Plate or anode chokes

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

johnheath

Well-known member
Joined
Jul 31, 2014
Messages
890
Location
Sweden
Hi... While looking into all other types of tube amplification schematics and reading comments and what not I sometimes hear about plate chokes... reducing hum and noise, frequency response and all.

Can someone tell me the advantage of using  plate chokes instead of plate resistors... if any?

And also I an curious about the very high values of those mentidoned chokes... somewhere between 150 - 250H?

An open and curious question from someone trying to learn the tricks of the trade :)

Sincerely

/John
 
here is a blurb that is not to technical-

Parafeed

The "parafeed" (parallel feed) uses a large choke to provide the DC voltage to the anode. This choke must be large in value, and provides a relatively high impedance at audio frequencies. The output transformer is capacitively coupled to the anode. Sometimes the capacitor is inserted between the anode and the transformer (as shown in the schematic above), and sometimes the transformer is connected directly to the anode, and the capacitor inserted in the ground lead.

The advantage of this kind of mechanism is the choke and the transformer can be individually optimized: the choke for saturation capability, and capacitance; whereas there is no DC in the transformer, so it can be made relatively smaller, providing both lower capacitance and leakage inductance with respect to its primary inductance. This is a long winded way of saying that it is possible to achieve wider frequency response.

There are four other advantages to this circuit. First, the capacitor inserted allows a low frequency "extension" due to the resonance between the transformers inductance and the capacitance value. The second is tha additional low frequency poles can be more easily controlled (as the capacitor and the inductance of the choke can be used as additional "degrees of freedom" in the design). The third is reduced "hum". In a normal transformer circuit, the hum voltage forms a voltage divider between the plate resistance and the load. For low plate resistances, the power supply ripple is coupled to the output. In the parafeed circuit, the power supply ripple forms a voltage divider across the choke, and is not coupled to the output. Thus, the power supply characteristics are more isolated from the output. The fourth advantage of this circuit is since the transformer has no high voltage DC on it, the transformer can be replaced with an "autoformer" (a single tapped winding) allowing the output autoformer to be further optimized.
 
I thought the primary reason (forgive the pun) for using an anode inductor is that you get twice the plate voltage swing and twice the output power of a circuit using a resistive anode load.

Cheers

Ian
 
generally speaking, the more iron you use, the better the amp sounds, add in some chicken head knobs and you have full mojo,  :D
more stuff>

http://tubecad.com/2014/09/blog0308.htm
 
Thanks guys

But know to the second half of my question…The high values of Henries?

Regards

/John
 
> The high values of Henries?

Do you not know how to calculate the impedance of a choke?

2*Pi*Freq*H = choke impedance.

You need to pick an impedance. 12AX7 is often loaded with 100K resistor, so we will aim there.

You need to pick a frequency. Pick 20Hz.

There's a way to work the formula backward. But I'm lazy and just throw numbers at it.

Pick 100 Henries.

2*Pi*20Hz*100H = 12,566 Ohms impedance.

Not big enough by a long shot.

12AX7 internal plate resistance is 60K.

50Hz may be a useful low frequency.

2*Pi*50Hz*100H = 31,426 Ohms impedance, 100H at 50Hz

Say you can find 250H.

2*Pi*50Hz*250H = 78,540 Ohms impedance, 250H at 50Hz

Some further thumb-math leads to 60K impedance, 250H, at 38Hz.

Bass response will fall-off below that point.

So in a voltage-amplifier we sure could need hundreds of Henries.

That will be a large hunk of wire. Large hunks of metal have large capacitance to the universe. Perhaps 300pFd. By a similar (though inverse) process, we get 8,846Hz for the treble roll-off of 60K tube with 300pFd stray C hanging off its plate.

12AU7 has lower internal resistance, about 10X lower, so the low-point moves lower and the high-point moves higher. Indeed the 12AU7 comes from the class of "general purpose" triodes, from days when "general" meant transformer (choke-like) loading. 12AX7 is "pinched" to give higher gain but at impedance too high to be practical for wide-range choke-loading (resistance loading 'only').

No "trick". Just basic math.
 
Thanks PRR… very informative as usual. I guess the trick is to know all this stuff ;)

So when using a tube with high internal resistance it would need a very big choke if it were to be used as plate choke and I guess it is due to space, weight and money. I actually saw this in the V76 preamp schematic shown to me during the discussion about the EF86 tube.

So… I automatically have to ask this: It seems like a plate choke is rarely used even for tubes with a lower internal resistance… so what would the benefit be of using a choke instead of resistors? I mean there must surely have been a reason to use it in certain circuits?

Regards

/John
 
take the V76 for instance, it has 2 chokes, about 250H and 400H,

first off, when you have current flowing in a wire, you have a magnetic field that extends into outer space, a steel core pull these field lines in from outer space  and concentrates them inside the core, more concentrated  lines of flux means more inductance which is what we want to load a tube with, we want to develop our signal across this choke so we can transmit it to the next tube without losing it to the pwr supply caps, so the choke is like an AC resistor, but has a lower DC resistance than a plate resistor, which means we get more of the pwr supply voltage to the plate of the tube, we can swing this voltage up and down with the tube and get more output,

400H is used in stage 2 of the 4 stage amp, it gives you more gain and less pwr supply hum at the critical first stages of mic amplification,

the second choke, 250H, allows you to use nickel alloy in the output transformer,
why? because without the choke, you have to run the B+ plate supply thru the OPT(output xfmr), a nickel core OPT does not like DC, it saturates the core, nickel saturates at 5-8 k gauss(measurement of flux density) , steel at 18 k gauss, DC=more gauss, so you have to lace the lams of the core in a way that lowers he inductance, all E lams going one way, all I lams stacked together to form an air gap between the E and I lams, this lowers the permeabillity of the core which means it will  not concentrate the magnetic lines of flux  as easily, you are increasing the resistance of the core, just like adding more resistance to a circuit, only instead of decreasing current like in a circuit, you are decreasing the ability of the core to pull in lines of magnetic flux, the core is more reluctant to be magnetized by the DC current of the plate supply, so to maintain inductance to load the tube, you have to increase the core size and use steel lams that do not saturate as easy as nickel,

now with the output tube plate choke of 250 Henries, (lower Henries because there is more current flowing in the output tube, which means bigger wire and thus, less room for turns and thus less inductance), also, the output tube has lower plate resistance which means we can use a lower value of inductance and still match the load to the tube, so now all the DC for the tube goes thru the gapped choke and not the OPT, so now we can stack the lams like this, one E lam going one way, the next E lam going the opposite way, this gives the core more perm, we can use less steel as we do not need to worry about saturation from the DC, so we can also use better steel, like a nickel allow, which will provide more inductance per turn than M6 steel, this means a smaller core, with less turns, this also means smaller wire in the OPT as it now only carries the AC signal, so less turns and smaller wire and smaller turn diameter due to the small core means less capacitance which means better high  frequency response, and we have way more Henries with the ungapped nickel core, so low end improves also, so now we can build a V76 OPT that runs flat from 5 Hz to 120K Hz!  so the purpose of the two chokes in the V76 is more gain, less noise, and better frequency response, now the output stage puts out more AC voltage than the inner stage 2, so we need to use an ally that will take the extra voltage, so we mix a little steel in with the nickel, that second stage core can be an 80% mu metal alloy, but we need to water that perm down a bit for the output stage,

but now we have a new problem, we have to use a lot of turns in the choke to get more inductance with the gapped choke core, about 21,000 turns for the 400H and about 17,000 turns for the 250H choke, that is a lot of turns, and a lot of capacitance, so we break the plate chokes into 3 sections which lowers stray C by a factor of 3 quadratic, or 3^2 which = 9 times less stray C, otherwise the stray C shorts out the hi end into the pwr supply cap, thus defeating the purpose of the choke in the first place,



 
Thank you Gus but I have searched and I can't find a decent answer or as good as those I have recieved here now.

Most of the threads here is about a very specific circuit or detail and just wanted the overall answer which I believe I have now.

But if you have valuable info to share please feel free to do so.

/John
 
Thanks again CJ - I guess transformers are a big intrest of you? =)

Anyway - I surely have a way better understanding of the prpose and use of plate chokes now.

Regards

/John
 
thanks for those links Gus! that is what i was lookin for but the site is gone i guess,

note the difference between the resistance of a resistor and the resistance (AC) of the choke,

a resistor will keep a pretty constant resistance over the frequency range in which it is used, but the choke will start out with say, 2 pi f L = 6.28 f L, pick 100 henries, 628 f is our resistance, so 6,280 ohms at 10  hz, 62K at 100 hz, 620K at 1000 Hz, then the core drops out due to hysterisis, so maybe 5 meg at 10KC and probably constant up to 100KC, if capacitance does not short out the signal,

so the choke goes from 6K to 5 Meg over the audio range, non linear, but as long as minimum resistance (or reactance for the purist) is maintained, response will be flat even with the no linearty of thew choke,
 
Interesting info… Am I wrong if I say that I think they don't use it in same  way as … "in those days" (old stuff)?

I have just ordered some chokes for my own investigation… can't wait to see if I can hear a difference.

I found the Hammond 156C, a 150H/8mA… small size so it will be possible to mount inside a 1U rack preamp.

Regards

John
 
> plate choke is rarely used even for tubes with a lower internal resistance

Used in EVERY Champ or Epi Jr gitar-amp. (The output stage has a 2-winding choke, more often called a transformer.)

The POWER efficiency of choke (transformer) loaded stage, for clean Sine, approaches 50%. For resistor loaded, only 6%.

So a stage which needs POWER output may opt for the choke.

OTOH, another tube section is often cheaper than the appropriate choke. There's still a bass cut from the coupling cap but this may be set quite low at low cost.

Choke-loaded may give higher voltage gain. But put your choke money in another tube, you get LOTS more gain. 1/2 12AU7 with choke, gain about 19. Resistor loaded gain is near 15. However both sections cascaded is gain of 15*15 or 225, FAR more than one section and a choke.

Other factors: in Broadcast, reliability is vital. Tubes fail. Simple chokes don't. (However a very fine-wire high inductance choke is not as reliable, though you can throw more money into sealing it.) A designer might feel that the choke design has better reliability, and that his customers will pay for it.

The output of a resistor-loaded stage is always less than the B+, often more like 40%. The output of a choke-loaded stage will exceed the B+, and 150% is not hard. If B+ is forced to be low, this matters. (Some of the classic German broadcast gear used B+ voltages which seem shy to jack-it-up American eyes. This may also have forced them to think about power economy.)

Frankly the choke has NO place in post-1937 audio, except power output stages, new-age tinkerers, and a few special corners like some German broadcast boxen.

Yes, get the little Hammond and play around. Aside from the obvious bass limit, you *may* find that a 50/60Hz power-smoothing choke has enough stray capacity to shave the top of the audio band. With a low-Rp tube, this may not be serious. But a low-Rp tube is almost certainly a low-Mu tube so gain will be low. Thinking outside that bottle, pencil a 2-stage amp (instead of tube+choke) and you will usually find better performance. Gain may become so high that you are "forced" to apply NFB, and many performance numbers get better yet. Yes, IM gets worse with small NFB, and high NFB suggests low gain which in 2 tube stages may mean the NFB loop loads the output heavily.

If there were One Perfect Answer we would be done already.
 
Thanks PRR - useful information as always.

Yes I know about the chokes in guitar amps. I just noticed that that for example the plate choke in the V76 had a huge value compared to what you find in a …Fender git amp… usually around 4H???

Anyway - I have been working my way around different circuit topologies based on tubes because I love the vintage tube sound if I may say so? Like everybody else I want good stuff when recording and for my playing style and voice and wanted sound I guess that I would have to pay big bucks for a decent transistor preamp sounding as nice as the pres that I (despite my lack of calculation skills) have managed to build.

I guess I just want to learn as much as possible but also I know that many great inventions are the results of …"oops, I didn't know that" derived from testing and trying… and that is something I can do =)

Gain is not an issue …rather the opposite as you assumed. It is basically my biggest concern … to keep the gain down without losing the bass and tube flavor. And if you say that a choke would increase gain a lot It might be useless. And another gain stage is also over the top :)

I will fiddle with the choke and see what I can come up with or if I have to ditch it ?

Thanks

/John




 
ps. A NOS 12AU7 is sold for at least 25 euros (roughly 28USD) and you can buy a Hammond choke for 17 euros (roughly 19USD) or less depending on size and values - I do not know if tubes are cheaper in the U.S than in Europe but Tube Amp Doctor in Germany have a very wide range of tubes in all different prices and qualities but a good tube is quite a lot of money still and it will require som extra parts like tube socket and at least a couple of resistors and a cap.

And a choke is possible to mount basically anywhere in the box or even on the outside... a tube socket has its limitations in where you can place it.

And a third tube would also increase current draw in heater arrangments and therefore also create a lot more heat which i believe is one of the major drawbacks with tube circuits of all kind... especially if you have a bundle of tube gizmos in a rack for instance.

Just a bit of information on how I think about it... absolutely no criticism at all :)

Regards

/John

/John
 
> chokes in guitar amps.

The B+ power filter choke is something else.

Output Transformers are Chokes (with added windings).

> around 4H???

Somewhere in there. Hammond 125ESE is a *really big* OT (choke) for a Champ and runs near 10 Henries. I can well agree that the usual suspects are few-Henry.

Now look at the tube conditions. Champ's output tube runs say 300V and 40mA. It runs "fat". Look at a voltage-gain stage. Since the load draws "no" (very small) current, we'd like to run the tube at low current. 1mA? 3mA? 10mA?? Obviously the load impedances must be 40X, 13X, 4X what we use for the Power stage.

The Champ and its OTs are for guitar. Doesn't go below 82Hz. Also they are small/cheap. They got a 8-inch speaker. Fairly hot midrange to maximize the sound from the small power stage. Such a speaker will punk-out for any real power below 100Hz even 120Hz. We *want* to shave that last octave so the tone isn't drowned in bass-mud. Also a lesser OT is cheaper.

But the general question of plate chokes often implies "full frequency". And because good chokes are never cheap, expectations run well below 50Hz, probably 20Hz. Compared to a 120Hz limit, that's another 5X more Henries.
_________________________

> to keep the gain down

Well, that's very different.

A resistor divider in a medium-level interstage comes to mind. Costs 24 cents and can be flat zero to 20KHz.
 
Thanks PRR

Yeah… sorry about the champ. I was thinking about the push-pull fenders… like the super or similar… they have a choke but in the filter section… often 4H. I was confused :).

I tried the plate-choke alternative and it was very susceptible to noise from transformer and heater supply among other things and it was not until I added some cable to the leads and placed it a bit away from the preamp that it was quiet. There was no possible way to mount it in the near vicinity of the preamp so I ditched the idea really.

I read about the trick to mount two plate-chokes back-to-back like a Gibbo humbucker but… no.

And as you say a divider may break the ice when it comes down to reducing gain if needed.

Regards

/John

 
CJ said:
you could try a solid state choke, i think they call it a gyrator or something weird like that,

Why? why? why would you do something like that?

I mean, if going for an active circuit a current source is what you would need, why sacrifice LF response even if it's free to go lower? The concept would be the same but the approach and the topology wouldn't . In fact it has a name, SRPP, where you make a current source with one triode to load the other... It could be done with a MOS or a BJT, I wouldn't go with a JFET but could be done in low voltage/low current applications, I don't know how the tube would behave in a situation where a JFET could survive. Sim time!

JS
 

Latest posts

Back
Top