Inductive (attenuator) based tone controls

stephie's place

Moderator: sbench

Post Reply
sbench
Posts: 296
Joined: Fri Jan 28, 2005 3:45 pm

Inductive (attenuator) based tone controls

Post by sbench »

The "other" frequency selective use for attenuators is inductive based tone controls. Attached is a LTSpice simulation of a set of tone controls based on this concept. The Treble consists of one tapped inductor (boost side) of 1mH+1mH+4mH+4mH+4mH+4mH (=100mH total) and 30 ohms. (lower resistance is better). The treble cut consists of 4 "discrete" inductors of 2.2mH, 2.2mH, 3.3mH, 3.3mH of the $2.00 "Digi-Key" variety.
The Bass consists of 1 boost tapped inductor of 16H+1H+1H+1H+4H+4H (=121Hy) and modeling 110 ohms resistance (not too critical). The Bass cut is another tapped inductor of 2H+2H+2H+2H (=32H). (100 ohms, but not critical)... I have added 200 ohms in series with this inductor "common" tap in the pix below. Full schematic in a bit, for right now, here's the frequency response plot with bass "flat" moving the treble, and then the treble "flat" moving the bass.

Steve
Attachments
tone.gif
tone.gif (20.16 KiB) Viewed 20218 times
sbench
Posts: 296
Joined: Fri Jan 28, 2005 3:45 pm

Post by sbench »

As promised, here's the schematic for the inductive passive tone controls.

-Steve
Attachments
ind_tone.gif
ind_tone.gif (9.33 KiB) Viewed 20214 times
dave slagle
Posts: 2085
Joined: Sat Jan 22, 2005 3:54 am
Location: NYC
Contact:

Post by dave slagle »

a few thoughts/questions...

why is the bass switching break before make, and the treble make before break?

I would think that a single inductor for each situation with all of the taps could be wired with a cut/boost switch to save on a coil and give a wider range of switching. a 2 deck 6 position per channel with a toggle should do it.

can you attach the .asc you used?

dave
sbench
Posts: 296
Joined: Fri Jan 28, 2005 3:45 pm

Post by sbench »

why is the bass switching break before make, and the treble make before break?
Ahhh. Note that the output is in series with the treble control... using a make before break keeps signal essentially present with only a slight burp during the switching cycte. The opposite is true of the bass control where notice that the inductor is directly on the "supply". During the momentary short, the larger inductance produces a bigger burp since you're essentially sending the output to +20dB or ground when the switch momentarily shorts, so break before make is more benign (signal travels thru the treble circuits, but only the emphasized or deemphasized bass passes thru the bass circuit. Ya know how some controls go clickity-click click when switching? didn't want that to happen.
I would think that a single inductor for each situation with all of the taps could be wired with a cut/boost switch to save on a coil and give a wider range of switching. a 2 deck 6 position per channel with a toggle should do it.
I had actually toyed with that idea. What disuaded me was that THAT was contrary to the way normal tone controls have come to be expected, and thus would be "surprising". I would agree that would simplify building the bass control... the treble is slightly a compromise, since the inductance values are so much different (that part must be related to the impedance ratio (10:1) and the cut has even less effect on a common core (thus, the separate inductors). That *could* be worked around, I'm sure, if I gave it more thought, but the "cheap" inductors hold up well in that application, so I didn't spend a bunch of little gray cells on it :roll:
can you attach the .asc you used?
Had the "other PC" off already when I read this. Hopefully it's rebooted now while I was answering the other Qs. Ahhh. Here tis now.

Steve
Attachments
tone3.asc
(9.28 KiB) Downloaded 753 times
dave slagle
Posts: 2085
Joined: Sat Jan 22, 2005 3:54 am
Location: NYC
Contact:

Post by dave slagle »

I had actually toyed with that idea. What disuaded me was that THAT was contrary to the way normal tone controls have come to be expected, and thus would be "surprising". I would agree that would simplify building the bass control...
OK but two discrete windings can reside on the same core, and it just so happens a 14 pin bobbin and a 12 position switch will give you normal tone controls.

in other words they do not need ot be on independant cores, just two independant windings, and since only one is used at a time the coupling between the two will not be an issue. (this saves us a core)
the treble is slightly a compromise, since the inductance values are so much different (that part must be related to the impedance ratio (10:1) and the cut has even less effect on a common core (thus, the separate inductors). That *could* be worked around, I'm sure, if I gave it more thought, but the "cheap" inductors hold up well in that application, so I didn't spend a bunch of little gray cells on it :roll:
thats interesting. do you think its the coupling from cut winding to cut winding that is the issue or couping from cut winding to boost winding?

could all of the treble cut windings couple to each other, or must they be discrete?

its easy enough to put them on the bobbin with the boost windings and give them a try. if it doesn't work just don;t use them.

dave
sbench
Posts: 296
Joined: Fri Jan 28, 2005 3:45 pm

Post by sbench »

OK but two discrete windings can reside on the same core, and it just so happens a 14 pin bobbin and a 12 position switch will give you normal tone controls.

in other words they do not need ot be on independant cores, just two independant windings, and since only one is used at a time the coupling between the two will not be an issue. (this saves us a core)
Brilliant! You're right, of course. Must be too close to the problem, but as obvious as that is in hindsight, it honestly never occurred to me. That's surely the way to go for the bass "inductor", two sets of tapped windings.

On the treble windings...
do you think its the coupling from cut winding to cut winding that is the issue or couping from cut winding to boost winding?

could all of the treble cut windings couple to each other, or must they be discrete?

its easy enough to put them on the bobbin with the boost windings and give them a try. if it doesn't work just don;t use them.
Coupling from boost windings to cut windings appears to be the issue. In that topology, the inductance from max boost winding to "flat" and the inductance from "flat" to max cut has to be the same ratio as the "attenuation" of the circuit. What that means is on separate cores, a 10:1 inductance ratio exists. On the same core, assuming perfect coupling, that's 100:1 Z ratio, so the tone control effect is horribly skewed. With separate cores, that's OK, and we *could* come up with a separate core for the cut. However, the values are such that off-the-shelf parts are effective for those inductors..

Wow, I think what we have so far is 2 cores and a couple discrete inductors for each channels total tone controls. Not bad!!!!!

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

Post by dave slagle »

Hey steve...
The Bass consists of 1 boost tapped inductor of 16H+1H+1H+1H+4H+4H (=121Hy) and modeling 110 ohms resistance (not too critical). The Bass cut is another tapped inductor of 2H+2H+2H+2H (=32H). (100 ohms, but not critical)...
I started playing with some numbers and the problems started showing up :-(

The first minor problem is how critical is the DCR of the bass boost winding? I find myself creeping upwards of 300 ohms or possibly more to keep the gap in the inductor large enough to assure reasonable linearity.

The second problem is the signal voltage across the bass controls. At the first few steps, things will be fine, but as you get to the extremes of the cut and the boost you will be using only a fraction of the winding which translates to high flux densities in the core. (which again means nonlinearity)

Luckily, what we need to do to keep the core linear and the flux down tend to follow the same pattern. The easy (and less expensive) solution is to take the DCR upwards and all will be fine since the increased turns will lower the flux and require a larger gap to get the inductance required.

The other solution is a bigger core. I know we had originally talked of nickel for this position, but after realizing the possibilities of high flux densities we may be better off looking at amorphous. It simply comes down to a tradeoff of core size vs. core type vs. dcr. Many will say use a big gap on an EI-100 M-6 core and be done with it but that's not my style :-)

Remember the old saying that goes something like 'quick cheap and good, pick any two and take the opposite of the third' ? We have that situation with size (cost), linearity & DCR.

Where do we put the compromise?

dave
sbench
Posts: 296
Joined: Fri Jan 28, 2005 3:45 pm

Post by sbench »

Dang... had this answered then realized wasn't logged in so starting over.........

Interesting questions...
The first minor problem is how critical is the DCR of the bass boost winding? I find myself creeping upwards of 300 ohms or possibly more to keep the gap in the inductor large enough to assure reasonable linearity.
Not a problem. I re-modeled it in LTSpice (attached TONE4.ASC - but note that the sim times are long due to the several step thrus I used)

I tried using 200 ohms for the 16H portion, 100 ohms for each 4H portion and 50 ohms for each 1H portion for a total of 500 ohms, and the resulting plot was prerfectly fine. (likewise, the bass cut is impervious to DCR).

Where the DCR *IS* important is the treble inductor, but that ought to be OK in any event.

The second problem is the signal voltage across the bass controls. At the first few steps, things will be fine, but as you get to the extremes of the cut and the boost you will be using only a fraction of the winding which translates to high flux densities in the core. (which again means nonlinearity)
Max practical signal level is 10 volts or lower. I presume that more turns as noted above will help that too.

Luckily, what we need to do to keep the core linear and the flux down tend to follow the same pattern. The easy (and less expensive) solution is to take the DCR upwards and all will be fine since the increased turns will lower the flux and require a larger gap to get the inductance required.
Agreed. Higher DCR does not seem to be a problem for the bass inductor, unlike the treble one that definitely wants to see low DCR.

Thus, the compromise is for DCR. I'd presume we can stay with the original size.

Steve
Attachments
tone4.asc
(6.37 KiB) Downloaded 776 times
dave slagle
Posts: 2085
Joined: Sat Jan 22, 2005 3:54 am
Location: NYC
Contact:

Post by dave slagle »

sbench wrote: I tried using 200 ohms for the 16H portion, 100 ohms for each 4H portion and 50 ohms for each 1H portion for a total of 500 ohms, and the resulting plot was prerfectly fine. (likewise, the bass cut is impervious to DCR).
cool...

i just realized that all of the boost positions have the 16hy leg in series with them so flux will not be an issue. The weakest link is the biggest cut position where L31 (2hy) is the only series inductor on the core. This suggests that that 2hy choke should be able to see 10V @ 20hz. with the current numbers in my head 5V@ 20hz puts it right at Bmax but we should still be OK since this is the max cut position the flux will be lower (right??)

I hopefully have it worked out so a small gap will be needed to get by the problems associated with setting a consistant butt gap.
Where the DCR *IS* important is the treble inductor, but that ought to be OK in any event.
yup...

Max practical signal level is 10 volts or lower. I presume that more turns as noted above will help that too.
yes... am I correct in assuming above that worst case is the max bass cut where L31 is the only series element on the core? Will that position still see the max 10V? What happens if that value of inductance goes down due to slight or severe saturation?
Thus, the compromise is for DCR. I'd presume we can stay with the original size.
I see no reason why not to give it a try, no sense chasing a problem that may not exist in reality :-)

dave
sbench
Posts: 296
Joined: Fri Jan 28, 2005 3:45 pm

Post by sbench »

yes... am I correct in assuming above that worst case is the max bass cut where L31 is the only series element on the core? Will that position still see the max 10V? What happens if that value of inductance goes down due to slight or severe saturation?
I think that's a good assumption. Interestingly, since this is LR, on the boost side, at most 10 volts is expected, as noted, but for the cut side, a MUCH lower max is expected. For maximum cut, at most 1 volt at 20Hz is likely to occur. In reality, much lower. (The idea here is for max cut, we are trying to get rid of excessive or unwanted bass, so the purpose is to produce a low level at low frequencies).

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

Post by dave slagle »

Hey,

I have a first version of each of the inductors done.

the treble has a DCR of 8.5 ohms and needs a nice sized gap to get the inductance you need. Setting the gap to get 100mhy end to end will get you all of the values you need. I wound them with some trifilar #33 wire i had and they measured the same at 100hz 1K and 10K which is a good sign the core is doing its job and resonance is a fair bit above 10K.

For the bass, I unintentionally ended up with numbers very close to what you said. The math suggested the exact numbers and practice brouhgt it in at a hair below 500 ohms. These chokes will be a bit harder to set the gap on since it will require a very small additional gap beyond the butt gap. With a butt gap i measured just over 200hy at 100hz and using the smallest spacer (.0005-silver) dropped me down to 80hy since the goal was 121hy it will just take some trial and error with to get the gap set properly. IT is interesting to note that when measured at 1Khz the inductance doubles and at 10K there is a negative inductance. This suggests to me that at 1K you are approaching resonance and by 10K you are past it. I will guess that the behavior at 20hz will be close to or the same as 100hz so again we should be in good shape.

I'll set the gaps as close as my patience will allow using 10K for the treble and 100hz for the bass) and ship them. Of course the gaps will be easily adjustable, but it will be fun to compare measurmeants.

dave

dave
sbench
Posts: 296
Joined: Fri Jan 28, 2005 3:45 pm

Post by sbench »

I have a first version of each of the inductors done.
W o n d e r f u l
the treble has a DCR of 8.5 ohms and needs a nice sized gap to get the inductance you need. Setting the gap to get 100mhy end to end will get you all of the values you need. I wound them with some trifilar #33 wire i had and they measured the same at 100hz 1K and 10K which is a good sign the core is doing its job and resonance is a fair bit above 10K.
That ought to be just right.
These chokes will be a bit harder to set the gap on since it will require a very small additional gap beyond the butt gap. With a butt gap i measured just over 200hy at 100hz and using the smallest spacer (.0005-silver) dropped me down to 80hy since the goal was 121hy it will just take some trial and error with to get the gap set properly. IT is interesting to note that when measured at 1Khz the inductance doubles and at 10K there is a negative inductance. This suggests to me that at 1K you are approaching resonance and by 10K you are past it. I will guess that the behavior at 20hz will be close to or the same as 100hz so again we should be in good shape.
The DCR ought to be OK. I'll see what additional options we end up with. The "R" associated with the L can be altered somewhat if we end up with something "critical", but I'll let the listening tests either confirm or deny that.

Somehow, the relatively low resonance seems disquieting, but it's probably no big deal. Again, the proof will be in the listening.

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

Post by dave slagle »

the tone control chokes and a pair of the#7 bobbins are going out today.

attached below are the measurements as i gapped them (or more precisely what the meter said)

the bass units had a precariously small additional gap and are a bit tempremental... giving them a whack changes things a bit. For a future version, if another few hundred ohms DCR (above the 500 and 300 now) isn't an issue, we could go for a more stable gap. (or a bigger core, but that is another can of manufacturing worms)

the treble units have a huge gap and are really close to each other. (as expected from the large gap) in a future version i'd like to get the flux up a bit by cutting the turns and this would also drop the DCR (8.4R) and the capacitance.

one interesting thing to note about the treble units, is when measured at 1K and 100hz the L creeps up a bit, but the L is stable with applied voltage (50mv 250mv and 1v) at each frequency. At first i thought it was an approach to a low Q resonance, but suspect my test rig is not entirely honest at 10K. If it is resonance, it can easily be fixed.

dave
Attachments
measurements from my B&K 885
measurements from my B&K 885
benchtone.gif (7.94 KiB) Viewed 20083 times
sbench
Posts: 296
Joined: Fri Jan 28, 2005 3:45 pm

Post by sbench »

the tone control chokes and a pair of the#7 bobbins are going out today.
Great!!!!!
the bass units had a precariously small additional gap and are a bit tempremental... giving them a whack changes things a bit. For a future version, if another few hundred ohms DCR (above the 500 and 300 now) isn't an issue, we could go for a more stable gap. (or a bigger core, but that is another can of manufacturing worms)
Actually the additional resistance should not be a problem at all. It will be interseting to see (and hear) how they perform. :D
the treble units have a huge gap and are really close to each other. (as expected from the large gap) in a future version i'd like to get the flux up a bit by cutting the turns and this would also drop the DCR (8.4R) and the capacitance.
I'll measure the overall response, of course. But, I'd guess that the lower turns will actually work better based on lower resistance and capacitance.

I'm kinda curious how the overall thing is going to sound.

Been doing some comparisons with different iron in phono stages and it's quite interesting how much difference the iron is making in the overall sound.

Steve
Post Reply