Hello everyone,
I have been doing quite a bit of reading on building a simply tube based buffer to experiment with. There is a great and at times confusing amount of information on this subject.
Easy Cathode Followers (Page 2): March's Circuit of the Month
Please see the Easy Pentode Circuit section.
If I add a 170v power source is this a functional circuit? Also would it be possible to wire in a switch to go from pentode to triode?
I have a Denon PMA-900V, I would like to use this in the processor loop so that I can take it in and out of the audio circuit.
I would appreciate any help and opinions.
I have been doing quite a bit of reading on building a simply tube based buffer to experiment with. There is a great and at times confusing amount of information on this subject.
Easy Cathode Followers (Page 2): March's Circuit of the Month
Please see the Easy Pentode Circuit section.
If I add a 170v power source is this a functional circuit? Also would it be possible to wire in a switch to go from pentode to triode?
I have a Denon PMA-900V, I would like to use this in the processor loop so that I can take it in and out of the audio circuit.
I would appreciate any help and opinions.
I'm fairly new to tubes, but why would you want to use a power tube like the 6BQ5/EL84 for this?
Surely a small signal pentode or triode would do fine, and save you a lot in the power supply section?
Do a search on "cathode follower" or "tube buffer" here; you should get plenty of results.
Gary
Surely a small signal pentode or triode would do fine, and save you a lot in the power supply section?
Do a search on "cathode follower" or "tube buffer" here; you should get plenty of results.
Gary
kinda depends what your source is and what you are buffering to.
for buffering a line signal before an amp (ss or hs) the EL84 is overkill imho - there are plenty of smaller cheaper signal level tubes that would fit the bill and would operate closer to their optimum operating point - just one twin triode tube would do it for two channels.
In answer to the question from the op - yep, it will work. I haven't done the math so I couldn't say whether is it optimal.
for buffering a line signal before an amp (ss or hs) the EL84 is overkill imho - there are plenty of smaller cheaper signal level tubes that would fit the bill and would operate closer to their optimum operating point - just one twin triode tube would do it for two channels.
In answer to the question from the op - yep, it will work. I haven't done the math so I couldn't say whether is it optimal.
Gary and Aard,
I believe that both of you are right in that the EL84 is overkill, the idea came from the link I provided. That is why I asked about being able to switch it from pentode to triode with a switch, as I would open up my experimentation options. I do not have my heart set on this tube, but I do like trying things that people "shake their heads" at. Aard as I stated in the first post I would like to use this in the processor loop of my PMA-900V which has a input of 150mV/30K ohm.
Please let me know what information you need and I will do my best to get it.
I am not looking for the "best" sound, as a matter of fact I am in search of tube distortion albeit the good kind. I am not a purist, I am simply looking to advance my very limited knowledge, have a little fun and to say "look what I made."
Thank you to both for your comments so far and please feel free to offer more.
Riki
I believe that both of you are right in that the EL84 is overkill, the idea came from the link I provided. That is why I asked about being able to switch it from pentode to triode with a switch, as I would open up my experimentation options. I do not have my heart set on this tube, but I do like trying things that people "shake their heads" at. Aard as I stated in the first post I would like to use this in the processor loop of my PMA-900V which has a input of 150mV/30K ohm.
Please let me know what information you need and I will do my best to get it.
I am not looking for the "best" sound, as a matter of fact I am in search of tube distortion albeit the good kind. I am not a purist, I am simply looking to advance my very limited knowledge, have a little fun and to say "look what I made."
Thank you to both for your comments so far and please feel free to offer more.
Riki
I don't like it. If you want to supply a pentode's screen from an active decoupler, a transistor would be a far better choice for that since the drain or collector is a good deal more insensitive to AC on the rail than any triode.
Secondly, if you're going to use a pentode for a cathode follower, the screen voltage must follow the cathode voltage, otherwise, it's not a pentode anymore. That means either an independent, floating screen supply or a series dropping resistor with capacitor coupling to the cathode. Pentode cathode followers are one case where you don't want to stiffen the screen supply.
If the screen-to-DC ground voltage is fixed, then Vsg moves with the cathode voltage. That's not what you want (it's more like ultralinear operation then). Vsg has to move with the cathode to make Vsg a constant.
Secondly, if you're going to use a pentode for a cathode follower, the screen voltage must follow the cathode voltage, otherwise, it's not a pentode anymore. That means either an independent, floating screen supply or a series dropping resistor with capacitor coupling to the cathode. Pentode cathode followers are one case where you don't want to stiffen the screen supply.
If the screen-to-DC ground voltage is fixed, then Vsg moves with the cathode voltage. That's not what you want (it's more like ultralinear operation then). Vsg has to move with the cathode to make Vsg a constant.
A buffer does exactly that - its buffers the output of a device (like a CD player) from the input of the following device (like a power amp). Usually, it is used to ensure a very light load is presented to a source while providing a low Z (and so by implication high current) input to the following device.
They normally do not amplify or alter the signal, just increase its ability to feed a heavy load.
A pre-amp typically amplifies the signal, and may ALSO (but not always) have a low Zout
OK, I'd go with a simple 12A*7 cathode follower buffer.
Easy to build, a wide range of easiliy sourced tubes with the same pin-out for tube-rolling, only one tube needed (appeals to the el cheapo in me!), simple to understand.
Miles is much more onto the technicalities than I, but for a simple to read and practical start, go to valvewizard's page on cathode followers. All the calcs you need are there. Bear in mind he is primarily interested in guitars, so mild distortion is his forte!
They normally do not amplify or alter the signal, just increase its ability to feed a heavy load.
A pre-amp typically amplifies the signal, and may ALSO (but not always) have a low Zout
OK, I'd go with a simple 12A*7 cathode follower buffer.
Easy to build, a wide range of easiliy sourced tubes with the same pin-out for tube-rolling, only one tube needed (appeals to the el cheapo in me!), simple to understand.
Miles is much more onto the technicalities than I, but for a simple to read and practical start, go to valvewizard's page on cathode followers. All the calcs you need are there. Bear in mind he is primarily interested in guitars, so mild distortion is his forte!
Last edited:
Whoa, that went so far over my head, I don't even think I saw it. 🙂
What is the active decoupler in the provided circuit? The circuit is made up of resistors and capacitors, I thought those were passive devices? The screen or is it a grid, closest to the plate is connected to the cathode.
I think I have a good bit more reading to do.
What is the active decoupler in the provided circuit? The circuit is made up of resistors and capacitors, I thought those were passive devices? The screen or is it a grid, closest to the plate is connected to the cathode.
I think I have a good bit more reading to do.
So, since I am using the processor loop I don't "need" a buffer for its intended propose? The work for this applacation because they do not provide any gain/amplification.
As for the Valve Wizard site, AC or DC coupled?
Can I build the circuit shown on Valve Wizard's DC coupled Cathode Follower and start experimenting?
Riki
As for the Valve Wizard site, AC or DC coupled?
Can I build the circuit shown on Valve Wizard's DC coupled Cathode Follower and start experimenting?
Riki
yes rick - build one. AC coupled is easier for you and no, as long as you use reasonable values of resistance and capacitance, it will do no damage.
They are simple (20 minutes soldering) to build, and easy to design. Try it - draw up what you think will work, then come back in and ask for positive critique.
Don't forget the power supply - you'll need 250-300V if you are wanting to use the 12A*7 series. You don't need a lot of current though - between 2 and 10mA per channel, so you can get away with a reallllly titchy transformer and lightweight filtering.
If you want to design a power supply go to Duncan's Amp Pages and download the PSUD-II power supply design software. My 10 year old can use this and it makes psu design a doddle. Also good for your learning.
Oh, and its free.
By now you should have a grip on my design imperatives - cheap, easy, cheap, simple to build, cheap, accessable and cheap.
They are simple (20 minutes soldering) to build, and easy to design. Try it - draw up what you think will work, then come back in and ask for positive critique.
Don't forget the power supply - you'll need 250-300V if you are wanting to use the 12A*7 series. You don't need a lot of current though - between 2 and 10mA per channel, so you can get away with a reallllly titchy transformer and lightweight filtering.
If you want to design a power supply go to Duncan's Amp Pages and download the PSUD-II power supply design software. My 10 year old can use this and it makes psu design a doddle. Also good for your learning.
Oh, and its free.
By now you should have a grip on my design imperatives - cheap, easy, cheap, simple to build, cheap, accessable and cheap.
I'm curious what you are looking to buffer from seeing the Denon's (presumed) built in buffer? Or does the Denon's loop have no buffer itself?
Actually jjman that has been the hold up for me, I don't want to buffer anything. My system is all SS, I like my system, I have the computer hooked to a LM-DAC3 V8, Hong Kong special, from there to the CD input of the PMA-900V and out to a pair of Infinity RS III's, the floor-standing ones, not the book shelve ones.
The thing is the Infinity's have that great EMIT, but over time the I find myself using the tone controls to turn down the treble.
I have had the pleasure of listening to a Citation II, yes I know that is on another plane of existence from what we are talking about here, but ever since that day I have wanted to learn about tubes. I learn best from doing, so it is better to think of this as an experiment for learning, then for building a device to solve a problem.
aard I agree with you, cheap, easy and simply to build. So I will do some reading on A/C coupled. I have spent so time at Duncan's before, never thought to download the software, I will have to do that. Which software is best for me to start with for simulating circuits with?
Thanks to all of you and I hope to be back soon with a design idea.
The thing is the Infinity's have that great EMIT, but over time the I find myself using the tone controls to turn down the treble.
I have had the pleasure of listening to a Citation II, yes I know that is on another plane of existence from what we are talking about here, but ever since that day I have wanted to learn about tubes. I learn best from doing, so it is better to think of this as an experiment for learning, then for building a device to solve a problem.
aard I agree with you, cheap, easy and simply to build. So I will do some reading on A/C coupled. I have spent so time at Duncan's before, never thought to download the software, I will have to do that. Which software is best for me to start with for simulating circuits with?
Thanks to all of you and I hope to be back soon with a design idea.
I use a Pioneer automotive EQ in the "loop" of my home stereo. Works very nicely especially when watching a live TV music program since they are often mixed poorly. I can't live w/o a multi-band EQ in my car and home.
My tube experience is limited to my 5 guitar amps so far. I've been resisting the temptation to jump in the tube pool for the home setup. I'm sure I'll give in some day.....
My tube experience is limited to my 5 guitar amps so far. I've been resisting the temptation to jump in the tube pool for the home setup. I'm sure I'll give in some day.....
A pencil and a calculator at this level.
You could go spice, but the reality is you will learn more with the data sheet including a set of anode curves, a sheet of graph paper, a pencil, and bottle of chardonnay.
The basic circuit is small and simple - if you can understand the transfer characteristics (gleaned from the anode curves), you are home and hosed, and will be able to do it again without even thinking about it.
Adding simulation software to the mix just adds another thing you have to learn before you can do what you actually want to do - build.
I look at it as a learning experience more then anything, plus my soldering skills could use the help. Worst case, I don't like it but I can use the materials to make something else.
aard: Very good point and that is the way I have been doing it. I have been working on a Magnavox 88 series amp, it needs new resistors. That is the reason I am trying to iron out this buffer, so I can order everything at once. Still working on the circuit.
aard: Very good point and that is the way I have been doing it. I have been working on a Magnavox 88 series amp, it needs new resistors. That is the reason I am trying to iron out this buffer, so I can order everything at once. Still working on the circuit.
Last edited:
Ok aard here goes, using the circut on Wizard's site for a A/C coupled follower.
Ht=280V
Av=.95
bias of -8v
Cin=22nF
Cout=1.1 microF
Rg=470K
Rb=470K
Ri=18K
If I am reading the manual for PMA-900V right the input load for the loop is 30K.
Do these values seem right?
Thanks
Ht=280V
Av=.95
bias of -8v
Cin=22nF
Cout=1.1 microF
Rg=470K
Rb=470K
Ri=18K
If I am reading the manual for PMA-900V right the input load for the loop is 30K.
Do these values seem right?
Thanks
lol - you ARE keen!
What tube are you designing for?
Numbers look in the ballpark, but bias seems a bit high - I would be picking to bias closer to -2v for a 12A*7 series tube.
edit- OK guessing you are setting up for an EL84, is that right?
Check Rb - 470 or 470k?
What tube are you designing for?
Numbers look in the ballpark, but bias seems a bit high - I would be picking to bias closer to -2v for a 12A*7 series tube.
edit- OK guessing you are setting up for an EL84, is that right?
Check Rb - 470 or 470k?
Last edited:
Sorry aard,
Going with a 12AU7/ECC82 as the tube. Over on Wizard, he states that "for maximum headroom we need to center-bias the stage bias. In this case a bias of about -8V would achieve this. On the other hand, if we don't expect to drive really low loads then we are probably more interested in very low output impedance for noise rejection (such as when driving long external cables), in which case we want to maximize gm by hot-biasing. In this case a bias of about -3.4V is chosen, giving about 50Vpk before clipping, into easy loads."
Since I am using this in the processor loop my cables will be rather short, 3-5 ft at most. I don't have a problem going with the -3.4V to help keep this as simply as possible for me. Could I use a pot so that I could adjust bias on the fly to see how it would effect the sound?
Good catch on the Rb, it is in fact 470 ohm.
Also I do not understand what Wizard is talking about when he talks about Heater elevation, I thought the heater voltage was 6.3V?
Going with a 12AU7/ECC82 as the tube. Over on Wizard, he states that "for maximum headroom we need to center-bias the stage bias. In this case a bias of about -8V would achieve this. On the other hand, if we don't expect to drive really low loads then we are probably more interested in very low output impedance for noise rejection (such as when driving long external cables), in which case we want to maximize gm by hot-biasing. In this case a bias of about -3.4V is chosen, giving about 50Vpk before clipping, into easy loads."
Since I am using this in the processor loop my cables will be rather short, 3-5 ft at most. I don't have a problem going with the -3.4V to help keep this as simply as possible for me. Could I use a pot so that I could adjust bias on the fly to see how it would effect the sound?
Good catch on the Rb, it is in fact 470 ohm.
Also I do not understand what Wizard is talking about when he talks about Heater elevation, I thought the heater voltage was 6.3V?
ok gotcha now.
Heater elevation is a trick that keeps the heater running at 6.3v but "lifting" it from an earth reference by referencing to a higher voltage. Don't worry if you don't click to it real quick - you will get it.
Habitually we reference everything to earth, ground or 0volts - we (wrongly) call it this interchangably.
If you are used to ss stuff though, you will understand + and - rails both equidistant from a 0V power and signal ground. So you will know that voltages are not always referenced to ground and ground may be ABOVE a voltage we are checking.
Took me ages to click to cos I'm used to automotive electronics where everything referenced a 0v earth.
I digress.
To ensure you keep the heater circuit within safe voltages of the cathode voltage, you often have to elevate the heater voltage. Another option is to float the heaters.
Floating is easy - just make sure that none of your heater circuit is connected to ground ANYWHERE. There you go - done. You just built a floating heater circuit. In your situation, this is probably the best solution EXCEPT for one thing - hum. A floating AC heater is an antenna sending out a hum signal at 60hz. Dammit.
To elevate the heater, you have to float it, and then attach both ends to a raised voltage source.
In your case, we will attach it to the top of the cathode resistor and the bottom of the bias resistor. Now the heater is running at 6.3V, but elevated to cathode voltage. We are safe.
Pictures soon.
A pot to adjust on the fly is ok BUT make sure that you don't leave enough room for the bias to allow the grid to go positive. Ugly results. I would make a bias resistance out of two fixed resistors in series with a pot across one or other of them. Now you cannot drop BELOW a specific bias voltage (say, -1.5V) but have the ability to crank in more to find out what happens.
Capiche?
Heater elevation is a trick that keeps the heater running at 6.3v but "lifting" it from an earth reference by referencing to a higher voltage. Don't worry if you don't click to it real quick - you will get it.
Habitually we reference everything to earth, ground or 0volts - we (wrongly) call it this interchangably.
If you are used to ss stuff though, you will understand + and - rails both equidistant from a 0V power and signal ground. So you will know that voltages are not always referenced to ground and ground may be ABOVE a voltage we are checking.
Took me ages to click to cos I'm used to automotive electronics where everything referenced a 0v earth.
I digress.
To ensure you keep the heater circuit within safe voltages of the cathode voltage, you often have to elevate the heater voltage. Another option is to float the heaters.
Floating is easy - just make sure that none of your heater circuit is connected to ground ANYWHERE. There you go - done. You just built a floating heater circuit. In your situation, this is probably the best solution EXCEPT for one thing - hum. A floating AC heater is an antenna sending out a hum signal at 60hz. Dammit.
To elevate the heater, you have to float it, and then attach both ends to a raised voltage source.
In your case, we will attach it to the top of the cathode resistor and the bottom of the bias resistor. Now the heater is running at 6.3V, but elevated to cathode voltage. We are safe.
Pictures soon.
A pot to adjust on the fly is ok BUT make sure that you don't leave enough room for the bias to allow the grid to go positive. Ugly results. I would make a bias resistance out of two fixed resistors in series with a pot across one or other of them. Now you cannot drop BELOW a specific bias voltage (say, -1.5V) but have the ability to crank in more to find out what happens.
Capiche?
ok, promised a picture, here it is.
Firstly, the schematic shows a 12A*7 tube. It has the on-the-fly bias adjustment you want via Rbadj.
A high value (330k-470k) resistor Rel taps from the cathode to the centre tap (pin 9) of the heater circuit. You can choose to apply 12.6V across the other two heater pins (4 and 5) or you can apply 6.3V across (4 and 9) and (5 and 9).
Make sure that NONE of your heater circuit is attached to earth ground 0V or any other such point!
Now your heater circuit is elevated to the same voltage as the cathode, less the bias voltage.
Open at this point for anyone else to chime in.
Firstly, the schematic shows a 12A*7 tube. It has the on-the-fly bias adjustment you want via Rbadj.
A high value (330k-470k) resistor Rel taps from the cathode to the centre tap (pin 9) of the heater circuit. You can choose to apply 12.6V across the other two heater pins (4 and 5) or you can apply 6.3V across (4 and 9) and (5 and 9).
Make sure that NONE of your heater circuit is attached to earth ground 0V or any other such point!
Now your heater circuit is elevated to the same voltage as the cathode, less the bias voltage.
Open at this point for anyone else to chime in.
- Status
- Not open for further replies.