Hey Guys,
All is good with the fils. I have the fil regs working and happily cooled with the CPU cooler. I can keep all the regs very cool with about 7 volts on the cooler fan, which gives very quiet performance.
Time to move the the neg bias supply. I would first like to ask what is the proper method to determine the current requirements of the supply when and if it goes into A2. I will used fixed bias for both the driver and the 845. Does anyone have a pref as to the type of pot to be used?
Also, if I were to use a choke input on the neg supply, would the increase in regulation compared to the cap input B+ supplies cause any drift in op point with changes in mains voltage, or would the difference be to small to worry about?
Thanks Guys,
Bryan
All is good with the fils. I have the fil regs working and happily cooled with the CPU cooler. I can keep all the regs very cool with about 7 volts on the cooler fan, which gives very quiet performance.
Time to move the the neg bias supply. I would first like to ask what is the proper method to determine the current requirements of the supply when and if it goes into A2. I will used fixed bias for both the driver and the 845. Does anyone have a pref as to the type of pot to be used?
Also, if I were to use a choke input on the neg supply, would the increase in regulation compared to the cap input B+ supplies cause any drift in op point with changes in mains voltage, or would the difference be to small to worry about?
Thanks Guys,
Bryan
Bryan,
I don't think choke input will be of any advantage here. The demand fluctuations will be minscule. Also, you will find it difficult to specify a choke for LC in this very low current application. Try CLC or CRC.
I don't think choke input will be of any advantage here. The demand fluctuations will be minscule. Also, you will find it difficult to specify a choke for LC in this very low current application. Try CLC or CRC.
Thanks John,
The reason I asked was that my transformer is going to give me 115VAC or 230VAC. I'm not too confident about the 115 wiring giving me enough voltage, so I'll use the 230 configuration. That said, I don't need the extra volts of a cap input. I can drop some voltage elsewhere.
As of now I am thinking of a simple schottky bridge - CLC filter with a couple voltage dividers and pots to dial in the bias. Very similar to whats "out there".
I have the old 10K 2Watt pots from the old project lying around. These seem to fit the bill. High power pots sure are expensive!!!
Any other thoughts or words of wisdom before I start drawing up the scheme?
Thanks All,
Bryan
The reason I asked was that my transformer is going to give me 115VAC or 230VAC. I'm not too confident about the 115 wiring giving me enough voltage, so I'll use the 230 configuration. That said, I don't need the extra volts of a cap input. I can drop some voltage elsewhere.
As of now I am thinking of a simple schottky bridge - CLC filter with a couple voltage dividers and pots to dial in the bias. Very similar to whats "out there".
I have the old 10K 2Watt pots from the old project lying around. These seem to fit the bill. High power pots sure are expensive!!!
Any other thoughts or words of wisdom before I start drawing up the scheme?
Thanks All,
Bryan
Bryan,
Use the 230V config. Drop the extra volts through a big resistor in the CRC network. The caps can be much smaller then. PSUD to the rescue😉
Use the 230V config. Drop the extra volts through a big resistor in the CRC network. The caps can be much smaller then. PSUD to the rescue😉
Hmm,
I guess chasing A2 is becomming more and more difficult.
With a bias supply which uses a big R, the impedence will probably be nowhere near low enough to provide any additional current under demand, no? This is assuming that I will go from the range on ~1-2mA grid current to around 100mA depending on how hard I drive things.... Big R is not looking good!
I also assume that a flat supply is necessary, as it is on the grid afterall...
Time to read up more....
At the end of the day I guess it may be wise to avoid the A2 dream for a little while, and keep things in A1.
I guess chasing A2 is becomming more and more difficult.
With a bias supply which uses a big R, the impedence will probably be nowhere near low enough to provide any additional current under demand, no? This is assuming that I will go from the range on ~1-2mA grid current to around 100mA depending on how hard I drive things.... Big R is not looking good!
I also assume that a flat supply is necessary, as it is on the grid afterall...
Time to read up more....
At the end of the day I guess it may be wise to avoid the A2 dream for a little while, and keep things in A1.
Hmm, If your really worried about the bias supply, you could try yet another regulator. Texas Instruments TL783 is a possible canididate.
Yes, you could also use a swinging choke input and a bleeder resistor (to make inductance requirements reasonable), but as mentioned specing (not to mention finding) these is complicated.
Im not sure the regulation on a choke input LC filter with a vanilla choke is going to be alot better than it would be with a Schottky rectifier capacitor input CLC filter.
If you have plenty of VA and dont mind even more heat, you could use a bleeder to swamp out the 2-100ma draw.
Yes, you could also use a swinging choke input and a bleeder resistor (to make inductance requirements reasonable), but as mentioned specing (not to mention finding) these is complicated.
Im not sure the regulation on a choke input LC filter with a vanilla choke is going to be alot better than it would be with a Schottky rectifier capacitor input CLC filter.
If you have plenty of VA and dont mind even more heat, you could use a bleeder to swamp out the 2-100ma draw.
A2
just build the bias supply like a B+ supply. CLC.
how you load it will affect how well your driver works.
I prefer IT's myself.
just build the bias supply like a B+ supply. CLC.
how you load it will affect how well your driver works.
I prefer IT's myself.
It seems like either cathode follower drive or IT drive are viable candidates for A2 operation. A cathode follower would allow you to use a relatively simple negative supply scheme, particularly if you use a CCS for loading the follower.
Kevin,
Care to elaborate a bit...?
From what I understand (and I admit I understand little) the current comes from the grid bias supply. For my application, I will be using an IT transformer between the driver and the output tube.
BK
Care to elaborate a bit...?
From what I understand (and I admit I understand little) the current comes from the grid bias supply. For my application, I will be using an IT transformer between the driver and the output tube.
BK
When an IT is used the drive current needed to enter A2 is supplied by the driver tube. Since grid current is drawn the bias supply must have a low output impedance for all audio frequencies or the bias voltage can shift when current flows. This can be accomplished in many ways. The simplest (and often least desirable) is a (or multiple) capacitor from the secondary of the transformer (the end not connected to the grid) to ground.
A regulated bias supply is also possible. This supply should be designed such that the bias voltage is present BEFORE the output tube gets B+. You need a seperate supply for each channel.
A regulated bias supply is also possible. This supply should be designed such that the bias voltage is present BEFORE the output tube gets B+. You need a seperate supply for each channel.
Exactly as tubelab describes with an IT, otherwise you can also try a CF direct coupled to the grid of the output tube. Negative supply would not have to be tightly regulated nor of particularly low source impedance - the current for A2 being provided by the plate supply of the CF..
Kevin
Kevin
A2??
I'm not sure I'd call the types of amp under discussion true A2 amps.
"A2" Amps using 211 or 845 etc are A1 amps designed to allow to pull some grid current on peaks. The grids are biased negative.
A2 amps draw grid current all the time and have positively biassed grids. The tubes are designed specifically to operate this way and they can typically produce much more power.
I use a 211 SET with DC cathode follower ( Ongaku style) and the arrangement allows it to produce a few emergency watts. I wouldn't call it an A2 amp. Same goes for an IT coupled 845 amp. About 20 watts full bandwidth before distortion rises.
cheers
I'm not sure I'd call the types of amp under discussion true A2 amps.
"A2" Amps using 211 or 845 etc are A1 amps designed to allow to pull some grid current on peaks. The grids are biased negative.
A2 amps draw grid current all the time and have positively biassed grids. The tubes are designed specifically to operate this way and they can typically produce much more power.
I use a 211 SET with DC cathode follower ( Ongaku style) and the arrangement allows it to produce a few emergency watts. I wouldn't call it an A2 amp. Same goes for an IT coupled 845 amp. About 20 watts full bandwidth before distortion rises.
cheers
The classic definition of a class A1 (or AB1) is that no grid current flows on any part of the input cycle. The definition for A2 or AB2 only states that grid current is allowed to flow. I have never seen it stated that grid current must flow for the entire input cycle.
I have the often quoted Morgan Jones book open to page 384 at this moment. It echos the classic definition.
If the definition for A1 demands that NO grid current flows, then what class do these amps operate in?
There are a few transmitting tubes that can be operated such that grid current flows for the entire input cycle. The 811A is the most common of these. If the 811A is operated at plate voltages above about 500 volts the grid will swing negative on peaks. In experiments with A2 on the 833A I measured peak grid currents of over 100 mA on positive peaks. This was at the 200 watt power level. The quiescent grid voltage was +30 volts. The grid swings negative on negative peaks and draws no current at that instant. This requires a serious driver design. I know that it offends many purists, but I use a mosfet source follower. I couldn't find a tube with a plate resistance of 2 ohms.
My 845 amp can produce about 37 watts at the onset of clipping without exceeding the tube specs. The grid begins to go positive above the 27 watt level. At 37 watts the grid reaches +55 volts. The driver must source considerable current at this level. I have not measured this current yet.
Why would I want my amp to be able to do this? Yes the distortion rises slightly as the power goes over 30 watts, and the third harmonic climbs, as the higher harmonics appear. This extra power is only used on rare music peaks and sounds a whole lot better than clipping. I have 10 watts of headroom that I didn't have before I added the A2 capability. It started out as an Ongaku clone, but morphed into something entirely different over time. I am satisfied with the amp design, but I am on my fourth power supply. The power supply IS in the audio path, and really makes the sound on these amps.
I have the often quoted Morgan Jones book open to page 384 at this moment. It echos the classic definition.
If the definition for A1 demands that NO grid current flows, then what class do these amps operate in?
There are a few transmitting tubes that can be operated such that grid current flows for the entire input cycle. The 811A is the most common of these. If the 811A is operated at plate voltages above about 500 volts the grid will swing negative on peaks. In experiments with A2 on the 833A I measured peak grid currents of over 100 mA on positive peaks. This was at the 200 watt power level. The quiescent grid voltage was +30 volts. The grid swings negative on negative peaks and draws no current at that instant. This requires a serious driver design. I know that it offends many purists, but I use a mosfet source follower. I couldn't find a tube with a plate resistance of 2 ohms.
My 845 amp can produce about 37 watts at the onset of clipping without exceeding the tube specs. The grid begins to go positive above the 27 watt level. At 37 watts the grid reaches +55 volts. The driver must source considerable current at this level. I have not measured this current yet.
Why would I want my amp to be able to do this? Yes the distortion rises slightly as the power goes over 30 watts, and the third harmonic climbs, as the higher harmonics appear. This extra power is only used on rare music peaks and sounds a whole lot better than clipping. I have 10 watts of headroom that I didn't have before I added the A2 capability. It started out as an Ongaku clone, but morphed into something entirely different over time. I am satisfied with the amp design, but I am on my fourth power supply. The power supply IS in the audio path, and really makes the sound on these amps.
Ive seen operation where grid current is drawn the entire input cycle referred to as 'A3' in some places. This is not cannon.
When I first started my web page (about 4 years ago) I was experimenting with what I called A3 operation. It was an attempt to avoid the sharp impedance change that occurs when you enter the positive bias region. I was using 811A's at low plate voltages, and enough bias such that the grid never went negative.
I gave this up when I discovered mosfet drive that didn't care whether current was drawn or not.
I gave this up when I discovered mosfet drive that didn't care whether current was drawn or not.
Hey Guys,
Sorry for any confusion with respect to the class of amplification I am shooting for. As tube;lab said, the headroom is what I am looking for. The amp will run in A1 but during peaks have the ability to move to A2 for the added demands of peaks in source material..... At lwast that was one of the original goals of this project. Time will tell where and what it becomes.
On another note, I have a brief question.
I have the fil circuits finished, the HV's and the bias is underway. I have not fired up the HV's as of yet for obvious reasons that the bias is not finished and tested. I have, however, tested all the fil regs and they are working fine. THe heat generated by the regs are sufficiently cooled with the CP cooler. One issue I have noticed is that the transoformers for the 300B fil circuit and the 845 fil circuit are running hot. The 845 is using a 12VCT/6A hammond transformer model 167Q12, and it gets warm, but not too uncomfortable to the touch. The 300B transformer is an 8.5VCT/4A trannie, and this gets VERY warm. It is also a hammond, model 166N8. This is uncomfortable to the touch, and I can only keep my fingers on the casing for about 5-6 seconds. The circuit should only be drawing 1.6A, and therefore the transoformer chould be well spec'd.
So what is "too" hot. At what point do I need to worry about the insulation being worn down, fire hazzard, or another type of malfunction?
All the fil circuits are schottky bridge - C -Reg - C- CMC -, and are set for variable voltage (should I need to change tube types in the future).
Thoughts?
Thanks,
Bryan
Sorry for any confusion with respect to the class of amplification I am shooting for. As tube;lab said, the headroom is what I am looking for. The amp will run in A1 but during peaks have the ability to move to A2 for the added demands of peaks in source material..... At lwast that was one of the original goals of this project. Time will tell where and what it becomes.
On another note, I have a brief question.
I have the fil circuits finished, the HV's and the bias is underway. I have not fired up the HV's as of yet for obvious reasons that the bias is not finished and tested. I have, however, tested all the fil regs and they are working fine. THe heat generated by the regs are sufficiently cooled with the CP cooler. One issue I have noticed is that the transoformers for the 300B fil circuit and the 845 fil circuit are running hot. The 845 is using a 12VCT/6A hammond transformer model 167Q12, and it gets warm, but not too uncomfortable to the touch. The 300B transformer is an 8.5VCT/4A trannie, and this gets VERY warm. It is also a hammond, model 166N8. This is uncomfortable to the touch, and I can only keep my fingers on the casing for about 5-6 seconds. The circuit should only be drawing 1.6A, and therefore the transoformer chould be well spec'd.
So what is "too" hot. At what point do I need to worry about the insulation being worn down, fire hazzard, or another type of malfunction?
All the fil circuits are schottky bridge - C -Reg - C- CMC -, and are set for variable voltage (should I need to change tube types in the future).
Thoughts?
Thanks,
Bryan
UL listed as a class A design (105C), using class B insulation rated at 130C. The smoke may come out if the windings exceed 130C.
Sometimes the peak current causes the transformer to get hot due to momentary saturation. I assume that you are using a full wave rectifier of some kind. half wave causes hot transformers and should not be used in high current applications.
If you are using full wave rectifier followed by a voltage regulator, and you have extra headroom in the regulator, try lowering the value of the input filter cap. This lowers the peak current that the transformer sees and often allows cooler operation.
If you are using full wave rectifier followed by a voltage regulator, and you have extra headroom in the regulator, try lowering the value of the input filter cap. This lowers the peak current that the transformer sees and often allows cooler operation.
You can also add a small amount of resistance or inductance (mH) before or after the rectifier to lower the pk current. Model the effects of doing this in each locations in PSUD if possible. (One case is ac, the other pulsed unipolar dc.) Also check for not so obvious wiring errors, inadequately rated capacitors, etc.
- Status
- Not open for further replies.