Measuring "The Metronome" ML TQWT with Dayton Audio UMM-6 measuring microphone using

[Hari Iyer]

https://i.imgur.com/nwSxwsr.jpg

Finally received the much awated Dayton Audio UMM-6 measuring microphone and could not wait to lay my hands over it. I first started by measuring the existing design before working on any corrections.

https://i.imgur.com/p0L04mW.jpg

As you can see from the FR curve there is a deep notch centered at around 10KHz and there was no major support received from the tweeter contributing to the FR balance. Needless to say this needs immediate attention and correction. Also this shows my inability to tune by ears only. Blessed are the ones who can understand FR anomolies in a speaker only by listening. When Prem and Sumanta had come over in the month of June, this is what they listened.

I quickly measured the impedance of the existing design and made some notes for correction and later performed some simulation to arrive at the final design. I first tried by directly connecting the Fostex without any compensation network and this is what i got.

https://i.imgur.com/DXuplbn.jpg

There was too much of mid-range energy as can be seen above that contributed to the midrange shout and needed some compensation for a more balanced sound. I added a BSC and played around with different values of capacitor for the tweeters to get the right balance of the mids and highs. Initial values caused some dynamic compression of the midranges and there was some listening fatigue over time.

I had to adjust for the final value of the capacitor more by ear and measurement rather than by simulation. The midrange now is less compressed, but i would like to be even more better than this which could be possible with more effort. The final FR is now this,
https://i.imgur.com/wrRWTEu.jpg

The more recent speakers which i heard which had these uncompressed midrange dynamics were the one owned by FM Bhagwan (Artos) and if my memory serves me correct then mpw (Merlin TSM). I am unable to recollect from my memory any other speakers which had this smooth uncompressed midrange dynamics due to time passage.

Here are the links to some measurements to feast your eyes.

https://i.imgur.com/FoG1Wzn.jpg

https://i.imgur.com/j7UHshp.jpg

https://i.imgur.com/OgcWyOP.jpg

https://i.imgur.com/xnDT3Wx.jpg

https://i.imgur.com/bAehJs9.jpg

 

[Hari Iyer]

Re: Measuring "The Metronome" ML TQWT with Dayton Audio UMM-6 measuring microphone us

Thanks to share your perspective. Usually no one measurements can show you the entire picture of how the speakers would sound. Hence you are required to measure all parameters and correlate each one of them. I in particular give importance to FR, Minimum phase, step and impulse, waterfall, CSD , Decay and spectrogram. All of them needs to be correlated for arriving at any valuable conclusion. Moreover none of these objective measurements can be synced with any subjective perceptions. For example.g. how could you correlate uncompressed midrange dynamics with any measurements. None exist. Having said that, it's very rare to find a very good speaker that sounds excellent but measures bad imo.

 

[kapvin]

Re: Measuring "The Metronome" ML TQWT with Dayton Audio UMM-6 measuring microphone us

Thanks to share your perspective. Usually no one measurements can show you the entire picture of how the speakers would sound. Hence you are required to measure all parameters and correlate each one of them. I in particular give importance to FR, Minimum phase, step and impulse, waterfall, CSD , Decay and spectrogram. All of them needs to be correlated for arriving at any valuable conclusion. Moreover none of these objective measurements can be synced with any subjective perceptions. For example.g. how could you correlate uncompressed midrange dynamics with any measurements. None exist. Having said that, it's very rare to find a very good speaker that sounds excellent but measures bad imo.

hello Hari,

nice to see you taking measurements. some thoughts. based on graph 2

1. that's a 10db scale so it is compressing some of your variations on y-axis. try changing the scale to 3dB to understand a bit more on the tonal balance of the speaker.

2. Is this an in-room response or are you gating for first arrival? the peak at 75hz would suggest the former. in which case you might find the mid is a little honky in the 2-4khz range, plus baffle step is undercompensated.

3. is the 75hz peak because of room resonance? does it change when you move the speaker / mike. else you may need to revisit tuning on the transmission line.

it's wonderful that you are measuring your speakers. though the social impact at home is high (making sure everything is quiet in a family home can be challenging) the returns are massive.

best wishes

 

[Hari Iyer]

Re: Measuring "The Metronome" ML TQWT with Dayton Audio UMM-6 measuring microphone us

Hi Kapil,
Thanks for posting after a long time. Its been a year when we shared some simulations for the Metronome using the mathcad.

1. that's a 10db scale so it is compressing some of your variations on y-axis. try changing the scale to 3dB to understand a bit more on the tonal balance of the speaker.

I have taken just a prototype measurement and am yet to finalize the design. I analyze using 3dB scale and post on 10dB scale else viewers will be shocked seeing the response and will create unnecessary impressions. Typically speaker marketing will post even higher 15dB scale so that their response look flat

2. Is this an in-room response or are you gating for first arrival? the peak at 75hz would suggest the former. in which case you might find the mid is a little honky in the 2-4khz range, plus baffle step is undercompensated.

Yes you are right, this is in-room measurement at 1meter mic distance on the tweeter axis. I have not gated the response and the 74Hz peak is the room mode which i get for all my other speakers too. Converting this to wavelength gives me exactly 15.2 feet which is the distance between the front and the rear walls on the longer walls. Only room treatment using thick absorbent or bass traps can resolve the issue which is ruled out due to wAF. I would have to either live with it or try some other solution like using a sub.

3. is the 75hz peak because of room resonance? does it change when you move the speaker / mike. else you may need to revisit tuning on the transmission line.

This is purely room mode and has nothing to do with the speaker/ mic position (though i have yet to measure them at my listening position). The TL has a reasonably medium stuffing till 2/3rd of the line and the scope of boom from the port would be remote. But the impedance curve do show the second peak around 80Hz though and the first peak is at 30Hz. The peaks are also well damped without any observed resonance in the impedance curve.

I am planning some more simulations / measurement later next week and after i have finalized the design will post them for your thoughts.

 

[Hari Iyer]

Re: Measuring "The Metronome" ML TQWT with Dayton Audio UMM-6 measuring microphone us

Last month i ordered a pair of new Ribbon tweeters to replace my current Audiopur RT. I strongly felt that the Audiopur were very bright and shrill in certain frequencies and not very linear.

My typical speaker building process is design --> simulate ---> build ---> listen ---> measure ---> tweak and then analyze. This process is called Analysis.

Now i am planning to do somewhat reverse like build ---> listen ----> tweak ----> simulate ----> measure and then analyze. This process is called Synthesis.

Analysis involves in analyzing given the drivers and crossover network what you would expect.Synthesis involves in what you want, and to get that, what you should be having as drivers and crossover network.

In my case i will not be changing the box or drivers (other than the ribbon tweeter) but have the liberty to modify the crossover to get what i am looking for. Since i will not be looking at what component values etc but just follow my instinct of tune by ear and later measure and analyze.

The RT is expected around next weekend and will try some ideas with the current Audiopur RT.

I have reviewed many top loudspeaker models curves (impedance, frequency response - anechoic and in room, CSD, Step response) and have documented them for the following speakers Vandersteen, Thiel Audio, Wilson Audio, Dunlavy, Voxativ from sterophile measurements. These are more for reference rather than anything else. It has been usually observed that good sounding loudspeakers also measures quite well for atleast 95% of the speakers. But the converse is not true ie. a well measured loudspeaker need not sound very good subjectively depending upon what measurements are been refereed to holistically.

Thanks for looking.

 

[jls001]

Re: Measuring "The Metronome" ML TQWT with Dayton Audio UMM-6 measuring microphone us

That's a very different approach from your usual style

Which new tweeter have you chosen?

As far as I know, Audiopur are car tweeters so may not be as refined as ribbons designed for hifi. Just a thought.

 

[Hari Iyer]

Re: Measuring "The Metronome" ML TQWT with Dayton Audio UMM-6 measuring microphone us

That's a very different approach from your usual style

Which new tweeter have you chosen?

As far as I know, Audiopur are car tweeters so may not be as refined as ribbons designed for hifi. Just a thought.

Most high-end RT are quite expensive and this time i have settled for an unbranded Chinese RT though it does not look that cheap. I have gambled here and hope the gamble plays well. Its always difficult to integrate any RT with any drivers as they are not easily predictable. I will be using the RT mostly after 8KHz or so where the fostex rolls-off. Also i will be rolling of the RT after 15KHz as i like a gradually rolled off response than a full flat response.

 

[Hari Iyer]

Re: Measuring "The Metronome" ML TQWT with Dayton Audio UMM-6 measuring microphone us

I finally integrated my new Ribbon Tweeter with the Fostex last week.The synthesis approach worked for the Metronome quite well and irony is there are only 4 components in the compensation network. Two are for baffle step (one inductor & one resistor) and one resistor for tweeter level matching and one capacitor for the RT filter. The most simplistic network for a two way system.

I tuned by ear for most of the time and once the final balance was closely achieved fine tuned them with simulations and measurements. This approach was faster too as the number of iterations was much lower than the analysis method.

Suprisingly most measurements are quite good and so is the subjective listening of this speaker. I have done some basic room treatment for removing the early first reflection points from my usual sitting position and added two absorbent panels at the left rear speaker to break some of the bass boom to my benefit.

Here are some measurements

Frequency Response and Minimum Phase: Flat from 300Hz to 12KHz. Below 300Hz the room will influence the measurement a lot and cannot be predicted without gating the response.
https://ibb.co/iiet3G

Impulse: Clean response without much ripples in the impulse with quick rise and fall times.
https://ibb.co/gbdWqw

Step: At this mic position (50 inches at tweeter axis) there is a minor detachment from where the woofer takes over the tweeter. Not sure if this is due to faulty mic position. Though does not qualify for phase coherency neverthless the impulse response is quite good given its a full range driver with Ribbon tweeter.
https://ibb.co/f92kHb

Decay: Some clutter seen in the 300Hz to 400Hz region and 1KHz to 1.2KHz, but again this is typical of a full range driver where mid-range is quite strong. Beyond 1.2KHz the decay is smooth and clean for my room.
https://ibb.co/cmF5Hb

CSD: No spike or resonating peak at any frequency and is very even, smooth and resonance free
https://ibb.co/bLAGOG

Spectogram: Minor boom observed at 37Hz and 74Hz but again this is room dependent. The rest of the spectrogram is very clean without any room contribution to the sound.
https://ibb.co/cDZROG

Thanks for looking,

 

[Hari Iyer]

Re: Measuring "The Metronome" ML TQWT with Dayton Audio UMM-6 measuring microphone us

For this weekend am planning another different approach for the crossover. I will try and reduce the electrical Q at resonance by damping the rising impedance at resonance by adding resistance across woofer terminal. Also am changing the cross over for the ribbon tweeters from a parallel to series crossover. The simulation showse better step and Square wave response than the parallel counterpart. Now the crossover will have just 2 component, one resistor and one inductor making it the simplest ever. The Q at resonance is lowered significantly and could possibly offer more lift at the bottom end. If successful, I can prevent my TL Sub project.

 

[Hari Iyer]

Re: Measuring "The Metronome" ML TQWT with Dayton Audio UMM-6 measuring microphone us

Finally could conclude my series network topology with improved performance. Some images as usual for the technically minded,

Frequency Response and Minimum Phase:

https://ibb.co/dGS7sw

Impulse Response:

https://ibb.co/dtSU5G

Step Response:

https://ibb.co/fpatXw

Square Wave Response @ 1KHz

https://ibb.co/bLdK5G

Cumulative Spectral Decay:

https://ibb.co/iADaeb

Impedance:

https://ibb.co/daixsw

The major improvement is in the network Q at resonance which is a low as 8 ohms providing a stable load for the amplifier to drive.

 

[Hari Iyer]

After a week of subjective listening to this speaker, i will now replace all the components with audiophile grades - as now i am liking what i am hearing from this speaker to my taste,

MOR resistors will be replaced with non-inductive type
Polyester capacitors will be replaced with Film capacitor type
22SWG inductors will be replaced with 18SWG inductors.

Also the PCB over past one year + has been abused by many trials and tweaks and that too will get replaced along with the internal cables.

 

[Hari Iyer]

Finally replaced base components with audio grade components and done subjective listening for 2 days now. The SQ is now elevated to an another level. Very much satisfied with the outcome. Resistors and capacitors are film type.

 

[jls001]

Yesterday Prem and self had the opportunity to fulfill a long overdue visit to Hari's place to listen to his latest creation - the Metronome speakers.

We used two sources, namely, his Technics SL-23B turntable running an Ortofon OM-30 body with an after-market Tonar replacement cartridge, and his Marantz CD5400 CD player. Since I'm not familiar with his sound I took a few CDs and LPs to have some sort of a baseline.

Initially we played a few tracks from three records, and I heard a couple of familiar songs from my childhood, bringing about a bout of nostalgia. The other one was totally unfamiliar. We followed up with a couple of my own LPs and it became clear that there was something not right about the way the upper bass and lower mids played. The decay in this bands sounded abrupt, as if they were being chopped off. This resulted in a break in the flow of music, making it sound staccato.

Prem identified the sound absorption panels placed directly behind the left speaker (left, as viewed by the listener) as the possible cause of selective attenuation of frequencies. So the lower of the panels was unscrewed from the wall. It immediately brought about an improvement, partly filling the gap in upper bass and lower mids. The upper panel was also removed, and things got even better.

However, the highs still sounded brittle. On closer examination, Hari explained that he had used two layers of turntable mat - 6 mm thick glass, topped with a thin layer of leather mat. I noticed that the stylus was sitting with a fairly negative VTA with these two layers, meaning the mat was too thick. We removed the glass mat and used only the thin leather. This immediately removed the nervous energy in the treble, possibly because a reflective mat was removed, which also helped change the VTA from negative to mildly positive. My guess is the SRA is more correct with the glass removed, leading to a better balance across the sound spectrum. We later played around with the original rubber mat alone, the glass mat alone, and the leather mat alone and concluded that the leather sounded best. The glass actually sounded nice if one was looking for a more lively sound. Rubber is fairly balanced and was nearly the equal of the leather, and would probably be same if they were of the same thickness. My educated guess is that the thinner leather mat imparted a more correct Stylus Rake Angle, but this is a guess till proven with further experimentations. There was no way to play around with SRA on the tonearm so changing mat thickness was the only way to vary SRA/VTA. The biggest change was observed in bass response, consistent with my observations on my experimentations with multiple analog setups. One can actually reduce or increase bass response simply by raising or lowering the arm at its pivot. Alternately, one can do that with different thicknesses of mats.

After this we played around with the sound absorption panels that we had removed, and placed them at various places like the first reflection points, and also in between the speakers. Hari already has panels at the first reflection points so enlarging the panel or doubling the thickness didn't really change things much, but placing the panel between the speakers did tame the reflections of higher frequency, bringing about better separation of voices and instruments, and making the sound more refined.

We also played with the actual placement of the speakers. We found that these speakers need reinforcement in the bass so they're best placed nearer to the front wall instead of further away. In fact his initial placement produced the most ideal balance of bass with mids and highs. There is a fairly narrow window of placement beyond which the quality of sound audibly breaks down.

After all these experiments how did the system sound? As mentioned in the beginning, I'm not familiar with the sound of this setup so it will be impossible for me to break down where the contribution of Hari's DIYed gears, namely, DCB1 buffer, JFET-based phono preamp, Burr Brown based PCM 1794 DAC, or even his Metronome speakers, begin or end. But I would describe the sound as fairly balanced across the entire spectrum. The ribbon tweeters certainly help the treble. Bass is about right for the room, though some may feel that it's a bit on the leaner side, but I think this has to do more with the restricted volume of the slim cabinets, as I've heard prodigious bass from similar Fostex wide banders (in more traditional Fostex recommended cabinets). This is a design choice on his part for a smaller footprint and it does come with its limitation. The midrange is full and vocals sound good. The highs can be on the hot side, especially on brighter recordings. If I may be allowed one suggestion for hardware improvement, I would say that he should experiment further with the roll off of treble, both amplitude of roll off (a wee bit more than current roll off), and roll off frequency (perhaps starting earlier). Image separation is decently good, though complex music can cause it to struggle to separate and dehomogenise the strands of the music. This could be due to any number of reasons, including everything in the chain or room interactions. Image depth is very much a function of how far we can place the speakers away from the front wall. Here we are unfortunately faced with a dilemma since when we placed it further away it loses bass reinforcement. Currently it is playing as deep as the plane of the front wall (the wall behind the speakers).

All in all, it was a very enjoyable and educative experience.

Thanks Hari for having us over. Thanks too for the sumptuous lunch at Viviana.

Wishing you all the best for further refinements and developments.

 

[prem]

Thanks Hari for having us over. Josh has pretty much said all.

For me personally, it was a very interesting experience. For the first time I heard the sound of different mats on a turntable. What I learnt was both material and thickness matter. Next time around would love to do a comparison of mats of different materials but of the same thickness so the SRA remains the same

 

[Hari Iyer]

A big thank you to Prem and Jousha for coming over and reviewing the speakers. Enjoyed their presence and also took some very important lessons about subjective listening from both Prem and Jousha. I was amazed by their knowledge of speaker placements and how to subjectively listen to speakers.

Last weekend, due to my over adventurism damaged the Fostex speaker cone beyond repair and now i will have to look for a new pair. I am considering other equivalent options and was reviewing if other full range drivers can be a fit. As i have now already tried the Fostex and what to expect out of them and curious how others will fit the bill. The challenge is to get something near the resonance of the fostex so that the same Metronome box can be put to use. Keeping the suspense and will post when i have purchase and start another post. Till then keep guessing.

Cheers

 

[Hari Iyer]

Today ordered a pair of new drivers for the Met which shall replace the FEs. Will post later after they arrive and fitted.