Just a short report - I finished building the B1 3 days ago and was listening to it for some hours.
Essentially I picked all components from my part-inventory, so some adaptations were required to make the best use of what I had in stock.
First about the power supply. I build a shunt regulator based on the concept of those "super-reg". 27-28V AC was rectified with 4 1N4001 diodes, then followed by 2 stages of RC filtering. An IRF9610 constant current source followed the RC stages, with a current set at around 18-20mA. The shunt regulation was performed by a TL431, while a PNP transistor was used as the pass element. The feedback network of TL431 gave 18V DC output, an also a unity gain at AC for stability reason. Circuit is attached here.
Then the FETs used were 2SK240 GR dual FETs. These are very old stuffs, running at about 5mA IDSS. Essentially these were 2 2SK170s packed in 1 package, hence they are pretty well matched and thermally-coupled. The other changes I made were essentially related to the availability of the components in my inventory: Output cap was 2x2.2uF instead of 10uF. The 9V voltage divider bypass cap was 3300uF instead of 15000uF. Also I omitted the powerline RC (R1, C1) and put a 1uF cap across the power/GND in each channel as decoupling.
For audition, my benchmark is my current setup: CD player --> 50K stepped attenuator --> j-PHL (PLH with input IRF610 replaced a pair of 2SK147 jFETs) --> Fostex FE168 sigma double-bass-reflective box. This is more or less the simplest you can get sound out from a CD. Replacing the 50k stepped attenuator with this B1 (also with a 50k stepped attenuator), I started to compare the sound of the 2 setups. Frankly I cannot hear any difference, maybe with the B1 the attenuator gives more robust control of the amount of the attenuation. This is no surprise as in my original setup, the input impedance of the j-PHL (which is about 14k) will interfere with the R-ladder and alter the attenuation to a certain extent. Nevertheless, the clarity, the way the music comes to you, are very much equal.
This is good news, as the B1 suppose should be transparent and do not add anything other than impedance matching. It seems to be doing its job nicely.
Lack of bass reported by some people here? - no such observation. I guess if we adhere to the basic rules and make sure no mistake made, B1 should live up to its promises.
Good luck!
Essentially I picked all components from my part-inventory, so some adaptations were required to make the best use of what I had in stock.
First about the power supply. I build a shunt regulator based on the concept of those "super-reg". 27-28V AC was rectified with 4 1N4001 diodes, then followed by 2 stages of RC filtering. An IRF9610 constant current source followed the RC stages, with a current set at around 18-20mA. The shunt regulation was performed by a TL431, while a PNP transistor was used as the pass element. The feedback network of TL431 gave 18V DC output, an also a unity gain at AC for stability reason. Circuit is attached here.
Then the FETs used were 2SK240 GR dual FETs. These are very old stuffs, running at about 5mA IDSS. Essentially these were 2 2SK170s packed in 1 package, hence they are pretty well matched and thermally-coupled. The other changes I made were essentially related to the availability of the components in my inventory: Output cap was 2x2.2uF instead of 10uF. The 9V voltage divider bypass cap was 3300uF instead of 15000uF. Also I omitted the powerline RC (R1, C1) and put a 1uF cap across the power/GND in each channel as decoupling.
For audition, my benchmark is my current setup: CD player --> 50K stepped attenuator --> j-PHL (PLH with input IRF610 replaced a pair of 2SK147 jFETs) --> Fostex FE168 sigma double-bass-reflective box. This is more or less the simplest you can get sound out from a CD. Replacing the 50k stepped attenuator with this B1 (also with a 50k stepped attenuator), I started to compare the sound of the 2 setups. Frankly I cannot hear any difference, maybe with the B1 the attenuator gives more robust control of the amount of the attenuation. This is no surprise as in my original setup, the input impedance of the j-PHL (which is about 14k) will interfere with the R-ladder and alter the attenuation to a certain extent. Nevertheless, the clarity, the way the music comes to you, are very much equal.
This is good news, as the B1 suppose should be transparent and do not add anything other than impedance matching. It seems to be doing its job nicely.
Lack of bass reported by some people here? - no such observation. I guess if we adhere to the basic rules and make sure no mistake made, B1 should live up to its promises.
Good luck!
salas said:
No, I have not heard any changes that are substantial that drive me to conclude there is a difference. The subtle difference that I might feel can well be attributed to the volume level variations. The overall impression is that both setups sound the same, in terms of dynamics or grip. I still consider this is a good indication, the B1 is really a neutral addition to my system.
jam said:
I think someone has.
"B3 Preamp. Maybe you’d like a B1 but you need some gain." from http://www.firstwatt.com/watts_new.htm
Looney
I just ordered my B1 PCB with 4 matching JFET's from Pass DIY for $45 including USA shipping. I am really excited about the prospect of building this. I am a complete newbie(having only completed a couple of kits), but this looks simple enough to be doable.
I plan to pair it up with a F5 when those boards become available through 8" Audio Nirvana high efficiency full range BR speakers. I will now have to reread this thread while I await my boards.
I plan to pair it up with a F5 when those boards become available through 8" Audio Nirvana high efficiency full range BR speakers. I will now have to reread this thread while I await my boards.
Nelson's B1 project has renewed my interest in simple jfet buffers and so I decided to build one. I am using symmetric suplies which eliminate the need for coupling caps. I matched the k170 fets for 10 mA. The higher Idss fet was used as current source. I employed low value source resistances to null the output offset voltage.
The sound I found was surprisingly reminiscent of a passive pre. This actually makes sense when you consider that the output resistance of B1 not much lower than a 10K pot in most volume settings.
When I lowered the values of the output resistances (R104 and 204 in Nelsons schematic), the dynamics increased impressively. I finally settled on leaving them out entirely. With R104/204 shorted, the output resistance is about 30 ohm which is similar to a normal gain stage (set by the output resistor). In fact I am thinking of trying the Borbely-White source follower to lower the output resistance even further. Bit worried about the lower psrr though.
Further I also lowered the input resistances (r102/202) to about 100 ohm. This also improves the sound a lot. Much more transparent. I am a but worried about lowering the value even further as I don't have means to observe parasitic oscillations.
All in all, I am positively surprised at the sound quality that can be achieved with such a simple circuit. I believe it will be difficult to better, if you don't need the gain and do not need to drive low impedance loads.
Thanks you Nelson,
Abo
The sound I found was surprisingly reminiscent of a passive pre. This actually makes sense when you consider that the output resistance of B1 not much lower than a 10K pot in most volume settings.
When I lowered the values of the output resistances (R104 and 204 in Nelsons schematic), the dynamics increased impressively. I finally settled on leaving them out entirely. With R104/204 shorted, the output resistance is about 30 ohm which is similar to a normal gain stage (set by the output resistor). In fact I am thinking of trying the Borbely-White source follower to lower the output resistance even further. Bit worried about the lower psrr though.
Further I also lowered the input resistances (r102/202) to about 100 ohm. This also improves the sound a lot. Much more transparent. I am a but worried about lowering the value even further as I don't have means to observe parasitic oscillations.
All in all, I am positively surprised at the sound quality that can be achieved with such a simple circuit. I believe it will be difficult to better, if you don't need the gain and do not need to drive low impedance loads.
Thanks you Nelson,
Abo
Re: Thanks................
For gate stopper in a capacitor-less B1 go down to 100R with caution, better be safe if you use it around in more than 1 checked system and stick with 220-270. The resistor quality is more apparent the higher you go. A 270R Riken will do better than 100R off the mill one. I stay with original 1K, but Kiwame carbon directly on Fet's input leg. It keeps harshness out of the picture with some sources and cables always. With lower Rin some sources are just cleaner, others just rougher. For output a 220R is not a problem and increases slam VS original 1K. I use 330R Kiwame, that sounded better than 220R cheap metal film.
I would optimize by ear for both Rin, Rout values and their quality if the B1 is servicing one steady system.
B1 is the easiest project for P-P, dead bug or perf board. Make one in your own style, and you know what? Shorter and P-P will give you slightly better detail. Nothing to lose.
ABO said:
Nelson Pass said:
I don't recommend this as a rule, as many installations will experience
parasitic oscillation with no output resistance. As an alternative I recommend
exploring lower values, such as 221 ohms.
For gate stopper in a capacitor-less B1 go down to 100R with caution, better be safe if you use it around in more than 1 checked system and stick with 220-270. The resistor quality is more apparent the higher you go. A 270R Riken will do better than 100R off the mill one. I stay with original 1K, but Kiwame carbon directly on Fet's input leg. It keeps harshness out of the picture with some sources and cables always. With lower Rin some sources are just cleaner, others just rougher. For output a 220R is not a problem and increases slam VS original 1K. I use 330R Kiwame, that sounded better than 220R cheap metal film.
I would optimize by ear for both Rin, Rout values and their quality if the B1 is servicing one steady system.
cadaverdog said:
B1 is the easiest project for P-P, dead bug or perf board. Make one in your own style, and you know what? Shorter and P-P will give you slightly better detail. Nothing to lose.
