I haven't done much soldering in recent years, so three switches with 24 pins on each, removal and replacing was good practice. Got me used to controlling my new temp controlled Hako. 6mm sounds right.
The HDX is certainly 26dBu possibly more. This rules out many Monitor Controllers. Disappointing, e.g. the otherwise very cool Drawmer.
Again can't remember exactly but the Massenburg is even hotter, maybe 28/30.
And the 1176 can push 32!
These levels are rare in Recording and Mixing, but Mastering brings them all out to play.

DD

Wow, I really can't fathom the +30dBu signal-strength world. Tons of headroom, I'd imagine. Must take like bipolar 30 or 42 VDC supplies to push those kinds of swings. They must be running discrete transistors to survive those kinds of conditions, as I don't know of any chip Op-Amps that can handle much beyond bipolar 18VDC with any reliability.

Speaking of which, and to get closer to the original topic (well, only slightly OT) I just read a couple of threads over at groupdiy-dot-com about some other products from the SM Pro Audio lineage, namely the PR8 and PR8E. Seems that for the low budget, but DIY minded value hunters those two mic pre packages may be a reasonable starting point. Simple Op-amp based designs that can be improved relatively easily due to their through-hole PCB's, and some relatively OK starting circuits. I've been thinking about setting up a (relatively) uniform array of octal pre's for the benefit of quick-and-dirty mobile or rehearsal full-band recording, and a few of the PR8E's for cheap would get me there for the cost of dirt and a little sweat (and allow me an Op-Amp evaluation test-bed, to boot - its been too long since I checked out the latest chip families.... as in decades).

So, I've plugged in the M-Patch 2 to my test rig, but haven't done any REW-based room pings with it yet. The reason I bought it was to allow alternating the mono-ish signal from a two channel 2i4 Scarlett source, yet switch the outputs to allow L, R, then L+R for my room sweeps (following some recommendations I read - were they yours? Thanks, by the way for the tons of data and discussion over the years).

I have heard some audio through it at minimum attenuation, and it seems decent enough. The headphone amp seems a bit anemic at the moment, but I hardly care as for my application it would just be used as an idiot-check for sweep integrity.

I have an old, ancient, Grampian Spring Reverb here. BBC design. It's output can drive a loudspeaker.
I would avoid that Mono switch. I never tested this but I never liked what happens when it is engaged. I think it just connects outputs together.
Also, as I don't power my unit, no LEDS..... The Mono switch got knocked when dusting or hoovering once. I started to have mysterious issues, until I eventually found the displaced switch.


DD

Never heard of Grampian, is it vintage 1960's? Amazing that it has the power necessary to drive speakers - is it a tube design?

In what way do you not like what the Mono switch does? I'll do an analytic resistance probe of the M-Patch 2 just to be on the safe side. As long as it leaves a couple of hundred ohms in series with the throughput I'll be satisfied it won't dead-short an output of the upstream device. The Block Diagram in the manual claims that the Mono Switch happens after the input Attenuator pots, so there is some hope that they thought it through.

The way I'm using it is a little different for the purposes of Room Pings: the single output from Ch.1 (because Ch.2 Out needs to be utilized for the REW timing loopback) of my mic pre will be fed into the M-Patch Main Left only (with the Mono switch engaged for all tests), and the M-Patch Out 1-L will feed the Left monitors in the system, but the M-Patch Out 2-R will be feeding the Right monitors. It was the only reasonable way I could think of to address both channels on independent switches from a mono source.

Hence my original question which you answered, about the switch logic of the M-Patch NOT being an XOR (vintage radio type push-button with prior selection cancellation), but rather a "classic" OR....TRUE, or in this case "on" if one or more inputs is TRUE (selected via switch).

Oh, and I got around to checking the math, a +30dBu signal would require a voltage swing of approx. 35 volts P-to-P, so, accounting for a couple of diode drops (such as transistors) the supply would need to be 38 VDC minimum (unless they played some jinky voltage-doubling game).

Grampian also made Ribbon Mics. Standard Mullard transistor output stage. No idea why anyone would want to drive a speaker with rev only. Classic rev tho, a hit like an EMT. The Massenberg pres have an outboard PSU, yes extreme high viktage. 1176 discrete op with a massive traffo. There is no need for loopback with REW. DD

I'd love to try out a Massenberg pre at some point. The closest to/cleanest pre I've got currently is an Earthworks 1022.

My understanding may be in error, but I thought the point of REW's timing ref (using loopback), was to account for in-room timing (propagation) delay, so that the accuracy of displayed results for ETC curves would be enhanced, for example. I'm aware that frequency response display in REW doesn't need loopback, but for time analysis?

RE: loopback statement/question above - That is to say that loopback is needed - so that the latency of the hardware + software + room propagation delay could be properly accounted for in the time domain?

You would love the Grampian circuit. A Hammond Spring, incandescent bulb in series as a Limiter!

I just looked at the Mass Pre. 27.4dBV at Clip indication. +/-28V supply.
I had Millennia before, they were the 30dBu.

REW and others used to use Loopback to keep timing absolutely consistent and correct. But now they all work fine with USB Microphones. No Loopback possible.
I think John and other have built in some sort of marker into the sweep and/or a recognition thing. Altiverb definitely has markers and recognition, it can play and record on different devices.

DD

Maybe the Grampian could drive a speaker because it was designed for an electronic organ, or similar? Funny about the incandescent bulb, reminds me somehow of my Morley Power Wah from the 70's, an optical (incandescent bulb) circuit with a concealment fabric flap attached to the pedal armature, playing "peek-a-boo" with a phototransistor!

Still, the GML pre has a lot of gain (and healthy supply) to achieve 27dBV! I didn't know that about the Millenia Media's pre's, but there reputation for cleanliness is probably linked to that larger than usual power supply.

So, I just re-read the available resources concerning REW (some hometheatershack-dot-com threads, including one from 2013 in which John was speculating how he would implement a substitute method for loopback in the case of using a USB mic).

In the recent Getting Started section of REW, I found out that apparently they are (well, John is), as you suspected using an Acoustic Timing Reference sweep prior to the main sweep (a high frequency only sweep for Acoustic Ref), but the caveat is that you can't use the Acoustic Ref (as a substitute for loopback) if you are doing multiple (averaged) sweeps. I am, so I still need to use my CSL calibrated UMM-6 and Scarlett 2i4 (with loopback). Either way, I don't have a USB mic currently. A USB mic might make certain things re: room pings easier, but perhaps less versatile. I'll have to see which "Rings True" (pun intended) as I continue to do my initial pings with the unit(s).

Also, I really should look at the M-Patch resistance measurement sooner than later, as something recently came up which should require me to know the Pan Law of the summing (mono) function of the test rig. Optimizing speaker placement vis-a-vis the distance of the base of the triangle to avoid a reduction of center image (or over-emphasis of it) when playing the game of optimization of image vs. smoothness of response (what a game of Tetris it is!).

The Grampian output is Wet only! The regular Line Out is on a DIN.
The Millennia are clean but I found sibilance issues. Couldn't avoid them even on dynamic mics. Perhaps the 6K input impedance.

I will check with John regarding Loopback. It would be good to know if one cannot do multiple Loops using USB Mics. I would have thought if a timing reference signal gets it right on one sweep, it should get it right every time. If not, it means USB Mics are severely compromised. We can't do ETC work in REW. Hard to imagine.

I think the Mono switch in the MP simply connects L and R together.
If properly summed we know what to expect. But I see that each output being somewhat shorted by typically very low Z of the other. Softened by resistors I suppose, but quite a load I expect. And what kind of intermingling does that do? It is not like real summing surely?

DD

I haven't had the chance to measure the MP's impedance yet, I've been working on a feature improvement for my custom speaker stands at the moment (well, for the last few days), and my analysis rig is installed at my studio a few towns away. I promise I'll post back with a real measurement of them soon.

In the meanwhile, regarding passive summing: Rane put out the following application note a few years back that covers some of the gotchas and a couple of recommended circuits to at least remove the possibility of bad things when two moderate voltage balanced units are mono-summed.
Why Not Wye?

The resistance values 470 Ohm and 20K ohm from the app note would provide only approx. -0.2 dB insertion loss, if I calculated it correctly.
More resistance scaling, (as they mention) decreases interaction (crosstalk specifically mentioned). However for our applications this is somewhat irrelevant (Common Mode Rejection Ratios issues aside for the moment) as the very notion of mono-izing L & R is to maximize "crosstalk" (without increasing distortion via an obnoxiously low load impedance) when switched in. A sensible resistance could be chosen that is low enough to not substantially increase noise, yet high enough to not utterly destroy noise immunity.
This still doesn't describe what the optimum signal levels should be reduced to (Pan Law, in effect) to enable a proper center image with no significant power change in the room. I understand that summing of correlated (same phase, frequency, and amplitude) signals will result in 6 dB increase, whereas summing uncorrelated (random phase, frequency, and amplitude..i.e.- noise) signals will only result in 3 dB.

For my purposes (given that the pinging signal will be a correlated waveform) if the equivalent resistances for the mono summing -6dB voltage divider are let's say 2k43 each input series resistor, with a 1k27 base resistor (tied to ground), and assuming a 10 k load at the final input destination (the plate amp input of the speaker), I'd feel pretty confidant that my driving devices (balanced single out) could handle an approx. 3k7 load with ease (worst case), for example. I just might need to cut a couple of traces and add the appropriate wire and 1/4 W resistors (maybe, if the sillies did indeed direct wire the mono switch). Of course my prior calculations all depend on the values of the parallel resistances in the attenuator pot(s)- hence I must measure!

Unless you know of something I forgot to look at?

The Moderate Voltage sources here are very hot, 24dBU and above, when Mastering.
I never liked the sound of the MPatch attempt at Mono and now I have literally cut it out. I looked at the REW getting started. There is a sentence saying not to use multiple sweeps with 'an acoustic timing reference'.
I still don't think there is any need or advantage to Loopback in your situation, but I will check with John.
BTW why do you want to use multiples? I just use the longest sweep and never encounter S/N issues. I use iMacs and all sorts of gear, without interface often. Y leads are fine for L+R, and I often just disconnect or turn off or down the unwanted speaker.

DD

Yes, it would seem that the Mastering grade devices have a fairly substantial nominal voltage if +26 dBu and above is common. I suppose I should try to find out the typical source impedance of those devices to attempt a mono switch mod that wouldn't load them down to the point of distortions.

Even though I'm not mastering at the moment, better to design for the heaviest case, especially if you are finding the MP's stock mono switch isn't "cutting it" (punny or...not) sonically. Still a resistive summing node should be the most transparent if chosen carefully, at least with less low end anomalies than commonly available x-formers could provide.

About my (relatively novice) use of REW: I've been doing extensive pinging of my studio, and am trying to develop a procedure and technique that is as efficient and transparent as possible, hence the use of the MP-2 vs. having to continually repatch or turn devices off and on. This became more important as I performed a "strip down" of the room, in which every major item returned was accompanied by a sweep. I performed over 30 condition sweep sets in one day, each with L & R individual pings , and after patching that much, and being compelled to log the conditions thoroughly I decided I needed to streamline the process as much as possible.

Regarding my choice of method within REW for the sweeps: I'm currently trying to adhere to a standard that Stuart Allsop proposed over at JL Sayers forum, in which he further recommends a dual sweep, possibly as a noise/variability contingency. I know that it may seem futile to attempt to standardize the testing, but it's a noble and worthwhile thing if data I'm collecting could be utilized by others in their own situations or rigs, and will certainly help me as I no doubt will need to reach out to other more acoustically astute individuals such as yourself on occasion.

The collective knowledge base of us users can certainly help the more scientifically based acoustic questions be validated properly, particularly if done using a known and repeatable procedure.

Anyway, I've got to finish up my speaker stand add-ons now (if curious, I'm designing and building a Panning Gauge that will enable me to accurately determine/repeat toe-in to +-1/2 a degree), I'll post back ASAP about the MP impedance.

Interesting thought that passive summing might be the best. But sure summing Mono in a plug in is pretty much 'perfect'?
I was in touch with JohnPM. He says there is no reason why a USB Mic would be inferior at timing measurements. However, the option of multiple sweeping is greyed out when REW see a UMIK.
I always use the standard longest sweep.
I recommend you keep an eye on sample rates, I think REW 'likes' 48K
Also your body needs to be well clear of the mic.

DD

It is probably true that software summing is capable of being more transparent, due to the lack of thermal noise, but I live in a mixed signal world (as do we all, really). I still have too much of an accumulated investment in analog devices of one kind or another that produce or modify real voltages (heck, I even still own a couple of signal generators), so having some provision at the speaker end of the equation for test capability of the most basic kind (polarity inversion, mono, for example) is of some use to me.

Regarding the use of REW with a traditional measurement mic and preamp such as what I currently own- one advantage is that I can run the analysis rig from an adjacent room (Shattered Saucer- my studio has a Sound Sausage fillable cable pass-through port installed that allows me to run a bundle of cables up to 4" diameter with minimal acoustic leakage). I don't think I could have that much cable length for a USB mic without running a USB port extender or "repeater".

As to the advocacy of multiple-sweeps, in an imperfect acoustic environment (at least regarding isolation), having an average of measurements should increase the signal-to-noise ratio for all kinds of noise sources, including the occasional bit of low level leakage from those sources outside the room.

If at all possible, I would like to do a measurement sweep someday with a calibrated version of a binaural dummy attached to a "ballistic gelatin" body to measure the room response with something closer to the ultimate absorber (us!), but until that day arrives, I'm aware that I need to remove my own absorption coefficient from the measurement results

Bruel and Kjaer warn that the human body near the mic can alter response around 400Hz by up to 6dB.
DD

I can easily believe that, and of course our hearing experience is affected by not just body absorption coefficient, but the Head Related Transfer Function of reflection/refraction which recent attempts to measure and model are illuminating!

Ok, this evening I only had time after integrating the Panning Gauges with their monitor stands and calibrating them to do a 1/2 hr. probing of the ports on the M-Patch 2 with an Ohm Meter.

I want to pop the lid again and verify a few things- I thought I remembered seeing some 1/8th Watt resistors in there, and I'd like to make sure of the way they are doing the monoizing. Probably with a voltage divider of some kind, but I need to increase the resistance across the switch somehow independently of the stereo side, maybe requiring a second switch or a more complex one to "tap the load ladder on a lower rung".

The good news is that the lowest DC resistance I got was approx. 2k5 (I think that might have been in mono, can't remember I was going so fast), with most loads to ground or series pass resistors in the 5k range. So I'd want to check the wattage of any fixed resistors and maybe suggest bumping the values up a touch for any Analog Mastering level duties (or potentially for the mono function) to be on the safe side.

For the purposes of my test rig I think it should be OK- not too much added voltage noise, nor excessive current flow for a lighter weight +4dBu device like my Scarlett.

Just a note to self- "self, read your own note"! Yes, the resistors are 1/4 W. And a bunch of them appear to be the "0 Ohm" jumper style as well around a few of the switches- based on one of the earlier posts in this thread, which bodes well for having the thru hole real estate to make the Pan Law mod if needed. Cool.

So, I haven't had the time to pop the lid yet, but I did do a room ping today in which I mono'd the test signal (to get the same signal feed to both monitors simultaneously), and I observed that the waterfalls definitely showed an increase in SPL with both signals being rung. The waterfall appeared to show a +6dB increase, which would be expected if the signals were being mixed without reduction during mono-izing, as you suspected Dan Dan. I will still pop the lid to verify the best way to go about the mod. Also, I listened to some playback with the mono switch engaged, and thought I heard an unpleasantness. What did you hear Dan Dan, a little extra grit, as in the onset of distortion?

I can't really remember. I've never really compared to software mono. it's just anytime I flipped I had a gut reaction to flip back. Maybe my mixes don't translate to Mono. Quite likely actually.
When you hit Mono your sweep is suddenly coming from two speakers right? +6dB from sonic summing I would expect?
DD

Well, if you had an adverse reaction, however subliminal it may have been- there's a possibility that your signal chain surrounding the M2 was being dragged down with too much load, as is expected when two unbuffered signals are being parallel summed (actually they are buffered in a sense - if you include the driving devices as part of the equation). From at least one view the impedance is being halved, and of course the separate output drivers are exposed to the virtual load of each other's output, with the possibility that when in opposite phase the load effectively doubles again (or halves if in phase) , if I understand things correctly.

Even if the resultant distortion only went up one 1/2 a percent, in the critical midrange, it might make you subliminally react with aversion. The other point I would make about this is that the higher voltage swings from your analog mastering chain would necessarily mean greater power draw from a heavier load, depending on the device's output impedance and current sinking or sourcing capability. With gear from a bygone era, the 600 Ohm nominal impedance (rather than some of the modern Op Amp designs with 100 Ohm source impedance) might not interface as well under these circumstances.

Yes, when I flip the mono switch, the signal is coming from both speakers (the way I have the MP-2 wired up). The +6dB rise is I'm sure a result of the acoustic summing- therefore I shall want to look into a voltage divider arrangement to compensate (mainly so it will be easier to compare the shapes of the waterfalls in a REW ping series without the annoying level shift, and without having to overlay a response to compare things rapidly).

When using the test rig as a music source, I would also like to know for example that aesthetic observations aren't being tainted by the SPL increase, and that I can depend on my perception of the phantom center to arrive at a proper spread between monitors as I set them up. Of course there's the issue of choosing which Pan Law formula to utilize- -3dB, -4.5dB, or -6dB, since music may be either correlated or uncorrelated. Maybe that's why certain mixer brands decided on -4.5, as a sort of compromise.

Speaking a little OT for a moment, have you ever tried the LEDR signal test for establishing monitor setup, such as for speaker spread and toe-in? Bob Katz was (might still be) a proponent of it for a while, though he wrote about it back in the late 80's/early 90's. I'm finding it difficult to find a source for the WAV's of that particular signal at the moment. I believe they're only available on out-of-print or hard to find CD's, or via a "streaming only" online source. Seems strange I'd never come across it before a couple of days ago while researching recommended procedures for setting up my mains.

I think you can offset REW graphs easily.
Haven't tried LEDR, but one of my ears is visibly angled differently and out further than the other......LOL
DD