First speaker build, feedback request for a low power tubeamp friendly build

[Basenboa]

First post, and first speaker project.. 🙂

I've been reading an learning a lot over the last past weeks about speakers and XO design and time has come for a first speaker project.

So, here's the story:
I bought a pair of Dual CL140 speakers in a second hand store some time ago because I liked the 70s look of them..
came home, hooked them up to my Marantz 2275 remote speaker output, they sounded ok but very little bass. A few weeks later my flatmate had a party while I was away for the weekend and her boyfriend turned up the volume knob and... blew up the speakers. Came home, apologies were made and accepted, he payed me back the 40eur/50usd that they costed me and I ended up with nice looking bookshelf cabinets with broken woofers laying around.
The cabinets have an internal volume of 15,35L (0.54cu.ft) and the woofers cut out hole is 175mm/7".

Now, I decided to use these cabinets to build the speakers for a chinese el34 DIY tube amp kit thats on my way atm, similar to this one:

https://www.diyaudio.com/forums/tubes-valves/182989-building-chinese-se-el34b-amp-diy-kit.html

I allready ordered it before reading here at DIYaudio that it is actually a cheap piece of junk, but since I ordered it for the fun of learnig and tweaking it in the first place it should meet its purpose.
Anyway, in the end it should serve daily to play music in the dining room/ kitchen.
Thats is a square 60m3/2100ft3 room.

The amp is stated to have a 13W*2 output, 4ohm or 8ohm speaker connections.

If I did my homework right it should pair well with a pair of speakers with a fairly high sensitivity (around 90dB/W), impedance drops should be avoided and the curve should be more or less flat.. ??
At first i searched for DIY speaker kits that matched needs, then decided it would be more fun and I would learn a lot more by starting my own project.

The speaker built should meet these requirements:

- Stick to the CL140 cabinets (15,35L internal volume)
- Woofer size should be 7,5" or 8" if possible (the cutout hole now is 175mm/7" and making the hole bigger is easier than smaller)
- Sensitivity of 90db/W (the more the better)
- Bass should go low enough
- Impedance friendly for the vacuum tube amp.

I pretty soon found out that its not easy to find a solution for every of the above requirements and that the solution might be found in a compromise, so this is what I came up with:
I used Dayton Audio drivers, just because they have FRD and ZMA files easily available that I can load into VituxCAD

RST28F 4ohm tweeter
DSA215 8ohm woofer (I tried many many woofers and got the best results with this one... low bass/flat FR/high sensitivity is not an easy combo)
A ported 15.35L enclosure with diameter 40mm, 140mm long port should give me an acceptable f3 of 56,6 Hz/ f6 43,6 Hz/ Fb 38,0 Hz

What are your thoughts, feedback of what I've got so far?
All comments, advice are very much appreciated.


Am I on the right track here?
Is the impedance curve acceptable? because I just cant get it flatter without messing up the frequency range.
Is that bump around 900Hz something to be worried about?
The overall sensitivity is around 89dB, will this be enough for the amp?





I attached the curves pics..
3rd pic is the enclosure SPL with the XO applied to it.

 

[giralfino]

- Stick to the CL140 cabinets (15,35L internal volume)
- Woofer size should be 7,5" or 8" if possible (the cutout hole now is 175mm/7" and making the hole bigger is easier than smaller)
- Sensitivity of 90db/W (the more the better)
- Bass should go low enough
- Impedance friendly for the vacuum tube amp.

One thing that is not really obvious to novices is that loudspeakers and drivers design is and exercise in compromises. For example you can have 2 of the following characteristics, low bass, high sensitivity and minimal volume, but not all three together. You have to choose what is least important for your and leave it.

Another thing that maybe you missed is that all drivers in a finite baffle exhibit a so called baffle step effect, that works in having gradually all frequencies under one frequency dependent on the baffle width reduced in level by 6dB. So if you work with a driver with a nominal 90dB sensitivity in free space, your speaker will have bass at 84dB. In real world situations the baffle step compensation can be only 3-4dB, so your overall speaker SPL can be in the range of 86-87dB, but not more. Your design omits the baffle step compensation (BSC), and you can do this way only if you put the driver in an infinite baffle.

Tied to the baffle step effect there is diffraction, another effect that depends on the baffle geometry. So if you use a finite baffle, you can't simply import frd files generated in free space into a crossover simulator. You have to manipulate them before in order to simulate diffraction and baffle step effects. VituixCAD has a tool to simulate them, but when you use it you also need to have VituixCAD at least calculate a new minimum phase. Last thing to do is to have an appropriate delay between drivers as the acoustical centers of the drivers won't be in the same point.

A last note on your design, impedance is not tube friendly. You need a parallel LCR filter before the crossover to have a really flat impedance. Have a look here for what I mean: Zaph|Audio - ZA-SR71

I think you can have something working out of the drivers chosen, but it won't easy as you think, and the result maybe won't meet your objectives. I still think that the better return on investment is a second hand speaker or a kit that can suit your boxes. You haven't stated the box dimensions, so I can't really suggest something, but as an example I would happily buy this kit, but it requires a larger box than your (I'm not affiliated in any way with the designer/seller): Suzie "Q" | Scan-Speak Discovery compact speaker

Ralf

 

[system7]

Hmm, certainly some compromises needed here as Ralf says.

Homework:

High Efficiency Speakers

Flat Impedance and Flat Power response design.

I used Jeff Bagby's values a lot.

Arpeggio Loudspeaker - diyAudio

I did find some stuff on the cabinets:

DUAL CL-140

15L, 25cm wide. Not untypical of a lot of 8" bass designs.

This is how the Visaton modelling turned out. Below. I show the effect of impedance correction, and the surprisingly improved response with a reflex driver, closed box and 6 ohm SET amp. Reflex, I would avoid.

 

[AllenB]

You haven't mentioned the output impedance of the amp, it may be small?

VituixCAD has a tool to simulate them, but when you use it you also need to have VituixCAD at least calculate a new minimum phase.

If you simulate diffraction, why would you undo it? Baffle diffraction is not minimum phase.

 

[waxx]

In general, high sensivity on low power needs a big speaker, especially if you want a strong bass response. And baffle step is an issue, so i would also go higher in bass response. In reality you loose about 4dB by diffraction, so you need a woofer that gives +4dB of the wanted sensivity. So i think you try to do the impossible breaking of the hoffman iron law. You need a bigger driver and box for what you want or use lower sensivity speakers so lower maximum volume. Choose your compromise.

And to be honest, i run now 88dB sensitive speakers from an 8w amp in a room of 3.5m x6.5m x2.4m H and it goes loud enough for all but home parties. I even don't use an active preamp for the moment (as it's still under construction), only a passive one (so only a volume knob and switch for source selection. So depending on how loud you want it, you can get away with fairly lower sensivity on that amp, certainly when you have a good active preamp that drives it at normal levels (2V input).

 

[frangus]

This is a good design for what you want: Suzie "Q" | Scan-Speak Discovery compact speaker

I have tried similar with a push pull EL34b amp and it was more than sufficient for my needs at the time and I enjoyed listening to it before my amp blew up. May fix it one day.

 

[bentoronto]

Much that's weird in this thread.

OPs project is driven by cabinets chosen for that great 70s look and by a single-sided tube amp that is sonic junk.

So starting with these "odd" constraints, folks are trying to derive fine engineering solutions.

B.

 

[system7]

Question mark over the valve amp. Is it SET or a feedback design? If it is SET, impedance is fixable.

I seem to remember I had a very good time with an Eminence Alpha-8A in a box like this... a real loud speaker! Awesome, in fact!

https://www.eminence.com/speakers/speaker-detail/?model=Alpha_8A

Think I ended up with a port stuffed with straws which is the famous aperiodic bass loading.

Tweeter was an Audax TW034, mainly because it goes loud enough to keep up with a 93dB woofer:

TW034

Another candidate is the Peerless/Vifa/Tymphany DX25TG59-04. Used in the 2Pi speaker on a mere 10uF capacitor with a 15R shunt. The woofer gets a 0.68 to 1mH coil at its simplest.

Peerless by Tymphany DX25TG59-04 1" Fabric Dome Tweeter

Worth consideration are the Monacor DT-300 and Morel CAT378 too, IMO.

If it doesn't work in that small cabinet, you still have decent drivers for a bigger box down the road. But in a small room, big bass is overrated.

This link may help:
What kind of Tweeters that match paper woofers?

 

[Basenboa]

First of all, thanks a lot for all the feedback.. it gave me a lot to read, dive into and to learn to understand. I really appreciate it.

I came up with something different now, an MTM design.. I'll have to CNC out a new front for it but luckily I know a guy with an amateur CNC machine.

So, do you guys think this is getting somewhere?
I do am a bit afraid about the reliability of the simulation software, it definately is an amazing tool, but I really hope its not to far of..

Tied to the baffle step effect there is diffraction, another effect that depends on the baffle geometry. So if you use a finite baffle, you can't simply import frd files generated in free space into a crossover simulator. You have to manipulate them before in order to simulate diffraction and baffle step effects. VituixCAD has a tool to simulate them, but when you use it you also need to have VituixCAD at least calculate a new minimum phase. Last thing to do is to have an appropriate delay between drivers as the acoustical centers of the drivers won't be in the same point.

- I now simulated baffle step and difraction, exported that FRD file and came up with a different crossovers taking into account the full -6db baffle step loss. The lower end is reasonably flat, but I think with the speakers positioned near to, or even against the wall I should end up with a slight downslope in the 100Hz->1000Hz range.

- I hope that I have a better basis now.. the idea is to play around with the woofers XO bypass resistor and tweeter series resistor values to tweak the frequency curve a bit once built. (simulation showed that changing these values have very littel impact on the impedance curve)

A last note on your design, impedance is not tube friendly. You need a parallel LCR filter before the crossover to have a really flat impedance. Have a look here for what I mean: Zaph|Audio - ZA-SR71

- This is what I really struggled the most with, keeping the impedance up together with the overall efficiency drove me nuts.. I think/hope this drivers/XO combo should be OK for the 4Ohm taps of the amp with EL34 tubes.

You haven't mentioned the output impedance of the amp, it may be small?

- Couldn't find it anywhere online.. its a chinese DIY kit, very little info to find about it

In general, high sensivity on low power needs a big speaker, especially if you want a strong bass response. And baffle step is an issue, so i would also go higher in bass response. In reality you loose about 4dB by diffraction, so you need a woofer that gives +4dB of the wanted sensivity. So i think you try to do the impossible breaking of the hoffman iron law. You need a bigger driver and box for what you want or use lower sensivity speakers so lower maximum volume. Choose your compromise.

- I think this new MTM design gives me the desired sensitivity of around 90db and also the possibitity to play around with bass exension. My idea is to make 2 back panels in mdf, a closed one (a rectangle in mdf is quickly made) and a second one that alows me to try ported with different lengths. Simulation shows that I should get fairly low F3 of 56Hz ported, which would be amazing, but I might prefer tighter bass... its just nice to keep the options open.

This is a good design for what you want: Suzie "Q" | Scan-Speak Discovery compact speaker

- Thanks for the suggestion 🙂, but my goal is really "learning by doing" here. Thats why I prefer the approach to start from scratch, bump into problems, learn, and find sollutions. I understand that for many here at DIYaudio the problems I faced so far with the first design is really basic stuff.. but for me its a new world that Im getting to understand now. Buying a speaker DIY kit would definately be easier, probably cheaper, and maybe sound better (I dont really want to know 🙂) but I wouldnt be forced to learn..

OPs project is driven by cabinets chosen for that great 70s look and by a single-sided tube amp that is sonic junk.

So starting with these "odd" constraints, folks are trying to derive fine engineering solutions.

- I actually wanted to build a pair of speakers for a while now, the broken Dual speakers and the DIY tube amp kit it ordered for fun gave me a reason why not to just pick up a diy kit. The amp might change one day and efficient bookshelves will allways come in handy.

Question mark over the valve amp. Is it SET or a feedback design? If it is SET, impedance is fixable.

Ok, this is new to me.. more to read into. The circuit diagram that comes with the amp has negative feedback, but I came across a few threats here with improved circuits and my intention is to play around with it.. I do know the amp does not come with a 43% tap on the OT so I can forget UL mode.. (attached the drawing)

I kinda hoped that the speakers could be somewhat "universal" i mean.. 4Ohm speakers, connect to 4ohm amp tap and dont need to worry.. is there something that I'm missing??

I seem to remember I had a very good time with an Eminence Alpha-8A in a box like this... a real loud speaker! Awesome, in fact!

https://www.eminence.com/speakers/sp...model=Alpha_8A

- That driver does look interesting, but I couldnt find FRD/ZMA files anywhere..

 

[Basenboa]

Good evening,
I'm about to order the parts unless I'm missing something here with this design..

I Would really appreciate if somebody could have a quick last look at it before spending my 💶 💶 💶
Thanks

 

[AllenB]

Do you have spare resistors in case you want the treble reduced?

 

[Basenboa]

Do you have spare resistors in case you want the treble reduced?

I was going to order a few different ones for fine-tuning, also for the woofer its XO bypass resistor.