Simple Attenuators - Design And Testing

KIL0

New Member
Joined
Aug 3, 2022
Messages
3
Reaction score
0
Hey there,
i was looking for a 16 ohms dummy load for a fender hot rod deluxe to use with my torpedo live (which is 8 ohms only) when I stumbled across this design and this thread!

First of all I need to say that this design with all the simulations and support is absolute great work! I'm an electronics dev myself and I highly appreciate all the time and effort that has gone into this!

I wasn't able to get through all of the by now 146 pages of this thread and the search function didn't get me where I wanted so please forgive me if I address a yet solved issue but I would like to discuss the following three things:
- Has anyone used the M2 as a dummy load without a cab with something like a torpedo? I guess that due to the simulation results it should be fine to connect an 16 ohm resistor to an switched jack to connect if there is no cab plugged in
- Is there any thought about a 4 ohms version of the M2 box? I'm actually a bass player by trade and I'm thinking about to get an Ampeg SVT 2 Pro. This amp only got 4 ohms output matching (it also got a 2 ohms setting, but this is irrelevant for this issue) but for smaller venues I want to be able to connect an 8 ohms cab to it as well (this issue has been talked through before by example of an 16 ohms cab with a 8 ohms M2 i guess).
- Can anyone link me an air coil they used or can give me some detailed specs about them? I'm actually able to wrap them on a machine at my company's plant, maybe even with a 3D-printed base to fit the height into a 1HE rack housing. Or has somebody even made the air coil themselves? In this case I'd be glad to get some experiences! ,)

Best regards from Berlin, Germany
 

Gene Ballzz

Well-Known Member
Joined
Feb 19, 2013
Messages
3,662
Reaction score
4,258
Location
Las Vegas, NV
@KILO

First, :welcome: to the forum!


That's a pretty ambitious number of watts to dissipate! :eek: I'm guessing the unit, as well as the resistors involved in the first "always on" stage to be massive and fairly large (or numerous)! Cooling will be a primary concern. That SVT 2 Pro is likely capable of putting out in upwards of 500 watts, when really cookin'!

Good, Detailed Planning Is Critical!
Gene
 

JohnH

Well-Known Member
Joined
May 31, 2008
Messages
5,041
Reaction score
3,260
Location
Wilton NSW
hi @KIL0 and welcome to our thread.

For the HRD, if you need a 16 Ohm load, you can build a 16 Ohm M2 and use it with either a 16 or 8 cab, or in principle, an 8 Ohm Torpedo. But when I looked up HRD's specs, It looked to have an 8 Ohm output rather than 16Ohm?, also ok with two cabs for a 4 Ohm load. But is yours a 16 Ohm version?

As a load box, it's safe to run an M2 with no speaker, so long as it's set to max attenuation. In fact, if you have the fixed Stage 1 plus at least the -14dB stage engaged, it's all fine with no cab.

I haven't tried IR loaders myself, but others have, I believe. If the line-out comes from the input, it captures the correct treble rise like a speaker, but there is not a bass-resonance peak there because M2 doesn't include it. So you might add a bit of low EQ to compensate. I tried some direct recordings and I found a bit of bass boost on the mixer, plus a sub 100hz cut did a good job.

But M2 gets the right response at the speaker, and its the real speaker doing it. So at the output, you do get the bass resonance plus treble rise, and its whatever is natural for the actual speaker. If you can keep a speaker plugged in, a line out from the output might be better.

M3 added a bass resonance circuit at the front end, but the parts to do it right get bulky and expensive (ferrite cored coil and big film caps). M2 works really well without this and I'm thinking it's not worth it.

You can build a 4 Ohm version, and basically coil and resistor values are x1/2 compared to 8 Ohm, then tweaked to fit with standard component values, ideally keeping ratios of values within each stage s consistent as possible. I have some values for a 4Ohm M2 if needed and others have built this.

But, if its really for a 500W amp, that is 5x anything attempted here before! I don't know how that'd look in practice, dealing with enormous heat.

Coils are easily bought. For the HRD, wind with 18 guage wire on an air-cored bobbin. They come from places that sell speaker cross-over components.

If you're winding one, you can start with an online calculator to work out turns etc based on bobbin dimensions.

I just bought a 1mH coil, which I plan to unwind to 0.9mH. I can measure this on my meter, and if you are in electronics, maybe you have access to inductance measurement gear too? In which case you can use an online calc to work it out reasonably close, and adjust with measurement.

With the online coil calculators, I find some of them don't provide a design that matches well with known real coils. But I'm finding this one is very close, (dimensions in inches though)


For my unwinding, I'm going to work out how much I need to unwind to go from 1mH to 0.9mH, first based on the calc, and then in reality with the meter. 8 Ohm M2 is designed with a 0.9mH coil, but many suppliers don't list this. Actually, it's not super critical but might as well get it at the optimum if possible.

cheers
John
 

KIL0

New Member
Joined
Aug 3, 2022
Messages
3
Reaction score
0
Hey Gene,
Hey JohnH,

thanks for your reply!

@Gene Ballzz There is something with the output power specs with tube amps that I never really understood. The Amp is rated as 300 W so I knew, that there would be a bigger array of parts in the first stage but I didn't thought of more than double the load to dissipate!
But for sure I planned with heat sinks and active controlled cooling with temperature probing.

@JohnH Actually I wasn't able to check the exact output configuration on the HDR due to me not having it here, yet. But due to it beeing single 12" combo I thought it would be 16 ohms ^^
If the HDR does 8 ohms, I would not need the M2 for this situation at all. But otherwise I'd be interessted to check between running the torpedo in front or in the back of the M2. I also read about the other variants (M3, ect), the work that went into this is great!

For the SVT 2 Pro I guess it wouldn't be necessary to dissipate the whole 500 W due to I never think I'll be able to run this amp at more than 50% but I will be taking a look on how much bigger it will get for sure! On the other hand I first thought about implementing a power measurement with the fan controller for measuring the real amp output.

Regarding the coil: This calculator looks really good, the results contain all necessary information! I need to check what wire gauges we got around but around 200 turns with AWG 16 for 0.9 mH would be absolute easy to manufacture.
Measuring the coils should be no issue as well. I guess I should have access to a measuring device at the plant, otherwise I've seen that there are lots of projects on hackaday ect. around as small saturday evening builds.

Thanks a lot so far!

Regards
 

JohnH

Well-Known Member
Joined
May 31, 2008
Messages
5,041
Reaction score
3,260
Location
Wilton NSW
I looked up that SVT! it's a BEAST! If you want to attenuate because of better tone up high, all the design theory should work but built with Stage 1 like an industrial substation. Speakon connectors instead of jacks, thick gauge for the coil and uprated resistors, probably divided in multiples, heatsink and fan cooling. It should work though.

(That assumes that, like a guitar amp, there are better tones at high volume, instead of just turning it down.)
 

KIL0

New Member
Joined
Aug 3, 2022
Messages
3
Reaction score
0
I calculated the coil and the resistors and if I use AWG10 in the coil this should work. If the amp really kicks 600 W (12,25 A @ 49V) AWG10 should be fine. For the resistors I would use the HS100 series from Arcol for the resistors, 10€ per 100 W of power consumption with 3 in parallel for R2A (33R) and R2B (22R) and for R1 maybe six 33R in parallel. This bank would have a size of around 300x200 mm without the air coil, so there would be lots of space for heat sinks and cooling fans.

@JohnH The main reason why I considered to build an attenuator for this amp was to be able to run it with 8 ohm cabs when not gigging with my own SVT 810 ,)
I guess I'll get the amp and start to measure out what currents I really get while playing and then think about getting this attenuator done. But in any case I'll go ahead and check if we got such thick wires for winding at the plant ^^

Regards
 

JohnH

Well-Known Member
Joined
May 31, 2008
Messages
5,041
Reaction score
3,260
Location
Wilton NSW
Unwinding an inductor coil: 1mH to 0.9mH

Nothing better to do on a Saturday morning, so I decided to unwind an air-cored inductor.

The 8 Ohm version of the M2 attenuator calls for a 0.9mH inductor. But many standard ranges only offer 0.82mH or 1.0mH. To be honest, its not really critical, a 10% difference in this value doesn't make much difference - its still near enough to the sweet spot. But these coils are very simple things, and its easy to take a 1mH coil and take off the tape, unwind it a few turns, and then secure it again and then trim the wires. But how much wire to remove?

Ideally, if you have a meter that can read inductance, you can measure it and unwind with the meter on, until the reading is reduced as wanted. While doing this, don't cut the wire yet, just let it be free of the coil and then its not part of the inductance. This is what I have done before, since I have an inexpensive but very useful multimeter that does L and C and a few other extra useful things. After the right value is achieved, then secure the coil with tape, trim the lead and scrape/solder-tin the new wire end.

But most meters don't have Inductance (L) ranges. So for this coil, destined for a future M2, I thought I'd study a bit deeper and note the outcome FYI. There are formulae for working out the inductance of a multilayer coil of given inner and outer diameters, and length of coil, and you can find these online as web apps, and also as mobile apps. So in principle, we can use one to work out the length of wire in a coil at one inductance value (given bobbin dims), and then at the reduced one and subtract to see how much wire to remove. I've been using this online one:


Unlike some others, I'm finding it does a good job of predicting a few real known coils.

Here is my coil, 1.0mH as bought from Wagneronline in Sydney. Its 18 gauge on a plastic bobbin, much as many others used i this thread. The brand is Dai-Chi:

coil1.jpg coil2.jpg

The bobbin has an inner diameter of 25.0mm (ie the surface where the wire is wound onto it) and a length of 21.5mm between the end flanges. Interesting how the mounting hole is at the opposite end to the leads, which may be unusual but allows mounting with just a short screw.

The first measurement read 1.04 mH:

measuring.jpg

I cant tell whether the extra 4% is the coil or the meter, maybe some of each. I wouldn't expect this meter to be super-accurate, though it does seems to measure known close tolerance parts within their spec. So, I decided to target unwinding a little conservatively, to 0.92mH.

Here's my results, both predicting how much to unwind using the web tool, and also what I actually did using the meter:

coil unwind 220806.gif

Based on my measured result and chosen target, the web app is saying to remove 1.2m, and in practice to get the result I removed 1.4m to get a 0.92 mH reading

The chart at left shows how the measured inductance reduced over thee steps of unwinding.

Had I not had the meter, and just assumed that the coil was 1mH and I was heading to 0.9mH, the app used this way would have predicted a 1.05m unwinding.

Conclusion:

Given we are making a small change to improve and the result doesn't have to be exact, given its acceptable with no change at all, using the web app to predict unwinding is ok. The stating point is to check that it reasonably predicts the basic coil. But having a meter that measures inductance is cool, and I like it!

Note the the specific results here are based on the geometry of my coil, and would be different if it was a different shape.
 

Chappy

New Member
Joined
Sep 7, 2013
Messages
18
Reaction score
13
Location
Canada
Effect of Screw (ferrous content) on Inductance

I was curious as to how much a normal machine screw placed through the centre of the coil would affect the inductance in my attenuator build. Results from my spare 0.91mH coil from Solen Canada measured on an LRC bridge.

Air core only: 0.94mH Q 8.6
1 screw: 1.045mH Q 5.4
2 screws: 1.135mH Q 4.4
3 screws: 1.225mH Q 4.1

I put my build together with one ferrous machine screw through the centre of the inductor to hold it in place. Looks like I need to find a nylon one and swap it out.
 

Attachments

  • Coil w Screw.JPG
    Coil w Screw.JPG
    585.9 KB · Views: 7
  • OneScrew 1.045mH.JPG
    OneScrew 1.045mH.JPG
    615.8 KB · Views: 7
  • NoScrew 9.45mH.JPG
    NoScrew 9.45mH.JPG
    674.2 KB · Views: 7
Last edited:

JohnH

Well-Known Member
Joined
May 31, 2008
Messages
5,041
Reaction score
3,260
Location
Wilton NSW
Hi @Chappy , thanks for posting those, And I totally LOVE that piece of test gear! I'm thinking that didn't come from eBay? I'm remembering a Whearstone Bridge circuit from the high-school lab in the 70's, but with a reactive branch? Cool anyway!

Those inductance shifts match what I found too. And there's also an opposite gremlin to investigate, a reduction in L due to proximity to a metal chassis causing eddy currents. Both best avoided, and sadly they both degrade the Q, even though L might stay similar overall.
 

Chappy

New Member
Joined
Sep 7, 2013
Messages
18
Reaction score
13
Location
Canada
Hi @Chappy , thanks for posting those, And I totally LOVE that piece of test gear! I'm thinking that didn't come from eBay? I'm remembering a Whearstone Bridge circuit from the high-school lab in the 70's, but with a reactive branch? Cool anyway!
John,

Actually it did come from Ebay. It is a US military AN/URM-90 test set. Built using top of the line components back in the late 50s. I had to wait quite a while for one to show up that was still in good condition. Fortunately I was the only bidder on this one, and I got it for a steal. I had to pay almost as much for the shipping as I did for the unit. I had to replace a few electrolytic caps that had dried up but otherwise it seems to work as advertised, giving very accurate results.

It is just as good, or better than the more popular General Radio bridges which show up quite often on Ebay. Also, at much lower cost as it is not as well known. Different types of bridges are switched into the circuit depending on the DUT. Wheatstone for R, De Santy for C, Maxwell or Hay for L. It uses a 0.1uF precision capacitor for reactance measurements. It has two internal DC voltages for resistance measurements, and a 1KHz sine signal for L, C, & AC resistance measurements. External signals can be used as well as external meters for showing the null.

Steve
 
Last edited:

Gene Ballzz

Well-Known Member
Joined
Feb 19, 2013
Messages
3,662
Reaction score
4,258
Location
Las Vegas, NV
So folks,

Having just built three of these units, I tested them extensively, in a sonic manner, as in no SPL meter! One was an M2/16ohm with an extra, tonally compensated 8ohm out and the other two were M2/8ohm with an added, tonally compensated 16ohm out. One of the M2/8ohm units has a fan installed! These made six of these units I've built! The first three were prototypes for testing, layout and reworks.

The test subjects were my Marshall SV20H head, a 16Ω 1960B, 4x12 with 98db Greenbacks and a 1965B 4X10 with 94db G10L-35s. I recorded a 1 minute-ish, cranking rhythm loop and for each test (except stepping through the stages) I let the loop crank at full attenuation for at least an hour or more. It should be noted that comparing a 98db speakers loaded cabinet to a 94db speakers loaded cabinet complicated the comparison, just a little bit.

Here are the sonic and heat results:
> Both the 8Ω and 16Ω units performed flawlessly, with no tonal artifacts and/or losses at all attenuation levels.
> On both units, the compensated out (for 16Ω on the 8Ω unit and 8Ω out on the 16Ω unit) retained all tone, but did indeed give just a tiny bit more attenuation (I'd guess around 1 or 2 db) than using a cabinet "native" to the unit's input impedance.
> All three units heated up pretty well (not burn your hand hot, but almost as hot as a fresh cup of coffee), but the one 8Ω unit with a fan, cooled down to just barely "luke warm" with the fan running!

My conclusion is that this attenuator design is "BY FAR" the best ever conceived, retaining all tone, feel and response throughout its range of operation! When properly ventilated and heat sinked, it requires no fan up to about 30 watts, fully cranked for long perioids of time. Anything above 35 watts and yup to 50 watts, fully cranked should utilize a fan!

Now, FWIW, the reason I've built/am building so many of these, is that every time I get one finished and folks hear/see it, someone wants to buy it! I use them all the time, so I need to have at least an 8Ω & 16Ω unit on hand, for my own use! Here ar pics of #4 & #5, and I'll get pics up of #6, once the fan guard/filters arrive and get installed! This is the formt I plan to use for all future 50 watt M2 builds, although, I've found some slightly more pleasing feet! Anyone interested in getting their hands on one of these, while avoiding the time/effort of building, can feel free to contact me privately, as custom features can be either added or deleted! I'm not a business and am not making these commercially, I simply want folks to be able to enjoy the liberation these units provide, for those not equipped to do a build themselves! They truly are the greatest thing since sliced bread!

Thanks 4 Looking!
Gene

M2/16Ω
IMG_0680.jpeg


IMG_0681.jpeg IMG_0682.jpeg IMG_0683.jpeg
 

JohnH

Well-Known Member
Joined
May 31, 2008
Messages
5,041
Reaction score
3,260
Location
Wilton NSW
Hi Gene! I'm really delighted to see these and its great news to hear about the testing. They look really good and I'm very happy to have developed the design together with you!

Thanks also to everyone who has trusted this design and provided encouragement and feedback.
 

Latest posts



Top