interstage transformer. It's been highly praised. We'll see how it sounds, whether the elimination of coupling caps, but the addition of a transformer is better or worse.

Stuart

_________________
Walt

If you like the KISS principle, hard to beat at DN2540 and a resistor for a CCS. Even the TO-92 package should be more than robust enough for your application...

Roscoe

_________________
I can explain it to you, but I can’t understand it for you.

Stuart,

Since you are committed to PP amps, the natural thing is to keep everything balanced (differential) to the power amp. No more phase splitter. It makes no sense to convert back to single ended out of the preamp and then split phase again at the amp. I know the OddWatt amps do the phase splitting in the output stage, so it would require a little rework to convert the SRPP driver into a differential, but hey, it is only resistors and perhaps another signal capacitor. But that is a system wide change and upsets everything including interconnects (right after you upgraded to RCA Audio Envy!). You should really look at Allen Wright's designs for a fully differential preamp/amplifier. Some really interesting and good stuff there.

All you are doing is executing in vacuum tubes what the Topping does with an opamp when it converts from balanced to SE after the DAC. The most basic circuit is a summing amplifier executed with a single opamp. Maybe the tube circuit will sound better, but maybe not. It is cheap to put something together with what Tom is suggesting. A pair of transformers is another solution that is easy to try without upsetting everything. Jensen and Cinemag makes some. I think I may even have a pair of Cinemag transformers that may work, I will have to dig around.

I have been toying with balanced in explorations of powerful versions of First Watt amplifiers (converting a stereo 25W Class A to a 100W monoblock), but have backed off because I realize that (1), I prefer tubes, and (2), a 30W tube amplifier does almost everything I need or want for the music I listen to.

David

Stuart, I wasn't saying the Unbalancer used transformers. It was just a general comment about the parts from THAT corporation compared to transformers. No doubt that transformers have the highest Common Mode Rejection Ration and can isolate higher voltages than ICs, but you have all the audio problems like saturation, leakage inductance, interwinding capacitance, hysteresis, etc. to deal with. The papers by Bill Whitlock mentioned in the data sheets are some of the best when it comes to describing the different approaches to both unbalanced and balanced. I have found in home situations in all the places where I have lived, I never needed to take advantage of the CMRR of balanced circuits. That is more important in industrial situations where there are long runs near many types of electrical equipment.

That works fine if I wanted balanced out. I may someday.

My first thought actually. I have lots. Dave B convinced me to try his solution. It works for him. I had all the parts.

BTW, you and I need to talk about Otaris.