-
We are looking to make improvements to the Classifieds! Help us determine what improvements we can make by filling out this classifieds survey. Your feedback is very appreciated and helpful!
Take survey
Simple Attenuators - Design And Testing
- Thread starter JohnH
- Start date
This site may earn a commission from merchant affiliate links like Ebay, Amazon, and others.
tmingle
Well-Known Member
Did you wind your own inductor?
The inductor came from qtasystems.co.uk - great service from them.
paradox419
Active Member
- Joined
- Jan 25, 2012
- Messages
- 123
- Reaction score
- 100
Parts were delivered today. This thing is everything it's claimed to be. My only change is going to be adding a bypass switch. You guys did an awesome job. Thank you.
JohnH
Well-Known Member
Component value substitutions
I've been doing a bit of OCD R&D. The base design, being M2, has component values shown on the diagram taken from standard values supplied by Arcol. If you are pony-ing up for brand-name resistors, particularly since you can probably get them with fast shipping, then these values apply.
The most common ranges for resistors, in general, go 1, 1.2, 1.5, 1.8, 2.2, 2.7, 3.3, 3.9, 4.7, 5.6, 6.8, 8.2, then multiples of 10, 100 etc times those. In the base designs for 8Ohm and 16Ohm builds of M2, I also use 30 and 75, which seem to be also available from Arcol suppliers. No changes are proposed to the base design:
But a lot of us are saving much $ (but with long shipping) buying resistors on ebay from China, and these work fine but generally don't come in quite the same range.
What you can find from a given ebay seller may vary, but a fairly common selection that Ive seen and bought from include:
1, 1.5, 2, 3, 3.5, 5, 6, 8, 10, 12, 15, 18, 20, 25, 27, 28, 30, 33, 50, 75, 80, 100, 120, 150, 200, 250, 300 etc
Here’s the link I look at: (Ebay from Australia)
25/50/100W 0.1-5000 Ohm Watt Shell Power Aluminum Housed Case Wirewound Resistor | eBay
They sell 25, 50 and 100W. Other links will sell similar ranges, but not always the same
So mostly pretty good, but there's a few notable differences for our design, particularly at 4.7, 5.6, 22, 39, 47 and 68 which are hard or not possible to obtain. (have seen the 22 at 25W though)
So how to adapt while keeping the design as optimal as possible? This design is based on keeping very close control of tonal balance and attenuation steps at each switched setting, and to do that, a series of very small decisions are made about optimum component values in each stage. Having worked those out, then we pick the nearest combinations that keep the balance using the ranges available. If we get close, then its close enough. So having decided on the best Ohm value to spec, then a tolerance of 5% is ok (no need to spec 1% tolerance). The ebay resistors that I've bought are within that tolerance.
We need to consider each attenuation stage, the values for the 8 and 16 Ohm builds and maybe also how best to do a 100W build based on the values required.
Stage 1 - Main Reactive Stage - 7db
This is the most important stage that sets everything up and receives the amp signal, absorbing most of it and setting tonal balance.
R1 = readily available in all ranges at 15 Ohm or 30 Ohm for the 8 and 16 builds, at 100W for a 50w amp.
If going for a unit to run with a 100W amp, I suggest to use two of 100W in capacity and spread them apart as mounted on the case.
So for an 8 Ohm 100W build, use two 100W 30 ohm in parallel to get 200W 15 Ohms. If its a 100W build for 16 Ohms, use two 100W 15 Ohm in series to get 200W 30 Ohms.
R2A and R2B for 8 Ohm build
These should be about 20% different in values for tonal balance, its somewhat critical and the base 22 (R2A) and 18 (R2B) is close to ideal. But 22 is usually not available from ebay. If its not, I'm now suggesting to use R2A = 20 and R2B = 18. The result is very close. Mathematically the 22 and 18 is better but you'd likely never hear it. Another recipe that is OK is 25 and 20, at the expense of a 1/2 db.
R2A and R2B for 16 Ohm build
This pair is harder to find on ebay and the standard 47 and 39 per the base spec aren't in the available ranges. There are big holes in the usual ebay ranges from 33 to 50 to 75. My suggestion is to combine two in series. eg R2A = 25+20 = 45 and R2B = 18+20 = 38. Each of these could be half the rating needed for R2A or R2B, ie 25W for a 50W build. This could be neat since they would then match the size of all the other resistors apart from R1.
Stage 2 -7db
Values needed for this are available in all ranges
Stage 3 -14db
This includes a 4.7 Ohm for R5 the 8 ohm build. The perfect value of this turns out to be about 4.8 Ohms, and so instead of 4.7, 5 ohms from ebay is a good choice, almost as close. It corresponds to 10 Ohm as shown for R5 in the 16 ohm versions.
R6 is available as spec'ed in all ranges
Stage 4 -3.5db
For the 8 Ohm build this includes 5.6 Ohm shown for R8 combined with 33 for R7. R8 is ideally around 5.3 so if 5.6 is not available, it is best substituted by 5 from ebay rather than 6.
The 16 Ohm version uses 75 and 12 here, which are available in all ranges.
That covers the main attenuation stages.
Out 3 for the 8 Ohm build into a 16 Ohm cab
R10, is spec'ed at 68 but 75 is OK too.
R11, if used is 10 Ohms so easy to find
Inductor L1
for 8 Ohm build, use 0.9mH, if available, or 0.85mH (1mH is OK)
for 16 Ohm build, use 1.8mH
Wire gage use 18 for 50W (19 gage from Madisound in the US is fine too). Use 16 gage for 100W.
I've been doing a bit of OCD R&D. The base design, being M2, has component values shown on the diagram taken from standard values supplied by Arcol. If you are pony-ing up for brand-name resistors, particularly since you can probably get them with fast shipping, then these values apply.
The most common ranges for resistors, in general, go 1, 1.2, 1.5, 1.8, 2.2, 2.7, 3.3, 3.9, 4.7, 5.6, 6.8, 8.2, then multiples of 10, 100 etc times those. In the base designs for 8Ohm and 16Ohm builds of M2, I also use 30 and 75, which seem to be also available from Arcol suppliers. No changes are proposed to the base design:
But a lot of us are saving much $ (but with long shipping) buying resistors on ebay from China, and these work fine but generally don't come in quite the same range.
What you can find from a given ebay seller may vary, but a fairly common selection that Ive seen and bought from include:
1, 1.5, 2, 3, 3.5, 5, 6, 8, 10, 12, 15, 18, 20, 25, 27, 28, 30, 33, 50, 75, 80, 100, 120, 150, 200, 250, 300 etc
Here’s the link I look at: (Ebay from Australia)
25/50/100W 0.1-5000 Ohm Watt Shell Power Aluminum Housed Case Wirewound Resistor | eBay
They sell 25, 50 and 100W. Other links will sell similar ranges, but not always the same
So mostly pretty good, but there's a few notable differences for our design, particularly at 4.7, 5.6, 22, 39, 47 and 68 which are hard or not possible to obtain. (have seen the 22 at 25W though)
So how to adapt while keeping the design as optimal as possible? This design is based on keeping very close control of tonal balance and attenuation steps at each switched setting, and to do that, a series of very small decisions are made about optimum component values in each stage. Having worked those out, then we pick the nearest combinations that keep the balance using the ranges available. If we get close, then its close enough. So having decided on the best Ohm value to spec, then a tolerance of 5% is ok (no need to spec 1% tolerance). The ebay resistors that I've bought are within that tolerance.
We need to consider each attenuation stage, the values for the 8 and 16 Ohm builds and maybe also how best to do a 100W build based on the values required.
Stage 1 - Main Reactive Stage - 7db
This is the most important stage that sets everything up and receives the amp signal, absorbing most of it and setting tonal balance.
R1 = readily available in all ranges at 15 Ohm or 30 Ohm for the 8 and 16 builds, at 100W for a 50w amp.
If going for a unit to run with a 100W amp, I suggest to use two of 100W in capacity and spread them apart as mounted on the case.
So for an 8 Ohm 100W build, use two 100W 30 ohm in parallel to get 200W 15 Ohms. If its a 100W build for 16 Ohms, use two 100W 15 Ohm in series to get 200W 30 Ohms.
R2A and R2B for 8 Ohm build
These should be about 20% different in values for tonal balance, its somewhat critical and the base 22 (R2A) and 18 (R2B) is close to ideal. But 22 is usually not available from ebay. If its not, I'm now suggesting to use R2A = 20 and R2B = 18. The result is very close. Mathematically the 22 and 18 is better but you'd likely never hear it. Another recipe that is OK is 25 and 20, at the expense of a 1/2 db.
R2A and R2B for 16 Ohm build
This pair is harder to find on ebay and the standard 47 and 39 per the base spec aren't in the available ranges. There are big holes in the usual ebay ranges from 33 to 50 to 75. My suggestion is to combine two in series. eg R2A = 25+20 = 45 and R2B = 18+20 = 38. Each of these could be half the rating needed for R2A or R2B, ie 25W for a 50W build. This could be neat since they would then match the size of all the other resistors apart from R1.
Stage 2 -7db
Values needed for this are available in all ranges
Stage 3 -14db
This includes a 4.7 Ohm for R5 the 8 ohm build. The perfect value of this turns out to be about 4.8 Ohms, and so instead of 4.7, 5 ohms from ebay is a good choice, almost as close. It corresponds to 10 Ohm as shown for R5 in the 16 ohm versions.
R6 is available as spec'ed in all ranges
Stage 4 -3.5db
For the 8 Ohm build this includes 5.6 Ohm shown for R8 combined with 33 for R7. R8 is ideally around 5.3 so if 5.6 is not available, it is best substituted by 5 from ebay rather than 6.
The 16 Ohm version uses 75 and 12 here, which are available in all ranges.
That covers the main attenuation stages.
Out 3 for the 8 Ohm build into a 16 Ohm cab
R10, is spec'ed at 68 but 75 is OK too.
R11, if used is 10 Ohms so easy to find
Inductor L1
for 8 Ohm build, use 0.9mH, if available, or 0.85mH (1mH is OK)
for 16 Ohm build, use 1.8mH
Wire gage use 18 for 50W (19 gage from Madisound in the US is fine too). Use 16 gage for 100W.
Last edited:
tmingle
Well-Known Member
I am considering these for my next build.
https://www.mouser.com/ProductDetai...=sGAEpiMZZMtlubZbdhIBIAkLiGKeVnd8E4X/S9lpCvc=
I just need to think how to best mount them. I am seriously considering making them part of the amps. I now have 2 DSL40's, a 40C & 40CR & have no worries about modding them. Actually, these mods(provide they are done neatly) may actually increase any resale value.
https://www.mouser.com/ProductDetai...=sGAEpiMZZMtlubZbdhIBIAkLiGKeVnd8E4X/S9lpCvc=
I just need to think how to best mount them. I am seriously considering making them part of the amps. I now have 2 DSL40's, a 40C & 40CR & have no worries about modding them. Actually, these mods(provide they are done neatly) may actually increase any resale value.
Gene Ballzz
Well-Known Member
They're gonna have a pretty large footprint!
Just Food 4 Thought?
Gene
Just Food 4 Thought?
Gene
tmingle
Well-Known Member
Yes, they do. I was thinking about laying them out on an aluminum plate & mounting it inside the amp.
I think about a lot of stuff though.
matttornado
Well-Known Member
- Joined
- Mar 14, 2012
- Messages
- 1,900
- Reaction score
- 1,510
Here's my 300 watt / 30 ohm resistor for my 100 watt build. Mounted on an aluminum plate along wit the others.
FYI: I got the large 300 watt R on ebay for about 30 bucks I think?
FYI: I got the large 300 watt R on ebay for about 30 bucks I think?
JohnH
Well-Known Member
I just put up this new M2 diagram version in Post 1, to include the chinese resistor substitutions as discussed above. No changes to the base design.
Marshall Arts
Active Member
JohnH: I'll humbly take my hat off for hours of very interesting read. Thank you!
If you don't mind, I might ask you a thing or two later on as I plan on making my own attenuator soon.
If you don't mind, I might ask you a thing or two later on as I plan on making my own attenuator soon.
JohnH
Well-Known Member
hi @Marshall Arts
Thanks, and you are very welcome.
Thanks, and you are very welcome.
Mjh36
Active Member
- Joined
- Nov 27, 2020
- Messages
- 157
- Reaction score
- 133
Hey I'm about to order supplies for the M2 8 ohm version. I've read this thread a few times and it's fantastic. I have questions because the "switched jacks" concept eludes me still even though I try to read about it. I understand it for switched DC jacks in pedals, but in amps I don't. I don't even know how to phrase questions lol! I did read that you don't need them in this attenuator build, but I'm also about to kit together a 20w amp head...
So in general, can someone give me info or a link on when and why I would use switched jacks in an amp build? I'm trying to understand it in terms of an amp's "speaker out" being "switched" for safety reasons if a cabinet is unplugged so it doesn't hurt your amp. I thought I read that somewhere. For some reason I can't grasp this subject. Where does the signal go when the jack is unplugged and is a "closed circuit" I guess? Thanks for this thread and trying to understand my question!
So in general, can someone give me info or a link on when and why I would use switched jacks in an amp build? I'm trying to understand it in terms of an amp's "speaker out" being "switched" for safety reasons if a cabinet is unplugged so it doesn't hurt your amp. I thought I read that somewhere. For some reason I can't grasp this subject. Where does the signal go when the jack is unplugged and is a "closed circuit" I guess? Thanks for this thread and trying to understand my question!
JohnH
Well-Known Member
Its true that on the main designs on this thread, im not actually using the switches on switched jacks, although some diagrams show them. I do use the ring connection on TRS ie 'stereo' jacks in some cases though, typically to ground something when a mono jack plug is pushed in.
Switched jacks can be useful to create a closed contact when there is no plug. eg, to ground an input when there's nothing connected, to reduce noise, or at the output, connect a dummy load if there is no speaker plugged in. Also. in home and practice gear, a headphone socket uses a switched jack to send the signal on to the speaker, when phones aren't plugged in. In these cases, the switch contacts are part of the same circuit that the plug is connecting to. But there are other switched jacks where pushing in a plug operates a switch that is completely independent, as so can do something on another part of the circuit.
Gene Ballzz
Well-Known Member
Hey Folks,
FWIW, I just got some nice WEICO brand, pre-tinned, 18 gauge wire from AES/CE DIST. After stripping, it can get gently flattened to easily allow two wires into the lug holes on those cheap resistors from chinesiawanoreanam! Pretty nice stuff and not terribly expensive.
https://www.tubesandmore.com/products/wire-weico-18awg-stranded-top-coat-pre-tinned-600v-50-feet
And @Mjh36 , as @JohnH alluded to, there are multiple types of switching jacks, for multiple purposes, some "normally open" and some "normally closed." The CLIFF UK style is pretty straight forward and only available one way (if switched version), but Switchcraft and others offer multiple options. When planning switching jack use, you need to REALLY wrap your head around what is going on, as quite often using one switch contact for one thing can change the operation of one and/or another of the switches. One way to think of a switched CLIFF style jack is is that each pair of contacts is actually two switches. One connection is the plug contacting the spring loaded blade and the other is that spring loaded blade getting disconnected from it's related lug. It can get get a little confusing, at best!
Just Sayin'
Gene
FWIW, I just got some nice WEICO brand, pre-tinned, 18 gauge wire from AES/CE DIST. After stripping, it can get gently flattened to easily allow two wires into the lug holes on those cheap resistors from chinesiawanoreanam! Pretty nice stuff and not terribly expensive.
https://www.tubesandmore.com/products/wire-weico-18awg-stranded-top-coat-pre-tinned-600v-50-feet
And @Mjh36 , as @JohnH alluded to, there are multiple types of switching jacks, for multiple purposes, some "normally open" and some "normally closed." The CLIFF UK style is pretty straight forward and only available one way (if switched version), but Switchcraft and others offer multiple options. When planning switching jack use, you need to REALLY wrap your head around what is going on, as quite often using one switch contact for one thing can change the operation of one and/or another of the switches. One way to think of a switched CLIFF style jack is is that each pair of contacts is actually two switches. One connection is the plug contacting the spring loaded blade and the other is that spring loaded blade getting disconnected from it's related lug. It can get get a little confusing, at best!
Just Sayin'
Gene
dougie tamson
New Member
- Joined
- Jan 16, 2021
- Messages
- 13
- Reaction score
- 10
So finally got the M2 8ohm version built and running some tests, sounding sweet 
will upload some samples to Soundcloud next week.
I used my 3d printed bobbin to wind L1 and used wire gauge for 50W amps. For the 100W capable coil I’ll print a bigger bobbin.
I added an extra heat sink over R1 and also a 12cm fan with an external mains power pack that has 3,5,7.5,9 and 12v dc switch which allows speed control for the fan.
I’ve run it with a 20W JCA22H head, the power to the fan was pulled, not needed, didn’t get warm.
I tried it also with a 4 x 6L6CG 120W amp with 2 valves pulled from the output stage so 60Watt (set the 4/8/16ohm switch to 4 on the amp and attached an 8ohm speaker load. I run this at full volume and full attenuation, got a bit warmer but was ok, I’ll take some temperature readings before and after next test.
Another test to do is to run the big amp with EL34 valves (there is a switch to select either type). This will give 50W with 2 valves pulled or 100W with the full set of 4.
I used a bigger case, 190x190x65cm, to take the fan. There is still room inside to double up the 15R 100W R1 as 2 x 30R in parallel giving 15R @ 200 watts.
I’ll get the new bobbin for L1 with heavier gauge wire done this week.
Doug.
will upload some samples to Soundcloud next week.I used my 3d printed bobbin to wind L1 and used wire gauge for 50W amps. For the 100W capable coil I’ll print a bigger bobbin.
I added an extra heat sink over R1 and also a 12cm fan with an external mains power pack that has 3,5,7.5,9 and 12v dc switch which allows speed control for the fan.
I’ve run it with a 20W JCA22H head, the power to the fan was pulled, not needed, didn’t get warm.
I tried it also with a 4 x 6L6CG 120W amp with 2 valves pulled from the output stage so 60Watt (set the 4/8/16ohm switch to 4 on the amp and attached an 8ohm speaker load. I run this at full volume and full attenuation, got a bit warmer but was ok, I’ll take some temperature readings before and after next test.
Another test to do is to run the big amp with EL34 valves (there is a switch to select either type). This will give 50W with 2 valves pulled or 100W with the full set of 4.
I used a bigger case, 190x190x65cm, to take the fan. There is still room inside to double up the 15R 100W R1 as 2 x 30R in parallel giving 15R @ 200 watts.
I’ll get the new bobbin for L1 with heavier gauge wire done this week.
Doug.
Last edited:
JohnH
Well-Known Member
Hi @dougie tamson , great stuff and thanks for posting. That's a serious looking piece of kit!
On heat and fans, I reckon that with the thick AL box and ventilation then we are ok up to around 50W without fan, and at a 100W one might be warranted depending on how hard it's driven.
On heat and fans, I reckon that with the thick AL box and ventilation then we are ok up to around 50W without fan, and at a 100W one might be warranted depending on how hard it's driven.
dougie tamson
New Member
- Joined
- Jan 16, 2021
- Messages
- 13
- Reaction score
- 10
Big thanks to John for his design.
Another plus for the bigger case was I drilled 2.5mm holes in the case then cut a tread with a M3 tap, I used the screws from the switches I bought, dual pole dual throw, each came with 3mm threaded terminals and 6 x 3mm screws, to fix the resistors to the case. I soldered the dual poles together with wire used for 32 amp UK mains socket runs (also used for all the internal wiring).
I was thinking that any heat build up might throw the resistance values off so went for the belt and braces approach.
Doug.
dougie tamson
New Member
- Joined
- Jan 16, 2021
- Messages
- 13
- Reaction score
- 10
Some sound clips from todays testing:
Strat PU position middle/neck position, middle has magnet poles reversed mod.
Clean channel gain maxed then master volume increasing with successive attenuation, ending up with the amp gain and master vol maxed out.
Strat PU position middle/neck position, middle has magnet poles reversed mod.
Clean channel gain maxed then master volume increasing with successive attenuation, ending up with the amp gain and master vol maxed out.
JohnH
Well-Known Member
Those clips sound great. I like them all, but you can hear how the amp is adding those power stage harmonics in the later clips, but other than that the tone is staying consistent. Its a good demo of why we might want an attenuator. The clean clip is a great example of fully clean too though.
if you're up for doing clips, and have a tolerant family, the other type of clip test is where you record at full volume unattenuated, again at -7db and then at max attenuation, all with amp settings the same. Probably have to set recording settings appropriate for each and then normalise the recording levels downstream. Then the ideal is that all clips sound the same and any differences are just due to the attenuator. BTW, what type of amp are you using? It definitely sounds to be of a design where the power stage can affect the tone and that it is very clearly heard. We haven't had many clip tests like that but there have been a couple.
Its an interesting point about how much do the resistance values vary with temperature. i went looking for data sheets and found this fairly comprehensive set from Arcol:
Microsoft Word - frontpagetemplate.doc (spiratronics.com) But i cant interpret the answer to that question from what I see there. But Im thinking that they must be at least somewhat stable over a reasonable temp range. its not too critical for the design.
I saw a few other things there though. One thing is did see was inductance values on standard resistors, which are typical a few micoHenries (say for a 10 Ohm 100W), so less that 1% of the inductances that are involved in a speaker and so can be ignored. Another interesting factoid was how to derate resistors if you use a heatsink but no heatsink compound. its not as much as Id expected, just 15%, but id still use it.
Of course, whether or not those data are anything like what applies to Chinese resistors is unknown, but im not too worried.
if you're up for doing clips, and have a tolerant family, the other type of clip test is where you record at full volume unattenuated, again at -7db and then at max attenuation, all with amp settings the same. Probably have to set recording settings appropriate for each and then normalise the recording levels downstream. Then the ideal is that all clips sound the same and any differences are just due to the attenuator. BTW, what type of amp are you using? It definitely sounds to be of a design where the power stage can affect the tone and that it is very clearly heard. We haven't had many clip tests like that but there have been a couple.
Its an interesting point about how much do the resistance values vary with temperature. i went looking for data sheets and found this fairly comprehensive set from Arcol:
Microsoft Word - frontpagetemplate.doc (spiratronics.com) But i cant interpret the answer to that question from what I see there. But Im thinking that they must be at least somewhat stable over a reasonable temp range. its not too critical for the design.
I saw a few other things there though. One thing is did see was inductance values on standard resistors, which are typical a few micoHenries (say for a 10 Ohm 100W), so less that 1% of the inductances that are involved in a speaker and so can be ignored. Another interesting factoid was how to derate resistors if you use a heatsink but no heatsink compound. its not as much as Id expected, just 15%, but id still use it.
Of course, whether or not those data are anything like what applies to Chinese resistors is unknown, but im not too worried.
Similar threads
