Just a quick question...
I was looking for RCA style RIAA circuits using 12AX7's, because I have some nice examples of those tubes. I'm thinking of making Eli's version, but I found John Broskie's take on it and thought it might be nice to try that.
However, there's something funny in there.
In the input stage, the cathode bias resistor is 470 ohms. JB shows the stage drawing about 1mA, and labeled the voltage at the cathode as 0.44V. OK, so maybe a little less than 1mA.
I've read in a lot of places (including good ol' MJ's Valve Amplifiers) that going that low in Vg-k with most tubes will result in the tube's grid drawing current. The input impedance goes way down, which would be a bad thing with a weenie little phono cartridge driving it.
In a simple 12AX7 input gain stage such as this, is a Vg-k below 0.5V to be avoided at all costs? (I think so, but I'm hoping for some advice from more experienced builders/designers.)
--
I was looking for RCA style RIAA circuits using 12AX7's, because I have some nice examples of those tubes. I'm thinking of making Eli's version, but I found John Broskie's take on it and thought it might be nice to try that.
However, there's something funny in there.
In the input stage, the cathode bias resistor is 470 ohms. JB shows the stage drawing about 1mA, and labeled the voltage at the cathode as 0.44V. OK, so maybe a little less than 1mA.
I've read in a lot of places (including good ol' MJ's Valve Amplifiers) that going that low in Vg-k with most tubes will result in the tube's grid drawing current. The input impedance goes way down, which would be a bad thing with a weenie little phono cartridge driving it.
In a simple 12AX7 input gain stage such as this, is a Vg-k below 0.5V to be avoided at all costs? (I think so, but I'm hoping for some advice from more experienced builders/designers.)
--
Last edited:
In general, yes. However, you do get some second harmonic cancellation from the opposing curves of the grid current and tube non linearity, so a case could be argued for Broskie's choice of biasing. Nevertheless, I makes me uncomfortable; I would recommend adjust on test.
Somewhere I have a copy of a paper published in the 1950s which suggests running the 12AX7/ECC83 in the region where grid current distortion opposes the normal anode distortion. It shows curves of second-harmonics being significantly reduced. Hooray!
What it also shows, but not commented upon as it was not realised back then how important it was, is the higher order distortions shooting up at the point where second reduces. This is exactly what you would expect to happen, as grid current produces a smoothish clipping on positive peaks. This is mainly second, but it has lots of higher orders too.
Unless you have a very low source impedance or very low signal amplitude (phono might qualify for the latter) I would keep the grid at least -1.0V wirh respect to the cathode. This means that the window of bias for the 12AX7 is quite narrow.
What it also shows, but not commented upon as it was not realised back then how important it was, is the higher order distortions shooting up at the point where second reduces. This is exactly what you would expect to happen, as grid current produces a smoothish clipping on positive peaks. This is mainly second, but it has lots of higher orders too.
Unless you have a very low source impedance or very low signal amplitude (phono might qualify for the latter) I would keep the grid at least -1.0V wirh respect to the cathode. This means that the window of bias for the 12AX7 is quite narrow.
I would agree with the previous two posts. Not to criticise Mr Broskie, but I am also uncomfortable. Firstly, how much telling 2nd harmonic distortion could be generated at the low signal voltage in operation here? But the high order products are another matter. And as said, a varying input impedance depending on signal amplitude?
(I would not use a 12AX7 as input tube to begin with, but that is another matter.)
(I would not use a 12AX7 as input tube to begin with, but that is another matter.)
Hi Johan,
The 12AX7 was designed as a low noise input tube from the start. You just have to use it properly, which many of us don't. Too accustomed to using the 6922 and other types for input. However the 12AX7 is available in other tube types as a single section too. These are more accepted for input amplification duty, and one wonders why.
The 6EU7 is a 12AX7 redesigned for lower noise and heater noise rejection. That is most definitely used as a low noise input tube. So basically, if the 12AX7 is used properly, it makes a solid choice for the input of a MM phono stage.
-Chris
The 12AX7 was designed as a low noise input tube from the start. You just have to use it properly, which many of us don't. Too accustomed to using the 6922 and other types for input. However the 12AX7 is available in other tube types as a single section too. These are more accepted for input amplification duty, and one wonders why.
The 6EU7 is a 12AX7 redesigned for lower noise and heater noise rejection. That is most definitely used as a low noise input tube. So basically, if the 12AX7 is used properly, it makes a solid choice for the input of a MM phono stage.
-Chris
Thanks for weighing in. It's good to have a little confirmation that I was generally on the right track.
Right, success of the first stage is more dependent on high gain and low noise than on low distortion. Distortion cancellation of 2nd harmonic at the expense of increased higher order harmonics sounds like the wrong direction for me, since I'm thinking of this as a way to use parts I have on hand to make a sort of vinyl 'warmifier.' I had made an RCA RIAA with 12AT7 cathode follower for my brother, years ago. When I heard it again recently, I was struck by how easy it was to listen through it. Not the most detailed/high-rez preamp, but very enjoyable. The adjective 'warm' came to mind. I want one. I'd also like to try this Aikido output stage.
So, it looks like if I use a higher voltage B+ supply, maybe 350V or even 400V, I could get the input 12AX7 biased up with Ip over 1mA and Vk at 1.2V or so, using an infrared LED in its cathode. That way I could get max gain out of that first stage and avoid the grid current region. That would be nice.
--
Right, success of the first stage is more dependent on high gain and low noise than on low distortion. Distortion cancellation of 2nd harmonic at the expense of increased higher order harmonics sounds like the wrong direction for me, since I'm thinking of this as a way to use parts I have on hand to make a sort of vinyl 'warmifier.' I had made an RCA RIAA with 12AT7 cathode follower for my brother, years ago. When I heard it again recently, I was struck by how easy it was to listen through it. Not the most detailed/high-rez preamp, but very enjoyable. The adjective 'warm' came to mind. I want one. I'd also like to try this Aikido output stage.
So, it looks like if I use a higher voltage B+ supply, maybe 350V or even 400V, I could get the input 12AX7 biased up with Ip over 1mA and Vk at 1.2V or so, using an infrared LED in its cathode. That way I could get max gain out of that first stage and avoid the grid current region. That would be nice.
--
You could take a simple step in the right direction by dropping the 40k in the supply rail to 15-20k (and possibly increasing the value of the 10uF decoupling cap). Raise the cathode resistor to 680. Basically, raise the anode voltage by 20V and the cathode by 0.2V. You should end up with about the same anode current, but much smaller grid current. You could even try swapping the 680 for a normal silicon diode.
Hi Chris (your post #5),
Thanks for info; did not know it was available as single tube.
Yes, not to create confusion. I am aware of the general low-noise specification. To explain, I work under the impression that noise is inversely proportional to Gm, so in that sense I myself am somewhat confused that particularly low noise can be achieved by the 12AX7, compared to triodes of Gm values > 10 m.mhos. Also, the relatively high Rp of the 12AX7 can make it difficult to drive some following circuits with highly variable impedance such as equalising networks and passive tone controls (although the 12AX7 output impedance can of course be incorporated in such designs).
But I am going OT; just to confirm that my stance previously mentioned was not intended as condemnation. I need to look up the 12AX7 noise specs again and compare. Thanks again!
Thanks for info; did not know it was available as single tube.
Yes, not to create confusion. I am aware of the general low-noise specification. To explain, I work under the impression that noise is inversely proportional to Gm, so in that sense I myself am somewhat confused that particularly low noise can be achieved by the 12AX7, compared to triodes of Gm values > 10 m.mhos. Also, the relatively high Rp of the 12AX7 can make it difficult to drive some following circuits with highly variable impedance such as equalising networks and passive tone controls (although the 12AX7 output impedance can of course be incorporated in such designs).
But I am going OT; just to confirm that my stance previously mentioned was not intended as condemnation. I need to look up the 12AX7 noise specs again and compare. Thanks again!
Last edited:
Looks great, really interesting toob there. But finding/paying for them might be a problem.
One version is 6AV6, in a 7 pin mini bottle.
http://www.mif.pg.gda.pl/homepages/frank/sheets/093/6/6AV6.pdf
All great suggestions. Generally speaking, a higher B+ voltage is going to help when using 12AX7 and 12AT7. They were designed to run from Ebb = 400V, as frequently seen in Fender guitar amplifiers, etc. It would be nice to be able to run the input 12AX7 with half the B+ at its plate, trying for that PSRR that JB claims happens at that operating point.
Fortunately, these tubes also run at low plate current, and I have a couple of leftover power transformers from old Dyna PAS2 preamps that would work fine. They're rated for 15mA. Four 12AX7 sections should draw no more than 5mA, and two 12AT7 Aikido sections should draw about 6mA, so 11mA total. I have some 12VCT transformers for heater duty, if needed. This is the kind of thing I won't want to spend a lot of money on.
Re: Noise -- You can definitely get lower noise using something like a 6DJ8 or triode-wired 6J9P in the first stage. But then you won't have quite as high gain, which impacts the signal to noise ratio. Besides, high transconductance always requires higher plate current, which I can't do because of the power transformer I'm planning to use. A good 12AX7 has enough gain and low enough noise to work OK as the first stage in a MM phono preamp. The one thing I've learned is that for most, the deal killer is the input capacitance, which when strays and cable capacitance are added can come to more than 300pF. However, that's covered too, because I have a Denon DL110 high-output moving coil cartridge to use with this. It's not effected as badly by input capacitance as most moving magnet carts are. I've found these cartridges sound best with a very small 1k as the input load resistor (input impedance). That will mitigate the Miller capacitance as well, right?
Also, it appears that some Shure MM's work well with a reduced load of ca. 35k and ca. 350pF input capacitance. So even those might work well with a 12AX7 input stage.
I just keep saying to myself, "Everything is a compromise..."
--
Last edited:
Aaargh!
Of course! In the RCA Tube Manual one is even referred there for characteristics of the 12AX7! Amplifier tubes in old AM radios just did not come to mind!
I find that for MM phono, for the input triode you can easily get away with running the 12ax7 at as little as 0.5mA
This opens up possibilities for use of shunted higher B+, allowing larger plate load and a bias of 1V or greater.
12ax7 can make a very nice phono stage but you must be ready to measure and tune each channel with an inverse RIAA circuit. I think the variability in the Rp's is the chief problem.
I never understood audiophools "rolling" 12ax7's in a phono stage...
Ian
This opens up possibilities for use of shunted higher B+, allowing larger plate load and a bias of 1V or greater.
12ax7 can make a very nice phono stage but you must be ready to measure and tune each channel with an inverse RIAA circuit. I think the variability in the Rp's is the chief problem.
I never understood audiophools "rolling" 12ax7's in a phono stage...
Ian
Hi Rongon
I just read your post #11. Yes, go for 400V. Feel free to make the plate load much higher.
Using only half a mA makes it easy, but you will likely find you need some way to regulate. I think I used a well sinked 24K Ohm, 25 Watt power resistor to shunt the supply. However, there are definitely more elegant solutions. Using a shunt was really just one of those temporary solutions that stuck for me...
Bias at 1V or maybe a little more to be on the safe side.
Ian
I just read your post #11. Yes, go for 400V. Feel free to make the plate load much higher.
Using only half a mA makes it easy, but you will likely find you need some way to regulate. I think I used a well sinked 24K Ohm, 25 Watt power resistor to shunt the supply. However, there are definitely more elegant solutions. Using a shunt was really just one of those temporary solutions that stuck for me...
Bias at 1V or maybe a little more to be on the safe side.
Ian
Last edited:
GE tube manuals have curves for the 12AX7 at low volts and less than 1.5mA. The commonly seen operating point of 100 Volts anode, -1.0 Volts grid and 0.5mA is in a quite linear region. Not easy to incorporate into a working circuit, but a useful if specialized device.
All good fortune,
Chris
All good fortune,
Chris
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.