After hijacking Rob's tread I'd like to give this one a try. As always comments and suggestions welcome, corrections too 
( schematic )
( load lines )
( power supply )
( step response )
An externally hosted image should be here but it was not working when we last tested it.
( schematic )
An externally hosted image should be here but it was not working when we last tested it.
( load lines )
An externally hosted image should be here but it was not working when we last tested it.
( power supply )
An externally hosted image should be here but it was not working when we last tested it.
( step response )
layout of components
Sawing and glueing it together in a (usually too small) 32x21cm wooden box, looking for optimal placement. All is to come underneath... tight fit.
- home brewn toroid: 5V & 5Vac 3.5A each
- small 30VA 182V power transformer
- two 10.5H 100 ohm chokes
- Elna Cerafine 2x 220uF 350V (little bit overkill)
- 15uF motor run cap (yes, MKV B25832 )
( layout 1)
( dito )
Sawing and glueing it together in a (usually too small) 32x21cm wooden box, looking for optimal placement. All is to come underneath... tight fit.
- home brewn toroid: 5V & 5Vac 3.5A each
- small 30VA 182V power transformer
- two 10.5H 100 ohm chokes
- Elna Cerafine 2x 220uF 350V (little bit overkill)
- 15uF motor run cap (yes, MKV B25832
)An externally hosted image should be here but it was not working when we last tested it.
( layout 1)
An externally hosted image should be here but it was not working when we last tested it.
( dito )
Looks like a fun project.
I've messed around a lot with the 27 some time ago, and I found that it really shines with a gyrator load.
Of course that would totally make the CCDA useless, but 15 euros worth of parts can make a really effective and simple regulator, which removes pretty much all worry about the PSU anyway.
In my experience the 27 doesn't have very good heater - cathode insulation, and I found it best to regulate the heater supply; it made an audible difference.
All in all it's a very good tube, if you can find a good condition mesh plate 27 it's the nicest sounding IDHT I've tried.
I've messed around a lot with the 27 some time ago, and I found that it really shines with a gyrator load.
Of course that would totally make the CCDA useless, but 15 euros worth of parts can make a really effective and simple regulator, which removes pretty much all worry about the PSU anyway.
In my experience the 27 doesn't have very good heater - cathode insulation, and I found it best to regulate the heater supply; it made an audible difference.
All in all it's a very good tube, if you can find a good condition mesh plate 27 it's the nicest sounding IDHT I've tried.
Thanks, that will be fun next experimenting with heaters. On the other hand I feel once heaters are fixed the power supply can't stay behind. Comes time, comes .... something else. Look before one leaps over board etc..
Some numbers...
It would be nice to know the bandwith, as the pre amp should ideally not limit the output stage by a factor of 2 to 3. The latter being estimated at 30kHz would take the pre at around 75kHz. Alright, in goes a fresh battery to the old Casio.
Where this current cancelling amp deviates from normal grounded cathode design is not apparent to me but I'll give it a try.
First there's the output impedance of the CD-player to the first stage. As I have no clue of its output impedance I'll guestimate 10k ohm. Cmiller= (6.5+1) x 3.3 = 25pF. Some 15pF for strays makes 40pF.
Cutt-off = 1/(2pi x 4.10e-11 x1.10e4 ) = 398kHz. That's a large number to count on one hand
Second there's the common grounded cathode stage.
rout= ra (the dynamic anode resistance) in parallel with RL (the 24 k ohm anode resister).
From the load line tanget I derive 200-60V over 16mA = 8.75 kohm. The result equaling 6412 ohm.
This 6K4 is looking in to the input capacitance of the cathode follower,
which is 3.3pF + (1-0.96)x3.5pF = 3.44pF. Allowing for some stray, say 7.5pF.
The cutt-off frequency = 1/(2pi x 7.5.10e-12 x 6.4.10e+3) = 3.3 megaHz. Wow, that must be a mistake
Last is the output impedance of the cathode follower at the combined capacitance presented by the interstage and the power amplifier. Because rout is lowish (round 1k5) this number is benign, meaning out of the way of the bottle neck of 398k.
Some numbers...
It would be nice to know the bandwith, as the pre amp should ideally not limit the output stage by a factor of 2 to 3. The latter being estimated at 30kHz would take the pre at around 75kHz. Alright, in goes a fresh battery to the old Casio.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
Where this current cancelling amp deviates from normal grounded cathode design is not apparent to me but I'll give it a try.
First there's the output impedance of the CD-player to the first stage. As I have no clue of its output impedance I'll guestimate 10k ohm. Cmiller= (6.5+1) x 3.3 = 25pF. Some 15pF for strays makes 40pF.
Cutt-off = 1/(2pi x 4.10e-11 x1.10e4 ) = 398kHz. That's a large number to count on one hand
Second there's the common grounded cathode stage.
rout= ra (the dynamic anode resistance) in parallel with RL (the 24 k ohm anode resister).
From the load line tanget I derive 200-60V over 16mA = 8.75 kohm. The result equaling 6412 ohm.
This 6K4 is looking in to the input capacitance of the cathode follower,
which is 3.3pF + (1-0.96)x3.5pF = 3.44pF. Allowing for some stray, say 7.5pF.
The cutt-off frequency = 1/(2pi x 7.5.10e-12 x 6.4.10e+3) = 3.3 megaHz. Wow, that must be a mistake
Last is the output impedance of the cathode follower at the combined capacitance presented by the interstage and the power amplifier. Because rout is lowish (round 1k5) this number is benign, meaning out of the way of the bottle neck of 398k.
I'd put the power supply and pre-amp in separate boxes. (Hum pick up) Will give you much more space to work and tweak.. A metal box with a wood wrap would reduce external electro-static pick up issues. (The tube has this issue too and can be shielded as I do with the 26 for mitigation)
Most 27/227/327 (globes) are quite microphonic so be prepared to deal with getting a number of them to select for lower microphonics. Note that the very expensive and rare Arcturus Blue globe mesh plate with top nipple sound much worse IMVLE than the garden variety RCA UY-227. (Sadly they look soooo cool. ) I've used 27s in a few power amp and line stage designs, but actually prefer the 26/226/326.
In addition the filament to cathode insulation is not very good in the 27 so I would use a separate transformer to heat the cathode follower which should be referenced to a voltage just a few volts higher than that found on the follower cathode.
Other devices to consider include the 56 (2.5V) or 76 (6.3V)..
Something to note here is that the feedback is essentially positive which IMO is not a good thing on a number of fronts from DC stability to potential effects on linearity. Decades ago I experimented with essentially the same circuit using the 12AU7A, and there is a point at which it may oscillate with injudiciously chosen component values between the cathodes of the two tubes.
Most 27/227/327 (globes) are quite microphonic so be prepared to deal with getting a number of them to select for lower microphonics. Note that the very expensive and rare Arcturus Blue globe mesh plate with top nipple sound much worse IMVLE than the garden variety RCA UY-227. (Sadly they look soooo cool.
) I've used 27s in a few power amp and line stage designs, but actually prefer the 26/226/326.In addition the filament to cathode insulation is not very good in the 27 so I would use a separate transformer to heat the cathode follower which should be referenced to a voltage just a few volts higher than that found on the follower cathode.
Other devices to consider include the 56 (2.5V) or 76 (6.3V)..
Something to note here is that the feedback is essentially positive which IMO is not a good thing on a number of fronts from DC stability to potential effects on linearity. Decades ago I experimented with essentially the same circuit using the 12AU7A, and there is a point at which it may oscillate with injudiciously chosen component values between the cathodes of the two tubes.
The challenge is, Kevin, to make a good functioning piece of equipment to enjoy for a long time, both visually and auditory. Athough tastes may differ I enjoy the sight of those pretty, slim boxes hiding technique where possible. In my own country there's a tendency for building with RVS and exotic woods for those willing and able to produce a statement. In Italy and Japan too the same craftmanship has lead to beautiful expressions of good taste. It's far more satisfying than bolting it all down in an industrial box.
So, first I'll have a go with the one-in-all box, although not ideal. I bear in mind your remark w.r.t. a metal shield (against RF I guess?). Would the anode be prone to pick up stray fields of computers and light bulbs?
The 26 is a beautiful direct heated triode but I experienced a lot of fuzz heating it. This 27 try is for a 'petite'; easy to place and smooth to the ear.
You suggested the use of a seperate transformer for heating the CFs. My idea is to power them in series from a seperate winding of the heater trannie and raise its potential 120 volts. As I understand you correctly, it would be beneficial to go even 10 volts higher? The Cunningham C-327 datasheet (pre WW2) recommends no potential difference between heater and cathode. If this practice is not followed, the heater may be biased preferably negative, but allowably positive, with respect to the cathode by no more than 45 volts.
I'm not familiar with the ins and outs of the circuit as Mr. Broskie has a talent of loosing me somewhere in the story, not necessary his bad. Is the feedback positive because of the mutual cathode resistor?
It's good to experiment else we'd all be listening to ipods on earplugs.
So, first I'll have a go with the one-in-all box, although not ideal. I bear in mind your remark w.r.t. a metal shield (against RF I guess?). Would the anode be prone to pick up stray fields of computers and light bulbs?
The 26 is a beautiful direct heated triode but I experienced a lot of fuzz heating it. This 27 try is for a 'petite'; easy to place and smooth to the ear.
You suggested the use of a seperate transformer for heating the CFs. My idea is to power them in series from a seperate winding of the heater trannie and raise its potential 120 volts. As I understand you correctly, it would be beneficial to go even 10 volts higher? The Cunningham C-327 datasheet (pre WW2) recommends no potential difference between heater and cathode. If this practice is not followed, the heater may be biased preferably negative, but allowably positive, with respect to the cathode by no more than 45 volts.
I'm not familiar with the ins and outs of the circuit as Mr. Broskie has a talent of loosing me somewhere in the story, not necessary his bad. Is the feedback positive because of the mutual cathode resistor?
It's good to experiment else we'd all be listening to ipods on earplugs.
An externally hosted image should be here but it was not working when we last tested it.
Got it in, almost all..
After sanding eight layers of blanc paint last week, I could not resist building it today.
Tubes were selected from a large stack. I adjusted the AVO to my operating point (Va 125V, Vg 6,6V) and picked an equal pair of mesh plates for the gain stage. For the c.f. I matched two solid plates to identical current and transconductance as their g.s. brothers in crime.
The heaters are powered in series per pair (5V 1.75A). The measured voltage over them differs for 0.08 to 0.10V, hopefully this does not influence emission too much...
Initial test was performed with a quick setup using "candy laces" and without the volume pot in place. I could not tell the difference between source and pre (apart from the volume that needed to be turned down). No humm or audiable noise, so far so good for the "all in one build". I measured 3mV AC per channel with inputs shorted. PS ripple was a little higher than simulated at 6mV but not too bad
Next tidying up of the wires is necessary and some iron work has to be done for RCA connectors, selector switch and the volume control. Coming to speak of it, is the Light Dependent Resistor volume control still in esteem? Popularity of solutions seem to come and go like fashion
After sanding eight layers of blanc paint last week, I could not resist building it today.
An externally hosted image should be here but it was not working when we last tested it.
Tubes were selected from a large stack. I adjusted the AVO to my operating point (Va 125V, Vg 6,6V) and picked an equal pair of mesh plates for the gain stage. For the c.f. I matched two solid plates to identical current and transconductance as their g.s. brothers in crime.
The heaters are powered in series per pair (5V 1.75A). The measured voltage over them differs for 0.08 to 0.10V, hopefully this does not influence emission too much...
Initial test was performed with a quick setup using "candy laces" and without the volume pot in place. I could not tell the difference between source and pre (apart from the volume that needed to be turned down). No humm or audiable noise, so far so good for the "all in one build". I measured 3mV AC per channel with inputs shorted. PS ripple was a little higher than simulated at 6mV but not too bad
Next tidying up of the wires is necessary and some iron work has to be done for RCA connectors, selector switch and the volume control. Coming to speak of it, is the Light Dependent Resistor volume control still in esteem? Popularity of solutions seem to come and go like fashion
An externally hosted image should be here but it was not working when we last tested it.
Very cool, I see you have managed to fit the whole thing into that little box, you are a braver guy than me..
Have you had the opportunity to listen to it more extensively, and if so what do you think?
Is the 3mV a typo or the artifact of the measurement? If the hum at the output were really this high it should be very audible after being amplified by the power amplifier. The actual test will be with the volume control and I/O installed.
The anodes are subject to pick up from any ungrounded metallic object with AC on it within a certain proximity - this obviously should not be a problem with the cathode follower, but may be with the first stage.
I would not allow the filaments to be more than 12V positive with respect to the cathodes as filament to cathode insulation breakdown becomes more likely with the bigger delta. 45V seems very optimistic to me.
I like stepped attenuators and last time I checked there were some reasonably priced ones on eBay. The optical attenuator is another option of course. (One I keep meaning to try since I have a kit here for one.)
Have you had the opportunity to listen to it more extensively, and if so what do you think?
Is the 3mV a typo or the artifact of the measurement? If the hum at the output were really this high it should be very audible after being amplified by the power amplifier. The actual test will be with the volume control and I/O installed.
The anodes are subject to pick up from any ungrounded metallic object with AC on it within a certain proximity - this obviously should not be a problem with the cathode follower, but may be with the first stage.
I would not allow the filaments to be more than 12V positive with respect to the cathodes as filament to cathode insulation breakdown becomes more likely with the bigger delta. 45V seems very optimistic to me.
I like stepped attenuators and last time I checked there were some reasonably priced ones on eBay. The optical attenuator is another option of course. (One I keep meaning to try since I have a kit here for one.)
Hi Disco,
Sorry to hear that, I use regulated supplies with my transformer coupled 26 line stage for similar reasons.
I think this topology though is sensitive to minor variations in supply voltage and will actually make it much worse as I noted in a previous post due to positive feedback. First try removing the cathode feedback and see what that gets you..
Good luck, keep us posted!
Sorry to hear that, I use regulated supplies with my transformer coupled 26 line stage for similar reasons.
I think this topology though is sensitive to minor variations in supply voltage and will actually make it much worse as I noted in a previous post due to positive feedback. First try removing the cathode feedback and see what that gets you..
Good luck, keep us posted!
mk2
For a renewed, stabilised build I simplified the circuit by loosing the common ground return (and positive feedback). I left the katodes unbypassed, like this:
The driver tubes heaters are biased 30V (and the CF 130V) by a so called thingy (M. Jones). Most akward was to stabilise the low voltage because things heat up rapidly when currents reach 7 ampere. I chose to combine the heaters of the driver tubes in series (5.0Vdc at 1.75A) and do the same for the CF. At this power level semiconducters heat up, so do power resistors. As I dislike large cooling fins outside the cabinet I went for forced air cooling. This is the low voltage PS setup:
For an initial test I hooked up one input and left out the volume control. All voltages are in ball park, no humm no noise, aOK. High voltage is tightly regulated by a single Salas Simplistic HV Shunt Regulator (V1), doing a fine job here at low current and low voltage. I don't think a second regulator will add much to the reproduction quality.
My 4W direct coupled (27 to AD1) SE needed the extra amplification of this pre. For starters I used a random set of 27 tubes. I must admit the overall gain in performance is impressive, compared to driving the SE with only the CD player - even to tin ears. Why? Probably because I'm a lousy designer allowing the SE to draw grid current when driven hard. This pre can deliver
I'm backed by the shack of a soul boss most turnin' stormin' sound o'soul
I am double u o, o, o and I'm still up here again
For a renewed, stabilised build I simplified the circuit by loosing the common ground return (and positive feedback). I left the katodes unbypassed, like this:
An externally hosted image should be here but it was not working when we last tested it.
The driver tubes heaters are biased 30V (and the CF 130V) by a so called thingy (M. Jones). Most akward was to stabilise the low voltage because things heat up rapidly when currents reach 7 ampere. I chose to combine the heaters of the driver tubes in series (5.0Vdc at 1.75A) and do the same for the CF. At this power level semiconducters heat up, so do power resistors. As I dislike large cooling fins outside the cabinet I went for forced air cooling. This is the low voltage PS setup:
An externally hosted image should be here but it was not working when we last tested it.
For an initial test I hooked up one input and left out the volume control. All voltages are in ball park, no humm no noise, aOK. High voltage is tightly regulated by a single Salas Simplistic HV Shunt Regulator (V1), doing a fine job here at low current and low voltage. I don't think a second regulator will add much to the reproduction quality.
My 4W direct coupled (27 to AD1) SE needed the extra amplification of this pre. For starters I used a random set of 27 tubes. I must admit the overall gain in performance is impressive, compared to driving the SE with only the CD player - even to tin ears. Why? Probably because I'm a lousy designer allowing the SE to draw grid current when driven hard. This pre can deliver
An externally hosted image should be here but it was not working when we last tested it.
I'm backed by the shack of a soul boss most turnin' stormin' sound o'soul
I am double u o, o, o and I'm still up here again
Last edited:
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
A couple of years ago I discarded multi channel audio because mono seemed to offer me all musical information I was looking for. For flexibility sake I designed this pre amplifier stereo with a mono switch in mind. To my surprise I today discovered an unintended output on the opposite channel. Because there was no unintended connection the cause had to be in the circuit. Looking at the schematic in #11 you see there's no separation possible between left and right channel when sharing one power supply. So, when you consider building this pre you should have two regulators feeding it. Unless you're not in to stereo that is
I roll from Sylvania ST solid plate to Cunningham Globe mesh plate, and found the sound is more musical and mellow.
Do you have same experience ? I just thinking that globe is better than st, due the mesh plate.
And I wondering what is the sound of Cunningham st and others, also other globes.
Do you have same experience ? I just thinking that globe is better than st, due the mesh plate.
And I wondering what is the sound of Cunningham st and others, also other globes.
bigbargain Imageshack went commercial taking down all pictures. Here's the last schematic:
Since I can't edit the old posts I'll upload 'm once more. Enjoy!
An externally hosted image should be here but it was not working when we last tested it.
Since I can't edit the old posts I'll upload 'm once more. Enjoy!
Last edited:
I'm a pretty big fan of the #27 as well, I'm using a IXYS 10M45S load, and it seems to work really well.
As for the heater-cathode insulation, I've experienced this issue as well. A pair that I had hummed pretty badly with AC, though I've been able to find quite a number that also don't seem to have this issue.. So I would say AC is still an option (it was for me anyways). The pair I'm using right now are totally hum free.
I was fortunate enough to buy a lot of 50 mesh plate globe envelope #27 tubes for $1.50 a piece. So I'm able to pick and choose.
post 1:
version 1:
version 2:
finished:
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
version 1:
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
version 2:
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
finished:
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.