Work status from JMH

**dar47** · 26 March 2016, 09:35 Saturday

Hope you feel much better soon and you need to take in lots of water so you don't get those again!

I'm just waiting for the SMPS1200 to come back in stock, April 20 and I will order and get their input board just to get started and have a listen. Ben has yet to model my case so this could be good timing. I think I will have lots of room for 2 mono's in 1 case but it is probably a good idea not to drill holes in the bottom of the case till you have completed your design. Do you have a back panel convention you stick to or will that be played with as well?

**JonMarsh** · 26 March 2016, 10:54 Saturday

This raises a question as to whether you guys would prefer hand wiring input connectors and what not, or go the way I would usually do, which is to use PCB mount Neutrik connectors and limit the hand wiring. The hand wiring would give you more assembly flexibility. Of course, I can always layout a board that can accommodate both- that is, laid out for board level connectors, but with headers for cabled inputs.

There will probably be more than a few topics like that where we should kick ideas back and forth and try to come up with the configurations that serves the widest range of needs.

BTW, I DO drink a lot of water normally, so I've aske my urologist about the possibility of running a metabilic panel to see if it there is something specific I could be doing differently. He was a little surprised at that question (I guess not the sort of thing patients are familiar with or ask about) but I said, "Hey, I'm an a engineer, or at least, I play one on TV, and a number of friends have been medical professionals." (True)

**dar47** · 26 March 2016, 15:34 Saturday

Why not if you or your provider will spring for the table!

I like Neutrik connectors. If your going to limit things to the SMPS1200 then I would think you lay out your ideal placement for mono cases and those with large cases can follow the layout. I would just need room for a trigger and display. If a trigger could be added to your board that would be cool.

**TEK** · 26 March 2016, 15:35 Saturday

Jon, I hope you soon get better.
Not sure that I catched that you are going into surgery again. Bad stuff those stones.
When it comes to water, the recent information here goes much more to "trink if you are thirsty", and the "drink a lot" trend seems to be fading off.
As with a lot else, it seems like a lot of stuff goes in cycles from helthy to unhealty to healty again - normally ending up with the basic point beeing to not overdo anything but make sure that you have a balanced intake.

Regarding your test strategy. I'm not sure if I got all, but I did not see a a point where you were going to do actual measurements of the nc500oem input board, the smps1200 and the nc500 - but It did seem like you vere going to simulate that setup.
Would it not be natural to compare actual measurements with the simulated values to verify that the simulation and real world fits together?

About connectors, I would like pcb mounted connectors - but if ben and/or darrell does not want that it's no big deal.

**dar47** · 26 March 2016, 16:10 Saturday

Of all the NC400 amp builds I noticed people are orientating them every which way. I'm sure limiting EMI and having a clean assemble would dictate a usual layout. When I see something like the M22 it's clean and thoughtfully done. I think Jon means those Neutrik connectors on board not 2 to 4 pin connectors. I don't want to make my own pin connectors but can. Bruno did make those nice power mains to board clamps that you can twist pairs to nicely for the NC400.

**JonMarsh** · 26 March 2016, 22:18 Saturday

Originally posted by TEK

Jon, I hope you soon get better.
Not sure that I catched that you are going into surgery again. Bad stuff those stones.
When it comes to water, the recent information here goes much more to "trink if you are thirsty", and the "drink a lot" trend seems to be fading off.
As with a lot else, it seems like a lot of stuff goes in cycles from helthy to unhealty to healty again - normally ending up with the basic point beeing to not overdo anything but make sure that you have a balanced intake.

Regarding your test strategy. I'm not sure if I got all, but I did not see a a point where you were going to do actual measurements of the nc500oem input board, the smps1200 and the nc500 - but It did seem like you vere going to simulate that setup.
Would it not be natural to compare actual measurements with the simulated values to verify that the simulation and real world fits together?

About connectors, I would like pcb mounted connectors - but if ben and/or darrell does not want that it's no big deal.

Yeah, it should go without saying that I will be measuring the "OEM Adapter" and NC500 module setup, too. That's a baseline, in essence. I might do that and upgrade the OpAmps to the higher grade TI parts (LME49860, I think) over the baseline LM4562 duals, and compare, though I'd expect only a small difference in the front end noise performance. Distortion? We'll have to see.

Becuase of the numerical imprecision with simulation, even if you crank ABSTOL and RELTOL down to low values, measuring/predicting distortion below a certain level is difficult- but simulation can be valuable for confirming things like capacitor ripple current, ripple voltages under load, other issues that ideally one would calculate but like to confirm to a certain confidence level prior to the first build. The first build boards will be setup for experimentation and flexilbity, anyway.

As I get further down this path I'll be posting info for you guys to check out and comment on. With regards to connectors, yes, I meant putting XLR's right on the input board, and for speaker output using something similar to what the 400 modules do OR even using female bannana's that are PCB mount. Generally, though, I'd prefer to see a screw type gas tight connection for wiring for the high current speaker output.

The thing about getting PCB's made in smaller quantities is that it costs about the same to make 10 pieces as 2 pieces; size does have an impact, but not as much as you would expect in low volume; a lot is about the setup and labor of fabricating the bare boards. I'm starting to put together component information such as for connectors and other parts. There's some stuff I have a lot of already, but will have to check voltage ratings and what not- for example, extra storage bypass caps, I did some volume buys last year because of my own module development plans.

I'm also kicking around some ideas still for the line preamp, including using a similar balanced transformer isolated front end- and have ordered some rotary encoders for doing a relay switched volume control. I think a big part of getting a preamp done well will be the power supply- isolation from noise, and also keeping all the modern noise pollution stuff out. Input transformer will help a lot with that, and I have some ideas about the power supply- maybe even a HF resonant supply with high common mode isolation from the power line, with shunt post regulators. But that's down the road away- right now it's just thinking, and collecting component info.

If/when I do get around to building something, I'll probably want to send it around to some of you guys to listen too, to make sure I'm not just kidding myself if I think I have something worthwhile. Have to see how that goes- I'm sure that's a ways off. Mostly I want to stay focused on moving the amplifier thing forward.

**JonMarsh** · 22 April 2016, 23:14 Friday

I'm feeling a lot better this week- to bad I had such a busy schedule!

I did get some other parts orders in for testing, including a new dual OpAmp part from TI that is designed mainly as a headphone amplifier, and able to drive very low impedance loads (think low input impedance to CLASS D modulator), but will be a b*tch to solder. But they really thought through a lot of issues compared with conventional OpAmps, including minimizing the impact of thermal distortion, having a very low noise figure, very low THD, and optimizing power supply PSRR. It might be a good candidate for an "upgraded" version of the basic adapter board- I'll have to see just how much performance can be wrung out of it, but the drive and input characteristics mean that I can cut the input resistor and feedback components impedance in half, which will cut noise (resistor noise becomes a significant factor at the level of performance that even the LM4562 works at-

Happy Friday Everyone!

**JonMarsh** · 23 April 2016, 11:29 Saturday

Oh, and it may even be a good candidate for the transformer isolated version, though I really lean towards the differential output LME49724 for that. It comes down to how well each responds to optimization and the resulting performance. Noise figure, input current and voltage noise, output drive (a big plus with this new part, as being designed for headphone amps it works OK down to 32 ohms!). Has an unconventional output stage design that tolerates high capacitive loads, too. Just introduced in December 2015, came across it as I'm trying to wrap up my review of candidate parts before setting up test circuits.

**JonMarsh** · 24 April 2016, 10:59 Sunday

More concept planning...

I'm using this thread just to keep you appraised of ideas and planning resulting from that, in case I come up with something that really strikes you guys as a bonehead idea, you can bring me back to earth. The intent is to share my design planning direction at a relatively early stage, to keep you informed, and give opportunity for feedback.

One key idea that has surfaced this weekend it to consider making this as modular as possible, which facilitates function block tests and also upgrades/configuration.

What do I mean by this?

Imagine it this way- there is a master system interface board component to which the power supply cable, the NC500, and the cables to I/O connectors plug into.

Into this board, there is an input buffer/driver module fulfilling the front end of the NCORE configuration- that could be available in different versions to support transformer or no transformer, a baseline LM4562 setup, or a baseline setup with the higher end TI OpAmps, the transformer coupled LME49724 full differential amplifier, or the transformer with the OPA1622, or what ever future iterations we might decide to spin. This board would have two aligned header connectors, one for power and one for signal I/O. It's possible to make this field replaceable, but more expensive- lower cost is configure at assembly, and solder.

There is a small header on the main board to mount a time delay output relay/DC protect board, and the actual optional relay stuffing is on the main board. There would be a pre-relay output for builds that don't use it, and post relay output for those that do, and a common ground return connection.

I anticipate possibly doing my own regulator boards, similar to the HrX Hypex boards, but with more options than 12V rails (unless a complete early review shows that suffices...) With 12V regulator rails, max buffer output is about 8V peak, or 5.6VRMS. We may want more than that with the SMPS1200A700. Gotta run numbers. The HrX boards are a bit current limited, fine for just a pair of LM4562, not so fine for parts with higher quiescent bias, or greater output current drive that we want to take advantage of using lower impedance networks in input gain buffer.

BTW, the HrX boards DO use a technique I've also used in the past; no zener diodes for reference voltages, but LEDs, because they are much lower noise developing a stable forward voltage drop reference.

The idea is that you DO want the buffer section to clip before the class D modulator clips (M22 does this), because the power stage clipping means loss of carrier frequency and some nasty audible range HF artifacts. If we all use SMPS1200A700, then everything can be designed with just one target.

BTW, another of the many things I like about the OPA1622 is that it was specifically designed for very benign clipping behavior (young folks do that with headphones cranked up you know...) and so using it as the pre-clipper is not something that worries me.

I'm currently favoring it and the differential LME49724 to evaluate for the transformer based buffer; I think it should be tried as an alternative for the direct low cost input, but becuase it has relatively high input bias current (about a uA) it really will work best with balanced impedances on the POS and NEG inputs, and that's difficult to assure going out direct to a preamp who's source impedance might vary from 10 ohms to 1kOhm! (Same issue with the LME49724, and to a lesser extent the ultra low noise LME49990). BTW, I have samples on hand for all of these already... Been working behind the scenes, getting stuff together.

Last, I plan to actually make a small PCB module for test/production for the OPA1622 which puts the 3x3mm part onto a DIP10 format, with local bypass and compensation on the mini PCB. They're going to be a bitch to assemble, but that's why I have two stereo assembly microscopes... I know my limitations. Might even resort to an outside board builder for those... But for a dozen or two, I figure I should just "man up".

All for now- should go get some work done this Sunday morning!

**dar47** · 24 April 2016, 12:30 Sunday

Okay I think I am seeing what your doing. I know the unlike the NC 400 which has the SMPS600 designed for it not the SMPS1200, is this why he didn't pre clip in the input? I agree just design for the SMPS1200 and I would think helping them with the current limit would be beneficial for say 2- SMPS1200's and 3 to 5 channels of NC500. Does it matter for just 2 mono's? I think if I understand correct the pre-amp or direct from DAC could affect off amp performance and selection this is why the differential input is your preference? Are you assuming a base line performance of the pre-amp / dac performance, my new moon NEO 350P has an output impedance 100 Ohms. It is differential so does that mean both +/- stays at 100 ohms? Is there a target performance for a straight from dac?

**JonMarsh** · 24 April 2016, 14:53 Sunday

A lot of questions here, Darrel, I'll get back to you in depth later in the week- BTW, the NC400 has the buffer opamps built in, the NC500 is JUST a class D modulator- it HAS to have a front end of some kind.

Both NC400's and NC500's have current limit built in, to protect the module- as the NC400 is really designed for slightly lower power supplies, it's also got a slightly higher current limit point and can deliver a bit more power at 2 ohms- if the supply will keep up with it.

Part of my preference for the 1200 is not just sheer output power, but the additional bulk capacitance and the stability that lends to the supply rails when you're driving music at high levels, especially full range. This is one of the reasons (IMO) that some report that the M22 sounds better that a pair of NC400 each running with an SMPS600. It's got 50% more bulk capacitance. OTOH, dual mono NC400's with an SMPS1200 should be better in principle than the M22, as regards supply stability.

Remember, neither the SMPS600 or 1200 or the M22 supply are regulated for voltage- they're similar to a conventional linear supply, just bridge input to caps, but at much higher voltage, so you pickup more energy storage per farad, as energy is related to the square of the voltage, but only proportionately related to the capacitance. (double the capacitance, get 2X the energy. Double the voltage on the same capacitance, get 4x more energy stored)

Will address your other questions later in the week when I can't work out in the family room, but can work on the computer...

**JonMarsh** · 28 April 2016, 23:08 Thursday

One other point to mention again- nCores and UCD's are self oscillating Class D amplifiers... they aren't driving in switching by a clock. When they get to the clipping point, they transition through a wide frequency range all the way down to the frequency of the clipping signal- that includes the range your tweeters work in. If you let the class D modulator clip by itself, it's very nasty sounding, and not particularly good for your tweeters.

Most of the time this isn't really a problem, because with the standard products they scaled the buffer voltage and the modulator gain so that they work together and the buffer clips first. if you run the setup on a lower than typical supply, then the modulator can clip first unless you modify the clipping point for the buffer. In all cases they should be matched to each other. When you run multiple channels on one supply, the power supply will be more heavily taxed, and lacking actual load regulation, may droop enough to clip the modulator stage, even though nominally running at high voltage at light load. So for multi-channel, it's a good idea to characterize how much supply droop, and adjust things accordingly. NAD does this.

**JonMarsh** · 28 April 2016, 23:18 Thursday

Here's a post I may not do that often- it's the schematic of one of the candidate buffer designs using the JT-10KB-DPC. There will be two other candidate designs (actually, this one can be tested with several op amp types- this is a very low noise, low input current series designed partly for headphone use and has high output current drive, wider bandwidth, and is significantly quieter than the LM4562, and the circuit is arranged to take advantage of that.

This is JUST the plug in buffer card- doesn't include the power or control interface or the proposed DC protection relay, or the low level power regulators.

There's also a version that I'm drawing up, using the full differential single chip amp. Some possible benefits that chip, too. Plus there will be a cheapie, just in case- no transformer, just your choice of opamps.

Now, the question is, apart from the end results, how much 'splaining do you guys want to hear? That takes time, too.... And I'd rather not post too much of it on the forum, but rather put PPT's or white papers together and just email them to you if you're interested in the reasoning and thoughts behind this.

Just let me know what you think... meanwhile, I'll keep churning out schematics and adding parts to the QPL and ordering stuff from Digikey, and get ready to go into layout soon.

The part I would really, REALLY like to try is the new OPA1622 (sort of a big brother to this OPA1611, but it's only available in one package- a 3 x 3 mm VSON-10 which would be an absolute stone cold bitch to solder manually, even with my setup with stereo microscope. I almost talked myself into using the small run assembly capability of Advanced Circuits to do just that, but then I put down the mouse and backed away from the computer to think it over some more... standard SMD like this I can hand solder without much difficulty (stereo microscope including a USB camera). But the VSON-10 package has got a lot of really small pins, and they're pretty much not accessible from the side- it's much like a QFN.

**dar47** · 29 April 2016, 18:45 Friday

White papers are fine as these look intense. This looks more like how many rabbit holes you want to go done, haha. I wouldn't think any more then 2 options for the dual mono and 1 for multi-channel. I myself would say just go for the killer mono option. looks like my case project isn't going to be cheap and probably this will be an amp I keep for a long time. Going to be neat see all the boards stack up!

**JonMarsh** · 30 April 2016, 09:43 Saturday

Well, the way I'm approaching this is via a statement that I, and one of our other engineers in our functional team at work (big Danish guy), insist on- "In God We Trust, all others we verify, even if they pay cash..."

So while one can tell a fair amount in general terms by calculations and simulation, I have as many as three or four theoretical candidate configurations for the transformer isolated version, and intend to at least build and test what I believe at this point are the top two (of three, one of the most promising ones on paper uses that impossible to mount and solder by hand chip, not available in a different package), and for the "conventional" input we'll test something that's pretty close to being electrically identical to the OEM adapter board (but with added regulators) but also test configurations with one or two alternate OpAmp configurations that look promising.

For a parable about why I think the upgraded system interface is important (I.E., the transformer isolated version) read this short essay on Stereophile's site...

Home Page

http://www.audiostream.com/content/open-and-tolerant-mbls-juergen-reis-listening-measurements-and-uncertainty#beLo414yb86ogXAt.97

Stereophile covers everything high-end and audiophile audio. Turntables and music servers, to solid-state and tube amplifiers and preamps, to loudspeakers.

**JonMarsh** · 30 April 2016, 09:46 Saturday

And I'll go the white paper route, reporting design issues and results. I'm pretty good writing that sort of stuff... (About 55-60 publications done for my day job- let me know if you need an example to help you with your insomnia!

**dar47** · 30 April 2016, 12:03 Saturday

Very interesting read and this can happen across all Engineering disciplines not just electrical. It's taken me 20 years to isolate efficiency of new equipment on a customer base of 650,000 and the changes on the effects it has on load growth and a network with warm heating season years. I have company VP's arguing when looking at just flow data in warm years saying we don't need to spend as efficiency is out striping load growth and we are using less gas. I say not so fast as there is a bottom to efficiency and you need to protect for those really cold years that I assume haven't went away forever. Demonstrating this to them to spend is the tricky part and these VP positions tend to change quickly. I'm low on the latter but I have only done this stuff for many years, haha. Hard part is getting money to spend on automating data collection and simulating fast enough as the new Engineers coming up don't tend to stick much longer then 5 years at anything. I'm old enough to value experienced people especially when they are still discovering new things.

Do you believe the amp is where you should do the most filtering as this is the lest thing in the chain or is there just as much need for the DAC and pre as well? I know I didn't worry a bought components dragging the sound down until the Ardents were in place and now the goal to free them up.

**JonMarsh** · 01 May 2016, 11:14 Sunday

Your comments about the work situation sound very familiar with me- I sometimes refer to dealing with my own and other management at work as "herding cats"- it's for their and our mutual good, but often they don't realize it. OTOH, I do have the ear of one of the guys on the managing board (that's my friend in Munich that I refer to now and then, who is actually an Indian) and he often has some useful ideas on how to proceed, but thinks it's more interesting to watch me stir things up from below than to directl intercede, as long as I can manage to get things done. Weird but interesting situation...

Now, I think everything should be designed with care- I'm particularly suspicious of a Class D amplifier with an SMPS, becuase there are just SO many possible things that could happen to generate common mode pollution as well as the more obvious and measured differential mode stuff. The thing is, a common mode issue is a loop, so stuff from the amp could be conveyed back to the preamp, and to othe parts of the system. As well, there are things with achieving common mode rejection that the differential mode front end buffer works pretty well on, but let's remember that it has a differential mode gain of "X" determined by the resistor ratios, but it always has a common mode gain of +1... So the buffer is always handling and passing one what ever comes in to it; in this case, it will present it to the balanced input of the Class D modulator.

Now, this is also a concern to me, if a lesser extent, with most preamps. Many balanced preamps, especially those who boast about their fully balanced input to output operation, pass on common mode signal with unity gain (if all works out well, no more than unity gain). Ironically, those that just have a balanced to single ended converter in the front end may likely do better, and resolve/attenuate the common mode signal. This is the intent behind the Mola Mola preamp, too. OTOH, these active balanced front end don't always really have all that great a common mode rejection, and common mode slips into the differential path, due to impedance imbalance and gain imbalance. Considering this, as you might expect, I do have a plan for a test preamp configuration, which will use a balanced receiver very similar, if not nearly identical to what I'm developing for the Class D amplifiers. And ET is working on something similar for his low power class A project.

In my estimation, or guess, it's likely that the simple level related HD of a high grade line transformer is less of a problem than common mode issues introduced to active circuitry.

Now, on a similar note, think back to "Munich Report Part One", and the final judgement on the preamps- on the bench, the Halcro will win the spec/measurement contest. But what is happening in the system? This I don't have a specific answer for, other than what both my Munich friend, his wife, and I hear very quickly just a few bars into the first of the reference cuts we were using for comparison- and believe me, in spite of him talking up the Chorale, I was not inclined to believe or expect that. So, the listening results were against my predisposition bias. I went into that test NOT expecting to hear what I heard....

Now, as regards the gist of what I'm doing with this buffer stage, the transformer is one of Jensen's very best, and selected also becuase it's PCB mount! It has a 4:1 input to output step down ratio. Also, it's a nice double screened version, with input side shields and output side shields for common mode shielding.

Why or how is that turns ratio of advantage?

First, let's remember the impedance transformation that occurs- at the square of the turns ratio. This transformer is designed for an ideal primary side impedance of 2-2.4k Ohms. This is a very ideal impedance for working into ultra low noise amplifiers- ones with Ein of 1-3 nV/root Hz. (That's the range of parts I'm looking at; if you have a higher input impedance, then the source resistance of the circuit dominates the noise signature! Note, this also applies to the inverting input- so feedback impedances must be low for optimum results, which all in all means you need an amplifier chip capable of outputting 30-60mA with low distortion).

So, if we run an effective load across the output of the transformer that's 2.4k, (the transformer is incidentally designed for optimum square wave response at that impedance) then the primary side input impedance looks like about 23k, which is a pretty easy load for any solid state preamp and most good tube preamps, and even the target value for most passive preamps.

Then we have the option of running a variable gain structure on the buffer amp (shades of AHB2) and choosing to support lower preamp output levels or higher levels. (Unbalanced converted to balanced, 2VRMS from an unbalanced preamp; using a balanced preamp which on each leg tends towards more optimum THD+N at 4VRMS per leg, I'd shoot for 8VRMS input balanced for full output. This further optimizes the noise structure of our front end preamp, as we can run it at lower gain, and take advantage of the ability to output up to 18-20VRMS, depending on the winner of the final configuration contest (on the AP and listening)

But, now here's a big deal- common mode gain isn't 1, or 0dBV, it's about -120dB, so we really should be able to break the common mode loop at the input to the amp. Or, at the input to the preamp-

For a preamp, I envision using completely switched inputs, so inputs are completely disconnected when not selected, and possibly having two receiver buffers- one setup for consumer signal levels, the other for pro signal levels, or having a single buffer and simply being able to select, and perhaps have this automatically controlled by signal relays depending on the input selected.

Well, I've probably rambled on too long; I'm at a Starbucks near the starting point of the Devil Mountain Run, which GF is walking today, having run it in the past (and having the medals to show for it with regards to women in her age group).

Hopefully will get to some speaker measurements this afternoon... There's a lot on the plate these days!

**TEK** · 11 May 2016, 19:33 Wednesday

Thanks a lot for the writeup Jon. When it comes to white-paper vs. writeups as you go I would really like to see the writeup as you go. However, the fact is that I do not have the necessary knowledge - not even remotely close to either - make any sence of or give valuable input in the design process. So I'm with Dar here, a (brief) whitepaper will be more than enough. I'll think I will stay on the outside of those rabbit holes ;-)

I also agree with Dar on having one "ultimate" setup for that is based on a mono amp design and one alternativ more low-cost alternative that can be used for a multichannel amp.I think that if we got those two options we would have the most important usages covered.

You mentioned a preamp. Is that something you are working on? On my behalf I'm playing with the idea of getting a separate DAC that could be feed the signals from the HTPC, from the DAC into a separate pre-amp that again goes to the mono-block.

Dar, you have a setup where you have separate a separate stereo and receiver setup. Could you say something about how you have connected your setup? I dont quite get how to build a solution where the receiver is totally out of the look when running in stereo mode.

**dar47** · 12 May 2016, 01:12 Thursday

Ya, even though this is HT guide over time I realize that 5.1 then 7.1 receivers or pre pro's is an ever ending OEM's plan to tell you we do video, we do music we do.... but even when they strip out the video (as it's in a cheap blue ray machine) amplification out of the pre-pro or pre outs out of a receiver with separate amps and just get down to switching and a pre-amp section for audio it is still a compromise. So can the new pre-pro with the best sound stand up to a separate DAC, pre amp and amp for music? No as it's trying to still do to much, (room correction, connectivity, networking, etc). I found that I stopped even thinking of the next big theater thing that causes a minimum $2000. upgrade and said I want a music only system in the theater area as well.

It's late here after spending the whole night trouble shooting my upstairs PC after installing a new CPU cooler and the ram not working in dual channel. Tomorrow I will list the theater gear and how I route the sources for both theater and music and the music only gear and how it's routed and the benefits.

**TEK** · 30 May 2016, 16:02 Monday

How is stuff coming along Jon?

**JonMarsh** · 03 June 2016, 08:27 Friday

I'm in PCB layout for the test boards, and have been ordering more parts as I get things figured out (connectors!)

Also, I just sold off some stock to fund a couple of PCB orders.

On the "obstacle" side, there is a lot going on domestically at the moment sucking up CPU cycles- we've been prepping for and just started yesterday to have a new roof installed; that's going to take until mid next week to finish, we're told; I've got to complete some arrangements for the installation of a standby generator switch over box with the main panel; (ironically, while that work is being done, we'll need to use the generator to keep the freezers and frig and a few other bits running, with some ad hoc monster extension cords)

W're doing a major upgrade of the main panel, which is needed to accommodate the solar install we're having done later this month... Which is going in at one power level for now, with headroom in all the gear to install another string of panels when/if a Tesla comes around in a year or two...

Good thing GF is retired, but there's a lot of stuff she needs my help with, so the last month leading up to this has been very busy, and June itself is going to be a monster before we're done. Oh, and did I mention the heatwave we're going through? Had to uncover and fire up the central air a month early; unfortunately it doesn't reach all the house (I.e., the mancave) so I'm working on some other solutions to extend my work day capability in there.

This is going to be a year for the books, by the time it's done... :W

**TEK** · 07 June 2016, 15:00 Tuesday

Jon, don't slip away from us. Do not go into the light (or outside to the roof), focus on the pcb, focus on the pcb!

**dar47** · 07 June 2016, 23:48 Tuesday

Haha, or the Dark!

Ya this adapter board feels so developmental, better arrangement and connections are defiantly needed. Far to many crimps and these cable connectors with different pins seam weak. Lets make this amp sing! Oh I mean Jon make this amp sing, haha.

**JonMarsh** · 21 August 2016, 17:24 Sunday

Getting back on track

Well, this has been one holy terror of a summer, of course, some of the problems have been of my own making, but some, like layoffs at work, confusion and gross tardiness in passing on critical materials from my manager to me, and other sundry items have all combined to make things rather more "entertaining" than I expected.

However, there is progress.

One issue I've been working on was the matter of schematic and PCB creation; I'm a long time user of Altium Designer, but on my own equipment, which the individual subscription license model we used to use worked fine with. While I was working at Apple a year and a half, they switched over to a network licensed model, on a regional basis, and converted all the subscription licenses, basically so that other people could use them if you weren't- i.e., to do any work, such as enter a schematic, you have to check out a license, then when you're finished working for the day, relinquish it. You can imagine how well that worked, when you're trying to cover more users with the same net number of licenses. Also, in my case, I'd accepted a low end laptop as my company computer years ago, with the understanding with my manager and IT that I would be doing my real work on my own computers. With a host of restrictive network licensing models being implemented in the last two years, that is not feasible for me on the POS computer I have, but do to segment margin pressures, my boss is adamant about not spending money on new computers; fortunately most of my colleagues here got upgraded to some decent gear before they started squeezing the lemon.

Solution? In the last year I've bought my own licenses to programs I need for work, and in June the Altium rep pitched me on a special deal they were having on end of fiscal year (end of June) pricing, which was $3K off the normal price, so I bit then, and have my own Altium subscription license. I tried several other program demos out, including Altium's cheaper Circuit maker, but I can make Altium sing and dance with ease, and have 12 years of footprints and projects in it, and even the cheaper "Circuit Studio" program from Altium (which wasn't that much cheaper than the special pricing I got on Altium Designer) couldn't import Altium libraries or documents. Not good, but things are OK now.

Also, the free version of TINA distributed by TI for their opamps and controllers is a version behind and fairly limited, and something came up at work which had getting a full version of the TINA simulator very desirable, so that happened last week, on my own nickel, at a negotiated 20% discount. (yes, I'm cheap at times- why do you think I buy as much stuff offshore these days?)

That means I can now simulate the input buffer/protection design properly and even combine it with a Class D module, either idealized, or a discrete one using my own design, to evaluate dynamics in TRANSIENT mode.

First stage was proving out the input buffer concept and feedback connectivity model I had in mind, to provide HF noise suppression and input filtering, and work with the 4:1 JT10 Jensen step down transformer.

What's tricky about a full differential amplifier is that as regards the external connectivity, it looks like a balanced input inverting amp, that is, two sets of inverting inputs which effectively drive virtual grounds *(the feedback summing points) and doing HF ultrasonic filtering is a little more tricky to get it effective, and to get at least 2 pole response. The above circuit actually has five HF poles, three from C2, C5+C6, and the feedback caps C1 and C4, but considering the two pole roll off of the transformer, has even more above 125kHz. The Ein is effectively dominated by the input noise of R8 and R9 and R4+R7; being fairly low resistance, it's fairly low overall, even with no input connected. The transformer step down has benefits in that the 4:1 ratio gives a 16:1 impedance ratio, so the effective differential input impedance looks like 4.8K X 16 = 76.8K, which shouldn't be a problem for any solid state preamp to drive, and most tube. BTW, L1 is just an inductor that models the primary side inductance of the transformer, so that low frequency roll off is modeled correctly

The transformer is fully shielded and the shield is connected to the VCOM system ground on the buffer board. Predicted response with balanced source impedances of 50 ohm is shown below.

The target response is very flat with very low phase shift from 10Hz to 20 kHz, with the ultrasonic roll off down -3 dB at 60kHz, what you want with a Class D design with output filter resonance at that frequency. Response is down 1 dB at 4 Hz; some might say that's a little lower than desirable for a Class D amp with SMPS because of the tendency for rail pumping with Class D and high power at very low frequencies, but the SMPS1200 has good protection for that.

I'm going to be doing some further transient mode simulation and simulation with a complete power stage, but at least I've got the tool setup worked out.

**dar47** · 22 August 2016, 22:18 Monday

So cool you can Transient model this, I have a similar problem with network licensees and to many casual users too but I usually make them buy 1 stand alone for my rig, as us super users got to have!

**TEK** · 25 August 2016, 18:19 Thursday

Hi

Super cool that you are back here again Jon. I was afraid the dard side had ruled you over and that you would not be seen again :W
This looks super interesting, even if I do not get a full grep on all aspects of the details :stupid2:

Waiting for the next episode

opcorn:

**dar47** · 24 November 2016, 16:04 Thursday

Ah, Jon seems you are tied up way more then usual. My Ardents are now upstairs in a new house and will need to do some treatments. I will start with a defuser behind my chair and some damping to soften at the sides. it's a small room and their out as far as they can go. This winter I may bring in a Moon 380D dac for trial and am still hoping you can find some time to finish this project so I can get the Tenor case built. In the mean time I have been gathering a shopping list of mods for the VFR and hope to tour down the West coast this coming summer. if I come close to you and you are in town it would be nice to stop in.

Image not available

**TEK** · 01 December 2016, 16:16 Thursday

Ahh Dar, thats seems like a very nice room for the Ardents.
Just figured out that I did not have notification turned on for this thread - so I missed this post.
Have you moved to a new location just to get the Ardents a better room :W

**dar47** · 01 December 2016, 21:24 Thursday

Ya, no the other house I built and had the great basement like yours. Now I am on my own after 37 years and we won't get into that scary story.

I can now put anything where I want and just 2 channel in the living room is perfect for this house. I even ordered one of those really nice Fjords chairs you guys are famous for. Thought the Stressless was steep in price though.

This place has a finish basement which I can make the dedicated theater in but I need to build a wall for the IB sub and the electrical is screwed up so a little long term project for that. I going to try and soften this room with some treatment and definitely need some diffusion on the back wall.

Just for fun while I waiting to build the basement out I put the surrounds and Ardent centre in my bedroom that is not big. Right at the end of the bed, looks awful but some great sound while lying down. I even measured the head of the bed. :lol::lol: Not going to hang the 70" tv in there though!

I think when Jon finds some time for the input stage and I try some other DAC's this living room will get even better. Doing a lot more listening and less building which is why we build this stuff Right!

**JonMarsh** · 02 December 2016, 19:26 Friday

Nice room, but sorry to hear about scary story... stuff happens. Sometimes you expect it, and it's like a train wreck in slow motion. Sometimes you don't expect it at all...

I think when Jon finds some time for the input stage and I try some other DAC's this living room will get even better. Doing a lot more listening and less building which is why we build this stuff Right!

This is the right attitude, but this year has not worked out well at all for me on that score- tons of work done and in progress, but had to take down the system in the family room for various reasons. The downfall was building the subwoofer for the HT, which then required moving the equipment rack over, which because an issue because it blocked some of her wall of cabinets on the side. Short version, it's all in storage right now. And she can get her cameras in and out again when she wants.

Maybe the living room, if I take out my part of the musical instrument stuff. We'll see.

There's also been a lot of BS going on at work, and in our group four marketing guys have quit this year, one just this week, which is 2/3 of the marketing group. The other two are on the East coast and pre-date our current manager of the last three years. Distance makes the heart grow less irritated, I think.

2016 has been one of those Chinese curse years, "May you live in interesting times". Of course, that beats dying in interesting times, I suppose... :W

**TEK** · 03 December 2016, 06:39 Saturday

Dar, very sorry to hear that you are on your own. After 37 years that have to take some time to adapt to.
Kindof thought that after soon 20-years on my part I should be quite safe to sail it through to the end, but I realize that I better continue to stay on my tows and keep sharp also for the next - hopefully - 40 year or so...

On the bright side - and one have to focus on that - you will probably get more time to listen to the Ardents and other selfish hoppies ;-)

Work status from JMH

Work status from JMH

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment