John Swenson's preamp! No more attenuators.
I'm so excited I just had to share it on diyAudio.com.
Originally posted over at AA.
http://www.audioasylum.com/forums/tubediy/messages/57358.html
Here are the details on the BDT preamp I built last weekend. This design used Beam Deflection Tubes, which were originally designed as color subcarrier demodulators in color TVs. To my knowledge this is the first use as a straight forward linear amplifier.
I had been thinking about these tubes as amplifiers since I first saw them in an RCA tube manual 6 years ago, but never got around to getting it built (I bought most of the parts though). At VSAC last year Allen Wright's talk on differential amps got me thinking about these tubes again and I just had to see if they worked or not. Since I'm a bit low on cash these days the entire project was made from components in the parts bin so some of the values and details may not be optimal, it was what was on hand.
As you can see from the schematic the tubes are a little strange, it starts out as a normal pentode, but the plate is split in two and two extra electrods are added called deflectors, these work similar to the CRT tube in an oscilliscope. When the same voltage is applied to both deflectors the electron beam coming from the pentode is evenly split betwen the two plates. When there is a difference betwen them the beam is deflected so one of the plates gets more current and the other gets less.
The result is essentially a long tailed differential pair in one tube, with the added benefit that the current can be adjusted by a DC voltage on the grid. This mean that when used as a stereo preamp you don't need any attenuator (volume pot) either pot or switched resistors or TVC or any such device. The input goes directly to the tube and one pot controls the gain for both channels. Another interesting advantage is that there is NO miller capacitance on the deflectors!
My original design was going to use a good line output transformer running PP from the two plates but that was way too much money for this first project. I wanted to see if these tubes had any hope as audiophile tubes before I spent any money on the project. Latter on I'll try some transformers and see how THAT sounds. As is if bought from scratch the whole preamp would cost about $50!
A few notes on the current circuit: the 10 ohm cathode resistors are not really neccessary, they are just there so I could easily read the total current through the tube. The 2K plate resistora are a compromise between gain and output impedance, as it is now the gain is about 1, which works well for a lot of systems. Higher resistance gives more gain but also higher output impedance. The voltage to the volume pot is a battery because that was the cheapest way to get a really good 9-12V for the grid. This is probably the most critical voltage in the design, the B+ has quite good PSRR but this does not. Cathode bias would not work because you need to vary the cathode current over a wide range. I will probably wind up with a voltage doubler off the 6.3V running a 12V regulator of some type for the REAL version of this. As is I have to change the battery every 4 or 5 days of continuous playing.
The PS is a voltage doubler (thanks PJ for turning me on to them!) driving a CRC filter since I didn't have any chokes. But it works very well as is. PSUD says its less than 2mv PP ripple which coupled with the very good PSRR means a VERY quiet circuit.
The sound: WOW, AWESOME, AMAZING!!! I've never heard a preamp this good before and I wasn't even trying to make something really good yet. The number one impression is smooth, but yet it is wonderfully detailed, hughly dynamic, very "punchy" when the music is so. PRaT is very high, nothing seems to be emphasized, its all just there. Its hard to describe because it just lets the music through.
Next up a 3W power amp using BDTs, but that will take some transformers so I have to wait till I have a little more money in hand.
John S.
I'm so excited I just had to share it on diyAudio.com.
Originally posted over at AA.
http://www.audioasylum.com/forums/tubediy/messages/57358.html
An externally hosted image should be here but it was not working when we last tested it.
Here are the details on the BDT preamp I built last weekend. This design used Beam Deflection Tubes, which were originally designed as color subcarrier demodulators in color TVs. To my knowledge this is the first use as a straight forward linear amplifier.
I had been thinking about these tubes as amplifiers since I first saw them in an RCA tube manual 6 years ago, but never got around to getting it built (I bought most of the parts though). At VSAC last year Allen Wright's talk on differential amps got me thinking about these tubes again and I just had to see if they worked or not. Since I'm a bit low on cash these days the entire project was made from components in the parts bin so some of the values and details may not be optimal, it was what was on hand.
As you can see from the schematic the tubes are a little strange, it starts out as a normal pentode, but the plate is split in two and two extra electrods are added called deflectors, these work similar to the CRT tube in an oscilliscope. When the same voltage is applied to both deflectors the electron beam coming from the pentode is evenly split betwen the two plates. When there is a difference betwen them the beam is deflected so one of the plates gets more current and the other gets less.
The result is essentially a long tailed differential pair in one tube, with the added benefit that the current can be adjusted by a DC voltage on the grid. This mean that when used as a stereo preamp you don't need any attenuator (volume pot) either pot or switched resistors or TVC or any such device. The input goes directly to the tube and one pot controls the gain for both channels. Another interesting advantage is that there is NO miller capacitance on the deflectors!
My original design was going to use a good line output transformer running PP from the two plates but that was way too much money for this first project. I wanted to see if these tubes had any hope as audiophile tubes before I spent any money on the project. Latter on I'll try some transformers and see how THAT sounds. As is if bought from scratch the whole preamp would cost about $50!
A few notes on the current circuit: the 10 ohm cathode resistors are not really neccessary, they are just there so I could easily read the total current through the tube. The 2K plate resistora are a compromise between gain and output impedance, as it is now the gain is about 1, which works well for a lot of systems. Higher resistance gives more gain but also higher output impedance. The voltage to the volume pot is a battery because that was the cheapest way to get a really good 9-12V for the grid. This is probably the most critical voltage in the design, the B+ has quite good PSRR but this does not. Cathode bias would not work because you need to vary the cathode current over a wide range. I will probably wind up with a voltage doubler off the 6.3V running a 12V regulator of some type for the REAL version of this. As is I have to change the battery every 4 or 5 days of continuous playing.
The PS is a voltage doubler (thanks PJ for turning me on to them!) driving a CRC filter since I didn't have any chokes. But it works very well as is. PSUD says its less than 2mv PP ripple which coupled with the very good PSRR means a VERY quiet circuit.
The sound: WOW, AWESOME, AMAZING!!! I've never heard a preamp this good before and I wasn't even trying to make something really good yet. The number one impression is smooth, but yet it is wonderfully detailed, hughly dynamic, very "punchy" when the music is so. PRaT is very high, nothing seems to be emphasized, its all just there. Its hard to describe because it just lets the music through.
Next up a 3W power amp using BDTs, but that will take some transformers so I have to wait till I have a little more money in hand.
John S.
Hi G,
I'm a dumbass. You know..essentially a cut and paste amp builder. So I can't pretend I understand how it works exactly either..... But I generally know a good thing when I see it. ;-)
And the absence of a volume pot in the direct signal path...and just one in the circuit appeals tremendously ...for a 6 channel Home Theatre preamp I have been dreaming of.
The low or non-existent miller capacitance is also a big advantage I think.
I think you may be right...that is another thing I never thought of.
But on the subject of impedance...the output impedance is quite high I think...I read an old post of John asking about a tranny of 40k primary impedance.
But I am going ahead and build it...to see for myself...
Cheers,
Bas
I'm a dumbass. You know..essentially a cut and paste amp builder. So I can't pretend I understand how it works exactly either..... But I generally know a good thing when I see it. ;-)
And the absence of a volume pot in the direct signal path...and just one in the circuit appeals tremendously ...for a 6 channel Home Theatre preamp I have been dreaming of.
The low or non-existent miller capacitance is also a big advantage I think.
I think you may be right...that is another thing I never thought of.
But on the subject of impedance...the output impedance is quite high I think...I read an old post of John asking about a tranny of 40k primary impedance.
But I am going ahead and build it...to see for myself...
Cheers,
Bas
I think the impedance will change. But it's a great idea.
We in the UK, and possibly Europe too, didn't have the "privelige" of valve chroma demodulators, as colour TV came later, and transistors were in vogue by then. Therefore most of us will never have come across these devices.
It's going to take a few days thought to get them straight in my mind
Thanks to Bas for "porting" this one, and especially to John Swenson
We in the UK, and possibly Europe too, didn't have the "privelige" of valve chroma demodulators, as colour TV came later, and transistors were in vogue by then. Therefore most of us will never have come across these devices.
It's going to take a few days thought to get them straight in my mind
Thanks to Bas for "porting" this one, and especially to John Swenson
Hi,
With a "regular" preamp the tube doesn't care what the setting of the volume control is for as long as it's not too extreme.
Output impedance of the stage won't change with volume setting either.
Not really, Miller capacitance only becomes troublesome with high mu tubes or with highly capactive Cag.
For a preamp it's nothing to worry about though.
It's bound to be since we're talking beam tetrodes and plate followers here.
Please do, Bas. Let us know what you think of it....
Just don't let'em beam ya up, hey.
Cheers,
With a "regular" preamp the tube doesn't care what the setting of the volume control is for as long as it's not too extreme.
Output impedance of the stage won't change with volume setting either.
Not really, Miller capacitance only becomes troublesome with high mu tubes or with highly capactive Cag.
For a preamp it's nothing to worry about though.
It's bound to be since we're talking beam tetrodes and plate followers here.
But I am going ahead and build it...to see for myself...
Please do, Bas. Let us know what you think of it....
Just don't let'em beam ya up, hey.
Cheers,
A balanced O/P is easy. Just add a coupling cap to the anode currently without 1.
Allowing for balanced I/P is easy too. The deflector that currently is directly grounded gets a 140 KOhm resistor between it and ground. Apply the balanced I/P signal to the deflector plates.
I'm in agreement with the earlier remarks about using buffering voltage followers, especially if the value of the load resistors is increased. Increasing the value of the load resistors raises both the gain and the O/P impedance.
Allowing for balanced I/P is easy too. The deflector that currently is directly grounded gets a 140 KOhm resistor between it and ground. Apply the balanced I/P signal to the deflector plates.
I'm in agreement with the earlier remarks about using buffering voltage followers, especially if the value of the load resistors is increased. Increasing the value of the load resistors raises both the gain and the O/P impedance.
Beam deflection tubes for audio
There were some articles published on using sheet beam or beam deflection tubes in audio back in the late 50s and early 60s, but they never caught on. Here is a variable gain preamp/amplifier (figure 6) from an article by J. N. Van Scoyoc in Transactions I.R.E. PGBTR circa 1960 pages 3 thru 7. (Sorry about picture quality, can't fit a better version.) Some other uses were for input signal selection, phase splitter, remote volume control, stereo balance control, tone control, audio compressor/expander. The beam defl. tubes will only produce up to a gain of about 6 due to the deflection feedback from the plate voltages. To get higher gain, use a triode or pentode cascode stage above it, (QST had an article on this). The tubes also work better (more linear) if de-magnetized.
Don
There were some articles published on using sheet beam or beam deflection tubes in audio back in the late 50s and early 60s, but they never caught on. Here is a variable gain preamp/amplifier (figure 6) from an article by J. N. Van Scoyoc in Transactions I.R.E. PGBTR circa 1960 pages 3 thru 7. (Sorry about picture quality, can't fit a better version.) Some other uses were for input signal selection, phase splitter, remote volume control, stereo balance control, tone control, audio compressor/expander. The beam defl. tubes will only produce up to a gain of about 6 due to the deflection feedback from the plate voltages. To get higher gain, use a triode or pentode cascode stage above it, (QST had an article on this). The tubes also work better (more linear) if de-magnetized.
Don
Attachments
Hello Bas,
are you going to give it a try? Any source for the tubes?
I have a trafo that might be good for the purpose, but if someone else is going to build it I might go and buy some of these tubes and give it a try.
Appealing is the easy way to achieve balanced operation, but what worries me is the output impedance... how tough the load can approximately be? 20k? 50k? >100k?
Cheers
Andrea
are you going to give it a try? Any source for the tubes?
I have a trafo that might be good for the purpose, but if someone else is going to build it I might go and buy some of these tubes and give it a try.
Appealing is the easy way to achieve balanced operation, but what worries me is the output impedance... how tough the load can approximately be? 20k? 50k? >100k?
Cheers
Andrea
Hallo Andrea,
Yes, am going to give it a try!
I got them from thetubecenter.com but they might be sold out by now. There has been quite a run on these tubes. But there should be plenty...warehouses will be scoured...and plenty will show up I think.
Anyone know of a European version for the 6AR8???
Cheers,
Bas
Yes, am going to give it a try!
I got them from thetubecenter.com but they might be sold out by now. There has been quite a run on these tubes. But there should be plenty...warehouses will be scoured...and plenty will show up I think.
Yes, I think your load should be >40k... But don't know for sure. Since John was looking for a 40k pp...I guess the load it would want to see is around 20k absolute minimum. Hopefully someone with real theoretical knowledge will post a definitive answer.
Anyone know of a European version for the 6AR8???
Cheers,
Bas
Bas, as I said earlier, we didn't have an application for these in Europe, so I suspect not. Of course I'd like to be proven wrong
WRT control range being limited - sand to the rescue
Since the control is a DC voltage, a comparator could be fed with the same voltage, and drive a relay to mute the sound at the bottom end of the control. I'm still convinced (gut feeling) that the output impedance will change with volume setting - higher at lower volumes, dispite the proclamation by Cardinal DeGrove
. We'll have to devise a test to prove this.Another thought is to try and run the thing in enhanced triode mode, applying signal to the screen grid. If it works, that'll drop the Ra.
Pardon me, but who cares? Just buffer the thing with a voltage follower. The buffer provides a stable, low, O/P impedance with good drive capability.
What's particularly appealing to me is the ease with which the circuit can be made inverting or non-inverting (flip/flop the deflector plates).
Hi,
Did I say it didn't change? Don't think so.
If it didn't it would work anyway...
Yet within the context of the high Zout of the device it's not relevant as whatever change occurs will be swamped by the Zout.
While looking for the "best" volumecontrol in the days passive preamps were being touted as the next best thing since sliced bread, I put this beam tetrode deflector idea forward for one of the technicians to look into.
It got discarded rather quickly because of, among other things, the relative cost (two tubes: one the beam tetrode, the other the buffer) compared to the solution we finally adopted for the preamp: a switched attenuator buffered preamp.
Within the context of a DIY project design criteria are rather different. The idea looks great in apps as shown in the pic by smoking-amp where you can couple the output directly to the next high Zin stage.
For a standalone preamp, you'd still need a buffer as Eli correctly suggests.
BTW, IIRC the idea was also reprinted in one of the Audio Anthology issues....
Little goldmines for a plethora of "not so succesful" but useful ideas, no matter what the Pope, the Cardinal or the Dean say about them....
Cheers,
Did I say it didn't change? Don't think so.
If it didn't it would work anyway...
Yet within the context of the high Zout of the device it's not relevant as whatever change occurs will be swamped by the Zout.
While looking for the "best" volumecontrol in the days passive preamps were being touted as the next best thing since sliced bread, I put this beam tetrode deflector idea forward for one of the technicians to look into.
It got discarded rather quickly because of, among other things, the relative cost (two tubes: one the beam tetrode, the other the buffer) compared to the solution we finally adopted for the preamp: a switched attenuator buffered preamp.
Within the context of a DIY project design criteria are rather different. The idea looks great in apps as shown in the pic by smoking-amp where you can couple the output directly to the next high Zin stage.
For a standalone preamp, you'd still need a buffer as Eli correctly suggests.
BTW, IIRC the idea was also reprinted in one of the Audio Anthology issues....
Little goldmines for a plethora of "not so succesful" but useful ideas, no matter what the Pope, the Cardinal or the Dean say about them....
Cheers,
fdegrove said:
Hello,
to my eyes (and to my wallet
) a 6AR8 costs much less than a decent stepped attenuator (few dollars vs 50+ for the attenuator, if you find cheaper ones just tell me..or better sell me!). Not as minimalistic (with a buffer) as attenuator+buffer but still a good solution IMHO.Can anybody suggest a suitable buffer (maybe with ECC88...)?
Cheers
Andrea
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.