Yes I did see some of his schematics. There is sand in all of them if you include rectifiers. ;-)
I will continue to consider the idea of common mode choke. How can I model it in PSUDII? I would like to keep ripple <1mV.
Those snowboarders are NUTS!!!!!!. My kids do it (not to that level Thank the Lord!). Me, I'd prefer to keep flying without a plane reserved only for that area in my mind... At least that way I can maintain my dignity! It also doesn't hurt as bad either
I will continue to consider the idea of common mode choke. How can I model it in PSUDII? I would like to keep ripple <1mV.
Those snowboarders are NUTS!!!!!!. My kids do it (not to that level Thank the Lord!). Me, I'd prefer to keep flying without a plane reserved only for that area in my mind... At least that way I can maintain my dignity! It also doesn't hurt as bad either
boywoder,
As tubelab said. It satifies both worlds. FETS have a VERY small gate current compared to the grid current of driving a kt88 directly, particulary at high output levels. Because of the low gate current, a higher impedance is possible. However, the output impedance of the FET source follower is low and is dictated by the source resistor. What the kt-88 sees as far as drive impedance is the source resistor which still must be <100k per the tube data sheet. IIRC
There is a limit as to how high of an impedance you can go to. It is set by the gate capacitance.
As tubelab said. It satifies both worlds. FETS have a VERY small gate current compared to the grid current of driving a kt88 directly, particulary at high output levels. Because of the low gate current, a higher impedance is possible. However, the output impedance of the FET source follower is low and is dictated by the source resistor. What the kt-88 sees as far as drive impedance is the source resistor which still must be <100k per the tube data sheet. IIRC
There is a limit as to how high of an impedance you can go to. It is set by the gate capacitance.
Okay here is the latest versions. I am trying to heed all of the advise that has been wonderfully given.
When it comes to the common mode choke, recognizing that the FET heat dissipation can be an issue, I decided against that particular solution for the time being. By not going the common choke path, I don't mean to imply or doubt for one second what George is saying true about that topology.
I feel that a seperate tranny under the hood $10.00 each, schottky diodes (I have a big tube of them Eva which I used on my Alephs), and a regulator will serve the same purpose. I have almost all of those parts in my inventory. I wanted to keep the C- as clean as possible which with the 30H choke and caps is very quiet. With the B1+ on a seperate circuit it will still be quite and I can fit the whole thing including the tranny on a small board under the hood. The other deciding factor was that I am out of room on top at this point using the monster caps.
If they (the monster caps) don't work well like George's experience indicates as a possibility, then I will have a little more room on top.
During the bread board stage of the amp, I plan on evaluating the temp of the FETs on the heatsinks. If the temp is an issue then in goes the b1+ rail supply before it becomes a problem. I want to test both setups anyway and I can do this without the tubes installed.
I have settled on a PT with the following specs. 760Vct @ 180mA / 85 V @ 100mA / 5V @ 3A. If need be I can drop a few volts with a series resistor.
When it comes to the common mode choke, recognizing that the FET heat dissipation can be an issue, I decided against that particular solution for the time being. By not going the common choke path, I don't mean to imply or doubt for one second what George is saying true about that topology.
I feel that a seperate tranny under the hood $10.00 each, schottky diodes (I have a big tube of them Eva which I used on my Alephs), and a regulator will serve the same purpose. I have almost all of those parts in my inventory. I wanted to keep the C- as clean as possible which with the 30H choke and caps is very quiet. With the B1+ on a seperate circuit it will still be quite and I can fit the whole thing including the tranny on a small board under the hood. The other deciding factor was that I am out of room on top at this point using the monster caps.
If they (the monster caps) don't work well like George's experience indicates as a possibility, then I will have a little more room on top.
During the bread board stage of the amp, I plan on evaluating the temp of the FETs on the heatsinks. If the temp is an issue then in goes the b1+ rail supply before it becomes a problem. I want to test both setups anyway and I can do this without the tubes installed.
I have settled on a PT with the following specs. 760Vct @ 180mA / 85 V @ 100mA / 5V @ 3A. If need be I can drop a few volts with a series resistor.
Okay here is the latest versions. I am trying to heed all of the advise that has been wonderfully given.
When it comes to the common mode choke, recognizing that the FET heat dissipation can be an issue, I decided against that particular solution for the time being. By not going the common choke path, I don't mean to imply or doubt for one second what George is saying true about that topology.
I feel that a seperate tranny under the hood $10.00 each, schottky diodes (I have a big tube of them Eva which I used on my Alephs), and a regulator will serve the same purpose. I have almost all of those parts in my inventory. I wanted to keep the C- as clean as possible which with the 30H choke and caps is very quiet. With the B1+ on a seperate circuit it will still be quite and I can fit the whole thing including the tranny on a small board under the hood. The other deciding factor was that I am out of room on top at this point using the monster caps.
If they (the monster caps) don't work well like George's experience indicates as a possibility, then I will have a little more room on top.
During the bread board stage of the amp, I plan on evaluating the temp of the FETs on the heatsinks. If the temp is an issue then in goes the b1+ rail supply before it becomes a problem. I want to test both setups anyway and I can do this without the tubes installed.
I have settled on a PT with the following specs. 760Vct @ 180mA / 85 V @ 100mA / 5V @ 3A. If need be I can drop a few volts with a series resistor.
So the 1.5mh and 1nf cap are forming what you are calling a "hash filter"? To me, that just mean not showing up at mom's house on Sunday evening for dinner. However, if you use an oil cap in the first position, can you avoid a hash filter? Seems too complex. Maybe I've just not heard the sonic benefits. Secondly, and personally, I'm thinking the advantages of the FET are being outweighted by the circuit complexity. Seems like you gain a few DB in headroom in exchange for a seperate PS, FETs, sonic impact in the circuits, etc... I think when I build my version of this it will be a garden variety fixed bias supply to which I can add the FET follower at some later date to see what it does. When I tried this in the TubeLab SE amp it did add bass weight from other SE 300b designs but I felt I gave up a lot in midrange tone, realism, etc. (no offense George, just my experience). I also wonder if some kind of simple protection circuit can be devised - a fuse is too simple. A circuit that protects the tubes and Fet seems to be in order.
No offense taken. When I tried the same amp with and without PowerDrive the differences ranged from minor (freedom from overload recovery on a 45 TSE) to dramatic (more power authority and better transient behavior on a KT88 amp). Loss of realism and tone wasn't noted.
I just noted another little thing. The math shows that a 1/2 watt resistor on the screen grids of the KT88's is sufficient. Experience has proven otherwise. I have blasted even 1 watt resistors on a Simple SE with EL34's by overdriving the amp with my guitar preamp. Definitely worse case, but resistors are cheap. I have been using 2 watt screen resistors in everything since then, and haven't blown one up since.
It's not the dielectric that matters. The large capacitance immediately following the rectifier is what matters. Large value cap. I/P filters are associated with tiny conduction angles. Part and parcel of a tiny conduction angle is a sharp, highly "triangular", ripple waveform. Fourier's Theorem tells us that such a waveform contains overtones of the ripple fundamental extending well up into RF. The capacitance associated with the many turns of wire associated with "typical" filter chokes shorts the choke out at RF. The "hash" filter kills the RF crud, before it gets to the choke.
Get the Physics right, before worrying about refinements!
Last edited:
George, I will trust your experience. A two watt will be fine.
smbrown. Your opinion is certainly valid for your build. In this one the Hash filter and FETS stay. Complexity is a personal experience and preference. The power supply is built around what the circuit needs to perform at its optimum. Some well spent time planning the layout will alleviate the complexity. That is my next step. Maybe there are different ways to skin the cat to achieve the same result, but I like and understand this way enough to have a reasonable chance of success.
With the Solid State and Hollow State amps that I have built (Only SETS so far in tubes) IMHO, the more quite and lowest impedance supplies win in the musicality department. Now if you wanted to comment on the merit of the monster caps I am using at least up to this point in design, I am not sure I would have a leg to stand on. They are "Over the Top, in fact way over the top". A good example of flying without an airplane...... I'll let my osc scope tell me how I land.
smbrown. Your opinion is certainly valid for your build. In this one the Hash filter and FETS stay. Complexity is a personal experience and preference. The power supply is built around what the circuit needs to perform at its optimum. Some well spent time planning the layout will alleviate the complexity. That is my next step. Maybe there are different ways to skin the cat to achieve the same result, but I like and understand this way enough to have a reasonable chance of success.
With the Solid State and Hollow State amps that I have built (Only SETS so far in tubes) IMHO, the more quite and lowest impedance supplies win in the musicality department. Now if you wanted to comment on the merit of the monster caps I am using at least up to this point in design, I am not sure I would have a leg to stand on. They are "Over the Top, in fact way over the top". A good example of flying without an airplane...... I'll let my osc scope tell me how I land.
First Stab at Layout
Parts have started to arrive!
Attached is my first attempt at laying this stuff out. I tried to keep symetric paths. There is no filament power yet but they will be twisted and run perpindicular to signal.
The pink path are leads / components in direct path of signal
Green is Gnd Buss
Dark Gray box are filament caps and power filtration
Lighter gray boxes are power rated resistors
brown is metal film resistors
Red is B+ or derivatives
Blue is B1+
Light Blue is C-
I didn't want to get too far if I am making any major mistakes. (Drawing this on a windows machine is tedious at best) I would like your feedback on suggestions to improve and/or errors in concept and then modify and detail the drawing as appropriate.
Thanks
Scott
Parts have started to arrive!
Attached is my first attempt at laying this stuff out. I tried to keep symetric paths. There is no filament power yet but they will be twisted and run perpindicular to signal.
The pink path are leads / components in direct path of signal
Green is Gnd Buss
Dark Gray box are filament caps and power filtration
Lighter gray boxes are power rated resistors
brown is metal film resistors
Red is B+ or derivatives
Blue is B1+
Light Blue is C-
I didn't want to get too far if I am making any major mistakes. (Drawing this on a windows machine is tedious at best) I would like your feedback on suggestions to improve and/or errors in concept and then modify and detail the drawing as appropriate.
Thanks
Scott
I have a fan on standby for magic smoke capture.
I decided to Set up the B1+ rail from the get go. The top plate for the amp was to be a heatsink. When I milled it for the tubes etc I didn't like the look. Without the heat sink, I won't be able to run the B+ rail across them safely. By using the second rail, power dissapation drops from 7Watts per device down to a max 1.2 Watt per device. Plus if it fails tube won't arc out as George pointed out.
I have to finish this thing now. My wife will go Ape if I bought all the parts and left them lay till next winter!!!!
I am cadding up the amp plate for bench testing now. My longest lead time is the xfmrs. It usually takes about two weeks for them to be delivered.
If there is any major layout issues please tell me.
I have been keeping up with what Johan is saying on the other Mullard kt-88 thread. At this point I am going to stick with the at7 and hope that the tubes I have provide enough swing. With the light load I am thinking that they will be fine.
I decided to Set up the B1+ rail from the get go. The top plate for the amp was to be a heatsink. When I milled it for the tubes etc I didn't like the look. Without the heat sink, I won't be able to run the B+ rail across them safely. By using the second rail, power dissapation drops from 7Watts per device down to a max 1.2 Watt per device. Plus if it fails tube won't arc out as George pointed out.
I have to finish this thing now. My wife will go Ape if I bought all the parts and left them lay till next winter!!!!
I am cadding up the amp plate for bench testing now. My longest lead time is the xfmrs. It usually takes about two weeks for them to be delivered.
If there is any major layout issues please tell me.
I have been keeping up with what Johan is saying on the other Mullard kt-88 thread. At this point I am going to stick with the at7 and hope that the tubes I have provide enough swing. With the light load I am thinking that they will be fine.
Last edited:
Scott,
As stated, you are buffering the LTP. Drop the LTP load resistors to 39 KOhms and up IB to 3.4 mA. With that setup, you buy some headroom for the 100 V. swing needed. I think said condition set is close enough to McShane's "ideal" to work well. Every design has compromises.
Last edited:
Using the FET drivers is NOT a compromise. SY will tell you that FET voltage followers sound better in this role than cathode followers.
The absence of a heater power draw doesn't hurt either. I'm a pragmatist. I use tubes because they do many, not all, audio jobs better than SS. Where "sand" works well, I say use it and move on. FWIW, I'd use a lump of galena if it worked well, but it doesn't.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.