CC = Cascoded
CCS = Constant current source biased
X = Supersymmetric (not that I have to explain that)
BZLS = (Well ... everyone knows what that means).
First, I must embarrasingly admit that the driving force behind completing this project was to provide myself a preamp with balanced outputs to work with my Supersymmetric GC amp project. My saving grace is that it wasn't the only reason, and over the next few months I plan on giving it some Mini-A's for company.
Some may ask "Why add the extra complications to the BZLS when you could have built an Aleph P1.7 for less effort?" The answer is that, while I am an engineer I'm not an EE, and I use my projects to teach myself. While Nelsons circuits are first rate designs, I don't learn much from the paint by numbers approach. So I try to throw a change into each project that I hope will produce an improvement, and those changes usually require me to learn in detail the workings of every part of the circuit so that I can interface my adaptations.
I should mention that much of what I have done has been derived from the previous efforts of Grey, Henrick and Alain Dupont, and of course none of this would be possible without the contributions of Nelson Pass. All I did was put the pieces together in a different way to make me think about it more. The new part that no-one that I am aware of had added so far was cascoding the differential pair, which I did to extend the bandwidth and to reduce distortion when working with single ended inputs or outputs.
For those interested, I have attached the schematic below. The only difference between what I indicate there and what I did in reality is that I substituted precision voltage references for the zener diodes in the CCS and cascodes.
CCS = Constant current source biased
X = Supersymmetric (not that I have to explain that)
BZLS = (Well ... everyone knows what that means).
First, I must embarrasingly admit that the driving force behind completing this project was to provide myself a preamp with balanced outputs to work with my Supersymmetric GC amp project. My saving grace is that it wasn't the only reason, and over the next few months I plan on giving it some Mini-A's for company.
Some may ask "Why add the extra complications to the BZLS when you could have built an Aleph P1.7 for less effort?" The answer is that, while I am an engineer I'm not an EE, and I use my projects to teach myself. While Nelsons circuits are first rate designs, I don't learn much from the paint by numbers approach. So I try to throw a change into each project that I hope will produce an improvement, and those changes usually require me to learn in detail the workings of every part of the circuit so that I can interface my adaptations.
I should mention that much of what I have done has been derived from the previous efforts of Grey, Henrick and Alain Dupont, and of course none of this would be possible without the contributions of Nelson Pass. All I did was put the pieces together in a different way to make me think about it more. The new part that no-one that I am aware of had added so far was cascoding the differential pair, which I did to extend the bandwidth and to reduce distortion when working with single ended inputs or outputs.
For those interested, I have attached the schematic below. The only difference between what I indicate there and what I did in reality is that I substituted precision voltage references for the zener diodes in the CCS and cascodes.
Attachments
Two transformers were the minimum requirement in order to supply the +70/-20 volts rails. I split the powersupply to dual mono after the transformers but before the rectifier bridges to minimize as much as possible the crosstalk without incurring too much expense. In the end I would not have had enough room in the chassis for the extra transformers required by true dual mono, so it was a good decision. I also provided separate star grounds to each channel, which are connected to chassis/AC ground via CL60 thermistors.
Attachments
Closer view of one channels PS board, which includes dual rectifier bridges using 1N4004 diodes, a CRC filter, and regulation before passing power out to the amplifier boards.
The result: with no input signal, the volume control set to maximum, the preamp connected to an amp with 35dB gain that is then connected to 89dB speakers, and your ears right next to the speaker, you can just hear a bit of white noise. I'm very pleased with the result.
The result: with no input signal, the volume control set to maximum, the preamp connected to an amp with 35dB gain that is then connected to 89dB speakers, and your ears right next to the speaker, you can just hear a bit of white noise. I'm very pleased with the result.
Attachments
The other complication I threw in was that I wanted to be able to switch between single-ened and balanced input and outputs just by turning knobs on the front panel, i.e. doesn't require insertion of jumpers into XLR connectors to enable single-ended operation. Thus the extra birdsnest if wiring between input selector switch (left) the balanced series attenuator (center) and the output selector (right). I think that decision will come back to bite me, but so far it works quite well.
Attachments
P3, A bias knob, cool. Adjustable from the front?
OK, I've decided if I make a CCS BZLS I'm going to have to have a bias knob on the front, sounds like too much fun.
OK, I've decided if I make a CCS BZLS I'm going to have to have a bias knob on the front, sounds like too much fun.
It´s very nice and constructed project you got there.
How do you define the gain in the circuit when you don´t use resistors in the sources?
Bets regards
How do you define the gain in the circuit when you don´t use resistors in the sources?
Bets regards
metalman,
I have a question regarding the current in the CCS zener,
with -20 volts minus 6.8 = 13.2 volts for the R18 = 10 K
that is 13.2/10000 = 1.32 mA I think a better value would be
around 5 to 10 mA to get better stability... so I suggest 2.6 K
plus putting a 20 mF across this zener would provide lesser
background noise.
Otherwise its perfect! very nice job. 🙂
Tell us about the sound...
Regards.
Alain.
I have a question regarding the current in the CCS zener,
with -20 volts minus 6.8 = 13.2 volts for the R18 = 10 K
that is 13.2/10000 = 1.32 mA I think a better value would be
around 5 to 10 mA to get better stability... so I suggest 2.6 K
plus putting a 20 mF across this zener would provide lesser
background noise.
Otherwise its perfect! very nice job. 🙂
Tell us about the sound...
Regards.
Alain.
Nope, the left one is an input select switch, the middle is a stepped attenuator for volume control on the output, and the right one is the output select switch.
The case is from Lansing Enclosures. Great quality and I was impressed with their service. For people like me who literally build this on their kitchen table (condo living = no workshop) it is a great option. I will be using them again.
The gain is 13dB, defined by the X-feedback resistors R15 and R16 on my schematic. The source resistors were omitted to maximize the effect of the supersymmetry, a suggestion Nelson posted in another thread.
That is not actually a Zener in my circuit but an LM329 precision voltage reference, and according to its data sheet it should be in its optimal operating range at ~1.5mA. I tossed back and forth on whether to use a zener with something similar to your suggestions, versus the LM329, and in the end I tossed a coin. I suspect I will find myself trying some alternate arrangements over the next few months to see what differences the various approaches create.
As to the sound, well the most I can say at this stage is that it really woke up the Susy GC amp I have been working on. I don't have any other "audiophile class" poweramps to work with at the moment. To my ears (which I suspect are overly proud of my own creation at the moment) it sounds amazing detailed without being harsh or dry, with a wonderfully fleshed out soundstage. I now know what people mean when they say they can hear the space between the instruments. My wifes first comment was that it sounds smooth. I am going to strong-arm a few friends into letting me plug it into their megabuck systems and see how it fairs.
Thanks everyone for all the compliments. I'll post a few more pics once I have the control shafts, knobs, and of course the Blue LED installed.
Cheers, Terry
metalman,
Very nice post. You are going to inspire me to get going again on my own BOSOZ. Could you give us more technically-challenged lurkers the details on your power supply? Thanks! 🙂
Ren
Very nice post. You are going to inspire me to get going again on my own BOSOZ. Could you give us more technically-challenged lurkers the details on your power supply? Thanks! 🙂
Ren
Ren,
Sadly, I don't have that schematic drawn up, but if you can wait a bit, I'll put it together and post it. Actually, I'll post two: the one I built, and the one I'd build if I were building another. Bear with me.
Cheers, Terry
Sadly, I don't have that schematic drawn up, but if you can wait a bit, I'll put it together and post it. Actually, I'll post two: the one I built, and the one I'd build if I were building another. Bear with me.
Cheers, Terry
Wow, great project. Are that different transformers?
edit: I now understand, 2 different voltage rails.
edit: I now understand, 2 different voltage rails.
Hi Terry!
Really nice work.
But acording to my book, you need some inputresistors at say 10K in order to get X. The feedback needs this input resistance in order to dampen the input signal and thus make the X efficient.
And what a nice job you have done on the XGC.
Really nice work.
But acording to my book, you need some inputresistors at say 10K in order to get X. The feedback needs this input resistance in order to dampen the input signal and thus make the X efficient.
And what a nice job you have done on the XGC.
DOH!!!... again,
Sheesh, that's twice in a fairly short period that I've pooched a posted schematic. That's what I get for rushing a drawing in order to get some posting done.
Better put a leash on my enthusiasm, and add a review step to my circuit drawing!🙄
Henrik's observation is absolutely correct, and I do indeed have 10k resistors at the inputs in my functioning circuit. I'll correct this and post a correction shortly. Thanks Henrik, both for the correction and the inspiration.😉
I'm just working on getting my extension rods and knobs installed. This project is starting to look like a real preamp!
Sheesh, that's twice in a fairly short period that I've pooched a posted schematic. That's what I get for rushing a drawing in order to get some posting done.
Better put a leash on my enthusiasm, and add a review step to my circuit drawing!🙄
Henrik's observation is absolutely correct, and I do indeed have 10k resistors at the inputs in my functioning circuit. I'll correct this and post a correction shortly. Thanks Henrik, both for the correction and the inspiration.😉
I'm just working on getting my extension rods and knobs installed. This project is starting to look like a real preamp!
OK, thanks to Henrik's keen eye catching a discrepancy in the earlier schematic, here is the correct schematic.
To those who Nelson has referred here who are interested in using this circuit as a single-ended to balanced converter, please note that you will need to connect the -In terminal to ground for it to act optimally as a converter. Also note that the 10K potentiometer at the outputs is meant as a volume control. Without attenutation, this circuit has 13dB of gain.
One last point overall is that P3 can be adjusted to set the bias current. I optimized this circuit to run best with ~32 volt drop across the 820ohm resistors R3 & R4, which should be 5W rated.
Cheers, Terry.
To those who Nelson has referred here who are interested in using this circuit as a single-ended to balanced converter, please note that you will need to connect the -In terminal to ground for it to act optimally as a converter. Also note that the 10K potentiometer at the outputs is meant as a volume control. Without attenutation, this circuit has 13dB of gain.
One last point overall is that P3 can be adjusted to set the bias current. I optimized this circuit to run best with ~32 volt drop across the 820ohm resistors R3 & R4, which should be 5W rated.
Cheers, Terry.
Attachments
- Status
- Not open for further replies.