I was always dredging up very old keyboard or mouse gear from the basement, even PS2 interfaces, for use in bringing up new microPC Linux boxes for some use like streaming media or split tunneling, etc. Or for when wireless fails. For just $30 I was able to get a mechanical USB KB & mouse... Cherry MX Red keys are fine; optical mouse is good; white is nice, as is backlighting.
Xfi OK Finally ~ Just A Bridge... I segregated COAX plants: terrestrial and Xfi separated directly at the house coax box, no diplexing ~ and then I removed routing from the XB8; it made all the difference. The Xfi "gateway" is just a modem (bridge mode) now. My first attempts to network via Comcast were frustrating as mentioned HERE. This frustration was such that I needed to minimize their admin footprint on my network. Net routing and admin duties are much better now. I also relocated the router near the modem and distributed Cat5e and WiFi from there. Not a great WiFi router location but serviceable.
Blodgett Router Up To The Task ~ WiFi 6e & 1Gbps... Once I got the router and modem assigned their duties and configured I went about the rest of the net. I reviewed the Cat5e run from the router to my main PC, a mac Studio in the study. I discovered an issue and replaced a patch cord. My mac's NIC now negotiates 1Gbps with the router. Wired was my main concern but performance is almost equal on wired and wireless in the study. I'm always checking my latency. While I'm loving the new bandwidth I still note a 50% latency increase from my "previous" ISP Century Link, which was 8msec to DEN. I have not "fired" them yet in COS, but I probably will now that Xfinity is stable and performant and the routing gear is much more manageable and adjustable.
HD HomeRun Reception & WiFi 6 Extended... I have two HD HomeRun OTA receivers (2 tuners in each) getting terrestrial HD VHF/UHF from a bow tie array antenna in the attic pointed at Cheyenne Mtn. They convert OTA reception into packet streams over Ethernet. I use these with my Plex DVR capability to get live TV and do recordings of shows on the 6 primary stations (ABC, CBS, FOX, NBC, PBS, CW). The HD HomeRuns are wired Ethernet only and I had to place a WiFi-Ethernet bridge near them where I attached them to "their" coax. No problem, and it acts as a WiFi 6 mesh extender too, which provides better coverage to the back of the house.
Performance Throughout... I measured WIFi 6 reception & performance throughout the house, at mid-day. Pretty decent even at the extreme locations farthest from the router and mesh gear. Between my main server and my mac I get a pretty consistent 1.6-1.8msec ping/latency and 1Gbps throughput... I could get more if I updated some switches to 2.5G or 10G but I'm good right now.
Next Steps... Next I'll work on secure remote access to resources on the Blodgett network. Likely via a Wireguard VPN.
Update 12Sep26 ~ New Arris S34 Modem... After discovering the Xfi XB8 was still using the radios and throwing WiFi SSIDs I did not need or want, despite my having disabled all of that, I bought the latest Arris DOCSIS 3.1 modem. I provisioned the new modem in just 15 min and it delivers fine performance. I even saw improvement in latency! But, I seemingly dropped in upload capacity by 66%! Maybe the channel config negotiated will be better with time ~ I need to look at the ATM/QAM log. I'll return the retired XB8 to Xfinity, else they'd likely charge me something. Speaking of charging me, I looked at my latest Xfi bill and they'd tacked another $100 for the tech call to my house to repair their network and modem. I call and railed, as all the problems were with the Comcast COAX network and Xfi gear, and they reversed the charge. Sheesh.
Update 21Sep26 ~ Cancelled Century Link... I cancelled Century Link's bonded twisted pair service to Blodgett before the next billing period. Xfi stable and offers 10X the bandwidth for the same price, though as noted latency is a bit slower. What a process: on the website, calling service center, receiving call back, etc.
Update 30Sep26 ~ Xfi Dead For 7 Days With No Apparent Reason; It's Back Up Now... On Thu 24Sep26 my viz into Blodgett went down; my new Xfininity ISP service to Blodgett went down hard after affirmed connectivity and performance for about a week. The harmonic convergence of my setting up a split Wireguard gateway from Summit to El Paso, Xfi performing "scheduled maintenance" in my area, and "storms in the area" like Xfi not dealing well @ my house, were too much for me to call a 'cause. I wasn't there so I let it lie. I bought new Starlink services to cover the outage and returned to Blodgett. No terrestrial Internet access - I moved everything to Starlink to cover Xfi's fail... it worked fine but was as latent as Xfi here and access required we be nearer the Starlink WiFi hotspot @ the front of the house ~ and that I keep the 12v system well charged.
Hmmm. I offered my opinion of Xfi performance thus far to their execs. I don't know if that notion landed or not but mail systems said not... hard to believe fully "not." Ultimately, talking to neighbors and weird "bots" from Xfi I determined "it's just me ~ only my service is down" and there are not outages in my area, and since I am @ Blodgett I should fix it or settle on Elon. I made an Xfi service appointment again ~ despite them charging me last time for fixing their own plant (I disputed and was rebated; too much investment for such an obvious contest). Infrastructure management can't be this goofed up can it?
Eric came today... a savvy youngster from Comcast Business. I do not know the source of my failed broadband; he blamed it on his company's own COAX plant. He claimed he replaced a number of terminations and even a splitter at my pedestal and showed me improved ATM/QAM signal levels on a phone supposedly connected to an instrument he attached to his COAX plant. He was methodical on his plant's signal but wanted inside the house for a review of that network ~ which has worked forever. Dang. I brought him inside and showed him my internal COAX net. He claimed the main line from my dmark to my "distribution point" didn't yield ~ however, retesting at the box was still lame as well... he didn't have to re-terminate my internal COAX really, but did on the side he could so do easily. We went back to the Arris S24 and he assessed that it was not activated on my account ~ uhhh; that's hard to believe as it worked well after I activated that same modem 7 days ago. Eric comprehended my confusion and continued to document my internal network with shots of my docsis 3.1 cable modem and my router. I am suspicious the whole event precipitating a call from a savvy tech was not orchestrated to get someone inside my net to document it all. Some call me paranoid; I pose it as secure and private. Eric spent 60 minutes out front of Blodgett and abutting the park documenting findings, closing the ticket and gosh knows what other surveillance items he was assigned. We are back on line.
Update 06Oct26 ~ Charge Me Again And I Will Leave... I noticed I had "another" $100 one-time charge for Eric's call to Blodgett. Yett per above he repaired only the Xfi side of the network. I contacted Xfi, fumbled through AI voice-bots and then AI chat-bots to finally get a live chat. The service rep reversed the charges. I also suggested my upload speed had been bound to a poor channel or something as I had dropped from 115Mbps up to just 40Mbps. The service person in the chat couldn't do anything but something happened as service has been updated to 115Mbps again. Good. Now all I need to figure is how to get latency reduced. Hmmm...
Blodgett Desktop Audio... I've never had a speaker setup for the Blodgett Desktop; it's always been headphones only. This had made sense with even open air headphones bugging other house members. My headphone system there is stellar Schiit Mjolnir2, Schiit Gungnir MB, and Mr Speakers EtherC Flow closed plannar magnetics or Sennheiser 600 open cans. Nonetheless, I like speakers better in general these days so why not toss a pair below my monitor for near-field listening in the study. I'll just need to pick my listening times properly.
GaN FET Class-D... I am watching the progress in the power, efficiency, and quality of "chip amp" audio amplifiers. I've owned, and used, and sold, and mothballed many of these inexpensive and efficient amps over time. These included Toppings, NADs, PopAmps, and other brands of "Class-D (digital)" amps and they were OK but nothing compared to my finer transistor or tube gear of classic linear amplification circuit designs. But I'm intrigued with the latest high speed class D variants in gallium nitride (GaN); they are even more efficient, even clearer & flatter, and still not expensive or large.
Class-D PWM Amplifiers... These convert analog audio input into high-frequency pulses using a technique called pulse-width modulation (PWM). Transistors (MOSFETs and increasingly GaN FETs) switch between power rails creating a pulse train (versus classic transistor amps using linear gain techniques). A low-pass filter is applied to eliminate very high frequencies and EMI from output. Efficiency is increased substantially in employing transistors fully on or fully off, 100%. Switching amps have been around since the 1930s but only relatively recently have gained the quality and performance status sought by audiophiles and music pros.
Mid-Fi Class-D... My dive on the category yielded numerous entries from $70 - $7000. The high end is not for me and I can't believe they'd be worth it compared to comparably priced classic SET, Class A or Class AB designs, much less SET tube designs for home audio. Many good products are raw boards for integration with powered speakers or in plate forms for PA systems, but I wanted a boxed variant. And, I prefer a unit with a pre-amp for volume control. The entry-level is littered with Chinese gear of varying sonic and reliability levels but one needs a bump in price to get GaN FETs and higher quality. After reviewing SMSL, Hypex, Purifi, atma-sphere, Nord, Buckeye, Crown, and others I decided on the Chinese entry-power level SMSL PA200. It's GaN stereo with 85W/ch into 8Ω with "no" THD for ~$400. Awesome on paper, and the reviews were 4.8/5.0 stars.
Install & 1st Impressions... The amp arrived quickly from Amazon. I'd retrieved a pair of "retired" NHT SB1 bookshelf speakers from the Nest that I repurposed to be the near-field transducers for this updated desktop system. Viola; it was all quickly installed and cooking. The audio path employs source flac files spun from my server through macOS via JRiver MC (or YouTube vids), then through my fine Gungjir R2R Multibit DAC over USB, and onto the SP200 amp from SE-out (single ended) via unbalanced RCA cabling; the last run to the speakers is via 16 gauge zip cord terminated with banana jacks. Listening demonstrated wanting low end so I need to work on a sub, but down to 100Hz it was stellar. Else the Blodgett Desk system is really clean and true and fun. Of course I measured the system raw at this point; measurements are taken @ listening position ~3ft away @ 1.5x the mains separation.
Subwoofer Trial... I own a closed SVS SB 1000 subwoofer deployed in the Blodgett Fam Rm theater. I borrowed it for a trial with the NHT SB1 mains for a 2.1 system to fill the low end. One failing of the SMSL PA200 is the lack of a sub-out... to solve this I added a combiner to the 2nd SE-Out of the Gungjir DAC. I did not optimize the sub location; I just plopped it on the floor beneath my desk. I tuned it a smidge and this is the Blodgett Desk Audio 2.1 SPL sweep. I twidled the sub crossover & output volume for good integration. So much nicer. The full range 2.1 off the Gungjir MB DAC is nice but I wanted to see if I could smooth out some peaks.
EQ Baby... I found other aberrations, beyond lack of low end, in the 2.0 sweep and so employed my miniDSP 2x4HD to flatten for the speakers and the room. I rue the moment I converted my sweet R2R built analog back into 24bit/48KHz via USB in the miniDSP, but I did. Raw was good, DSP EQ was maybe better. Time will tell. I eliminated the Gungjir MB and used the miniDSP directly via USB from macOS, as DAC and PEQ, and drive both the mains and the sub exclusively from the miniDSP, with less perturbations and fine results. I measured again with REW, then created PEQ/biquad filters and saved the file ~ I loaded the file into proper slots in the miniDSP. Here are the Gungjir raw (orange) and miniDSP EQed (green) final integrated sweeps. Nice, a new fine mid-fi audio, system is born.
Update 04Sep26 ~ New Kanto Subwoofer & New Sweeps & Filters... The trial with the borrowed subwoofer worked very well but I returned it to its rightful place in the Blodgett Fam Rm system. So, I bought an entry level sealed sub (preferred for tighter musical bass) with decent reviews, a Kanto SUB8VMB for $350. It's smaller than the SVS but fine for the desktop audio low end. I again spun the dials testing crossovers and gains, measuring them each time. I ultimately chose a 100Hz crossover point and created a new filter set for the sub itself. And, to be most accurate, I remeasured the NHT SB1s and built filters off the new measurement for them and reloaded into the miniDSP 2x4 HD. The new Blodgett Desktop audio system sounds great. It is used sparingly in waking hours but useful in preparing monthly Music Appreciation entries when in COS.
Remeasuring NHT SB1s ~ Here are the filters and then the raw vs EQed SPLs; clear improvement but displaying the need for a subwoofer.
Kanto Sub Integration ~ Here are comparisons of Kanto Sub8 and mated miniDSP crossover settings for the sub, with the baseline NHT SB1 wanting low end. I settled on the Kanto set at 70Hz crossover (not really as the integrated xover is not very effective) and a 100Hz low pass filter in the miniDSP (more effective). The Sub8 can get loud but gets pretty distorted. The gain on the sub's Class D amp is probably ~25% here to produce 75dB, as my SB1s are doing. At final setting THD is ~11%! And, it loses steam from 20Hz-30Hz. I'd go to the SVS if I could but it is 2X the price of the Kanto. There was a little bump at 100Hz I could tame in the miniDSP and so I built a biquad for it.
Update 24Sep26 ~ Blodgett Desk Audio Center Channel Upgrade; 3.1 Baby… I had an unused NHT center channel speaker and an open output on the miniDSP 2x4HD. I bought a cheap $40 Nobsound class-D amp based on the TI TPA3116 chip, which on paper will drive sufficient 40w in 8Ω for this application (to be sure though I also got a $15 24V/6A power supply as an upgrade to the included 19V/4A wart included with the amp ~ the chip can handle it). I’m running the amp for full range, not just for a passive subwoofer. Rather than buy an old Dolby PLIIx receiver to get a “dynamically derived adaptive steering” CC, I just went old school matrix tech and implemented my own center channel derivation from 2 channel PCM on my in-place miniDSP signal processor. I began with a simple summed signal (L+R)/2 for the CC, and then I tried a difference scheme R-L (harder to devise the correct settings to get this, as the 2x4HD can’t simply invert output). The difference scheme sounded weird and insufficient. The basic summing scheme sounded harsh on the cymbals at center and a smidge boomy on the kick drums at center. I therefore applied a 125Hz high-pass filter as well as a gentle 10Khz rolloff low-pass filter and even a broad bump in the mids 1KHz-4KHz; CCs are supposed to do voices and dialogue after all. Sounded good.
Measuring The New CC... The system sounds great with a more consistent near-field sound-field than the pure stereo “phantom center” one thrown prior. With these NHT speakers and my new 3.1 matrix setup, the Blodgett Desk Audio is now kindof special, beyond the headphone gear. The primary downside of this setup is my manually needing to dial the gains on the two independent amplifiers in order to achieve proper balance. I can do that by ear or by using Sonic Tools on my iPhone to measure SPLs... easy enough. I did measure the center channel itself. That is, after the Xfi fiber pull construction outside the house died down a bit from 108dB measured inside Blodgett... sheesh. It's loud; they are tearing up the yard, and my broadband service has been down for a week.
Here the green is the CC, the yellow shows the mains L & R, and the blue is the Kanto sub. Great result with the 3.1 scheme. I was able to return the 24v power supply as the 19v one is plenty sufficient.
Update 09Oct26 ~ Something Different This Way Comes... I bought another tiny class D amp based on the TPA3116, a $40 Fosi, this time stereo. I split the main L&R stereo at line level between repurposed upper NHT SuperZero Xμ speakers (on the new Fosi) and the primary NHT SB1s on the SMSL. The new 3+2.1 front array sound-field is great as my eyes and ears are at the center of the monitor and the sound surrounds that stage. I repurposed unused stands for the uppers and rebuilt them by removing a section to make them short enough to sit atop the lower speakers... they have security screws for the speakers and cable management to run the speaker wires. While the sound-field is great, by splitting, I am applying DSP designed for the lower SB1s to the uppers as well, so the uppers are not flat and come with a bit different timbre. I do have another unused pair of SB1s and might experiment with those as uppers. But frankly, right now the is sound killer, and the lighter uppers seem a bit safer.
I measured just the uppers alone at listening height (green) and the applied PEQs for the SB1s do make them sing bright some. While it sounds fine actually, swapping the uppers for SB1s is a righteous experiment yet to be performed. Just info, the fuschia shows mains, the mustard shows CC, and blue is the sub.
Update 13Oct26 ~ SB1 Flying Uppers Evaluated Much Better... As alluded above, I grabbed another pair of NHT SB1s to trial in the "flying uppers" spot in the Blodgett Desk Audio system. The bigger SB1 flying uppers present a better balanced sound-field than the smaller satellites. This is most likely from a more appropriate PEQ set up specifically for them (which I could not do for the prior satellites, due to the line level circuit I used). It is a very large array visually, but better. I remeasured of course and the first graph shows sweeps of the alternative "flying uppers;" the SB1s are shown in green and the SPL is flatter. The 2nd measurement shows each of the speakers, or pairs, measured separately. The array seems complete and better to me now.
Service Experience... I contracted Comcast in late Jul26 for internet service and after having never found a line into the COAX house panel for my self-install, after some hours on the phone and with a useless "AI Assistant," an outside guy pulled another line. This is all because of observations HERE. The alert he'd be coming was one day after arrival ~ Kala was surprised. I returned to COS and hooked up the XB10 modem they'd sent and I got nothing ~ activation is only possible via an app on a phone (what?) and my LEDs just flashed no signal. I called and worked through the AI chat bot again and ultimately landed a HUMAN chatter. In an additional hour I was able to convince her that I needed a tech call ~ so we scheduled it 3 days out. I was bombarded with confirmations and obligations for the dispatch (who had to be where, etc). Despite confirming texts on the day of and time of the appointment - NO ONE FROM XFINITY SHOWED ALL DAY but howdy I was here all day - Kala reminded me why I dumped them last time. After another two days of negotiating with AI bots, one's only inbound choice, finally a human intervened and my situation (and angst) were seemingly comprehended; another physical tech appointment was established 3 days out. He arrived and we tussled as it was quickly determined I had no signal at the house termination ~ the prior call pulled new COAX and terminated on my side but did not connect it to the Xfinity pedestal. Why not and why didn't Comcast know the finish status of the job and their physical plant? I am flabbergasted that my account records with parse-able chat records and diags from central do not yield understanding or progress at anticipated pace. Each tech is fine but the integrated system of service is bad. Ugh. So back to today's tech call: he arrived on time with the right tester and connected my run to the pedestal ~ we have signal! He eliminated all of my diplexers (combiners-dividers allowing over-the-air (OTA) UHF/VHF and lower frequencies to co-exist on the same coax); this is fair enough for a clean test of Xfinity's ATM/QAM service for broadband. My XB10 from Comcast still would not connect! He replaced the modem down to an XB8 version which finally connected inside the house! Though... there were exchanges about the efficacy of many moves as we sorted our way through the house and it's COAX connectivity. So, without diplexer capabiltiy I now have an new high bandwidth broadband internet connection but "no" OTA connections for my TVs to receive the six key OTA HD signals from Cheyenne Mtn. Hmmm. There were many attempts to have me expose my internal WiFi or network to the tech's app but in the end, to close the job, I said I would complete activation. Too, at this point of even exposing my name to Xfinity, my cell spam and text spam has increased ~10X today; they specifically are telling me how great they are; selling my vibe are ya babe? Suitably completing the circle ~ it is impossible to supply righteous feedback to Xfinity ~ dark patters lead to obtuse no-wheres... good for you fucknuts. I'm online now but seek to escalate my concerns and alter my net to reduce your intrusions. Why do I need to feel this way; why do I need to change out after moments under service? Golly you are abhorrent in so many ways.
Network Reconfiguration... After dismissing the technician via talks with his supervisor (he was fine and got it working) I ultimately reconfigured the internal COAX and Cat5E networks. Unfortunately we do not have a modern wiring closet for either COAX routing, and the only Cat5E runs that exist are stuff I installed after the fact. I only have 2 COAX runs from the exterior box to other primary "TV spots" @ Blodgett, the Fam Rm and the Theater Rm. My Ethernet runs only from the lower level office to my upper study and from the lower level office to the Blodgett Theater. After a dive on reliable circuits for diplexing VHF/UHF (30MHz-3GHz) with DOCSIS 3.1 (54MHz-204MHz), either expense or interference got in the way. So, I segregated the COAX net with one end getting Xfinity and one end getting OTA. But dang it... there is no location @ Blodgett where I can get both a great Cat5e connection AND the OTA COAX connection (without diplexing) ~ and frankly this is key for my use of HomeRun receivers converting OTA (COAX) to Ethernet streams for my Plex server to consume and create live streams for remote use efficiently, or for my DIY HDR. I could blow through a concrete wall but I don't wanna.
Wrangling Xfinity... My un-diplexed network is "up" with one TV area getting OTA and the other getting Xfinity. Sure, wireless is available everywhere and I updated the Blodgett AP to WiFi 6E so there are not a lot of sacrifices at the end of the day. My best and most secure net design separated the Xfinity gateway from my in-home WiFi. I shut down the all the radios on the "gateway" ~ though I need to confirm this a wireless RF analyzer yet. If one doesn't carefully unsubscribe from Xfinity hot spots and eliminate all your personal WiFi networks as well, Xfinity will know "all" and sell that data to brokers. I'm aghast I let Xfinity be DHCP server for my entire network, despite the separation of gateway and wireless AP duties. They probably do network discovery algos as well, and sell this information - I know they will parse every unencrypted DNS ask unless I hide it, or search if they can get it anyway... I am just sad where privacy and "net neutrality" has degraded toward. More importantly at this point is security, and I think I am on decent footing. I may yet replace AP firmware, or "de-router" the Xfinity modem. This may be only V1 of my net redesign as I can imagine moving routing as well to the current "AP" and replacing Xfinity with a 3rd party DOCSIS 3.1 modem, while the router/AP becomes open source firmware. I can't imagine the unwashed getting any of these things right.
Performance Assessment... I've not done a stress test over time on the Xfinity infrastructure. Results in Wildernest showed big congestion issues there. But here are the before on Century Link, the after on Xfinity on my WiFi6 AP, and after wired on a 1Gbps link on Xfinity performance characteristics of the new net @ Blodgett. Bandwidth is vastly improved and latency is substantially weaker... latency REALLY counts and is noticeable... I lost 50% in response time in adopting Xfinity (for now). Big pipes are fine but most initiated access (a new page) to the net is slower by 50%!
I cloned my prior WiFi characteristics/IDs when setting the new kit. That is for convenience. Perhaps I also should reset/change all that too. Hmmm.
