XR6 Frequency Response

[CruzBay]

Replied to your call with a PM, the AA2 hasn't been turned on for as long as I've been out of the stereo business, probably sometime in the early 1990's, so bad caps is really likely. Considering that the AA2 I have (well, you have, now) was probably built in the early 80's.

In the PM I specified that if a little caig labs R-5 didn't bring it back, to just write it off and return it. I forgot who else knew about these things, and how relentless Greg is with doing things right.

Sorry the thing isn't working out of the box, it did fine 20 years ago. (When the last MQ-107 I did)

Interesting that 1) Greg has already gotten the schematics, and that the device is so simple.

If you throw any parts at it, I'll make you whole on them.
Thanks for your advice, c_dk.

Mike

 

[CruzBay]

Geez, just now looking at the AA2 restoration thread!!

 

[Rich Andrews]

If this were my project, I would invest in a modern audio analyzer. It would pay for itself in time spent. There are some software based ones that look pretty good for under $300.

 

[Gregory]

First results with remote SLM

It takes a lot of preparation but not much time to change Sig Gen frequencies and record number. Here are the results of the first Remote SLM output method, where voltages are recorded from over 12 feet behind the tripod and behind an absorptive sofa.

Test Conditions were:

• Power Amp output: 0.1 Watts (0.89 volts across 8 ohms)
• Upper Mids installed: original rebuilds w/o damping pads
• MQ104 Bass setting: #4
• SLM Mic distance: 1 meter

Here is the chart with 10 dB added (to convert from 0.1 to 1 Watt) and 1/3 octave SLM cal factors. This chart has measured ambient sound level drawn in.


Latest XR6 response with MQ bass set to "4" and UM divers missing damping pads. This is the first Remote SLM measurement and worksheet.


There's a reflection from the RIGHT wall.. and I've lost the tweeter energy for some reason. The MQ104 has boosted bass compared to measurements in 2010. Another issue is.. the Sig Gen needs to be on the higher shelf to look directly into its dial.


Original 2010 measurement

Previous measurement with UMs after rebuild with damping PADS and plot formatted identical (dB scale) to the one above from today. This is the FIRST test run over all 1/3 octave frequencies and with new worksheet that processes remote SLM data.


Earlier reference data with rebuilt UM divers WITH damping pads. This is chart has same y-axis scale as chart at top.


Yep.. a lot of time being spend here. I couldn't put Clay off any longer.


Wrap-up

It only took for 3:00 till 5:15 pm to set up, adjust power levels and step through ALL the frequencies, measuring both L & R each step. That time includes creating a new "worksheet section" with a NEW chart, typing in the measured values and applying the math, then clipping out the charts to post.


The reason I think Mike loaned out his AA2 was to spare everyone this kind of embarrassment!


-Greg

 

[Rich Andrews]

It takes a lot of preparation but not much time to change Sig Gen frequencies and record number.

With a good analyzer the whole measurement process could be reduced to a few seconds. In less time that it takes to make one measurement, the whole spectrum could be done.

 

[Gregory]

If that truly is what the response of those speakers is, it looks terrible. What do they sound like? Does it sound like something is wrong?

You really need to be able to look at the phase vs frequency response of the speaker system to really understand and narrow down what is actually happening there with the mid range.. It looks to me like there may be some kind of phase issue in the mid but it's really only speculation without more data.... There are also some extreme inconsistencies in the response throughout the range which one would really hope is just the room.... If all I had was the data you have there I would reverse the polarity of those mid drivers and see what that does to the curve..... You can then eliminate phase as an issue if it makes it worse..... Yes, Rich Andrews is right, there are ways to do this that would take seconds to measure, but at least you have he patience to work through using this method.....

Well yes.. I'm started to get worn out, but glad the remote SLM method was implemented. It's now available for use in the electronic "toolbox." The frequency response plots of Mc speakers (of old) were never disclosed, published or spec'd in the owner manuals. Roger may have some charts on his website.. certainly in his files.

This was the first of the 2012 sound measurements, and only took 2-1/2 hrs including setup, test, number crunch and post. Later in the evening I noticed the SLM response Cal was added, not subtracted (you must subtract the SLM response) so the math had to be fixed and the graph re-posted. Too much in a hurry, and cranking a new worksheet.

Photo of speakers in room

I threw together the photo below (sorry tripod in use) with some markups. The XR6's are about 1" apart and Mic 1 meter away. There are sectional sofa pads used as absorber on the flanks. See on the RIGHT where I think the REFLECTION (and beating in response) is coming from. The biggest concern is.. where did the tweeter energy go???


Square room extension with speakers in the center. Several absorber pads (sectional cushions - 17 available) are placed on the flanks. Furniture is arranged symmetrically.


I think the reflection seen on the RIGHT is from the wall behind the lamp. Will be running this again, and later with Clay's UMs. He has been very cooperative and patient.


Test data and worksheet

This activity is for everyone's benefit and enlightenment.. even to see the screw-ups. The standard disclaimer is.. is you see something that's not right, then that helps.. and you can go and do it better next time. Here's the worksheet with yesterday's measurements and SLM conversion.

Attached is: XR6 Frequency Response v061 - SLM output.xls​

You are welcome to look it over and point out issues, suggestions or mistakes.. even put it to use in your room.


Wrap up

Was really excited to get a hold of an AA2, but it wasn't working and that was a big disappointment. Still.. it was good to get the EXTERNAL VOLTAGE method together. Will let you know how long it takes to set 31 Frequencies and record L & R numbers on a clip board.


-Greg

 

[c_dk]

Having measured hundreds of speakers/rooms over the last 34 years I can honestly say I never saw anyone listen at 1 meter or with speakers 1 inch apart. Readings at 3, 4, or 5 meters with speakers in real world locations would give more information about actual performance.

My only concern has been the claim that these domes were being rebuilt in a flawed manner and therefore had to hold back recommendations to those looking for help with repairing these speakers.

I have had enough...I called and talked to the rebuilder himself who has been doing them for the last 12 years with the same back stock of parts Mac has had on hand since then.

If you have a bad 036-035 dome get it done soon while they still have the parts to do it as they always have.

 

[Gregory]

Having measured hundreds of speakers/rooms over the last 34 years I can honestly say I never saw anyone listen at 1 meter or with speakers 1 inch apart. Readings at 3, 4, or 5 meters with speakers in real world locations would give more information about actual performance. My only concern has been the claim that these domes were being rebuilt in a flawed manner and therefore had to hold back recommendations to those looking for help with repairing these speakers. I have had enough...I called and talked to the rebuilder himself who has been doing them for the last 12 years with the same back stock of parts Mac has had on hand since then. If you have a bad 036-035 dome get it done soon while they still have the parts to do it as they always have.

I'm already stated several times that the objective is NOT to EQ the room. NOT NOT NOT!

Hope that's clear. You say you called the rebuilder? well I know who he is but am letting John handle it. That's who I paid (Mc) to rework the drivers and QA them. And now you have more "secret information?" What action can be taken, or decision made from your statement after talking to speaker shop? and if the driver is bad? Please explain in clear terms.


-Greg

 

[c_dk]

Gregory this is your soap box but.....

I have been called on the phone, emailed and PMed by owners about these drivers. I have deferred while you have done your faulty tests.....secret message: get them rebuilt while you still can because when the original parts run out IT WILL BE DONE.

 

[Gregory]

So I'll go on with the faulty tests

That is to say "debugging the test" as the process is still being developed. The method itself is correct, and it is what's used at big corporations. The difference is.. the corporation has $500,000 worth of equipment, and I have $500 of equipment.

So the game plan is:

1) Determine why there is reduced tweeter energy,

2) Run the XR's again while placing more absorber on the RIGHT wall, and raising the Sig Gen higher, to eye level for better set-ability/repeatability,

3) Get a repeatable set of numbers,

4) Substitute Clay's UMs and run again and compare UM pairs,

5) Determine is the UM driver is faulty, and if it's necessary to send them out a third time for service.​

However no-one seems to know what the reference or baseline response of the XR6's was when designed. So, screw around with them for two years and post a whole lot of stuff on AK.


Objective

So.. the objective is to determined "If you have a bad 036-035 dome" as was the objective in determining if all the other drivers were working suitably. Simply and clearly stated: determine if the speakers are working properly (and match L to R) after a major restoration. Then make a decision on what to do next. At this point.. I believe the production response of the XR6 is still a company secret.


-Greg

 

[Rich Andrews]

When I did a EQ using solid tones recently the results were miserable. Moving the microphone 2 inches horizontally caused the response curve to be all over the place. A 3 second sweep wasn't much better. I quickly went back to pink noise where mid and upper frequency room resonances were not an issue and microphone positioning was not critical.

 

[Gregory]

After having a think about this, and one of the comments made by c_dk, it is possible that you may be too close to measure the true response of the XR6 as a "system". For the interaction of the drivers to represent a desired response, it is possible the speakers were developed at a minimum listening distance..... Just thinking out loud here.... Why not repeat the test at say 10ft? My estimation is you will have a smoother and more realistic response...... Yes you might have more room in it, but the room looks to be anything but live. I would also recommend placing one speaker at a time in a marked position and take those other speakers away, it's likely the other XR 6 is working as a passive radiator while you are running the tests, definitely in the low frequency anyway..... The other problem you are having is you will be charting the response of the uncalibrated microphone on the old realistic meter.....

All good points.

Looking back at the XR6 activities, they were tested at a distance of 3 meters in late Oct 2010.. see pic below. Will have to dig up the results from that run.




As for placing one speaker at a time, that's a logistics issue. It's more convenient to run both, L and R switching back-and forth with the preamp MODE control. One of the downsides is moving the cabinets in and back out, and the repeatability of setting the analog Sig Gen dial. There will be errors either way. Let me think about it some. I like the idea of simul-run L & R on the same chart with same Sig Gen adjustment. The driver array is straight vertical and therefore the patterns are symmetric, yes?

The other speaker being a passive radiator is possible.. there will be coupling but the question is how much? and what effect will it have? or how much error? Moving all the speakers out of the way.. you say? Ok.. will mull that over too.


The SLM Mic cal is in the spreadsheet math.. yes? The two functions used are Cal #2 (analog steps) and Cal #3 (1/3 octave steps) as seen below. Is that what you mean? Those data get SUBTRACTED from the measured numbers.. correct? IOW if the Mic has a certain response.. then you subtract it in dB.




And thanks for the useful comments.


-Greg

 

[Gregory]

When I did a EQ using solid tones recently the results were miserable. Moving the microphone 2 inches horizontally caused the response curve to be all over the place. A 3 second sweep wasn't much better. I quickly went back to pink noise where mid and upper frequency room resonances were not an issue and microphone positioning was not critical.

That was one of the 2010 experiments.. and when moving the Mic vertically there were BIG swings in sound level.




The AudioControl C101 has a pink noise generator and Mic with selective bands and LEDs. This is a good way of taking a "quick snapshot" of the room at any one Mic position. See the contour? It's what I found with stepped CW tests.. heavy upper bass and reduced upper midrange.




Kind of like a sanity check......... and oh look.. a RABBIT!!



-Greg

 

[Gregory]

Just thinking out loud here.... Why not repeat the test at say 10ft? My estimation is you will have a smoother and more realistic response...... Yes you might have more room in it, but the room looks to be anything but live.

to address your question above.. there's a response test sheet on the webspace from 10/26/2010 that describes the run at 3 meter distance.

Arrached: XR6 Response octaves rev016.xls

There are five tabs at the bottom of the SS (spreadsheet) and the last one has a photo of the Mic and speaks

That worksheet has other features (on the tabs) such as a model of the C33 150Hz EQ control, and you can enter a % knob rotation in the cell. There's also a model of the MQ104 Wide response at 200 Hz and a control cell. These factors are lumped into ONE Cal array (with the SLM factors, but no bass roll-off) which is graphed in the lower left. The speaker is still heavy in the upper BASS area, and weaker in the mids. Don't be too harsh now.. it's an early version of the worksheet.

I may have an explanation for that (heavy in the bass, weak in mids)... which comes from the text of the AA2 Owners Manual. Must get that into a "post-able" form.

How do they sound?

How do they sound? well the midrange, vocals specifically sound constrained and colored in some way. What is still needed is a C33 cap update..... if I could only get off back-side and order the parts.


-Greg

 

[c_dk]

I did post a while ago a XR-6 installation and provided graphs and such.

Using the AA2 we were able to determine a MQ101 failure severely limiting bass not boosting it.

Running the response curves indicated cap fatigue which we corrected by restoring the crossover.

Using the recapped speaker pair response curves negotiated an improved speaker location with the LRs interior designer/owner.

Because of the still asymmetrical installation were able to see common room effect flaws which were different for each speaker.

Then did the final correction with a MQ104.....just like we always did, when the customer would let us, even back in 1978.

These speakers have always been able to make great sound in real world living rooms when they are the correct ones for the room and are set up correctly.

 

[Gregory]

I did post a while ago a XR-6 installation and provided graphs and such.

Maybe you could point out where those graphs are. I don't remember seeing them and cannot find anything under this subject. There's no need to harp on the 100s of environmental equalns from the past unless you can provide some relevant data that can be used to VALIDATE what Clay and I have found.

The latest and best data that CORRELATE well are Clay's SLM measured XR16's (same three upper drivers?) and my SLM measured XR6s. Both of us had UM drivers rebuilt by Mc's vendor. Clay was astute enough to do a before and after UM rebuild test.

Here are the plots overlaid.. data acquired on different systems, in different rooms and at different frequency points:

• Clay has BEFORE and AFTER traces
• I have LEFT & RIGHT taken side by side
• My XR6s do not have an MQ in circuit
• They are all similar with room envelope

Clay ran "before" tests, but my cones were flinging rubber surround material and one UM dome was OPEN CKT, so I didn't power them.


Spec An snapshot, 3 meters side by side

See the following general spectrum display contours which are similar to Stepped CW results.

Photo check sheet October 23, 2010 @3 meters. Spec-An display and room pic.




Spec An snapshot, 1 meter in corner

Photo check sheet December 29, 2010 @1 meters in corner. Spec-An display and room pic. Mic attached to tripod.




Validation data

What do you have for speaker response data to validate? .. not asking about mods or different methods for which I have no software or test gear. Thanks


-Greg

 

[Rich Andrews]

Not to throw a monkey wrench into the works, but what does the off axis response look like?

 

[kevzep]

Not to throw a monkey wrench into the works, but what does the off axis response look like?

I dont think its throwing a monkey in the works, I think its a valid point actually.......

Im still interested to see what happens when the polarity is reversed on the mid range driver.....

 

[c_dk]

I have posted in past threads, latest 8 months ago, information gathered in what I consider a typical Mac service call. I included a verbal description and also graphs of what we found and what we did to solve the client's audio system problems.

This happened to be for XR16s with cap issues and XR6s with MQ101 and cap issues.

I think this type of service call should be typical whether it is 1978 or 2012.

 

[c_dk]

Never measured them in a anechoic chamber like Mac did, only in listening location in client's homes which OP does not want to replicate in any manner.

I have to repeat, even back in the day MOST did not like these speakers, they were designed to have a laid back deep in the concert hall type sound with a typical Row 18 +4 dB at 100 Hz, 0 at 1000 Hz, and -3dB at 10K type of curve.

They needed to be a a large room, away from side walls, to sound their best.

We had Celestion Ditton 66s on the sales floor for those who wanted a Row 10 type sound.