Tapped inductors for PLLXO

Design and use of Chokes for PS, anode and filament use.
Post Reply
nic
Posts: 12
Joined: Thu Jan 19, 2006 8:59 am
Location: London UK

Tapped inductors for PLLXO

Post by nic »

Hi Dave ,
I thought this would be a good place to start :) I'm looking for some custom tapped chokes for use as 2nd order LC HP and LP filters, to be used at the input of the power amps after the line stage . Just wondering what the posibilities were , would like to keep overall input impedence of the amps at 50K and have taps for fine tuning 550-700Hz . It's not something I've ever seen , so pardon my naivety :) Got any ideas Dave ?

cheers

Nic
dave slagle
Posts: 2085
Joined: Sat Jan 22, 2005 3:54 am
Location: NYC
Contact:

Post by dave slagle »

hey nic,

i do tapped inductors all the time for this purpose. I'm not quite sure what you mean by the 50K number, does this mean that you need inductors in the 13hy range?

a sketch or LT spice doc would make the whole thing much clearer.

dave
nic
Posts: 12
Joined: Thu Jan 19, 2006 8:59 am
Location: London UK

Post by nic »

Hi Dave ,
First of all apologies for letting this one fizzle out . It's taken a major rethink and I've still got the latest project sitting half-finished on the bench . I'm unsure if plan A is feasible where the PLLXO is between the AVC and linestage valve for each stage . Plan B is the PLLXO at the input of the power amp , trouble is that the bass amp uses a Sowter 1:1+1 so I'd guess this would need a linestage in between . What do you think ? Also is it worthwhile keeping the input impedence of each dedicated amp/linestage as high as possible ?

cheers

Nic
dave slagle
Posts: 2085
Joined: Sat Jan 22, 2005 3:54 am
Location: NYC
Contact:

Post by dave slagle »

hey nic,

i'm not sure i follow. At some point you need to define an impedance to make a crossover work. A simple block diagram would put us in the right direction. Adding the important numbers from the filter perspective will help.

dave
nic
Posts: 12
Joined: Thu Jan 19, 2006 8:59 am
Location: London UK

Post by nic »

dave slagle wrote:hey nic,

i'm not sure i follow. At some point you need to define an impedance to make a crossover work. A simple block diagram would put us in the right direction. Adding the important numbers from the filter perspective will help.

dave
Hi Dave ,
What can you do for an amp with 50k input impedence ? Sowter list gapped EQ inductors on their site (M6 core , mumetal can , max 35H with 5 taps) . Thats the sort of thing I'm looking for

cheers
nic
Posts: 12
Joined: Thu Jan 19, 2006 8:59 am
Location: London UK

Post by nic »

Here's a schematic of the thing . Any good ?
Attachments
proposed linestage/crossover
proposed linestage/crossover
316axover.JPG (15.1 KiB) Viewed 16742 times
dave slagle
Posts: 2085
Joined: Sat Jan 22, 2005 3:54 am
Location: NYC
Contact:

Post by dave slagle »

now we are cooking with gas.

I'll capture the curves of the 316A and see where we end up in spice. I can make the inductors, but first we need to see how the "ideal behaves" so we at least know where to start. Then with an idea how the concept behaves, we can start talking about the direction to go for inductors.

on an aside, i interpolated the curves for the 316 in photoshop a while back and applying the math to my interpolations will be interesting. I can also measure a few of the samples i have on hand to see the actual behavior in the area you plan on operating to see how things line up.

I'll attach the curves so you can give a ballpark operating point in case i actually get in the mood to actually trace curves.

dave
Attachments
316A.gif
316A.gif (48.39 KiB) Viewed 16708 times
nic
Posts: 12
Joined: Thu Jan 19, 2006 8:59 am
Location: London UK

Post by nic »

Thanks Dave ! The schematic is not set in stone but the 316A looks ideal and is a stunning looking valve , other valves such as 801/10Y and 101D are also available . As an aside , how well do you think the filament arrangement will hold up ?

cheers

Nic
dave slagle
Posts: 2085
Joined: Sat Jan 22, 2005 3:54 am
Location: NYC
Contact:

Post by dave slagle »

nic,

i'm still trying to get my head around your circuit. where will the output be taken from?

What B+ will you be using and do you have ballpark numbers for the inductors?

dave
nic
Posts: 12
Joined: Thu Jan 19, 2006 8:59 am
Location: London UK

Post by nic »

Hi Dave ,
Outputs are taken from the 316A anodes , B+ around 300V . I did not include any inductor values , I'm still not 100% sure what values would be best suited driven from the AVC

cheers
dave slagle
Posts: 2085
Joined: Sat Jan 22, 2005 3:54 am
Location: NYC
Contact:

Post by dave slagle »

got a model done of the 316A in spice and 1/2 way into entering it it hit me like a ton of bricks what you were doing. I guess i'm a bit obtuse at times :-)

My thinking is the load on the autoformers is a needed evil so make it high to make lessen its effects. the added benefit of this is your caps will get smaller in size.

I finished the model and got the circuit "Running" and everything looks well with a 50K, 3nf and 22hy values. see attached image for the response for each setting of a butt gapped autoformer. Sure the filament bias is a bit hack in the model but it seems to work.

Earlier you asked about the filament situation and i still haven't gotten my head around the situation and the following should prove that. My initial thought was the left tube will operate as expected, but the tube on the right will see the other tubes filament (and any signal present) as part of its cathode load. This made me want to reverse the filter order so any extra cathode Z would show up in the high pass. Then i wondered if the signal across the filament would also show up as feedback making me feel the current arrangement is best since you will not inject the LF info back into the HF signal. Then of course the CCS might make all of this discussion moot. did i mention i don't have my head around this yet?

my crude fake of filament bias in spice didn't show any issues but that is a big fwiw.

let me know if oyu can use LT spice and i'll post the .asc file. I will add it is really easy and steve bench has posted a wealth of info on how to get it running.

dave
Attachments
the autoformer is not shown but jut to the left of the 3n cap
the autoformer is not shown but jut to the left of the 3n cap
Picture 3.png (16.26 KiB) Viewed 16704 times
Picture 2.png
Picture 2.png (28.93 KiB) Viewed 16704 times
nic
Posts: 12
Joined: Thu Jan 19, 2006 8:59 am
Location: London UK

Post by nic »

Hi Dave ,
Looks good , I've never got my head round Spice though . The filament bias arrangement is designed to make use of available chokes , I don't think I could stomach shoehorning in another four filament chokes . The cathode resistor is only an ohm , plus the filaments would have a sub 1 ohm impedence so i doubt it will be a problem . It would probably be best for me to build a prototype and investigate the filament supply further , if all else fails there's always other valves such as 10Y which can be subbed in with a parallel rather than series filament bias arrangement . Just out of interest , how close would you be able to make the inductor taps ? I ran the numbers for a Linkwitz Riley aligment with 20hz increments and found the following :

680Hz 23.4H 0.00234uF
660Hz 24.1H 0.00241uF
640Hz 24.8H 0.00248uF
620Hz 25.7H 0.00256uF

Is it possible to have more finer taps at 10Hz increments ?

cheers
dave slagle
Posts: 2085
Joined: Sat Jan 22, 2005 3:54 am
Location: NYC
Contact:

Post by dave slagle »

chances are these would be done on the autoformer sized core and the PC mount bobbins have 14 pins so you do the math :-) The biggest problem will actually be setting the gaps identically so that 10hz steps match in frequency from channel to channel. setting a .001 inch gap to 1% precision is not an easy task.

how wold you handle the caps?

dave
nic
Posts: 12
Joined: Thu Jan 19, 2006 8:59 am
Location: London UK

Post by nic »

dave slagle wrote:chances are these would be done on the autoformer sized core and the PC mount bobbins have 14 pins so you do the math :-) The biggest problem will actually be setting the gaps identically so that 10hz steps match in frequency from channel to channel. setting a .001 inch gap to 1% precision is not an easy task.
I did not think of that , would coaser steps be easier to match ? Then I guess the resistor and cap could be adjusted :)

how wold you handle the caps?
One higher value cap with another lower value in parallel . The parallel cap would be fitted into one of those ZIF IC sockets . Used this scheme before with a simple 1st order and it allows 'hot' swapping of caps in-circuit . It's a pity the cap value is so high otherwise I'd use air ganged variable capacitors

cheers
dave slagle
Posts: 2085
Joined: Sat Jan 22, 2005 3:54 am
Location: NYC
Contact:

Post by dave slagle »

I was speaking more to the 10hz increments matching. Coarser steps would be more difficult to match. With 10hz steps you have a better chance of precision since wost case, you can never be more than 5hz off.

So why not use a single cap (russian teflon?) for your lowest value and then a variable air trimmer for the adjust?

I almost have to wonder out loud about upping the impedance values to 100K or more to make the cap value and air trimmer more appealing. you know the range of easily available caps and the inductors can be anything you want so why not look at the area around 100K and see what fits.

dave
Post Reply