"As much as I like the concept of BOZ, I don't think it was
as good as the balanced version, and that seems to be
the prevailing opinion out in the world. Of course neither
of those is a neutral as the B1."
Papa Nelson, based on the past tense in the above quote, I conclude you are not engaging in one of your cryptic comments on "works in progress." Is there a link I may have missed in my preliminary investigation into the BOZ?
By the way, I just engaged in a modest "tube rolling" exercise as a result of the B1 and my Conrad Johnson MV-55. I now have a greater appreciation of the "micro sonic" aspects" that the tubies are always going on about in connection with certain tubes. But I noticed that the lower level of more refined 12AX7's suggests a tad more gain than the B1 produces. Is the "balanced BOZ" available for prime time?
Again, thanks for you incredible and gracious contributions to us duffers!
as good as the balanced version, and that seems to be
the prevailing opinion out in the world. Of course neither
of those is a neutral as the B1."
Papa Nelson, based on the past tense in the above quote, I conclude you are not engaging in one of your cryptic comments on "works in progress." Is there a link I may have missed in my preliminary investigation into the BOZ?
By the way, I just engaged in a modest "tube rolling" exercise as a result of the B1 and my Conrad Johnson MV-55. I now have a greater appreciation of the "micro sonic" aspects" that the tubies are always going on about in connection with certain tubes. But I noticed that the lower level of more refined 12AX7's suggests a tad more gain than the B1 produces. Is the "balanced BOZ" available for prime time?
Again, thanks for you incredible and gracious contributions to us duffers!
Hi,
in a single channel B1 the two transistors are in series. They pass the same quiescent current.
The B1 is a little bit different from most other jFET followers (=common drain) in that the B1 follower runs at 100% of it's Idss. Most jFET followers run at between 50% and 90% of their Idss.
To allow the B1 follower to run at 100% of it's Idss you must pass that amount of current through the Follower.
This is achieved in the B1 by feeding the current through a Constant Current Source/Sink (CCS).
A jFET CCS is simply one jFET with Gate connected to Source, for an Nchannel jFET feed +10Vdc to the Drain and -0Vdc to Source and Gate.
By definition the jFET CCS will pass it's Idss while it's temperature is Tj=25degC. Don't worry about temperature, that's part of the clever bit in the PASS implementation.
The Follower needs a current flow of 100% of it's Idss and the CCS passes it's own Idss. Choose two jFETs with the same Idss and job's done.
It's the upper (Follower) and the lower (CCS) devices that need to be selected for equal Idss.
Now we move to a stereo version.
The second channel will sound more like the first channel if it uses two almost identical jFETs. But I think this is a secondary effect. I think you can select a second pair that are close to the Idss of the first pair and still get good stereo. I accept that a pair of 6mA devices might sound different from a pair of 11mA devices, but when it comes to pairs of 9.0mA or 9.9mA, I think you would be struggling to hear any difference in performance.
Last edited:
Thanks for that explanation Andrew - much appreciated.
So in that case, if I have understood you correctly, in the case of the transistors I have measured, it would be best to have my pair of 9.8s for one channels CCS & follower, and for the other channel my 10.3s for both?
(as opposed to the two other scenarios I gave previously:
1) Pair of 9.8's for CCS, and pair of 10.3's for the follower transistors
2) 9.5 (CCS) and 9.6 (follower) for one channel, and 9.8 (CCS) and 9.9 (follower) for the other?)
Thanks,
- John
So in that case, if I have understood you correctly, in the case of the transistors I have measured, it would be best to have my pair of 9.8s for one channels CCS & follower, and for the other channel my 10.3s for both?
(as opposed to the two other scenarios I gave previously:
1) Pair of 9.8's for CCS, and pair of 10.3's for the follower transistors
2) 9.5 (CCS) and 9.6 (follower) for one channel, and 9.8 (CCS) and 9.9 (follower) for the other?)
Thanks,
- John
yes ! Hurrah.
There is more.
What if the two jFETs are not exactly the same Idss?
in this situation one could run the Follower at 99% of Idss or swap it with the CCS and run it at 101% of Idss.
I don't know which is better. I feel, from the literature, that 99% is better for the jFET Follower. This results in a tiny negative voltage from Gate to Source of the Follower. Whereas 101% of Idss results in a tiny positive voltage from Gate to Source.
This next part only applies to the DCB1:
You can check the output offset after assembly to find if you have zero offset (=exactly equal Idss) or +ve offset or -ve offset. If the offset is large then something has gone wrong. Damaged one or more, badly measured one or more, mixed up the pairs during assembly.
Excellent - we have a result, and I've learned a thing or two on this cold morning as a bonus 
I'm building the original B1 Buffer PCB at present, though do have a DCB1 PCB on order also so will take into consideration what you said about selecting for zero offset when it comes to building that one. I may well install some sockets so I can test them quickly.
However one project at a time - should be able to get the B1 up and running by the end of the day!
Cheers & thanks for the help!
- John
I'm building the original B1 Buffer PCB at present, though do have a DCB1 PCB on order also so will take into consideration what you said about selecting for zero offset when it comes to building that one. I may well install some sockets so I can test them quickly.
However one project at a time - should be able to get the B1 up and running by the end of the day!
Cheers & thanks for the help!
- John
Thanks for the insight. I built a B1 with no jFET matching. It still sounds excellent. I use BL JFETs
Yes I've read several times on here that matching isn't 100% necessary for good results.
I'm probably getting wayy to OCD about this haha!
It just 'looks' wrong to me to have one channel with two 9.8 value Jfets, and the other with a pair of 10.3 idss Jfets - wouldn't one channel be slightly louder than the other? But then again Andrew is an expert and I am nothing more than a 'build by numbers' DIYer who attempts to learn as much as my excuse for a 'brain' will allow, but much of it unfortunately I find quite hard to grasp not having much of a head for numbers and maths.
I was wondering about keeping the 9.8s in one channel, and having a 9.9 and a 9.8 in the other so all 4 are more closely matched...
Probably over-thinking this though
- John
I'm probably getting wayy to OCD about this haha!
It just 'looks' wrong to me to have one channel with two 9.8 value Jfets, and the other with a pair of 10.3 idss Jfets - wouldn't one channel be slightly louder than the other? But then again Andrew is an expert and I am nothing more than a 'build by numbers' DIYer who attempts to learn as much as my excuse for a 'brain' will allow, but much of it unfortunately I find quite hard to grasp not having much of a head for numbers and maths.
I was wondering about keeping the 9.8s in one channel, and having a 9.9 and a 9.8 in the other so all 4 are more closely matched...
Probably over-thinking this though
- John
To my ears the DCB1 takes papas excellent design to another level and has replaced a TVC pre in my system. Its a keeper; can't turn it off!
no, the gain is not determined by device parameters. This Follower has a gain of ~1times (+0dB). The topology sets the gain not the device, nor the current that the device runs at.
One channel will not sound louder (have a different gain) even when using 5mApair vs 12mA pair.
No, volume depends on voltage output. Depending on input impedance of the next device, it will suck more or less current. Voltage should be the same that you give to the input of each buffer, since its a unity gain preamp.
These IDSS point out the maximum output current a jFET can supply. Am I wrong?
Maybe...
No, just thank me for giving you the homework...maybe you can come out of the dark cave now.
a jFET used as a CCS will be limited to no more than it's Idss
A jFET used as an amplifier will try to amplify at any current you ask it to.
JC reminds us many times that a jFET will continue working well even when Vgs goes positive and as a result Id>Idss.
But, there is a limit. The hard limit would be when the jFET blows up.
A lower limit would be set by excessive distortion.
I used to think that jFETs worked best when biased to ~70% of Idss thus giving an operating range of +-30%Idss, but B1 proves otherwise.