So, what started out as a quick little writeup about crossovers...turned into a bit more...about 60 pages more.
The DIYRM-A is a Do-It-Yourself Reference Monitor....much along the lines of an LS3/5a. Can be used a bit more broadly and works better with modern music.
The writeup can be found at:
goo.gl/XboqRW
The document covers:
Introduction
Design Goals
Driver Selection
Enclosure Design and Construction
Design Measurements
Crossover Design
Listening Notes
Crossover Power Handling Testing
Conclusion
They are excellent for a variety of applications from near-field monitoring, two channel stereo where space is a concern and surround sound applications, again where space prevents the use of larger designs.
I hope you enjoy.
Scott "SpeakerScott" Hinson
The DIYRM-A is a Do-It-Yourself Reference Monitor....much along the lines of an LS3/5a. Can be used a bit more broadly and works better with modern music.
The writeup can be found at:
goo.gl/XboqRW
The document covers:
Introduction
Design Goals
Driver Selection
Enclosure Design and Construction
Design Measurements
Crossover Design
Listening Notes
Crossover Power Handling Testing
Conclusion
They are excellent for a variety of applications from near-field monitoring, two channel stereo where space is a concern and surround sound applications, again where space prevents the use of larger designs.
I hope you enjoy.
Scott "SpeakerScott" Hinson
Last edited:
So I got into a bit of a debate/question answer session over PM regarding the number of components in the crossover. One of the concerns was that there was no way that a crossover with this many components would be "linear".
By definition a system with linear components, no matter how many there are is still a linear system. This is a pretty fundamental tenant of electrical engineering rigorously supported by a bunch of mathematics I vaguely remember knowing at one point in my life.
Physical components will always have some amount of non-linearity caused by a junction of dissimilar metals, hysteresis of materials etc. In the DIYRM-A crossover the single most suspect component is the 1.5mH laminated core inductor. The lamination will suffer from both hysteresis and saturation resulting in some amount of distortion.
I can't actually remember seeing if someone has ever tested an iron-core inductor for distortion under multi-tone conditions, so I thought I'd give it a shot.
*NOTE: I did this test using the FFT function of an oscilloscope...which has either an 8-bit or 10-bit front end...I will need to re-do this test with a device with enough resolution to actually measure the distortion induced, since this setup did not have enough capability to actually measure the distortion. That said...(spoiler alert) the distortion was lower than I expected.
So I set up a circuit with a 5mH ERSE air core inductor from the same line as the 1.5mH used in the DIYRM-A, in series with a 4 ohm resistor. I used my Carver TFM-25 amplifier to put a two-tone test signal, total signal amplitude of 17.5VRMS across the resistor of 20Hz and 200Hz, equal magnitude.
The 17.5VRMS across the resistor equates to 4.375ARMS. Into an 8 ohm load that current equals 153 Watts continuous power.
That's a lot for a speaker intended for nearfield use.
No...that's a $%#@$ lot of power for a speaker intended for nearfield use. Well more than you should ever put into a 6" woofer you're sitting within 1-10 of.
Under test I could feel the inductor core vibrating due to magnetostriction, or the property of magnetic materials to change shape under a strong magnetic field.
No single harmonic of the 200Hz or 20 Hz signal came within 40dB of the applied signals. 40dB down is 1% distortion...please trust that at these signal levels the woofer will show well higher than 1% distortion...as will nearly every 6" woofer.
So...in summary, I have even less concerns regarding using a iron core inductor than I did before. To get the equivalent series resistance in an air core part would require a large inductor that would cut into the interior volume of the speaker and be tough to fit on that back panel.
Attached picture of the oscilloscope screen during testing.
By definition a system with linear components, no matter how many there are is still a linear system. This is a pretty fundamental tenant of electrical engineering rigorously supported by a bunch of mathematics I vaguely remember knowing at one point in my life.
Physical components will always have some amount of non-linearity caused by a junction of dissimilar metals, hysteresis of materials etc. In the DIYRM-A crossover the single most suspect component is the 1.5mH laminated core inductor. The lamination will suffer from both hysteresis and saturation resulting in some amount of distortion.
I can't actually remember seeing if someone has ever tested an iron-core inductor for distortion under multi-tone conditions, so I thought I'd give it a shot.
*NOTE: I did this test using the FFT function of an oscilloscope...which has either an 8-bit or 10-bit front end...I will need to re-do this test with a device with enough resolution to actually measure the distortion induced, since this setup did not have enough capability to actually measure the distortion. That said...(spoiler alert) the distortion was lower than I expected.
So I set up a circuit with a 5mH ERSE air core inductor from the same line as the 1.5mH used in the DIYRM-A, in series with a 4 ohm resistor. I used my Carver TFM-25 amplifier to put a two-tone test signal, total signal amplitude of 17.5VRMS across the resistor of 20Hz and 200Hz, equal magnitude.
The 17.5VRMS across the resistor equates to 4.375ARMS. Into an 8 ohm load that current equals 153 Watts continuous power.
That's a lot for a speaker intended for nearfield use.
No...that's a $%#@$ lot of power for a speaker intended for nearfield use. Well more than you should ever put into a 6" woofer you're sitting within 1-10 of.
Under test I could feel the inductor core vibrating due to magnetostriction, or the property of magnetic materials to change shape under a strong magnetic field.
No single harmonic of the 200Hz or 20 Hz signal came within 40dB of the applied signals. 40dB down is 1% distortion...please trust that at these signal levels the woofer will show well higher than 1% distortion...as will nearly every 6" woofer.
So...in summary, I have even less concerns regarding using a iron core inductor than I did before. To get the equivalent series resistance in an air core part would require a large inductor that would cut into the interior volume of the speaker and be tough to fit on that back panel.
Attached picture of the oscilloscope screen during testing.
Attachments
Way back in 1988 Dutch mag Elektuur published distortion tests on core inductors. Then modern developments such as Corobar and Ferrobar cores from German distributor Intertechnik showed that distortion with these materials was no issue (<0,4%) with currents related up to 1000W on 8Ω. Nor with laminated cores by the way. All loudspeakers show far bigger distortion figures under such powers.
Is that article online? I'd love to see it, I did a few quick Google searches and came up empty.
Scott
JenswoldJ,
Do you mind if I answer the questions publicly? I think many people might be interested in the answers....
Scott
Very nice writeup! Thanks for putting this design together - I really like Dayton RS series drivers as they represent some great performance and value.
I am surprised how well the foam diffraction control works - speaker is not to be used without the grill then.
I am surprised how well the foam diffraction control works - speaker is not to be used without the grill then.
Dense wool felt would probably work even better, but that would have been another line item adding cost, shipping expense and a third vendor. I really wanted to keep this speaker to a two store project.
And yes...this is one of the few speakers designed from the start to be listened to with a grill on it, all frequency response measurements and crossover voicing was done with the grill on the the speaker.
Scott
I got the following questions about the speaker/design, thought I would answer them publicly since they are good questions:
Okay...so as far as the crossover questions. I apologize for the confusion...but there are two schematics for the crossover, one in Figure 30, one in Figure 48.
Figure 30 is the crossover schematic as I actually built it. Figure 48 is the schematic if all of the paralleled-series components are reduced to a single part. You'll see three resistors in the location you mention in Figure 38.
The reason I used the parts I did was two fold. First (and minor) is cost...you can buy a 10 pack of 1watt or 2watt resistors for less cost of a single 10 watt non-inductive crossover resistor at parts express. Since the highest value of non-inductive 10 watt resistors is 82 ohms, you would have needed two per speaker to get to the right value (20 ohm + 82 ohm). That's ~$2.80 per speaker...where the paralleled three 1 watt resistor cost $.43 per speaker. A minor savings but a savings.
The second, and more important reason, space. This is a crossover with a pretty good number of parts, all of which are required to achieve the phase response, tonal balance and frequency response shaping required by the drivers in this application. It's also a pretty tiny speaker, the crossover barely fits on that back panel as it is...you have to follow my construction placement closely to get it to fit. You might actually have trouble getting it to fit if the 10 watt resistors
As to the sound...I have not heard the Continuums, so I can't comment on them. I have great respect for Jeff's crossover chops, and the driver compliment is pretty good. If the LS3/5a sound is what you are specifically chasing then go Continuum, as I have heard they are a very good facsimile with improvements (due to the improved driver technology and excellent crossover work) in detail and resolution.
As much as I love the LS3/5a I find it loses it on EDM, electric guitar and synth especially as the volume goes up, even if you high-pass to keep out low bass notes. Your mileage may vary, but I believe I've been able to take out most of that while still retaining the detail in classical, jazz and acoustic music that make so many people love them so much. You still need to watch the low frequency excursion (can't change physics) but the speaker was specifically voiced with a wide variety of musical choices.
I can't/won't tell you which ones to build, I'm sure you'll be happy either way, since I believe both are high quality designs.
Scott
Okay...so as far as the crossover questions. I apologize for the confusion...but there are two schematics for the crossover, one in Figure 30, one in Figure 48.
Figure 30 is the crossover schematic as I actually built it. Figure 48 is the schematic if all of the paralleled-series components are reduced to a single part. You'll see three resistors in the location you mention in Figure 38.
The reason I used the parts I did was two fold. First (and minor) is cost...you can buy a 10 pack of 1watt or 2watt resistors for less cost of a single 10 watt non-inductive crossover resistor at parts express. Since the highest value of non-inductive 10 watt resistors is 82 ohms, you would have needed two per speaker to get to the right value (20 ohm + 82 ohm). That's ~$2.80 per speaker...where the paralleled three 1 watt resistor cost $.43 per speaker. A minor savings but a savings.
The second, and more important reason, space. This is a crossover with a pretty good number of parts, all of which are required to achieve the phase response, tonal balance and frequency response shaping required by the drivers in this application. It's also a pretty tiny speaker, the crossover barely fits on that back panel as it is...you have to follow my construction placement closely to get it to fit. You might actually have trouble getting it to fit if the 10 watt resistors
As to the sound...I have not heard the Continuums, so I can't comment on them. I have great respect for Jeff's crossover chops, and the driver compliment is pretty good. If the LS3/5a sound is what you are specifically chasing then go Continuum, as I have heard they are a very good facsimile with improvements (due to the improved driver technology and excellent crossover work) in detail and resolution.
As much as I love the LS3/5a I find it loses it on EDM, electric guitar and synth especially as the volume goes up, even if you high-pass to keep out low bass notes. Your mileage may vary, but I believe I've been able to take out most of that while still retaining the detail in classical, jazz and acoustic music that make so many people love them so much. You still need to watch the low frequency excursion (can't change physics) but the speaker was specifically voiced with a wide variety of musical choices.
I can't/won't tell you which ones to build, I'm sure you'll be happy either way, since I believe both are high quality designs.
Scott
Last edited:
Scott,
Thank you for sharing your hard work and took the time to write down your thought process for us.
Your design reminds me of the now discontinued "Dayton Audio RS621 Speaker Kit" (PE 302-963). The DA kit used the aluminum version of the tweeter and woofer from the same RS series. The Continuums also uses the RS28A-4. Your thought on selecting the soft dome RS28F-4?
I have been looking to build a mini-monitor for the computer, but was not able to decide on the Continuums because of the lack of design information, particularly the crossover. The DIYRM-A may be the one I will end up with.
Thank you for sharing your hard work and took the time to write down your thought process for us.
Your design reminds me of the now discontinued "Dayton Audio RS621 Speaker Kit" (PE 302-963). The DA kit used the aluminum version of the tweeter and woofer from the same RS series. The Continuums also uses the RS28A-4. Your thought on selecting the soft dome RS28F-4?
I have been looking to build a mini-monitor for the computer, but was not able to decide on the Continuums because of the lack of design information, particularly the crossover. The DIYRM-A may be the one I will end up with.
Sorry for taking so long to get back to you....I'm glad you liked the writeup and design. I think you'll be VERY happy with them in the computer application. I've listened to them at distances from ~18" to about 48" and been ecstatic with the sound. In a large room, at longer distances they still sound great but lose a little of the magic since they are struggling to fill the space with sound.
Scott
I have been looking for a direct comparison of the DIYRM-A vs Continuum, but couldn't find any. It is unlikely that I can hear both any time soon. From the response curves and your description, I saw that you gave up efficiency for a more extended bass. It may make the DIYRM-A a better fit for my cramped computer room. The detailed write-up is another important factor to me. I fully trust Jeff Bayby who helped me a lot in my current project, but I still do not like the idea of building a black box.
However, cost is not the issue. I add all the components at Parts Express shopping cart. The total cost is within $5 of the Continuum bare bone kit at Meniscus ($311).
Continuum, Pr
I can get the 1/2" Baltic Birch Plywood in 24"x30" sheet at Rockler. They have a local location. Price seems reasonable.
Baltic Birch Plywood-24 Inch X 30 Inch - Rockler Woodworking Tools
Now, only if I can convince my wife that I need to spend $400 in parts to build another wonderful sounding mini-bookshelf............
I am at the final stage of finishing my main system bookshelfs (for replacement of a deterionated Barry Hughes DIY Bookshelf due to old age, HiFi Answers, July 1978). They sound wonderful so far. The drivers are Peerless SDS 830875 and RS28A-4.
Last edited:
I'd love to hear a set of Continuum's myself (anyone know of a set in the Austin, TX metroplex? I'll bring my DIYRM-A and a bottle of wine for chance to compare....)
As far as the plywood I didn't use baltic birch, though the cabinet will only be improved by it's use. (Sometimes folks get antsy about material changes...in this case I'm not at all concerned.) I used the 5 layer project panels pre-cut to 2' x 4' at Home Depot.
You are absolutely correct I traded off efficiency for extension, there were multiple ways to do it and the least expensive was the addition of mass to the cone. Any other way (adding electronic equalization either active or passive) was much more expensive than a few inches of 12 gauge copper wire and a bit of paint/glue.
I actually used some EDM in the crossover voicing, and even with the limited bass output from the sealed box and small(ish) woofer, the results are quite pleasing as long as you don't expect concert levels in a large room.
Scott
When modeling a speaker with a program that uses TS parameters, is it as simple as increasing Mms value to simulate adding mass to the cone? One should see inproved bass extension and reduction in sensitivity?
In order to see the effect of increased moving mass, one has
to lower the Fs. The formula says that Fs depends solely on
Cms and Mms.
Fs=(0,5/pi)*sqrt((1/(Cms*Mms)); Cms(m/N); Mms(kg)
to lower the Fs. The formula says that Fs depends solely on
Cms and Mms.
Fs=(0,5/pi)*sqrt((1/(Cms*Mms)); Cms(m/N); Mms(kg)
Not really, if you enter the wrong MMS into some programs they'll dumbly calculate a response prediction even if the parameters are not physically realizable.
I have a spreadsheet in the dropbox folder that re-calculates the T/S parameters for added mass.
Fs, Qes, Qms and Qts and sensitivity change as Mms changes. Cms (and VAS) is a result of the drivers suspension which doesn't change when you add mass.
Adding mass at the apex of the voice coil/cone joint also does not change the frequency response of the driver except for the sensitivity shift. It's critical to get it right at the voice coil, otherwise you'll change the stiffness and breakup modes of the cone.
Again, the spreadsheet changes everything for you to figure out the resulting parameters after adding mass.
Scott
Many, many moons ago, I did something similar (adding mass) using a different approach. I built a pair of Bailey transmission line speakers (the non-resonant line with long fiber wool stuffing). I used the Phillips AD10100 10" woofer. When I received the shipment, the units turned out to have much higher resonance frequency Fs than the spec, 41 Hz instead of 35 Hz. I mixed some high temperature epoxy (moisture resistant) and started coating the woofer paper cone with thin layers of epoxy until I got the 35 Hz Fs. It worked very well for me both sonically and mechanically.
For subwoofer use where you're counting on the driver to act as a piston this is perfectly reasonable. I couldn't use this method with the DIYRM-A since I couldn't be sure that everyone's units would turn out the same.
Scott
Correct, which is easily determined by measuring impedance
and using standard formulae for calculating parameters.