It's not an intuitive circuit. I took me many years to get
a proper handle on it, assuming that I have.
a proper handle on it, assuming that I have.
Hi vossie
As I see your circiut, which looks nice and simple, you might have some DC at the inputterminals.
The Base-emitter (diffpair) current draw will give you a couple of volts over the feedbackresistors, wich will be seen either at the input or the outputterminals. I am not quite shure if you can get around this problem with BJT´s without capacitors, but I have tried to change the diffpair to some ZVN3310 fets, an then it looks as if the DC problem is gone.
I have only tried this in the simulator.
With your approach, it´s possible to make some verry compact design at lover voltage level.
I have attached the Simetrix files.
Best regards
As I see your circiut, which looks nice and simple, you might have some DC at the inputterminals.
The Base-emitter (diffpair) current draw will give you a couple of volts over the feedbackresistors, wich will be seen either at the input or the outputterminals. I am not quite shure if you can get around this problem with BJT´s without capacitors, but I have tried to change the diffpair to some ZVN3310 fets, an then it looks as if the DC problem is gone.
I have only tried this in the simulator.
With your approach, it´s possible to make some verry compact design at lover voltage level.
I have attached the Simetrix files.
Best regards
I think it's fair to point out that the circuit shown is slightly
simplified, particularly regarding those elements which will
ensure the proper biasing of the DC output.
simplified, particularly regarding those elements which will
ensure the proper biasing of the DC output.
Thanks Henrik for your input.
I have mailed Nelson privately with a comment and some further details regarding this schematic.
I will wait for his reply before I comment further on the circuit.
Hope to hear from you soon Nelson.
Vossie
I have mailed Nelson privately with a comment and some further details regarding this schematic.
I will wait for his reply before I comment further on the circuit.
Hope to hear from you soon Nelson.
Vossie
For what it's worth, this is essentially the same circuit that I executed with tubes. I took the feedback from after the DC blocking caps at the outputs.
It's about as simple a practical circuit as you can do as an X. Yes, you can do a single differential as X, but you pay a penalty in gain or output Z.
Grey
It's about as simple a practical circuit as you can do as an X. Yes, you can do a single differential as X, but you pay a penalty in gain or output Z.
Grey
GRollins said:
To throw in my 2 cents, this is the limit you get from the devices
themselves, essentially their transconductance versus the output
load. If you can get a high enough transconductance figure out
of the diff pair, it goes a long way toward solving the problem.
Actually, now that I've had supper and rolled it around in my head for a while, the gain question might not be so important anyway. I nearly always use phono as input (can't afford a good CD player and refuse to put up with the sonics of the cheaper ones [anyone got a decent CD player they're throwing out with the trash?]--besides, at this point I'd like to get one that would do SACD and DVD-A), and need a little extra gain. Those who listen to high level sources (i.e. CD) need little gain, if any, and can afford to burn off all but 6-12 dB as feedback if they want. That should be well within the capability of a MOSFET design. With enough feedback, you can bring the Zout down, though higher levels of feedback tend to carry a penalty.
Want more gain and/or feedback? Use bipolars.
Tubes, however, are pretty much going to have to stick with the general topology posted above in order to bring the Zout down. There is, however, one loophole. If you were to go for a massively parallel design, using multiple differentials in parallel with each other, you could bring the Zout down to any arbitrary level that you might wish and get more gain in the bargain. Come to think of it, the same trick would work for solid state devices.
Hmmm...lemme think about this for a spell...I may have talked myself into something, here.
Grey
Want more gain and/or feedback? Use bipolars.
Tubes, however, are pretty much going to have to stick with the general topology posted above in order to bring the Zout down. There is, however, one loophole. If you were to go for a massively parallel design, using multiple differentials in parallel with each other, you could bring the Zout down to any arbitrary level that you might wish and get more gain in the bargain. Come to think of it, the same trick would work for solid state devices.
Hmmm...lemme think about this for a spell...I may have talked myself into something, here.
Grey
Some further comments
I waited for Nelson to reply on a mail I sent to him off the forum, on which he did - thanks for that.
The reason for all the "cecrecy" was that the circuit is taken from a USA patent (granted in Oct 1981 so most probably obsolute now). According to Nelson
X is GENERALLY defined as: "In general I would qualify X as having a symmetric circuit build around differential inputs where the feedback is symmetric so as to create a cancellation of errors between the outputs, nulling distortion and noise seen by a balanced load."
Generally speaking the X topology must then be dependant on what type of input the circuit accepts (single-ended = not X / balanced = X). This could mean that the X-BOSOZ (and a few other circuits on the forum) do not stricktly qualify as X in practise as very few of them are driven balanced - hmmmm?
Being very familiar with Nelsons patent it states that the circuit could be driven single-ended as well (in which case the signal is reproduced balanced throughout. All patent drawings shows balanced input however.
If we compare apples to apples then the ONLY difference between the circuit shown in this threat (patent 42296383) and Nelson's X topology circuits as demonstrated in this forum is the way in which it accepts input - some are identical as very few are driven balanced!?
If we however compare most of the circuits shown on the forum with Nelson's patent drawings there are differences in some cases. In practise they comply more with the old patent shown in the first post.
In fact, the x-communication happening in this circuit is 100% the same as Nelson's patent, but so the same is happening in a few of the other patents of old. Does this make all of them X - I don't think so?
Take a look at the patent mentioned, noting the input section - single-ended. All that was changed before presenting the circuit on this forum was the circuit was drawn to represent a balanced input. We have allready put the X-stamp on this circuit which leaves a lot open for discussion ...a very interesting one at that.
There are 101 ways to arrange components in a circuit to perform X. If we argue and say that all topologies arranged as a differential amplifier with negative feedback taken from the outputs back to the differential inputs qualify as X then Nelson can finally start to making good use of those 4 lawayers of him...
Be nice please...
I waited for Nelson to reply on a mail I sent to him off the forum, on which he did - thanks for that.
The reason for all the "cecrecy" was that the circuit is taken from a USA patent (granted in Oct 1981 so most probably obsolute now). According to Nelson
X is GENERALLY defined as: "In general I would qualify X as having a symmetric circuit build around differential inputs where the feedback is symmetric so as to create a cancellation of errors between the outputs, nulling distortion and noise seen by a balanced load."
Generally speaking the X topology must then be dependant on what type of input the circuit accepts (single-ended = not X / balanced = X). This could mean that the X-BOSOZ (and a few other circuits on the forum) do not stricktly qualify as X in practise as very few of them are driven balanced - hmmmm?
Being very familiar with Nelsons patent it states that the circuit could be driven single-ended as well (in which case the signal is reproduced balanced throughout. All patent drawings shows balanced input however.
If we compare apples to apples then the ONLY difference between the circuit shown in this threat (patent 42296383) and Nelson's X topology circuits as demonstrated in this forum is the way in which it accepts input - some are identical as very few are driven balanced!?
If we however compare most of the circuits shown on the forum with Nelson's patent drawings there are differences in some cases. In practise they comply more with the old patent shown in the first post.
In fact, the x-communication happening in this circuit is 100% the same as Nelson's patent, but so the same is happening in a few of the other patents of old. Does this make all of them X - I don't think so?
Take a look at the patent mentioned, noting the input section - single-ended. All that was changed before presenting the circuit on this forum was the circuit was drawn to represent a balanced input. We have allready put the X-stamp on this circuit which leaves a lot open for discussion ...a very interesting one at that.
There are 101 ways to arrange components in a circuit to perform X. If we argue and say that all topologies arranged as a differential amplifier with negative feedback taken from the outputs back to the differential inputs qualify as X then Nelson can finally start to making good use of those 4 lawayers of him...
Be nice please...
That's 7 lawyers, if you include my sister 😉
My view is that a patent validates a circuit for marketing
purposes, and it is rare that it does a good job of protecting
on encroachment.
More important from a practical standpoint is the tradename
and trademarks, which generally present more airtight
arguments.
The best precedent for an X type circuit is still the Hadley,
although it doesn't appear that he fully appreciated what he
had. (no pun intended).
The issue of a single-ended input is fairly moot. Any balanced
input amplifier can amplify the difference between ground and
some other signal.
Again, I haven't seen the schematic, although I have read the
claims. Perhaps someone will post it.
My view is that a patent validates a circuit for marketing
purposes, and it is rare that it does a good job of protecting
on encroachment.
More important from a practical standpoint is the tradename
and trademarks, which generally present more airtight
arguments.
The best precedent for an X type circuit is still the Hadley,
although it doesn't appear that he fully appreciated what he
had. (no pun intended).
The issue of a single-ended input is fairly moot. Any balanced
input amplifier can amplify the difference between ground and
some other signal.
Again, I haven't seen the schematic, although I have read the
claims. Perhaps someone will post it.
- Status
- Not open for further replies.