Simple Attenuators - Design And Testing

[iefes]

Quick update: Tonight I installed a 10k 2W resistor in series with a 10k pot (only had a 16mm on hand but will change for a 24mm after I did some testing) and tapped the signal from the pot off into the interface --> garageband with Torpedo Wall of Sound. Signal is tapped directly at the input jack of the attenuator. I hooked up an 8 Ohm dummy resistor at the output. This way I could switch between the purely resistive and reactive loads and could hear the difference on my headphones.

Reactive definitely sounds brighter/livelier. I didn't expect such a difference. I mean if I wouldn't have heard the tone with the reactive load in direct comparison I probably wouldn't notice that something is missing, but with the reactive load it does sound really different in the highs/presence. I couldn't hear a difference when switching in the subsequent attenuation stages though.

Will do more testing but didn't have more time today.

Thanks again, so much fun to play with this!

 

[JohnH]

Hi @iefes ,on the dummy load resistor:

Personally, I wouldn't see a need for it and would prefer the simplicity of being able to get a dummy load setting with no added parts and switches. But if it was added, Id put it at the speaker end of the attenuator, then it wont matter how the switches are set.

 

[iefes]

Yep, simplicity is a good point. I think I'll leave it like it is for now. I can easily attach my 8 Ohm 150W dummy resistor (wired in an enclosure with jacks) and have the option to switch between reactive and resistive loads by using the bypass switch. In the reactive case the dummy resistor will sit at the end of the attenuator-chain and when bypassed, the dummy will take up all the power.

I build tube amps from time to time (currently working on a ~40W SLO and a Jim Kelley clone for a friend) so the option to switch between resistive and reactive loads may be nice for testing purposes (check voltages, adjust bias, frequency sweeps etc.).
And then there's this great attenuator! I've played it a little more today and was really happy with the sound. I like it a lot more than the Weber MiniMass I had tested. I'll have to do a comparison to the TAD Silencer of a friend of mine at some point.
I may have to add that I'm not really a Marshall player. Most of the time I play vintage Fender style amps and that's what I used to test the Attenuator. I tried the Blackface AB763 Channel and the 6G3 Channel of my Hybrid-Deluxe build and got really nice results with both of them. I tried it with and without negative feedback and liked the response with a bit of NFB better, but I don't think this has to do with the attenuator.

 

[JohnH]

Just another option to condider:
If you want to create a resistive load for testing or comparison, you can just shunt the coil with a switch. It doesn't change the basic resistance values seen by amp or speaker except for a part of the resistance of the coil itself

 

[Mjh36]

Thank you JohnH and GeneBallz for answering my question about switched jacks on Page 82. I think I get the concept finally. Recently got my parts for the build so I just have a few more questions.

With switched jacks on the attenuator I can:

1. Ground the input jack hot lug to reduce noise etc.?
2. Ground the hot lugs of the speaker out jacks AND line out jack when no cable is plugged in? Just for whatever reason, safety, etc.?
3. Basically it's ok to ground every jack tip when nothing is plugged in to that particular jack? Not necessary but doesn't hurt?

For example, say I use the amp > attenuator input > line out with full attenuation and no speaker plugged in. The 8 and 16 ohm jack tips will be grounded. All good?

While it seems unnecessary, I still want to do it and just want to make sure I'm not messing something up. I think it doesn't matter in this case if a jack tip is grounded or not, because nothing would be plugged into that jack anyway? I also think I worry too much!

Other than that I hope to post pics soon!

 

[JohnH]

Hi @Mjh36

On page 82, we were talking generally about switched jacks and their uses, which were mainly about amps.

But for the attenuator, there's no need to ground the tips in any of the attenuator jacks, it's not needed and not really desirable.

But in any case, good luck with the build and we'll look forward to seeing it!

 

[Gene Ballzz]

Thank you JohnH and GeneBallz for answering my question about switched jacks on Page 82. I think I get the concept finally. Recently got my parts for the build so I just have a few more questions.

With switched jacks on the attenuator I can:

1. Ground the input jack hot lug to reduce noise etc.?
2. Ground the hot lugs of the speaker out jacks AND line out jack when no cable is plugged in? Just for whatever reason, safety, etc.?
3. Basically it's ok to ground every jack tip when nothing is plugged in to that particular jack? Not necessary but doesn't hurt?

For example, say I use the amp > attenuator input > line out with full attenuation and no speaker plugged in. The 8 and 16 ohm jack tips will be grounded. All good?

While it seems unnecessary, I still want to do it and just want to make sure I'm not messing something up. I think it doesn't matter in this case if a jack tip is grounded or not, because nothing would be plugged into that jack anyway?

The whole attenuator circuit is still connected to the jacks, even with nothing plugged into/out of those jacks. Connecting the jacks' positives to negative creates a dead short! Remember, there is no actual ground, per se. And according to design, the jacks are best being Cliff UK style and isolated from the the chassis of the attenuator.

Best To Ya,
Gene

 

[iefes]

I would like to stress that it is probably best not to switch any of the tips to ground when nothing is plugged in.
1.) It does not have any impact on the noise as there's no active high impedance circuitry involved in the circuit.
2.) If used as a load you don't want the output jack's tip to be grounded. It won't be a problem as long as all stages are switched in, if you switch one of them off for whatever reason, in place of the 8 ohm speaker there'll be a short between the hot and ground wires which could lead to a lower reflected impedance and thus high current in your amp.
3.) You could ground the line out signal but it's not necessary, there won't be any effect on noise at all.

Regarding 2.) it's not a big issue, however, I would try to keep sources of errors to a minimum. In case you forget to switch on all the subsequent resistive stages and have the output jack grounded (no speaker cable plugged in), your amp will see an impedance of 6 Ohm instead of 8 (simulated with LTSPice for the 8 Ohm Version of the attenuator). That's not too much of an issue, however, it doesn't have any advantage, so I wouldn't do it.

Simply put: Don't switch tips to ground on unused jacks.

Edit: Sorry for the long post, I probably over-discussed this slightly

 

[Mjh36]

I would like to stress that it is probably best not to switch any of the tips to ground when nothing is plugged in.
1.) It does not have any impact on the noise as there's no active high impedance circuitry involved in the circuit.
2.) If used as a load you don't want the output jack's tip to be grounded. It won't be a problem as long as all stages are switched in, if you switch one of them off for whatever reason, in place of the 8 ohm speaker there'll be a short between the hot and ground wires which could lead to a lower reflected impedance and thus high current in your amp.
3.) You could ground the line out signal but it's not necessary, there won't be any effect on noise at all.

Regarding 2.) it's not a big issue, however, I would try to keep sources of errors to a minimum. In case you forget to switch on all the subsequent resistive stages and have the output jack grounded (no speaker cable plugged in), your amp will see an impedance of 6 Ohm instead of 8 (simulated with LTSPice for the 8 Ohm Version of the attenuator). That's not too much of an issue, however, it doesn't have any advantage, so I wouldn't do it.

Simply put: Don't switch tips to ground on unused jacks.

Edit: Sorry for the long post, I probably over-discussed this slightly

Thank you I appreciate every bit of it. Still learning a lot, and I'm in that phase of knowing a little, yet wanting to implement everything in every build. Always feeling I need to build stuff extra bulletproof. So I'll maybe do the switched to ground input jack on the amp build when I start it. But for the attenuator I'll just keep it stock and simple.

 

[iefes]

Yes, on an amps input it is totally recommended. Otherwise you'll have a floating high impedance node picking up all the noise when no cable is plugged in.

Good luck with the build!

 

[dbishopbliss]

The bass resonance that occurs on tube amps is IMO, a quirky side effect that happens due to high output impedance. Its not really fundamental to the character of the tone, and only occurs at a few of the very lowest guitar notes. Its nowhere near as important as the general rise at higher frequency.

I've tried to provoke a recording into demonstrating what might be missing by not having the bass resonance in M or M2 designs, with low drop D playing and pushing the amp, and I just cant find any audible tonal difference compared to full unattenuated volume.

But we do have circuit M3 worked out, which needs another big coil, and particularly a big low-loss bipolar cap. It would double the $ for parts though, and these are bulky too. I'd try without!

Has anyone actually built the M3 design for comparison to the M2? Would a 100uf Motor Run capacitor work? I have one that I used in the power supply of a tube amp I built but its sitting around now.

 

[JohnH]

Has anyone actually built the M3 design for comparison to the M2? Would a 100uf Motor Run capacitor work? I have one that I used in the power supply of a tube amp I built but its sitting around now.

i don't know of any builds of M3, and I haven't done it myself. If i did, Id use a big polypropylene cap, or maybe two in parallel, to get a bipolar cap with very low losses. Those motor caps might work OK though, so long as they are bipolar (I think they usually are) - so try at risk! 100uF would be a good value for a 16 ohm version.

M3 is theoretically better at capturing the bass resonance and showing it to the amp. But in all the testing Ive done without it, Ive not seen or heard any measured or audible issues with not having it. i currently think that it is not essential for capturing the main character of the amp tone and dynamic response. That resonance is more of a quirky outcome of speaker and amp design, which only affects a few of the very lowest notes. In M2, we capture it with the real speaker response, though the amp doesn't see it.

If you try it for science, Id be very interested particularly if you can configure it so the bass circuit can be shunted, which will convert M3 to M2.

 

[dbishopbliss]

i don't know of any builds of M3, and I haven't done it myself. If i did, Id use a big polypropylene cap, or maybe two in parallel, to get a bipolar cap with very low losses. Those motor caps might work OK though, so long as they are bipolar (I think they usually are) - so try at risk! 100uF would be a good value for a 16 ohm version.

M3 is theoretically better at capturing the bass resonance and showing it to the amp. But in all the testing Ive done without it, Ive not seen or heard any measured or audible issues with not having it. i currently think that it is not essential for capturing the main character of the amp tone and dynamic response. That resonance is more of a quirky outcome of speaker and amp design, which only affects a few of the very lowest notes. In M2, we capture it with the real speaker response, though the amp doesn't see it.

If you try it for science, Id be very interested particularly if you can configure it so the bass circuit can be shunted, which will convert M3 to M2.

I am building an attenuator for 8 ohms so I guess that motor run won't work for me. They are not polarized so I guess that makes them bipolar.

I was just looking on Parts Express and saw they have a bipolar 200uF/100V electrolytic cap for $2.35. I'm not sure if it qualifies as having "very low losses" but it is intended for use in crossovers so I suspect it would be appropriate for these purposes. The DF is 5%. You could also get two Audyn 100uF caps for under $33. The DF for these is 0.0003% so I suspect that is "very low". I think I will go ahead and give the electrolytic cap a try since I can get that and the inductor for under $20 shipped. Please confirm that 9mH is the correct value for L2.

Speaking of shipping. I decided to buy my power resistors off of the Ebay vendor mentioned earlier in this thread (zhany-m). I just got the tracking information and it says the estimated delivery is August 15. Does it really take 10 weeks to ship resistors from China to the US? If so, I may have to go ahead and buy the more expensive resistors from Mouser and just build a second one when they arrive.

 

[JohnH]

Yes 200uF and 9mH should be right for the resonance circuit in an 8 Ohm M3. It gives a resonance in circuit at about 120Hz, similar to greenbacks in a 1960A closed cab. Open-back cabs are more like 80 Hz, needing larger values if you want to match them.

In the attenuator, the amp will see this resonant circuit, and so will a line out if you take it from the input. A speaker attached to the attenuator output will do its own thing though, responding with its own natural resonant frequency, as it does in M2.

I honestly dont know about the minimum rating spec on the caps. I looked at this older thread about load boxes, where they use the big poly caps for the resonance:

Aiken's Reactive Dummy Load. | The Gear Page

In the load-box circuits, almost the entire signal is going back and forth through the cap. In our circuit, there'd be somewhere in the range 1.5 to 2A of reversing current through the cap, from a 50W amp input. So the cap will heat up based on the 5% DF factor, which implies a internal series resistance of around 0.5 Ohm in the cap. (as I understand it, Im happy to be corrected by anyone who knows more).This also takes a bit off the resonant peak. But I reckon it will probably work for a careful test, but id suggest to monitor the cap for over-heating.

 

[dbishopbliss]

I went ahead and ordered the 200uF cap and 9mH choke. I also ordered resistors from Mouser. I have a gig coming up and I would like to be able to turn up my amp without overpowering the rest of the band or the audience. I will just make a second and third attenuator when the parts arrive from China. This will allow me to do a side by side comparison to see if I can hear any difference.

 

[musicheals]

Hi,
I'm new at this forum and I came across this thread. I have to say, it is such a pleasure to find so many good information here. Especially about John's attenuator, which I heard only good responses. Of course I am thinking to build one. I want to build the M lite for 8 ohms. In the past I made a air break, which is a good one to my opinion. But I'm sure, John's is much better. So here's my question:
I got a ferrite ring coil with 10mH (1.32mm wire diameter). My idea is, to unwind it, until I can measure 0.5mH. I would like to use the unwinded wire for winding the 0.33mH on the same ferrite core. Is this possible or do I need a seperate air coil?

Thanks a lot and kind regards
musicheals

 

[JohnH]

Hi @musicheals , thanks for your interest in these attenuators.

There's a couple of aspects where what you describe may work, but is outside of what we know does work.

The main coil is designed to be based on an air core, so that it can never saturate. If it does then it could lose its inductance, right at the time where we want it when driving hard. Nothing bad would happen, but it might not sound as good.

Also, with M and M lite designs, which have two coils, putting them on one core will couple them, changing their performance in ways I can't predict.

M-lite was a simple version of M. But you could also just build M2 which can be even simpler and with the same performance, using just one coil. You can leave off the 3rd output jack and associated parts if you don't want it. For an 8Ohm build, the coil is 0.9mH. If you want to try your coil, you could unwind it to that value, and see. You'd probably never know if its different from an air-core one though, since the difference is probably a subtle effect and you wouldn't be able to A/B them.

 

[musicheals]

Thanks a lot John for your quick reply!
So I think it's best, not to use the core. I got a few empty plastic bobbins and I will use the wire of this big coil (10mH) and wind it myself by hand. This will be fun I think

Thanks a lot
musicheals

 

[JohnH]

That's great, very handy if you have a meter to measure inductance. I've got L and C ranges on mine and although its only a $50 unit, it's very useful for checking coils and pickups.

Your wire sounds like a good thickness, maybe 16ga? If you go to places like Madisound.com you can see the type of diameters and final ohms that they get for equivalent coils, then you can check there's enough wire in your donor coil. There's also a few online calculators for multilayer coils, which could help with a check on wire length needed.

 

[musicheals]

....when I look at the gauge/diameter list, it should be gauge 17. Is threre an advantage for having a stronger wire?

Yes, it's very comfortable to have a multimeter that measures L and C. Once I made a coil for a wah wah and it gave a great result.
Next I will order the resistors, if possible from one seller (if I will find one in China maybe)

Thanks a lot
musicheals