47 ohms on the plate is a significant % of the plate Z.
Unbypassed cathode resistors of 1kohm WITH fixed bias??
Why 6080s??
Do you have a usable output transformer.
You know a LARGE part of how a given guitar amp sounds has to do with the "overload" characteristics of stages. This includes the input stage and output stage. Your amp is likely to be somewhat different sounding than any commercial guitar offering.
If ur serious about building using these tubes, I'd suggest that you breadboard it up and try different things before committing to a chassis.
Unbypassed cathode resistors of 1kohm WITH fixed bias??
Why 6080s??
Do you have a usable output transformer.
You know a LARGE part of how a given guitar amp sounds has to do with the "overload" characteristics of stages. This includes the input stage and output stage. Your amp is likely to be somewhat different sounding than any commercial guitar offering.
If ur serious about building using these tubes, I'd suggest that you breadboard it up and try different things before committing to a chassis.
jazbo8, thanks for Mcprounds schematic.. interesting.
The cathode resistors are 1 ohm current sense and for bias adjust and balance only.
The 47 ohm plate resistors function as load sharing and keep the paralleled tubes from driving current into each other in case they are not well matched, Vox used 100 ohm on the AC100.
Why 6080's? Indirectly heated cathode triodes, low plate resistance, readily available and cheap.
According to my composite loadline calculations a 3K-3K OT should work for the paralleled tubes [though that seems high].
I was thinking of one of these 50VA step down toroids. AS-0512 - 50VA 12V Transformer - AnTek Products Corp
The dual pri would allow connection in P-P
Mark
The cathode resistors are 1 ohm current sense and for bias adjust and balance only.
The 47 ohm plate resistors function as load sharing and keep the paralleled tubes from driving current into each other in case they are not well matched, Vox used 100 ohm on the AC100.
Why 6080's? Indirectly heated cathode triodes, low plate resistance, readily available and cheap.
According to my composite loadline calculations a 3K-3K OT should work for the paralleled tubes [though that seems high].
I was thinking of one of these 50VA step down toroids. AS-0512 - 50VA 12V Transformer - AnTek Products Corp
The dual pri would allow connection in P-P
Mark
Last edited:
So you're making a 50 watt amp?
Also these toroids don't do well with even a little DC offset. Some folks seem to do ok with them, so maybe the problem isn't as bad as some have suggested.
Yeah, 1k is how it looked for sure.
You should not need "load sharing" resistors with the cathode resistors set a bit higher in value, although that is not my preferred method. Wonder what plate Z you're figuring on with the 6080. I'd have thought a pair in parallel ought to be on the order of 300 ohms or less??
Vox used what tubes? Not familiar with that particular amp...
The other thing is that you may or may not want to consider separate cathode resistors as this provides a cheap "fuse" method, and if one happens to short or runaway, you still get to run with a single tube. Which during a gig may be better than no amp at all?
_-_-bear
Also these toroids don't do well with even a little DC offset. Some folks seem to do ok with them, so maybe the problem isn't as bad as some have suggested.
Yeah, 1k is how it looked for sure.
You should not need "load sharing" resistors with the cathode resistors set a bit higher in value, although that is not my preferred method. Wonder what plate Z you're figuring on with the 6080. I'd have thought a pair in parallel ought to be on the order of 300 ohms or less??
Vox used what tubes? Not familiar with that particular amp...
The other thing is that you may or may not want to consider separate cathode resistors as this provides a cheap "fuse" method, and if one happens to short or runaway, you still get to run with a single tube. Which during a gig may be better than no amp at all?
_-_-bear
Ok, quick Gargoyle search and I see, those are EL34.
Afaik they are not for "load sharing" but more for parasitic suppression.
In the case of the EL34 with something like a plate Z in the 5kohm range, these 100ohm resistors are a small percentage of the impedance and suck up little power. In your case, unless I am forgetting these tubes each have a plate Z around 300ohms, so in parallel ur around 150ohms. Even ur 47 ohm resistor is a large % of the load value. I'd expect a lot of power wasted in this situation.
Afaik they are not for "load sharing" but more for parasitic suppression.
In the case of the EL34 with something like a plate Z in the 5kohm range, these 100ohm resistors are a small percentage of the impedance and suck up little power. In your case, unless I am forgetting these tubes each have a plate Z around 300ohms, so in parallel ur around 150ohms. Even ur 47 ohm resistor is a large % of the load value. I'd expect a lot of power wasted in this situation.
Yes, DC offset in toroids bad... so dual bias/balance should take care of that.
You may be right about not needing the "load sharing resistors" and in fact separate low value cathode resistors [<10 ohms or so] would be more effective but they will also hurt gain and mu is already low.
All that stuff is easy enough to tweak on the bench.
Here's how I arrived at OT impedance:
Bias:
Vgq =100 Vq=200V Iq=65ma-- just under max diss of 13W
Looking at the curves:
Max plate current of 125ma intersects Vgk = 0 gridline at 35V.
this gives a loadline which intersects 0V axis at 135ma
200V/135ma= 1481 ohms for 1 tube... 750 for 2 tubes-parallel
750 X 4 for anode to anode = 3K
Like I mentioned that seems low to me too.
and maybe my math is fuzzy.
Thanks for your input, it helps to refine the circuit.
Mark
Where is the amp DC coupled? An interstage trasformer won't work with DC and i also see an input cap.
Tesla88, Thanks for your observation, you are of course correct!
I probably should have titled 'transformer coupled" there is however no caps between driver and output grids to cause bias excursion and blocking dist when in grid current.
I may remove anode resistors and add some resistance to each output cathode to help with load sharing but keep fixed bias.
Thanks for the suggestions.
Mark
Last edited:
Agreed a higher value for the cathode resistors is to make the bias partly cathode bias and have the cathode resistors drop 7.5V minimum per the data sheet. Resistors in the plate circuits are usually used with higher gain tubes like screen grid tubes to discourage high frequency oscillation, which should not be happening in 6080s.
I would treat them as they are treated in regulated power supplies so each cathode resistor is large enough to equalize the idling current to some degree. There is no reason with those to ever drive grid current as they can make their full plate current with g1=0V and only 65-70V on the plate. If DC balance is a concern in the toroid there are ways such as a differential pair but it is feedback then and more parts. Or use a quite larger toroid.
Here is a list of resources that should help.
From this discussion, it seems important to have:
1.) DC coupling
2.) Self-balancing for the toroid (or use a much bigger toroid that won't care?)
3.) Minimal variation of balance regardless of varying tube characteristics
The 6080 has an issue in that the sections of any one tube, and the tube to tube variations are much greater than tubes like the 12AX7. Some kind of extreme unction is in order unless the user does not mind adding a balance meter and a 10 turn pot and fiddling.
Ok what is certain is the end diagram to approach those goals won't be as simple as the original one. There are times when it is good to use feedback, including DC feedback. It could be possible to have any feedback be mostly DC, and not so much affect the audio signal. I have to point out that the 6080 is not a very linear tube, so I really see more benefit than harm in adding what is good to make the amp self-balancing.
1.) DC coupled amplifiers with cathodes driving grids:
American Gothic Handbook of AM
be aware it is a 264MB file.
pages:
38
53
129
138
140
Self balancing circuits can be done a few ways.
One is to treat the 6080s as if they are the pass tubes in one of those variable voltage DC supplies. Circuits are everywhere for those. Change it so that the current is what is set, and so that except for idling current, it is not set to a fixed value and one side mirrors the other. Those power supply circuits are hackable to act as current regulators with error voltage sensing etc. just like the usual power supply. A drawback is that unless the audio signal is included in the error amp system, only the DC will be balanced.
2.) Another maybe better method is to set up a self-balancing DC coupled circuit like a scope or chart recorder uses.
This is described in "High Fidelity Circuit Design" by Crowhurst and Cooper, 1956. Page 278. It's an evolution of a more basic audio circuit in the same book around page 111. It ought be downloadable on pmillett.com and readable on https://archive.org/details/HighFidelityCircuitDesign .
3.) A slightly different treatment is given in "Understanding Hi-Fi Circuits" by Crowhurst, I didn't see this book online but have it here.
4.) There is a wealth of self-regulating circuitry in MIT series volume 18 - "Vacuum Tube Amplifiers" at https://www.jlab.org/ir/MITSeries.html
5.) There is a good example of a balanced DC coupled chart recorder amp here. Don't be put off, the frequency response seems only limited by the chart recorder pen, not by the circuit. Sanborn Vacuum-Tube DC General Purpose Amplifier
One thing that is more complex is the op-amp using vacuum tubes, but op-amps are great for controlling other circuits. An example of an old and respected one the K2-W is diagrammed here: Operational amplifier models : Operational Amplifiers - Electronics Textbook and to make the most use of it, the main site The Philbrick Archive should be consulted.
The point is not to make this overly complex, but I hope these references help improve understanding so as to make the amp more successful and easy to use.
There are paper reprints of some of these well worth having. Shop around for the originals and reprints because prices vary crazily.
From this discussion, it seems important to have:
1.) DC coupling
2.) Self-balancing for the toroid (or use a much bigger toroid that won't care?)
3.) Minimal variation of balance regardless of varying tube characteristics
The 6080 has an issue in that the sections of any one tube, and the tube to tube variations are much greater than tubes like the 12AX7. Some kind of extreme unction is in order unless the user does not mind adding a balance meter and a 10 turn pot and fiddling.
Ok what is certain is the end diagram to approach those goals won't be as simple as the original one. There are times when it is good to use feedback, including DC feedback. It could be possible to have any feedback be mostly DC, and not so much affect the audio signal. I have to point out that the 6080 is not a very linear tube, so I really see more benefit than harm in adding what is good to make the amp self-balancing.
1.) DC coupled amplifiers with cathodes driving grids:
American Gothic Handbook of AM
be aware it is a 264MB file.
pages:
38
53
129
138
140
Self balancing circuits can be done a few ways.
One is to treat the 6080s as if they are the pass tubes in one of those variable voltage DC supplies. Circuits are everywhere for those. Change it so that the current is what is set, and so that except for idling current, it is not set to a fixed value and one side mirrors the other. Those power supply circuits are hackable to act as current regulators with error voltage sensing etc. just like the usual power supply. A drawback is that unless the audio signal is included in the error amp system, only the DC will be balanced.
2.) Another maybe better method is to set up a self-balancing DC coupled circuit like a scope or chart recorder uses.
This is described in "High Fidelity Circuit Design" by Crowhurst and Cooper, 1956. Page 278. It's an evolution of a more basic audio circuit in the same book around page 111. It ought be downloadable on pmillett.com and readable on https://archive.org/details/HighFidelityCircuitDesign .
3.) A slightly different treatment is given in "Understanding Hi-Fi Circuits" by Crowhurst, I didn't see this book online but have it here.
4.) There is a wealth of self-regulating circuitry in MIT series volume 18 - "Vacuum Tube Amplifiers" at https://www.jlab.org/ir/MITSeries.html
5.) There is a good example of a balanced DC coupled chart recorder amp here. Don't be put off, the frequency response seems only limited by the chart recorder pen, not by the circuit. Sanborn Vacuum-Tube DC General Purpose Amplifier
One thing that is more complex is the op-amp using vacuum tubes, but op-amps are great for controlling other circuits. An example of an old and respected one the K2-W is diagrammed here: Operational amplifier models : Operational Amplifiers - Electronics Textbook and to make the most use of it, the main site The Philbrick Archive should be consulted.
The point is not to make this overly complex, but I hope these references help improve understanding so as to make the amp more successful and easy to use.
There are paper reprints of some of these well worth having. Shop around for the originals and reprints because prices vary crazily.
opcom, that's fantastic!!!
That's for all the reference links, that's what I'm looking for.
As far as cathode resistance, it looks like around 120 ohms on each cathode at a proposed bias of 65ma per tube would do it.
I have provided for static DC balance in the toroid but I like what I think you're saying about DC feedback, considering non linearity and mismatched sections in 6080's that the dynamic balance could be all over the map once signal is applied, if I understand that correctly. I'll look into it
Thanks for the help.
The point of this project is not so much to build the killer guitar amp, but to implement these "hard to use" tubes in a creative design and learn something more than RC coupled pentodes with tons of NFB
I've got some reading to do.
Mark
That's for all the reference links, that's what I'm looking for.
As far as cathode resistance, it looks like around 120 ohms on each cathode at a proposed bias of 65ma per tube would do it.
I have provided for static DC balance in the toroid but I like what I think you're saying about DC feedback, considering non linearity and mismatched sections in 6080's that the dynamic balance could be all over the map once signal is applied, if I understand that correctly. I'll look into it
Thanks for the help.
The point of this project is not so much to build the killer guitar amp, but to implement these "hard to use" tubes in a creative design and learn something more than RC coupled pentodes with tons of NFB
I've got some reading to do.
Mark
Here is the updated circuit.
I think this bias balance servo will work. just have to calculate loadline for PNP's.
The MPSA92 is 200V device so should be OK.
The 2200uf caps are to increase the time constant to sub audio so it wont affect signal but respond to average shift over time [I hope] The 120 ohm cathode resistors are to balance current through both sections and stabilize runaway as was mentioned by bear and opcom in previous posts.
[/quote]
I think this bias balance servo will work. just have to calculate loadline for PNP's.
The MPSA92 is 200V device so should be OK.
The 2200uf caps are to increase the time constant to sub audio so it wont affect signal but respond to average shift over time [I hope] The 120 ohm cathode resistors are to balance current through both sections and stabilize runaway as was mentioned by bear and opcom in previous posts.
Hi Guys
MPSA92 is 300V; you must be thinking of MPSA93.
You might want to add base-stops for the BJTs. You should also add a resistor across the bridge to prevent cross-conduction failures in standby, and to act as a bleeder when power is turned off. Note that standbys are not required at these voltages or power levels.
The sensitivity of toroidal OTs to DC imbalance is quite over-played. A simple passive bias-set network is fine and the tubes will stay in adjustment for a long period of time. If you use active bias control it is a good idea to add a current clamp to protect the tubes from runaway.
The usual practice when paralleling multisection tubes as you have done, is to use one section of each tube on each half of the circuit. Where wiring one complete bottle to each side is simpler, the split approach spreads the heat more evenly over the two tubes.
If the target grid voltage is something around -120V, then the bias supply should be designed to provide much more than this. You want to be able to turn the tubes off and 120V won't do it unless Va is quite low.
Menno vander Veen used the 6AS7 in a push-pull cathode-biased amp to produce about 7W (the Maurits amp). The OT is 8kaa and Rk=1k5, bypassed; Va=390V. The limiting part of the circuit was the driver which used local feedback to attain a lower drive impedance, but the signal swing was less than what would have pushed the output to full capability.
Have fun
MPSA92 is 300V; you must be thinking of MPSA93.
You might want to add base-stops for the BJTs. You should also add a resistor across the bridge to prevent cross-conduction failures in standby, and to act as a bleeder when power is turned off. Note that standbys are not required at these voltages or power levels.
The sensitivity of toroidal OTs to DC imbalance is quite over-played. A simple passive bias-set network is fine and the tubes will stay in adjustment for a long period of time. If you use active bias control it is a good idea to add a current clamp to protect the tubes from runaway.
The usual practice when paralleling multisection tubes as you have done, is to use one section of each tube on each half of the circuit. Where wiring one complete bottle to each side is simpler, the split approach spreads the heat more evenly over the two tubes.
If the target grid voltage is something around -120V, then the bias supply should be designed to provide much more than this. You want to be able to turn the tubes off and 120V won't do it unless Va is quite low.
Menno vander Veen used the 6AS7 in a push-pull cathode-biased amp to produce about 7W (the Maurits amp). The OT is 8kaa and Rk=1k5, bypassed; Va=390V. The limiting part of the circuit was the driver which used local feedback to attain a lower drive impedance, but the signal swing was less than what would have pushed the output to full capability.
Have fun
Last edited:
The above bias calculation assumes class A operation, but you are running class AB1 (I agree that you forget about running AB2 for the reason already stated above). For class AB, the important load line is class B which is 1/4 of the Ra-a. So you need to re-do your loadline and pick a different OPT.
Thanks guys,
Sruth, thanks for the circuit mod suggestions, very helpful.
I assume low value for base stops,<2K? Splitting the triode sections makes sense, I was going to go the lazy route and wire them together but I will split them.
I had shown the original circuit with passive bias and balance but I kind of like the idea of the active control. I will look into current clamp, perhaps I can incorporate it into the bias control.
The Maurits amp with Va at 390v is interesting since specs show Va max at 250v.
jazbo8,
I have done some load line plots, let me know if it looks correct.
Here's my observations:
Class B load line is a-a/4 and always intersects Va at Ia=0
Class A load line is a-a/2 and intersects Vq at the operating point.
For class A operation the class A and class B load lines intersect at Vg =0 gridline.
Class AB1: A and B load lines intersect at some point >Vg=0 [more neg grid line]
It looks like the ideal class A operation would be:
Raa 5.5K, Va=205v, bias 62ma. limited by Ia max
However, instead of trying to use an oddball OPT I can use a common 22848, 4k and use 8K load line:
8K aa, Va 250v, bias 50ma, class AB
These load lines are for one tube so actual impedance would be half.
I am not going with the toroids after all and I have already wound a power transformer to original voltages but it wont be too hard to rewind the HT to optimize Va or use mosfet VR in similar designs in the future.
Chassis fabrication has started so it won't be too long.
Sruth, thanks for the circuit mod suggestions, very helpful.
I assume low value for base stops,<2K? Splitting the triode sections makes sense, I was going to go the lazy route and wire them together but I will split them.
I had shown the original circuit with passive bias and balance but I kind of like the idea of the active control. I will look into current clamp, perhaps I can incorporate it into the bias control.
The Maurits amp with Va at 390v is interesting since specs show Va max at 250v.
jazbo8,
I have done some load line plots, let me know if it looks correct.
Here's my observations:
Class B load line is a-a/4 and always intersects Va at Ia=0
Class A load line is a-a/2 and intersects Vq at the operating point.
For class A operation the class A and class B load lines intersect at Vg =0 gridline.
Class AB1: A and B load lines intersect at some point >Vg=0 [more neg grid line]
It looks like the ideal class A operation would be:
Raa 5.5K, Va=205v, bias 62ma. limited by Ia max
However, instead of trying to use an oddball OPT I can use a common 22848, 4k and use 8K load line:
8K aa, Va 250v, bias 50ma, class AB
These load lines are for one tube so actual impedance would be half.
I am not going with the toroids after all and I have already wound a power transformer to original voltages but it wont be too hard to rewind the HT to optimize Va or use mosfet VR in similar designs in the future.
Chassis fabrication has started so it won't be too long.
Since this is a guitar amp, the load lines that you showed are still pretty conservative, the output power would be pretty low (unless that's what you want). BTW, the Ia max rating of 125mA is for continuous sine wave signal, which is not really applicable for MI use. Why don't you try biasing it with the 70% rule and using a lower Ra-a and see what you get...
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.