Hi,
talking about school education or lack of it, just before I left school, for a couple of lessons we got to experience the use of mechanical calculators, yes the things you had to wind the handle just like you see in war time films , digital calculators had not yet reached the market, if my memory was better I might even remember if they had been invented before I left school.
On another tack, during that final year we had access to a terminal that was connected to Ferranti's mainframe using Fortran to play with some numbers and logic.
Maybe that's why my keyboard skills are so poor but it becomes a challenge not an excuse.
Back on track. The Bryston output circuit uses both sides of the driver feeding complementary Sziklai and darlington connected output quads. There are no series connected pairs to help overcome Vce limitations.
I am surprised that a simple voltage divider attached to the collectors of the series connected pairs manage to equalise the voltage division. I would tend to agree with the other posters' reservations on this point.
talking about school education or lack of it, just before I left school, for a couple of lessons we got to experience the use of mechanical calculators, yes the things you had to wind the handle just like you see in war time films , digital calculators had not yet reached the market, if my memory was better I might even remember if they had been invented before I left school.
On another tack, during that final year we had access to a terminal that was connected to Ferranti's mainframe using Fortran to play with some numbers and logic.
Maybe that's why my keyboard skills are so poor but it becomes a challenge not an excuse.
Back on track. The Bryston output circuit uses both sides of the driver feeding complementary Sziklai and darlington connected output quads. There are no series connected pairs to help overcome Vce limitations.
I am surprised that a simple voltage divider attached to the collectors of the series connected pairs manage to equalise the voltage division. I would tend to agree with the other posters' reservations on this point.
Thanks for your comments.
If you imagine for a moment every thing in stasis ie stood still the say that the top pair of transistors are passing 1 amp you will find that the emitter resistors will place 0.5 volt each this tend to saturate the small differance of the transistor characteristics.
The 10 k resistors are to balance the voltage across the transistors when they are not! conducting !ie cut off
I tend to have a visual memory we this kind of thing and as I have not been able to talk to others re this it is difficult for me to communicate.
This was and was never intended to be a perfect solution re dividing the rails perfectly but on my prototype it did not seem to matter
Where this scores is when you have a 80 volt supply and only 60 volt transistors!
Rather than going for the whole hog of 120 volts!
To be Honest most of my building and development has been from scavanged parts
I will post a full schematic of my prototype before the week end.
And I promise to try to improve re punctuation etc. and feel that you are correct re this as I did not consider fellow members that use English as a second language. I suppose this is because we English are an island race
Regards
Trevor
If you imagine for a moment every thing in stasis ie stood still the say that the top pair of transistors are passing 1 amp you will find that the emitter resistors will place 0.5 volt each this tend to saturate the small differance of the transistor characteristics.
The 10 k resistors are to balance the voltage across the transistors when they are not! conducting !ie cut off
I tend to have a visual memory we this kind of thing and as I have not been able to talk to others re this it is difficult for me to communicate.
This was and was never intended to be a perfect solution re dividing the rails perfectly but on my prototype it did not seem to matter
Where this scores is when you have a 80 volt supply and only 60 volt transistors!
Rather than going for the whole hog of 120 volts!
To be Honest most of my building and development has been from scavanged parts
I will post a full schematic of my prototype before the week end.
And I promise to try to improve re punctuation etc. and feel that you are correct re this as I did not consider fellow members that use English as a second language. I suppose this is because we English are an island race
Regards
Trevor
Earliest 4B were like the Leach (actually, SWTPC). Later they went to the phase splitter driver, but still used the series connection (SWTCP).
http://www.bryston.ca/BrystonSite05/pdfs/LegacyPowerAmplifiers/4BE-Schematic(MA94).pdf
The latest ones use the phase splitter driver with all paralleled outputs (same as previous post).
http://www.bryston.ca/BrystonSite05/pdfs/SSTAmplifiers/4b+7bSST-SSB-SCH-6_20041013.pdf
http://www.bryston.ca/BrystonSite05/pdfs/LegacyPowerAmplifiers/4BE-Schematic(MA94).pdf
The latest ones use the phase splitter driver with all paralleled outputs (same as previous post).
http://www.bryston.ca/BrystonSite05/pdfs/SSTAmplifiers/4b+7bSST-SSB-SCH-6_20041013.pdf
Well now that Ihave seen the Bryson schematic and thats my second suprise for you all gone.
From the drawing I can see that the main advantage is in reducing the dissipation in the driver transistor as we are using the driver current twice. Once in the emitter of the driver and once in the collector of the same driver! this saves on driver current and reduces driver dissipation. But this does nothing re the voltage capabilties of the output devices.
As I have an obviouse interest in this i wonder if anyone out there would try the series arrangement and see how it works for them. .
regards for now Trev
From the drawing I can see that the main advantage is in reducing the dissipation in the driver transistor as we are using the driver current twice. Once in the emitter of the driver and once in the collector of the same driver! this saves on driver current and reduces driver dissipation. But this does nothing re the voltage capabilties of the output devices.
As I have an obviouse interest in this i wonder if anyone out there would try the series arrangement and see how it works for them. .
regards for now Trev
Hi djk,
First link is dead for me. Darn!
Trevor,
Ahhh, you read the "bryston blurb" where they list the advantages of this circuit. You may be right in that this amp does not use stacked outputs. I just repaired one and can't remember.
The Marantz 510/M used a stacked output arrangement. That would have been around 1976. The later 2500 receiver used the same heatsink assy. with high voltage devices used in the normal fashion.
-Chris
First link is dead for me. Darn!
Trevor,
Ahhh, you read the "bryston blurb" where they list the advantages of this circuit. You may be right in that this amp does not use stacked outputs. I just repaired one and can't remember.
The Marantz 510/M used a stacked output arrangement. That would have been around 1976. The later 2500 receiver used the same heatsink assy. with high voltage devices used in the normal fashion.
-Chris
Some numbers
Let's look at some numbers. Let's suppose we want the Vce to be maintained within 1V of half the rails. Because the resistive dividers have a parallel resistance of 5k, 1V change requires 200uA. For the power transistors to have Ic within 200uA of one another their Vbe must be kept within just 100uV. I think this is impossible in practice, especially over a range of Ic.
If you look at the datasheets for the power devices you'll find the Vbe for npn and pnp devices can be quite different for the same Ic. For example, The MJ21193/4 have Vbe of 0.7V and 0.73V at 1A Ic (as shown in the graphs). This 0.03V difference across 0.5 ohms is a current difference of 60mA...or 300V across 5k. I think similar differences will exist with all power device pairs.
I think this is an ingenious approach but I just not sure how this can ever work reliably as a Vce balancing circuit.
Let's look at some numbers. Let's suppose we want the Vce to be maintained within 1V of half the rails. Because the resistive dividers have a parallel resistance of 5k, 1V change requires 200uA. For the power transistors to have Ic within 200uA of one another their Vbe must be kept within just 100uV. I think this is impossible in practice, especially over a range of Ic.
If you look at the datasheets for the power devices you'll find the Vbe for npn and pnp devices can be quite different for the same Ic. For example, The MJ21193/4 have Vbe of 0.7V and 0.73V at 1A Ic (as shown in the graphs). This 0.03V difference across 0.5 ohms is a current difference of 60mA...or 300V across 5k. I think similar differences will exist with all power device pairs.
I think this is an ingenious approach but I just not sure how this can ever work reliably as a Vce balancing circuit.
Thanks for the input.This Idea as I have allready stated I have used this idea in the past and it has worked. I shall build another prototype,I am drawing the schematic now and will post this very soon. Then I shall make a few measurements to prove the point.As i have said the point of the 10k resistors is to balance the mid point when the transistors are cut off! i.e. not passing any current,But from memory the prototype was pretty well balanced with the points being botom collectors 0.25
output 0.5 and topcollectors being 0.75 of the applied voltage on the rails.
I couls easily adjust any switch on variations on the output transistors by altering the ratio of the emmitter /collector resistors very slighty to one another.
However on the prototype all rseistors were of the same value!
When passing 30mA all the values were as i would have hoped for!
Please any opinions are very welcome. as i have learned more in this short discusion than being on my own,its all food for thought.
Regards
Trev
output 0.5 and topcollectors being 0.75 of the applied voltage on the rails.
I couls easily adjust any switch on variations on the output transistors by altering the ratio of the emmitter /collector resistors very slighty to one another.
However on the prototype all rseistors were of the same value!
When passing 30mA all the values were as i would have hoped for!
Please any opinions are very welcome. as i have learned more in this short discusion than being on my own,its all food for thought.
Regards
Trev
Hi djk,
As for the first generation. I remember I had to reverse engineer one to create my own schematic, years before Bryston released them. At that time, everything was a secret. They were very upset with me at Bryston at the time.
-Chris
I have them too. I just don't care to look at them much. They are not that interesting really.
As for the first generation. I remember I had to reverse engineer one to create my own schematic, years before Bryston released them. At that time, everything was a secret. They were very upset with me at Bryston at the time.
-Chris
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.