Choke primer

[mhuss]

Again, merely a practical decision. Most of the power supply current in an amplifier is consumed by the output stage -- if you put the choke after the output stage, you can use a much smaller choke.

Also, most guit-amps with chokes are push-pull designs -- this topology has a nice side effect of canceling most hum present on the power supply rails. Traditionally, SE amps were the "student" models, e.g., Champ, etc., and budget was a severe consideration.

An SE amp would benefit (RE hum reduction) by choke-filtering the entire amp, and with a low-power amp, the choke could be reasonable in size.

--mark

[Ears]

Again, merely a practical decision. Most of the power supply current in an amplifier is consumed by the output stage -- if you put the choke after the output stage, you can use a much smaller choke.

Also, most guit-amps with chokes are push-pull designs -- this topology has a nice side effect of canceling most hum present on the power supply rails. Traditionally, SE amps were the "student" models, e.g., Champ, etc., and budget was a severe consideration.

An SE amp would benefit (RE hum reduction) by choke-filtering the entire amp, and with a low-power amp, the choke could be reasonable in size.

--mark

Thanks again Mark, very much appreciated. You have confirmed much that I already suspected, but since entire amps are so seldom choke filtered I wondered if I had missed something not too obvious, such as response to large transient demand. I will now probably benefit by looking at some hi fi low power SE supplies. One can get a little paranoid about asking silly questions in here and since no one had replied I began to wonder if my query was labeling me as dumb.

[mlp-mx6]

I encourage you to ask such questions! Most questions elicit answers that we ALL benefit from.

[Ears]

I encourage you to ask such questions! Most questions elicit answers that we ALL benefit from.

<smile> (I can't get the little faces to work) One final thing about my previous post, I'm not talking about choke input filters, (I have reasonable understanding of their characteristics), but CLC filter for the entire system.

[mhuss]

A the iron for a choke input filter is *much* larger -- no one has used them since large value capacitors became available. They do provide the best regulation, though.

--mark

[dehughes]

A the iron for a choke input filter is *much* larger -- no one has used them since large value capacitors became available. They do provide the best regulation, though.

--mark

Choke input filter....this is something the early AC15s used yes? I'm pretty sure Brian at Top Hat is doing that with the Supreme 16 amp he's building...

[Ears]

A the iron for a choke input filter is *much* larger -- no one has used them since large value capacitors became available. :wink: They do provide the best regulation, though.

--mark

Choke input filter....this is something the early AC15s used yes? I'm pretty sure Brian at Top Hat is doing that with the Supreme 16 amp he's building...

And the late Fred Nachbaur in this monster amp :http://www.dogstar.dantimax.dk/tubestuf/dzart-1.htm

[Ears]

Whoops, here is the index page of dogzilla http://www.dogstar.dantimax.dk/tubestuf/dzindex.htm

[mhuss]

I suspect Fred was using the choke input partially to limit the monster turn on surge into that big bank of caps.
edit: A choke input is also handy when you need to bring down the B+ a bit; a fair amount of voltage is typically dropped across that input choke.

One thing I love about all of FN's designs -- he never followed a path someone else had already taken!

In Bad Cat's case, I suspect it's more of a tradition/mojo thing vs. a reasoned engineering decision. And if you're charging $$$$$ for an amp, the parts economy model can be a little different.

--mark

[PRR]

> a choke input filter .... They do provide the best regulation, though.

I believe this is semi-true with vacuum rectifiers and costly caps, not at all true with sand rectifiers and big cheap caps.

The cap-input pulls current in short spikes. The choke-input (when working right) pulls current all the time. Rectifier resistance matters more as duty cycle decreases.

The choke input has the BIG flaw that, below a certain current determined by cost, its output voltage soars to 1.5 times the nominal voltage. So it tends to want over-spec caps, or heavy bleeders, or a staggered-start scheme.

> Is there any reason apart from expense

Cost is ALWAYS an issue. OK, you can afford a choke, but is there a better way to spend that money?

> what exactly is the merit behind taking the plate supply from the 'front' of filter and taking the screen post filter?

High-current chokes cost more than low-current chokes.

Screen is 10 times more sensitive to ripple.

> Surely, the smoother the plate supply the better.

So get a 300V battery (25 car batteries), charge them in the morning, disconnect and play from them all night. The DC is pretty pure. The weight and cost is 900 pounds and $900.

Your well ran dry. You can pump from a lake but the water is dirty. You can filter it, but filters only do so many gallons and then you have to replace them. A finer filter does fewer gallons.

Watering your lawn: you can put dirty water on grassy dirt. Just screen the large trash and pump it.

Washing the dog, flushing toilets: you don't want to get the mutt dirtier, but he's just gonna get dirty again. Grit makes toilet valves stick and bowls look dingy. Use a coarse filter so the water looks clean.

Drinking: you want more filtering. But since the lawn and the dog have their own clean-enuff water, your real-clean water needs are small. Use a fairly expensive short-life filter to get 99% of the dirt and germs out.

Washing burns and bedsores: such water must be VERY clean to avoid infection. You need a hyper-expensive filter or a distillation system. But (I hope) you don't need a whole lot of water for such tasks, so the large expense per gallon is tolerable.

Sure, you could spend $5/gallon for super-pure water and put it on your lawn. But for what that costs, you could forget the lawn, go live in a hotel in Hawaii.

There is always such a thing as "too good".

[Ears]

Thanks PRR. I did enjoy the analogies. All the info I needed was in:

>High-current chokes cost more than low-current chokes.
>Screen is 10 times more sensitive to ripple.

To which I add/deduce: Taking supply prior to choke is sufficient (obviously, or they wouldn't be doing it) and that there are NO issues I have missed due to back emf or something.

<smile>

[Wild Bill]

Thanks PRR. I did enjoy the analogies. All the info I needed was in:

>High-current chokes cost more than low-current chokes.
>Screen is 10 times more sensitive to ripple.

To which I add/deduce: Taking supply prior to choke is sufficient (obviously, or they wouldn't be doing it) and that there are NO issues I have missed due to back emf or something.

<smile>

Choke input filters are a LOT different than the usual cap input configuration! With a choke, your output voltage is .9 x the rms voltage of the transformer secondary feeding it. It's rock solid and far better regulated than cap input, which means less hum. You can also draw an extra 50% more current from the transformer winding than with cap input, which will either let you use a trannie a bit on the small side (and cheaper!) and/or it will run cooler.

What's the catch? Well, in order to work you must be drawing a minimum amount of current at all times. The old rule of thumb formula is:
Lcritical = Voltage / Current in ma.

Say we have an amp and the power supply is running 400 vdc. If we are running say 50 ma as the total idle current then we need a choke of at least 400/50=8 henries.

What happens if the choke isn't big enough? The circuit snaps out of choke input mode and acts like a cap input!

This would mean that if we were expecting 400 vdc which is .9 x the rms applied then we would see 400/.9=444 times 1.41 to get the peak rms voltage = over 625 volts!

Consider that until the tubes warm up and draw idle current there's nothing being drawn from the choke input filter. Until the idle currents reach 50 ma you'll see cap input style peak voltage. You better have filter caps rated to handle it!

Tube rectifiers help a lot for choke input. Until the rectifier tube warms up there's no high voltage! By the time it appears the other tubes should be starting to draw at least something!

In the Golden Years of tube amps it was quite common for designers to put some bigass bleeder resistors across the power supply. This did 2 things. It would discharge the filter caps when you were working on the unpowered amp so you didn't risk getting shocked. It also gave some minimum current draw even if the idle current hadn't had time to begin happening.

Something like an 8000 ohm resistor would bleed 50 ma all by itself, at 400 volts. That would eliminate all fear of snapping out of choke mode and seeing sky high voltages. It would also dissipate 20 watts! You'd need some safety factor so you could use two 4000 ohm units, each at 15 or 20 watts to give a safe total.

Remember that you need a choke of at least 8 henries but it must be beefy enough to handle the peak current of the tubes when cranked to full output! Mind you, average power is much less than peak and it's average power we have to worry about. A choke is just a coil of wire around some iron so what burns it out is heat from the current. Peaks of current cause heat but between peaks the coil can cool off a bit. Still, for a pair of 6L6's I'd want a choke of at least 150 ma rating.

Such chokes will be big and expensive. They would take up probably as much real estate as the power transformer. That's why choke input is not popular in guitar amps. Leo Fender and the other guys were too cheap! Hifi amp designers love it because of the better regulation. Still, most of us at one time or other have had a power tranny in our junk box that had a HV winding just too damn high in voltage for use with the usual output tubes. A choke input filter would let us use that tranny after all!

Again in the Golden Years there was a kind of power choke in common use called a "swinging choke". These things were great! The inductance would vary over a range. At low currents it would be very high and it would reduce at high current draws. You might see a rating of 20 henries @ 10 ma and 5 henries @ 150 ma. This was perfect for choke input filter schemes! The low current rating made sure you were always going to be staying in choke input mode even as the tubes were warming up. The max rating meant that the choke was not nearly as large, which meant cheaper!

Sadly, no one seems to make swinging chokes anymore. I think it's because tube circuits are such low volume that it's easy to just charge us more for a big honking choke rather than tool up a line of swinging models.

Anyhow, this is probably more than you wanted to know!

Wild Bill

[novosibir]

Anyhow, this is probably more than you wanted to know!

No, it's too less

Can you tell me some more about the swinging choke? How does it work? How is it different in construction to usual chokes, so that it obtains the described behaviour?

Is it a similar behaviour like usual chokes, which drop down in inductivity, when the current reaches the designed current rating?

BTW: Welcome on this board, Bill! I've always very much appreciated your posts over at Ampage/Firebottle - which mostly were very helpful and 'deeper digging' into the stuff

Larry

[El_Martin]

Hi Larry!

I guess they react a bit like usual chokes since they ARE "poorly built" chokes

Chokes normally have airgaps with their E I laminates of the core having the same orientation. U knew that for sure
Swinging chokes are more built like transformers with their alternating orientation of the laminates.

Saturates easier, inductance is more current dependant.

Is it THAT easy? I doubt..

Ciao
Martin

[Wild Bill]

Anyhow, this is probably more than you wanted to know!

No, it's too less

Can you tell me some more about the swinging choke? How does it work? How is it different in construction to usual chokes, so that it obtains the described behaviour?

Is it a similar behaviour like usual chokes, which drop down in inductivity, when the current reaches the designed current rating?

BTW: Welcome on this board, Bill! I've always very much appreciated your posts over at Ampage/Firebottle - which mostly were very helpful and 'deeper digging' into the stuff

Larry

"Vee Gates?" my friend! It's been a while! I was thinking of you these past few days. I've got another Dclone on the go and I had printed out some advice you had given me some time ago.

I haven't been on Ampage much, simply because of time! I discovered a Canadian forum board http://www.guitarscanada.com that needed some help with techie stuff. Next thing I knew I was spending so much time on that board that I just didn't have much time for any others! We all have only so much disposable time in each day.

I don't know a lot about construction of swinging chokes. Other folks seem to be helping with that. I do know that the gap is important. Also, while it's true that all chokes drop a bit in inductance at rated current with a swinging choke it's much more extreme. Note I suggested a typical rating of 20 H @ 10ma that drops to 5 H at 150ma, or whatever. You get a huge inductance when you need it, during warmup. Once idle current starts to happen the choke quickly drops down to a smaller value that still should be enough to keep the circuit acting as a choke-input filter.

I found my last few swinging chokes at Fair Radio Sales in the USA. They're great for old military surplus stuff. Haven't seen anymore that are suitable for our needs in quite some time. I do wish some of the folks that offer retro-engineered OT's would consider offering swinging chokes. They are much cheaper and smaller than a choke rated at full current with extra inductance to work at startup currents.

Still drinking Spaten Pils once in a while!

Wild Bill