The No.1 Website for Pro Audio
DIY stepped attenuator questions
Old 15th October 2018
  #1
DIY stepped attenuator questions

Hello folks,

I've been searching for ages and can't find the information I'm looking for to build this: A passive attenuator with 6 dB steps for balanced line level. Input and output impedancy should be constantly 2,5k.

Now what switch do I need for each balanced channel? 2P6T for 6 steps? I can't find a source on what parts exactly I would need and how to wire up that switch.

Thanks a lot for any help!
T
Old 15th October 2018
  #2
2P6T

2 poles 6 positions should be the switch for an H-pad I guess. Shorting. Like the Grayhill - 56sdp30-01-2-ajs.

Is that ok?
Old 15th October 2018
  #3
Lives for gear
 

There's virtually no reason to use an H-pad for connecting line-level balanced audio . . . all the build-out resistors do is add noise and increase the effects of cable reactance. So for your application, a U-pad is a better choice.

There are three ways to make a U-pad variable: Switch only the shunt resistor (one switch pole per channel required), switch the series resistors (two poles required), and switch both (three poles required). The main difference between them is that switching the shunt resistor puts the lowest noise at highest attenuation, switching the series puts the lowest noise at zero attenuation, and switching both puts the highest noise at 6dB attenuation, wherever that happens to be in the range you work out. The best choice between the three is a balance of complexity and your requirements, and how much difference there is between them depends on the resistance value of the attenuator as a whole.

I'm not sure why you chose 2.5K as a value . . . but incidentally, one does NOT match the attenuator impedance to either the source or destination impedance. You usually want the maximum resistance the input sees as being the lowest value you feel the driving equipment can handle without increased distortion. 2.5K is a pretty safe bet with virtually all pro gear, but a bit on the low side for consumer or some MI-oriented stuff. If you know the preceding equipment is comfortable driving a lower-impedance load, you can take advantage of this to reduce the noise by designing a lower-impedance attenuator.
Old 15th October 2018
  #4
Quote:
Originally Posted by kirkus View Post
There's virtually no reason to use an H-pad for connecting line-level balanced audio . . . all the build-out resistors do is add noise and increase the effects of cable reactance. So for your application, a U-pad is a better choice.

There are three ways to make a U-pad variable: Switch only the shunt resistor (one switch pole per channel required), switch the series resistors (two poles required), and switch both (three poles required). The main difference between them is that switching the shunt resistor puts the lowest noise at highest attenuation, switching the series puts the lowest noise at zero attenuation, and switching both puts the highest noise at 6dB attenuation, wherever that happens to be in the range you work out. The best choice between the three is a balance of complexity and your requirements, and how much difference there is between them depends on the resistance value of the attenuator as a whole.

I'm not sure why you chose 2.5K as a value . . . but incidentally, one does NOT match the attenuator impedance to either the source or destination impedance. You usually want the maximum resistance the input sees as being the lowest value you feel the driving equipment can handle without increased distortion. 2.5K is a pretty safe bet with virtually all pro gear, but a bit on the low side for consumer or some MI-oriented stuff. If you know the preceding equipment is comfortable driving a lower-impedance load, you can take advantage of this to reduce the noise by designing a lower-impedance attenuator.
Thanks Kirkus! The attenuator is meant to be after the mic pres (TAB V376 and Siemens V276, transformer balanced out) and before the ADC Metric Halo LIO-8. I guess the mic pres should see a constant load of maybe 2,5k or even lower? Cables are short between the units. So no need for higher impedance.

Most practical attenuation would be around 12dB for most of the applications. 6 steps of the switch would be from 0dB bypass to -30dB.
What you describe sound as if the 2 pole choice might be the right one.
Old 16th October 2018
  #5
Lives for gear
 
emrr's Avatar
I'm trying to imagine what situation requires that much attenuation AFTER a mic preamp. Those appear to be well into distortion a little after +20, so not huge headroom.

One question is what loading effects exist are or necessary with those, do they measure fine into a 10K load, or do they need a specific load to be flat on top? Often that is an undocumented value differing from quoted load which must be measured in person. From that answer, I would begin to chose the best method.
Old 16th October 2018
  #6
Lives for gear
Another consideration: if the pre and ADC are very close to each other, why do you need balanced wiring? Yes, many devices will give you a tiny bit better dynamic range (not because of the balancing, but because of the higher voltage)... but you're adding an attenuator anyway.
Old 16th October 2018
  #7
Lives for gear
 

Quote:
Originally Posted by emrr View Post
I'm trying to imagine what situation requires that much attenuation AFTER a mic preamp. Those appear to be well into distortion a little after +20, so not huge headroom.

One question is what loading effects exist are or necessary with those, do they measure fine into a 10K load, or do they need a specific load to be flat on top? Often that is an undocumented value differing from quoted load which must be measured in person. From that answer, I would begin to chose the best method.
These are excellent questions . . . I'm assuming that the OP is selecting vintage transformer-coupled units for color and character . . . and in that case, I can see that being able to shift the gain structure between the preamp and the A/D would greatly increase the flexibility in this regard.

Many vintage transformer-coupled output stages do ring when lightly loaded, and I've found 1K to be low enough to alleviate the problem with anything unless it has very poor quality iron . . . in which case there's nothing that really does the trick. It's also a value that any "electronically balanced" studio-grade output stage should be able to drive with no sweat up to clipping, even though some might misbehave slightly at higher levels if loaded with anything lower than 600 ohms.

Quote:
Originally Posted by Jay Rose View Post
Another consideration: if the pre and ADC are very close to each other, why do you need balanced wiring? Yes, many devices will give you a tiny bit better dynamic range (not because of the balancing, but because of the higher voltage)... but you're adding an attenuator anyway.
Balanced wiring can eliminate the effects of common-impedance coupling of leakage currents between the units, which isn't really a length-related phenomenon. Transformer-coupled output stages have the same headroom, output impedance, and noise floor regardless of whether they're driving an unbalanced or balanced input stage . . . but an active balanced input stage will have 6dB better headroom if there's no common-mode voltage. It is cheaper and easier to wire an unbalanced attenuator, and if this was a 47-position switch that would make a big difference . . . but for a 6-position attenuator, I don't think there's much reason to set it up as unbalanced.

Quote:
Originally Posted by tobi-mobile View Post
Most practical attenuation would be around 12dB for most of the applications. 6 steps of the switch would be from 0dB bypass to -30dB.
What you describe sound as if the 2 pole choice might be the right one.
See attached picture; sorry for the crude art and iPhone photo. This is a basic series attenuator, split in half to use two sections of a switch in a balanced configuration. Resistor values are taken from the E96 (1%) table. It keeps a constant approx. 1K load on the driving electronics for transformer damping, and worst-case noise will be better than -128 dBu (20K bandwidth, -6dB attenuation). It's drawn in the way you'd expect the contacts to be laid out for a 6-position, single-deck, two-pole rotary switch as viewed from the back.
Attached Thumbnails
DIY stepped attenuator questions-1k-attenuator-6db-steps.jpg  
Old 16th October 2018
  #8
Quote:
Originally Posted by kirkus View Post
These are excellent questions . . . I'm assuming that the OP is selecting vintage transformer-coupled units for color and character . . . and in that case, I can see that being able to shift the gain structure between the preamp and the A/D would greatly increase the flexibility in this regard.

Many vintage transformer-coupled output stages do ring when lightly loaded, and I've found 1K to be low enough to alleviate the problem with anything unless it has very poor quality iron . . . in which case there's nothing that really does the trick. It's also a value that any "electronically balanced" studio-grade output stage should be able to drive with no sweat up to clipping, even though some might misbehave slightly at higher levels if loaded with anything lower than 600 ohms.

Balanced wiring can eliminate the effects of common-impedance coupling of leakage currents between the units, which isn't really a length-related phenomenon. Transformer-coupled output stages have the same headroom, output impedance, and noise floor regardless of whether they're driving an unbalanced or balanced input stage . . . but an active balanced input stage will have 6dB better headroom if there's no common-mode voltage. It is cheaper and easier to wire an unbalanced attenuator, and if this was a 47-position switch that would make a big difference . . . but for a 6-position attenuator, I don't think there's much reason to set it up as unbalanced.

See attached picture; sorry for the crude art and iPhone photo. This is a basic series attenuator, split in half to use two sections of a switch in a balanced configuration. Resistor values are taken from the E96 (1%) table. It keeps a constant approx. 1K load on the driving electronics for transformer damping, and worst-case noise will be better than -128 dBu (20K bandwidth, -6dB attenuation). It's drawn in the way you'd expect the contacts to be laid out for a 6-position, single-deck, two-pole rotary switch as viewed from the back.
Exactly, I use four V376 and four V276 for colour to slight distortion plus eight Metric Halo mic preamps (ULN-8) for clean gain.

I am very thankful for the technical insight! And the drawing: Wow, you made my day! I will try this with an Elma 01 2x6 switch. Would this need shorting or non shorting mechanics?
Old 16th October 2018
  #9
Lives for gear
 

Shorting switches are preferred for this application - with the non-shorting type, the sound will drop out in the in-between space as you turn the knob, and you'll sometimes get more noise and popping as well. But when you're not making adjustments, they will perform identically.
Old 16th October 2018
  #10
Use good resistors. Dale CMF50 are 1/8th watt sized and fit easily.
Old 17th October 2018
  #11
Quote:
Originally Posted by Jim Williams View Post
Use good resistors. Dale CMF50 are 1/8th watt sized and fit easily.
Thanks, Jim! This would be my choice due to availability and price situation:
  • Grayhill 56SD30-01-2-AJS
  • RPR Audio Metal Film Resistors 0,5W PRP9372

The Elma 01 seem to have very long shafts which are not practical for this purpose. Grayhill quality should be perfectly fine. Last thing I'm looking for is a comfortable way to install the two DB25 on the rear of the 1U housing.

I will post the final result once it's built.

Old 17th October 2018
  #12
Check those resistors with a magnet. If they stick, don't use them. The Dales cost about $.25~.50 each but are well worth it if you intend to blow $ on an expensive switch. For those in the higher priced seats, the Vishay SOIC bulk foils are the best on the planet but cost around $15~$20 a piece.
Old 17th October 2018
  #13
Check those resistors with a magnet. If they stick, don't use them. The Dales cost about $.25~.50 each but are well worth it if you intend to blow $ on an expensive switch. For those in the higher priced seats, the Vishay surface mount bulk foils are the best on the planet but cost around $15~$20 a piece.
Old 17th October 2018
  #14
Gear Addict
 
samwinston123's Avatar
 

Be aware that those 56 series switches are TINY. Ive had trouble stuffing them with 1/4 watt resistors, let alone 1/2 watt. The 71 series are easier to work with at about the same price.

Quote:
Originally Posted by tobi-mobile View Post
Thanks, Jim! This would be my choice due to availability and price situation:
  • Grayhill 56SD30-01-2-AJS
  • RPR Audio Metal Film Resistors 0,5W PRP9372

The Elma 01 seem to have very long shafts which are not practical for this purpose. Grayhill quality should be perfectly fine. Last thing I'm looking for is a comfortable way to install the two DB25 on the rear of the 1U housing.

I will post the final result once it's built.

Old 17th October 2018
  #15
Lives for gear
 

Quote:
Originally Posted by tobi-mobile View Post
The Elma 01 seem to have very long shafts which are not practical for this purpose.
Many off-the-shelf pots and switches available in-stock from suppliers have long round shafts, and it's no sweat to cut them to the length you need - simply mark the length you want, clamp the waste end of the shaft in a vise, and slice them with a hacksaw or cut-off wheel in a die grinder. The whole switch/pot can be wrapped in masking tape to prevent filings from finding their way inside. You can then use a collet or setscrew style knob, and orient the markings whatever way is suitable.

If you go manufacturer-direct on the order, then there are myriad options in shaft length, style, and configuration of flats, etc. etc., but custom options face significant minimum orders. I'm not saying you shouldn't use the Greyhills . . . but I wouldn't let just the excessive shaft length keep you away from a part you prefer based on other parameters.
Old 18th October 2018
  #16
Lives for gear
 
ruffrecords's Avatar
Grayhill 56 series are a real PITA to wire. I would recommend Grayhill 71 series. They are much easier to wire and they have quarter inch shaft versions (which the 56 series do not) which means a much better choice of knobs.

Cheers

Ian
Old 18th October 2018
  #17
Quote:
Originally Posted by ruffrecords View Post
Grayhill 56 series are a real PITA to wire. I would recommend Grayhill 71 series. They are much easier to wire and they have quarter inch shaft versions (which the 56 series do not) which means a much better choice of knobs.

Cheers

Ian
So that would be this one - 71ED30-02-1-AJS


2 deck, 1 pole, 6 positions
Old 18th October 2018
  #18
Gear Addict
 
samwinston123's Avatar
 

That’s the 4mm metric shaft version. If you want a 1/4” shaft, which is a more standard size, then it’s 71B instead of E. Also, if you’re only using 6 positions, then it’s cheaper to use a 2 pole/deck switch: 71BD30-01-2-AJS

Quote:
Originally Posted by tobi-mobile View Post
So that would be this one - 71ED30-02-1-AJS


2 deck, 1 pole, 6 positions
Old 18th October 2018
  #19
Quote:
Originally Posted by kirkus View Post

See attached picture; sorry for the crude art and iPhone photo. This is a basic series attenuator, split in half to use two sections of a switch in a balanced configuration. Resistor values are taken from the E96 (1%) table. It keeps a constant approx. 1K load on the driving electronics for transformer damping, and worst-case noise will be better than -128 dBu (20K bandwidth, -6dB attenuation). It's drawn in the way you'd expect the contacts to be laid out for a 6-position, single-deck, two-pole rotary switch as viewed from the back.
These seem to be the closest values I can get:
255
127
63,4
34,8
17,8

The PRP Audio Metall-Films seem to be recommended by audio guys. They're in stock. I'll get them.
Old 18th October 2018
  #20
Lives for gear
 

Quote:
Originally Posted by tobi-mobile View Post
These seem to be the closest values I can get:
255
127
63,4
34,8
17,8

The PRP Audio Metall-Films seem to be recommended by audio guys. They're in stock. I'll get them.
Subbing to the values you indicate will present a load of 1030.8 ohms to the mic preamp, and give you steps as follows (based on a source impedance of 30 ohms, and a load of 10K from the ADC):

-5.93 dB
-11.74 dB
-17.34 dB
-23.31 dB
-29.43 dB

Personally, I think that this will be very much more precise than you'll ever need, but you be the judge.

The PRPs are fine resistors.
Old 13th November 2018
  #21
Lives for gear
 

Quote:
Originally Posted by tobi-mobile
I have all the parts now and going to put it all together. One question came up: You said you calculated this based on 10k load from the DAC and 30 Ohms from the mic pre. Actually the Mic pre is 300 Ohms.
(V276.pdf - Google Drive)

What difference does that make? Do you have an online calculator for this?
I'll admit my German is super-rusty, especially with long precise compound words . . . but doesn't "Ausgangsscheinwiderstand" refer to the output source impedance? This is the pertinent figure to use when calculating the effects of the attenuator's load on the overall gain. I think the specs that are merely labeled "Ausgang" are describing the levels available when there's a load of 300 ohms presented by the receiving equipment. FWIW 45 ohms is about what I'd expect from a 1970s discrete-transistor transformer-coupled output stage.

In any case, if the preamp really did have a 300 ohm output impedance (instead of 45 ohms), it would result in a level that was about 1.9dB lower from ALL the steps, INCLUDING the "0" position . . . the accuracy of the steps would be virtually unchanged. This is because the series attenuator design presents a constant load impedance to the driving source, which is desirable in your application as it will keep the damping of the preamp's output transformer constant at all control settings. In practice, this voltage loss from the loading is wholly negligible, and hugely, vastly preferable to the noise penalty of a higher-impedance attenuator and the possibility of ringing from the transformer.

As for an online calculator . . . I don't use one or know of any, as I do all of this sort of stuff in Excel. Attached is a version of the spreadsheet I use for series attenuator designs, set up for your 6-position switch of -6dB each. I was only thinking of my own use for it, so forgive the clumsiness of the layout. All of the changes and inputs are made on the "worksheet" page, and you can make it for however many steps you like simply by inserting rows above the final "off" position, and then dragging down the second row (row 15) of formula cells (columns C through P) to all the rows (except "off"). The "nearest value" column comes from looking up the calculated target resistance to the nearest value from the back sheet, which contains a table of standard E96 1% resistor values -- if you want to see the results with other values, simply overwrite the values in this column with your own.

Note that the spreadsheet is for a single-switch "unbalanced" design, so for the "balanced" version one simply halves the target impedance, and doubles the value of the last resistor . . . think of it as two series attenuators in series with each other. The noise will then be about 3dB worse in reality than what the table says, given that the calculations will refer to only half of the attenuator . . . but for a 1K attenuator, it's still going to be much below the noise floor of any equipment you'll ever connect to it.

I find I'm a bit embarrassed to realize that when I posted the original values, I entered them in a hurry and didn't catch that I'd set up the load impedance entry field to take values in kiloohms, rather than ohms, which affects the calculations a bit. So with the values you posted, your steps will be:

-6.27 dB
-12.26 dB
-17.32 dB
-23.25 dB
-29.34 dB

I wouldn't think this level of inaccuracy would make any difference in your application, but my apologies for any inconvenience it may have caused.
Attached Files
File Type: zip Series Attenuator 6x-6dB.zip (41.5 KB, 18 views)
Old 14th November 2018
  #22
Lives for gear
 
Richard Crowley's Avatar
If the purpose of this exercise is subjective effects, then what exactly is the need for controlled, constant impedance?
And for precisely 6db stepped attenuation, for that matter?
If this were some kind of laboratory test and measurement rig, that would be a different matter.
Old 14th November 2018
  #23
Lives for gear
 

Quote:
Originally Posted by Richard Crowley View Post
If the purpose of this exercise is subjective effects, then what exactly is the need for controlled, constant impedance?
And for precisely 6db stepped attenuation, for that matter?
If this were some kind of laboratory test and measurement rig, that would be a different matter.
I think that the reasons for having a constant load on a transformer-coupled line output are quite well-established, and the choice of value is based on the best compromise between loading loss, and low source impedance and low Johnson noise.

As far as the precision required in the attenuation steps . . . that's an application question for the user to decide. It takes the same amount of effort to solder together an attenuator regardless of the precision of the resistors, so why not choose the values that best fit the original design goal? And if one deviates from the closest values based on parts availability . . . isn't it a good idea to calculate the effects of this choice before building the device? This is especially true if one finds an error in the arithmetic that's been used to make a parts-purchasing decision.

Personally, I don't think it ever makes sense to simply be lazy with the design and optimization process, especially if it's based on my own value judgement on how somebody else may choose to use the device I'm designing.
Old 14th November 2018
  #24
Lives for gear
 
Richard Crowley's Avatar
Quote:
Originally Posted by kirkus View Post
I think that the reasons for having a constant load on a transformer-coupled line output are quite well-established,
Yes, as you accurately enumerate, to optimize the best performance from the transformer. But THIS application is to exploit the shortcomings of the transformer. Which appears to be exactly the opposite goal.

Quote:
It takes the same amount of effort to solder together an attenuator regardless of the precision of the resistors, so why not choose the values that best fit the original design goal?
Yes, if you are going for laboratory precision measurement objectives. But again, that appears to be the farthest thing opposite of the desired outcome. Why not simply use a potentiometer if you want controllable attenuation. And another potentiometer to apply variable loading to the source output transformer. Then you can "tune" the effect until you like what you hear.

Constant impedance and stepped attenuators just smacks of audiophool fluff with no apparent benefit to the ultimate goal.
Old 14th November 2018
  #25
Lives for gear
 

Quote:
Originally Posted by Richard Crowley View Post
Constant impedance and stepped attenuators just smacks of audiophool fluff with no apparent benefit to the ultimate goal.
What it "smacks of" is not my concern . . . the design problem was for a switched attenuator of 6dB steps for the output of a vintage mic preamp, not a potentiometer level control, nor a variable-transformer-loading control.

With a bit of imagination, I could see that repeatable settings could be a boon, and also that the ringing of an unloaded output transformer is perhaps not the most subjectively desirable attribute of these mic preamps. Perhaps that if the steps are precisely calibrated, this makes it easier to do a valid subjective comparison by adjusting other parts of the gain structure to keep the overall level constant.

I guess that I trust that the OP has his reasons, and it's not an unreasonable or difficult device to build. But if I just make it about myself, I'd say that . . .
Helping people with audio projects they want to build = fun.
Convincing them to settle for something different or crappier = not so much.
Old 15th November 2018
  #26
Lives for gear
 
Richard Crowley's Avatar
Quote:
Originally Posted by kirkus View Post
What it "smacks of" is not my concern . . . the design
Helping people with audio projects they want to build = fun.
Convincing them to settle for something different or crappier = not so much.
Helping people see alternatives to what they think they want. = fun.
Helping people break out of conventional thinking. = fun.

Nobody is trying to convince anyone to "settle" for anything.

The OP's stated goal is "colour to slight distortion". Maintaining optimal, fixed load impedance is simply antithetical to that goal. Do we encourage the OP to build stepped attenuators with constant impedance, or do we interpret his stated goals of experimenting with non-conventional operation?

If we wanted to avoid "crappier" we wouldn't be using our grandfathers old audio gear and trying to create "colour to slight distortion". We will apparently have to agree to disagree on what the OP is trying to do, and what are some means to achieve that.
Old 15th November 2018
  #27
I know how these old preamps sound and react in different situations. I had lots of them wired up with faders. The range up to 30 dB reduction was what I used in front of an ADC. Going directly into the ADC doesn't give me the flexibility to achieve these results and the load of the ADC is too high, to make the preamp shine.

I simply love the sounds I get from this setup. That's all the reason it needs I think. Why argue about sound preferences?

Now with a mobile setup I don't have space for fader packs and the stepped attenuator will give me the load I need and repeatable settings.

Quote:
Originally Posted by Richard Crowley View Post
Helping people see alternatives to what they think they want. = fun.
Helping people break out of conventional thinking. = fun.

Nobody is trying to convince anyone to "settle" for anything.

The OP's stated goal is "colour to slight distortion". Maintaining optimal, fixed load impedance is simply antithetical to that goal. Do we encourage the OP to build stepped attenuators with constant impedance, or do we interpret his stated goals of experimenting with non-conventional operation?
Old 15th November 2018
  #28
Quote:
Originally Posted by Richard Crowley View Post
If we wanted to avoid "crappier" we wouldn't be using our grandfathers old audio gear and trying to create "colour to slight distortion". We will apparently have to agree to disagree on what the OP is trying to do, and what are some means to achieve that.
That's quite some paternalistic nonsense. Feel free to dislike how V276 and V376 sound when overdriven with different loads. I like it.
Old 15th November 2018
  #29
Thank you

You are absolutely right. Impedance is stated as <45 Ohm so everything's fine. The values are more precise than this application actually needs, but I do agree on: once you build something like that, try to do it right.

I will post pictures of the unit soon. Building it right now.

Thanks again kirkus for your help!

Quote:
Originally Posted by kirkus View Post
I'll admit my German is super-rusty, especially with long precise compound words . . . but doesn't "Ausgangsscheinwiderstand" refer to the output source impedance? This is the pertinent figure to use when calculating the effects of the attenuator's load on the overall gain. I think the specs that are merely labeled "Ausgang" are describing the levels available when there's a load of 300 ohms presented by the receiving equipment. FWIW 45 ohms is about what I'd expect from a 1970s discrete-transistor transformer-coupled output stage.

In any case, if the preamp really did have a 300 ohm output impedance (instead of 45 ohms), it would result in a level that was about 1.9dB lower from ALL the steps, INCLUDING the "0" position . . . the accuracy of the steps would be virtually unchanged. This is because the series attenuator design presents a constant load impedance to the driving source, which is desirable in your application as it will keep the damping of the preamp's output transformer constant at all control settings. In practice, this voltage loss from the loading is wholly negligible, and hugely, vastly preferable to the noise penalty of a higher-impedance attenuator and the possibility of ringing from the transformer.

As for an online calculator . . . I don't use one or know of any, as I do all of this sort of stuff in Excel. Attached is a version of the spreadsheet I use for series attenuator designs, set up for your 6-position switch of -6dB each. I was only thinking of my own use for it, so forgive the clumsiness of the layout. All of the changes and inputs are made on the "worksheet" page, and you can make it for however many steps you like simply by inserting rows above the final "off" position, and then dragging down the second row (row 15) of formula cells (columns C through P) to all the rows (except "off"). The "nearest value" column comes from looking up the calculated target resistance to the nearest value from the back sheet, which contains a table of standard E96 1% resistor values -- if you want to see the results with other values, simply overwrite the values in this column with your own.

Note that the spreadsheet is for a single-switch "unbalanced" design, so for the "balanced" version one simply halves the target impedance, and doubles the value of the last resistor . . . think of it as two series attenuators in series with each other. The noise will then be about 3dB worse in reality than what the table says, given that the calculations will refer to only half of the attenuator . . . but for a 1K attenuator, it's still going to be much below the noise floor of any equipment you'll ever connect to it.

I find I'm a bit embarrassed to realize that when I posted the original values, I entered them in a hurry and didn't catch that I'd set up the load impedance entry field to take values in kiloohms, rather than ohms, which affects the calculations a bit. So with the values you posted, your steps will be:

-6.27 dB
-12.26 dB
-17.32 dB
-23.25 dB
-29.34 dB

I wouldn't think this level of inaccuracy would make any difference in your application, but my apologies for any inconvenience it may have caused.
Old 16th November 2018
  #30
done.

Here it is. Took me more time than I thought. Details like getting the right DB25 PCB, the housing, the Grayhills, layout of the panel...

As you can see the design is made to fit the alignment of the Metric Halo knobs so that the channels correspond in their layout. I like ergonomics to be intuitive. I chose a 9,5" housing to let some space to the sides. The Metric Halos get quite hot so the more space in between both units the better.

@ kirkus : Thanks again for the drawing! It was not only super helpful, but made a nice logo on the front :-)

I made some tests before and I can say that the -6dB and -12dB settings seem to be my standard more or less. I tend to leave a good amount of headroom at the ADC in, so even the -12dB setting doesn't make the preamps overdrive, but nicely and subtly saturate. Of course, -30dB is extrem, but I love to have options on different styles of distortion in my productions. From stomp boxes to plugins. Now with these preamps and those heavy settings I have some more. Cool stuff!
Attached Thumbnails
DIY stepped attenuator questions-_c1_00006.jpg   DIY stepped attenuator questions-_c1_00009.jpg   DIY stepped attenuator questions-_c1_00010.jpg   DIY stepped attenuator questions-_c1_00011.jpg   DIY stepped attenuator questions-_c1_00012.jpg  

Post Reply

Welcome to the Gearslutz Pro Audio Community!

Registration benefits include:
  • The ability to reply to and create new discussions
  • Access to members-only giveaways & competitions
  • Interact with VIP industry experts in our guest Q&As
  • Access to members-only sub forum discussions
  • Access to members-only Chat Room
  • Get INSTANT ACCESS to the world's best private pro audio Classifieds for only USD $20/year
  • Promote your eBay auctions and Reverb.com listings for free
  • Remove this message!
You need an account to post a reply. Create a username and password below and an account will be created and your post entered.


 
 
Slide to join now Processing…
Thread Tools
Search this Thread
Search this Thread:

Advanced Search
Forum Jump
Forum Jump