Simple Attenuators - Design And Testing

[dougie tamson]

Hi dougie, thanks for posting that, its awesome work!

Are you able to measure your inductance or is it based on a calculation?

i like the standoffs for the screws. Ive been thinking we should generally space the inductors off the case (eddy currents, more on that to follow soon) and this does that.

Hi John, my meter doesn't do inductance. These are calculated values but audio specific, site I used diyaudioandvideo dot com Calculator/AirCoreInductorDesigner/

Size wize they are similar to the dimentions of the ones for sale.

For some future research I was thinking of including an option to have a sliding or threaded iron bolt to adjust the inductance up/down by screwing it in and out to taste.

Now I can start building

Doug

 

[dougie tamson]

Here are the calculated values for winding the coils for an M2v when using my 3d printed bobbin:

For the 16/8/4Ω version:

L1
Inductance 0.91 mH
DC Resistance 0.36 Ohms
Wire Gauge 18 AWG
Wire Diameter 40.3 mils (1 mil = .001 in)
Coil Length 0.8 in
Coil Inner Diameter 0.6 in
Coil Outer Diameter 1.49 in
Average Turn Diameter 0.99 in
Wire Length 55.99 feet
Copper Weight 0.28 pounds
Turns 215
Levels 10.83
Turns/Level 19.85
L2
Inductance 0.61 mH
DC Resistance 0.29 Ohms
Wire Gauge 18 AWG
Wire Diameter 40.3 mils (1 mil = .001 in)
Coil Length 0.8 in
Coil Inner Diameter 0.6 in
Coil Outer Diameter 1.41 in
Average Turn Diameter 0.93 in
Wire Length 45.24 feet
Copper Weight 0.22 pounds
Turns 185
Levels 9.32
Turns/Level 19.85


For the 8/4/2Ω version:
L1
Inductance 0.45 mH
DC Resistance 0.25 Ohms
Wire Gauge 18 AWG
Wire Diameter 40.3 mils (1 mil = .001 in)
Coil Length 0.8 in
Coil Inner Diameter 0.6 in
Coil Outer Diameter 1.33 in
Average Turn Diameter 0.9 in
Wire Length 38.93 feet
Copper Weight 0.19 pounds
Turns 166
Levels 8.36
Turns/Level 19.85
L2
Inductance 0.3 mH
DC Resistance 0.2 Ohms
Wire Gauge 18 AWG
Wire Diameter 40.3 mils (1 mil = .001 in)
Coil Length 0.8 in
Coil Inner Diameter 0.6 in
Coil Outer Diameter 1.24 in
Average Turn Diameter 0.85 in
Wire Length 31.47 feet
Copper Weight 0.15 pounds
Turns 142
Levels 7.15
Turns/Level 19.85


Doug

 

[dougie tamson]

Here's a zip of the 3d model I created to make the bobbin, result is better when printed in 2 parts then glued together (with acetone = nail varnish remover).

 

[dougie tamson]

Deep dive...

http://info.ee.surrey.ac.uk/Workshop/advice/coils/air_coils.html

 

[JohnH]

Thanks for those references.

When I last got a meter, I got one with inductance, it was about $50, for good amateur quality. It works well enough to show that inductors that I bought at a certain value and % range are within spec, which gives me confidence in both the meter and in the coil.

It would be very interesting to compare calculated vs measured. Its also interesting to see how things like steel bolts increase the value. The next thing I want to test with it is how close proximity to an aluminium panel can reduce the inductance (and add resistance) by eddy effects.

 

[dougie tamson]

My main take-away from that reference was the cross sectional shape of the windings to give the most effiicient inductance/weight of copper, if you look at the commercially available coils,the coil cross section is mostly square.

 

[JohnH]

yes, I just learnt that too from your link!

 

[dougie tamson]

yes, I just learnt that too from your link!

Would be helpfull if you could post the dimensions (in mm) of the copper windings for your coils?
The wire gauge, internal diameter (ie 1st coil winding), outside diameter (last coil) and the height/thickness excluding the plastic.

I may revise my bobbin, the one I made might be a bit closer to the one on the right of the diagram in the reference!

Doug

 

[Osman Qureshi]

Great that all sounds fine. That line-out recipe is intended to be very versatile in terms of giving whatever signal level you need, from a wide range of amp powers. There may be times when you'll have to set it very low to suit your next device without overloading it, but also if you put it after the attenuator, with speaker attached, you should get a good signal even at max attenuation, if you turn the knob up high. Once you have tested it with how you want to use it, you might want a different resistor for the front end connection, but this is good to start.

On the component powers, it depends. But lets take a worst case, say you are driving an amp at 100W, and if this was a 16 ohm attenuator. (is it 8 or 16?). The signal voltage at the amp output = attenuator input would be (100 x 16)^0.5 = 40V. The power taken into your line-out circuit of 2.2k+10k would be 40^2/12200 = 0.13W. Of this, 10/12.2 x 0.13 = 0.11W goes into the pot. So there's still a good factor of more than 2 against its 0.25W spec.

Good luck with the build!

Thanks John. It’ll be an 8 ohm build so the 0.25 watt pot should be more than sufficient. I won’t get to building it until the weekend. Will revert with results.

 

[JohnH]

Would be helpfull if you could post the dimensions (in mm) of the copper windings for your coils?
The wire gauge, internal diameter (ie 1st coil winding), outside diameter (last coil) and the height/thickness excluding the plastic.

I may revise my bobbin, the one I made might be a bit closer to the one on the right of the diagram in the reference!

Doug

Sure, I'll get a couple of examples to compare with in the next day or two.

 

[JohnH]

Coil tests

hi @dougie tamson

I don't have a 0.9mH coil around to be a direct match. But I have a 0.4mH one, which would serve as a way of comparison of theory vs actual inductance.

in free air, it measures 0.40mH, which is what it was bought at.

wire gauge 18
inner diameter = 15.5mm
outer diameter = 31.5mm
height of coil = 20mm
Measured resistance 0.2 to 0.3 Ohm (after allowing for meter leads)

There's roughly, 9-10 winding layers with about 20-22 turns per layer

I checked on that calculator:
Air Core Inductor Designer / Calculator (diyaudioandvideo.com)

Putting in the bobbin inner diameter, height and wire gauge, I get this:



The overall dimensions, once converted mm to inches, check out very closely, but i think im seeing more turns though. DCR is consistent but i can only measure to the nearest 0.1 Ohm at best, so not accurate.

Then i checked how it might change due to things through it or near it,

A 25mm M3 bolt right through it increased L to 0.6mH, This is due to high magnetic permeability of steel.

Placing it close to an aluminium box face reduced inductance reading from 0.4 to 0.39mH. With the bolt, 0.6mH was reduced to 0.57mH. This is an eddy current effect. Its like a transformer with a single turn secondary, or like an induction cook top. A current is induced in the aluminium plate that magnetically opposes that in the coil. I think this shifts the phase, making it less inductive and more resistive, hence absorbing energy. This effect may vary with frequency, which is not captured by my meter test.

 

[dougie tamson]

Coil tests

hi @dougie tamson

I don't have a 0.9mH coil around to be a direct match. But I have a 0.4mH one, which would serve as a way of comparison of theory vs actual inductance.

in free air, it measures 0.40mH, which is what it was bought at.

wire gauge 18
inner diameter = 15.5mm
outer diameter = 31.5mm
height of coil = 20mm
Measured resistance 0.2 to 0.3 Ohm (after allowing for meter leads)

There's roughly, 9-10 winding layers with about 20-22 turns per layer

I checked on that calculator:
Air Core Inductor Designer / Calculator (diyaudioandvideo.com)

Putting in the bobbin inner diameter, height and wire gauge, I get this:

View attachment 85408

The overall dimensions, once converted mm to inches, check out very closely, but i think im seeing more turns though. DCR is consistent but i can only measure to the nearest 0.1 Ohm at best, so not accurate.

Then i checked how it might change due to things through it or near it,

A 25mm M3 bolt right through it increased L to 0.6mH, This is due to high magnetic permeability of steel.

Placing it close to an aluminium box face reduced inductance reading from 0.4 to 0.39mH. With the bolt, 0.6mH was reduced to 0.57mH. This is an eddy current effect. Its like a transformer with a single turn secondary, or like an induction cook top. A current is induced in the aluminium plate that magnetically opposes that in the coil. I think this shifts the phase, making it less inductive and more resistive, hence absorbing energy. This effect may vary with frequency, which is not captured by my meter test.

Excellent info, thanks John, just what I needed.

I'll measure the 1mm wire I bought with a micrometer, I'm assuming the 1mm spec is for the copper only and doesn't include the insulation layer. I might need to add small amount of space to my bobbin to allow for this.
Even if it's only 0.05mm, after 20 turns that's an extra 1mm of space required to fit them all in.

Reminds me of the first guitar I built, when building the neck (a tele copy) I used a nice sharp pencil to mark the lines to saw the fret slots, after I glued the fret board to the neck and strung it up, the intonation was off and beyond any adjustments at he saddle. My mistake was measuring each successive fret cut line from the previous, even with a sharp pencil to mark the line, there was enough error in the pencil line width compounded 22 times for a 22 fret neck that the scale was way off between the open and 12th fret. Solution was to use a fine craft knife to mark the lines and mark with reference to the nut... I removed the fret board and made a new one.

Doug

 

[acromarmot]

Hi @JohnH @Gene Ballzz @TheOtherEric and others!

As mentioned earlier I'm trying to include the attenuator in my current amp-build. As tube positioning, housing etc. has already been done long before I came across this thread I have a hard time to place the coil.

Do you have any input on that?

Thx in advance!

 

[acromarmot]

Obviously I cannot add photos here. Dunno why but here're some...

http://cloud.ibak.de/index.php/s/TW496XmMlzlEwbO

Cheers
acromarmot

 

[ThePanda]

The last items I need to build John's attenuator are the inductors, Instead of buying I made my own
First I 3d printed a bobbin, I used '3d Builder' feee app from Microsoft, the whole thing is made up of a series of cylinders of various sizes, added to and subtracted from each other. I included 4 off-set screw holes to mount it.
The bobbins are made of ABS (same stuff Lego bricks are made of).

I then wound the 1.8mH coil with the following spec:
1mm wire, 278 turns, internal diameter is 15mm, 40mm OD, 20mm bobbin height.

I'll wind the 0.9mH one next on the empty bobbin, this will be the same spec but 215 turns.

Pic of empty bobbin and a completed wound inductor attached.

Doug

View attachment 85337

That is really cool. What did you use the wind the coils?

I just realized this is my first post.

Hello everyone. I was lurking here and learned a lot from this thread.

 

[JohnH]

Hi @JohnH @Gene Ballzz @TheOtherEric and others!

As mentioned earlier I'm trying to include the attenuator in my current amp-build. As tube positioning, housing etc. has already been done long before I came across this thread I have a hard time to place the coil.

Do you have any input on that?

Thx in advance!

Its hard to make a clear suggestion. Things to watch out for include that it probably sends out a field of magnetism when running, which could put feedback into input circuitry, This would be greatest on the coil axis. Id be inclined to try mounting it at high level, from the upper part of the heatsink fins, as high as possible above the small transformer, with its axis horizontal, maybe using zip ties?

 

[dougie tamson]

That is really cool. What did you use the wind the coils?

This low cost winder:
https://www.ebay.co.uk/itm/Manual-C...rentrq:92ab37ff1770a9e423170f7bfff3b07c|iid:1

 

[TonyK]

Great that all sounds fine. That line-out recipe is intended to be very versatile in terms of giving whatever signal level you need, from a wide range of amp powers. There may be times when you'll have to set it very low to suit your next device without overloading it, but also if you put it after the attenuator, with speaker attached, you should get a good signal even at max attenuation, if you turn the knob up high. Once you have tested it with how you want to use it, you might want a different resistor for the front end connection, but this is good to start.

On the component powers, it depends. But lets take a worst case, say you are driving an amp at 100W, and if this was a 16 ohm attenuator. (is it 8 or 16?). The signal voltage at the amp output = attenuator input would be (100 x 16)^0.5 = 40V. The power taken into your line-out circuit of 2.2k+10k would be 40^2/12200 = 0.13W. Of this, 10/12.2 x 0.13 = 0.11W goes into the pot. So there's still a good factor of more than 2 against its 0.25W spec.

Good luck with the build!

Hi @JohnH, Thanks so much for your clarifications. I'm in no rush to do this yet, but have noted your thoughts and once finished with some house revoking, I'll do this as a little side project.

Oh, the day after you commented on my idea of getting a Vintage Modern, one came up for sale in a city 200km from me. I called to go see and buy about an hour after the ad was posted, but the amp had already gone! Of course it was in very good condition and being sold for only 200 euros! That would have been an outstanding buy. Next time...

 

[halcyon]

I just found the forum searching on DIY attenuators, having need for one and the skills to assemble, but lacking design chops.

WHAT A GREAT THREAD!

I have some parts and am wondering If can get by for a while building a resistive version with what I have prior to going to reactive at a later date.

I am concerned about power handling. The amp is 85W full up, but I can knock it down to 2 of the tubes for ~40W. I also don't need to crank it for the sound I want, maybe master output at 3-5. So question about a possible alternative resistive box:

I don't need small dose attenuation, rather a larger drop in db. So, how would a primary stage with 200W of resistors into a 100W L-pad with capacitor bleed work out? Would the primary stage act as a soak in this setup before the pad?

 

[JohnH]

Hi @mountainhick , thanks for your message. L pad attenuators, although common, and they do work, will suck down your treble tone at low volume, unlike what we are doing here. What you propose will work, but our design will work better!.

If you are OK with your amp at 40W, you can build design M2 in whatever ohms you need and it will bring your amp down to late-night home volume with no tonal change and full responsive feel. You can build it a a pure resistive version too, which will get the basic tone but lose the feel. But the reactive coil is only about $10 to buy.