Not sure but if it does it should have the benefits of Parafeed and Ultra path circuits. It certainly lowers the impedance that the AC signal sees.
http://smg.photobucket.com/albums/v224/GRaider/?action=view¤t=minimonkeyparafeed.jpg
http://smg.photobucket.com/albums/v224/GRaider/?action=view¤t=minimonkeyparafeed.jpg
R3 was shorted when I looked at the schematic before but it seems to have been corrected now.
There is no B+ for U1 - where is the 150v on its plate supposed to come from? And why is the same point in the circuit supposed to be at -24v on the grid of U2? Frankly, with errors like that, there could be other problems. I wouldn't trust the rest of the schematic.
kevinkr said:
Would that work here too? http://electra-print.com/images/drdamp&ps2.jpg
If so, any thoughts on why Jack wouldn't have done it in the first place?
ray_moth said:
Yes I corrected the circuit as R3 was shorted. The B+ for U1 comes from the top of the 6.4K Resistor. The cathode of the 12B4A is sitting at 174 volts. The "rest of the schematic" is fine. I was just wondering about the connection of the output transformer. Since it is AC coupled it should work as shown taking its signal off of the plate and the cathode instead of off the plate and back through the power supply.
dsavitsk said:
It would work. Thorsten designed a amp that uses the "Ultrapath" on both stages but I have since learned that some of the power supply ripple is "injected" into the signal through the cathode of the output tube. I don't plan on using a regulated supply so I decided against the Ultrapath connection.
I've been thinking about this since last night, and realized a couple of things since I last posted on this design.
I noted the lack of ultrapath cap from the cathode of the output tube back to the supply and while I understand your objection to injecting supply ripple into this node a low ac impedance is needed here for the output stage to provide a reasonable amount of voltage gain.
A technique I use is what I call a hum cancelling connection, it is more or less an old Western Electric technique to buck hum in the output stage. Basically you have a conventional cathode bypass cap and an ultrapath cap - the ratio between them is 1/mu+1 so with a typical dht the ultrapath cap will be about 1/5 the value of the cathode bypass - this actually will null almost all of the ripple in the output stage, and also confers the sonic benefits of ultrapath imo.
Once you make the above change you can also try an ultrapath connection between the bottom of the choke and the cathode of the driver tube if you are so inclined.
I noted the lack of ultrapath cap from the cathode of the output tube back to the supply and while I understand your objection to injecting supply ripple into this node a low ac impedance is needed here for the output stage to provide a reasonable amount of voltage gain.
A technique I use is what I call a hum cancelling connection, it is more or less an old Western Electric technique to buck hum in the output stage. Basically you have a conventional cathode bypass cap and an ultrapath cap - the ratio between them is 1/mu+1 so with a typical dht the ultrapath cap will be about 1/5 the value of the cathode bypass - this actually will null almost all of the ripple in the output stage, and also confers the sonic benefits of ultrapath imo.
Once you make the above change you can also try an ultrapath connection between the bottom of the choke and the cathode of the driver tube if you are so inclined.
Some further thoughts, and sorry for muddying the waters still further. I think I need to bolt my head on straight before I post on this subject :
The hum cancelling connection I mentioned in my last post is appropriate only for conventional SE (non-parafeed) connection.
Because the ripple voltage in a parafeed amp is developed across the choke and not the primary of the output transformer the parafeed has a much higher PSRR than the conventional SE - you really wouldn't want to inject ripple from the power supply into the cathode circuit in this case.
So given the above you would not want to use the ultrapath connection at all unless you had a ripple free supply. In this case I would just add a cathode bypass cap from the top of R2 to gnd. Leave C1 connected where it is.
I use 100uF motor run caps for this application, and I would probably also shunt it with a smaller solen or similar to improve high frequency performance.
Edits: for clarity, typo.. etc.
The hum cancelling connection I mentioned in my last post is appropriate only for conventional SE (non-parafeed) connection.
Because the ripple voltage in a parafeed amp is developed across the choke and not the primary of the output transformer the parafeed has a much higher PSRR than the conventional SE - you really wouldn't want to inject ripple from the power supply into the cathode circuit in this case.
So given the above you would not want to use the ultrapath connection at all unless you had a ripple free supply. In this case I would just add a cathode bypass cap from the top of R2 to gnd. Leave C1 connected where it is.
I use 100uF motor run caps for this application, and I would probably also shunt it with a smaller solen or similar to improve high frequency performance.
Edits: for clarity, typo.. etc.
kevinkr said:Some further thoughts, and sorry for muddying the waters still further. I think I need to bolt my head on straight before I post on this subject :
The hum cancelling connection I mentioned in my last post is appropriate only for conventional SE (non-parafeed) connection.
Because the ripple voltage in a parafeed amp is developed across the choke and not the primary of the output transformer the parafeed has a much higher PSRR than the conventional SE - you really wouldn't want to inject ripple from the power supply into the cathode circuit in this case.
So given the above you would not want to use the ultrapath connection at all unless you had a ripple free supply. In this case I would just add a cathode bypass cap from the top of R2 to gnd. Leave C1 connected where it is.
I use 100uF motor run caps for this application, and I would probably also shunt it with a smaller solen or similar to improve high frequency performance.
Edits: for clarity, typo.. etc.
Thank you Kevin for the depth of your replies. I thought the whole point in the Ultrapath connection was to eliminate the need for the AC signal to return through the power supply. It seems as though that is the case with my diagram also. What am I missing?
There is no low impedance path for ac cathode currents to flow in the cathode circuit of this design, the ultrapath connection would normally provide this. In this instance it will degrade the inherent psrr of the parafeed topology, but you still need a low impedance at the cathode, so a cathode bypass cap would seem to be in order..
You know after thinking about this circuit for a while I realized that since the gian of the 5687 is going to be near it's mu that this amp will easily clip at the output stage. Maybe a Sakuma style circuit would be better. A 12B4A driving a 12B4A. With a input of 2 volts the gain stage would swing 24v PP into the output stage which would then swing 144v PP into the load. Unfortunately I need a gain of about 12 in the gain stage to utilize the useable bandwidth (48v PP) of the output stage. Of course I could just ditch the plate choke, Increase the B+ and use a plate resistor on the 5687 which would bring the gain down. In fact that may be the best bet. Thoughts?
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.