Wanted to create a reference phono to decide bacic questions like AC heating vs DC heating. SSHv2 vs RCRC in power supply.
The schematic is this basic passive RIAA with cathode follower output but when measuring real life carbon resistors and Philips mustard capacitors, I had to adapt to something that also gave a flat response, while all components were available in pairs.
This is how the RIAA looks like:
The power supply for anode is my tailormade Salas SShv2.
The case was an AN fake(?) iron housing. After measuring it, I had to add an iron shield for the transformer to totally eliminate 50Hz hum also from the left channel.
The heater is DC or AC. Has centering with potentiometer to minimize hum and DC elevation(50V) or GND switch.
Sound is good, being evaluated (see the orange LED case where new phono is housing):
After measuring with my Antiriaa turned out, above 2kHz I have an +0.5dB increase, however in Ltspice curve was flat. R52 was increased to 48K and result is super flat. Details here:
Problem description: A day before the 2023 Munich High End Show, me and my friends first time met “Djabe”‘s “Art in Tone” album and we noticed immediately the horizontaly dancing” tonearm. Their previous albums didn’t hit my attention, but I really liked the music on this one, highly recommend it. On the named show it was played on several stands because it was a great, new and really enjoyable test music. It was obvious that the record produced ~0.1% extra wow, which was caused by the hole displacement. In this recording, you can clearly see from far, how much amplitude the tonearm has from left to the right at 0.8Hz (45RPM), causing a measurable WOW. Video: https://www.youtube.com/shorts/oAVfN-q2b8o We saw the same phenomenon at ~5 exhibitors on that day.
At the same show, I got presented the DS-Audio’s ES-001 gadget which, in case spending 5 thousand EUR, helps center the record up to the limit of the holes “headroom”.
The fact that such tools exist shows that centering of vinyl is still a possible issue, which is hardly noticed, but rather accepted as poor result of old technology.
The DS-audio principle gave me the idea to mount a microscope (Tomlov AF 4K for 214USD with stand) that can be just pulled in seconds above the record and observe the inner groove, to check the proper center. The result is here https://www.youtube.com/shorts/ymg6UVStewI . Music here: https://www.youtube.com/watch?v=DBzokiOygiY Anyway, now that I have this gadget, I’ll of course look again at how to lower the previously measured 0.035% wow, but I can already see the potential for halving it. (Regarding the title: Yes wow, and not flutter as the frequency is below of the misplacement caused wobbling is below 1Hz! Flutter is higher frequency wobbling.)
For a long time, I have considered designing and building an RL12P35 tube-based amplifier.
Why? The German army took my grandparents’ house in the last days and used it for radio communication. We were the poorest family in the village, but the house was the last in the street at the highest spot. Before the russian army arrived, they flew without the radio equipment and left it with some RL12P35 spare tubes. So the tube has some emotional value for me to do something with these parts.
Design goal: Tube level NFB, 2 such tubes.
Option1: Single ended push pull arrangement. I simply tok the AG9015, but changed the ECC83 to an ECC83 in SRPP pair. It works somehow, even G2 is around the 200V wish, but the simulation shows distortion above ~6V Vp output. Ltspice file here.
Option2: Parallel Push Pull (Circlotron), whose theoretical drive is similar to SEPP. Here I copied this circuit, but replaced the output pentode. Simulation is close to useless… , probably to do with simulation method. Ltspice file here.
I purchased Idler for my Elac 22H and for my Dual 1229 from audiovault.ca
The webpage’s professionalism is surprisingly high. Such niche market normally works via individuals, who are accessible via email, while order process is very manual. But not this shop. Probably the owner is a tech freak and very talented in many areas. Beyond touchless shopping, questions, emails were answered extremely well (I accidentally purchased the wrong Dual part, so I had to ask for a fix…).
The parts received were well packed (Made in China sticker, so likely a supplier behind), contained new carbon washers, a transparent fixing ring and a very clear short instruction, where you can’t go wrong. Also had a small towel, unsure reason, but as it clearly advises to not touch the wheel rubber part, I used this to grab the idler.
The Idler has copper sinter bearing, which is an improvement from the factory steel one, what was always a design flaw. Quality had high promises on the webpage, I really liked the Mitutoyo clock measure with 0.01 tolerance, where I would have been glad to see 0.001 precision results.
I received ISO48 and ISO68 viscosity oils, which I both tested. For me, clearly the ISO68 was the better performing in my wow measurement. My factory Idler was surprisingly good, so likely not much improvement for me, but the 0.035% W&F RMS is extremely good and is a testament to how good the audiovault idler is. More on the updated W&F measurement here.
The size of the wheel is slightly smaller than my factory one, where I earlier had to reduce frequency to 49.37Hz to get exact 33 1/3RPM and 49.55Hz in 45RPM mode. With Audiovault I was setting the frequency to 49.99Hz for precise RPM. So this is also an improvement compared to the factory one.
While wow measurement, I noticed for the first time peaks in the spectrum, which is coming clearly from the motor. Run frequency*2 (so 100Hz or 72Hz, when I downclock motor) and from the motor wheel (18.2Hz as I run it in 45RPM mode, with ~36Hz, which is ~25Hz if run from 50Hz @33RPM mode). This is likely because of the way less rubber in the Idler, which design filters less. So way less dampening! If you can reduce the voltage of the motor, you can live with this note. The Elac motor rumble is extreme also in factory condition; the Dual is way better on this in factory condition, but helps both if less motor vibration is caused by the unnecessary high voltage.
Summary: The part is very well done, I highly recommend taking it, especially since there is no real competition, while priced fairly!
Note: Tuning Idlers are not new to me. Had Lenco L75, PTP, with Audiosilente parts. Their Idler was a nightmare. They gave 5 different Idlers to their fancy Idler wheel. Could select between bad and bad till one of them somehow worked. Had to develop technology to put the round Idler on the wheel without even a minor twist which otherwise led to horrible wow. Hard to believe that there were so many freaks like me to go that mile. While!!! After some years of use, audiosilente started to degrade in quality and cause wow. Recently I saw, that audiosilente changed the rubber part, so now advertising it as “NEW”, so I hope they really fixed the flaws. Of course really interested, how long will the audiovault last. The factory one can last 50years easily, so is a nice target.
For W&F measurement, the very first thing is that you need tools. A phone is a useless device as it causes extreme imbalance, which can’t be tolerated while measurement. First, you need a record that has 3000 or 3150Hz on it prerecorded. Example DIN 45 545 where both sides on the full-length is dedicated for this measurement, with this I could achieve the best measurement (the outer track has a 0.01% adder, while the inner 2-3records is better as I observed). Also nice, my favorite vinyl Tacet, which can also help in many other measurements and is top to set antiskating ( has ~0.02% “built” in Wow!). Also good, the very well (over)engineered test record from Hungary (this has also some wow adder, but likely as it is on the outer edge, like the Beuth).
Above of the record, you need something to evaluate the tone. Hardware solution EMT 424 is nice, but hard to keep such old machine alive and calibrate it, while results are very limited compared to software solutions. Such software could be the WFGui, which can measure W&F, but is really bad when it comes to the question, why is it that bad?
I personally like the Nakamichi T100 inspired NAK-T100 . Made a video, which we can analyze and learn about possible steps while reading data measured with an Elac 22H turntable.:
On the above picture nothing much interesting. You need to select 3150Hz, which then matches with what is on the recording. You finetune the speed of the turntable mechanically or by setting the motor frequency with turntable power. The deviation meter will be in the middle of NAK T-100. The bottom graph shows the deviation in time and draws it continuously. The aim is to get this as flat as possible. But how? Here, the second instrument helps in understanding:
As visible, 0.121% is the weighted RMS W&F. The 2 type of contributors are:
-Wow, that is a slow rotation speed change of the turntable. Like how cat says “meow”. Such slow changes are caused by the plate imbalance (33 1/3 is ~0.5Hz ) or by slow motor speed changes < 0.5Hz, by belt weakness…
-Flutter is a fast change. This is more annoying in the tone. The sample video above demonstrates this clearly, a fast vobling tone on the 3150Hz. The source of such is typically the drivetrain as its moving parts are rapidly enough to cause such(wheel, idler, motor ). But we may also think on a tonarm resonance on 10Hz induced by self resonance (good to measure and know tonearm resonances!).
Good idea to change measurement method to Quasi-Peak, which is more sensitive for Flutter. Increasing sensitivity 0.5%–>0.1% helps to better visualize the spectrum, where we would see 3 peaks:
-0.5Hz, ~2Hz, ~10Hz
Measuring through the wheels, knowing the turntable plate RPM(33 1/3), we can calculate all RPM’s or the frequencies of the drivetrain parts. Here harmonics (double) are also interesting to watch for.
Now it is obvious! 0.5Hz is caused by the plate imbalances. The ~10Hz turned out to be caused by incorrect Idler height, where it was touching the next (45RPM) speed wheel just slightly and causing a flutter close to 10Hz. This frequency made the Idler as suspect. Was 8Hz when zooming in. Likely was touching the motor wheel twice/every turn if you see the relation of ~3.4Hz what I calculated for the Idler.
So, analyzing the spectrum, comparing to drivetrain frequencies can help to identify the reason for the bad result!
As a result of fixing the mentioned Idler height, here is the good measurement. Of course, I was cheating, with lowering the voltage, as running on 230V causes way more motor rumble, which then lowers W&F with ~0.02%. To cheat more, decided to run the motor on ~37Hz while setting the drivetrain to 45RPM mode. The W&F in this setting dropped further to 0.05%. As visible, the plate wow remained. Now purchased Mitutoyo clock measures to analyze how I can make this better.
I expect that while improving, it will be hard to tell if W&F or it is rumble what we measure. So I also sourced DIN 45 544 vinyl, and may(500EUR!) to source a measure adapter, which is a different way to measure rumble(while more sensitive to not measure groove noise, it is seeing vibrations that the vinyl is filtering out).
The current Elac 22H result is 0.035%!, with the audiovault Idler. This is very close to the DD results, where it is interesting that WRMS is shared in spec , to hide flutter, which I think is a likely DD problem.
Audiovault helped me with fingerpointing on the oil variations (viscosity, ISO48, ISO68). Is critical, that the sintering bearing gap is well filled up, so axis is super stable. I used ISO68 on all Elac bearings.
Beyond all the above: With all records ~0.01-0.02% of wow (0.5Hz) can be added due to the wrong positioning of the hole. Unfortunately, the mastering vinyl loses its hole and a new one is punched while vinyl manufacturing, which is not always exact to the groove’s center. With the centering tool (385EUR) you can make measurements repeatable, but if the punching is bad, you have that bad constantly. Due to price, I am making my own interpretation here. Best solution is DSAudio ES-001, which detects the inner groove fluctuation, but hey!!! 5kEUR for 0.01% Wow? Actually, they lie when they claim flutter, as this tool reduces only the 0.5Hz range wow, nothing else, but that works really well.
After my heavy 1219 modification, I created a similar mod, but keeping the shell of the original player with a 1229, which is nearly the same, but has a neat stroboscope separately flashing by turntable power.
Here are some pics. Soundwise no diff from original Dual mod. The wow measurement is quite good 0.05%. Video.
I recently purchased an Elac Miraphon 22H turntable. The reason was the nice PAPST motor ( HSZ 14.50 4 900 ), which is an 3 phase 4-pole hysteresis motor, so a synchronous one, where I should take less attention on sleep compared to asynchronous motors. I also use this time turntable power to run it, connected with the 2-phase variant, like this, to reduce vibration to the lowest possible. The turntable modification is similar to my earlier tuning on the Dual. Throw away the useless automation, the tonearm as well. Leave the drive as is. After gluing the empty holes, I applied leather. Thorough cleaning after complete disassembly, with protegum (bearings, gear switch, ect), also for rubber parts (avoid painted surfaces!). Silicon fet for gear switch, while sinter and ball bearings with oil. The 10H, 50H has a different plate bearing, more noisier, similar to Dual, so easy to modify to a Magnet bearing. The 22H is way quieter in factory condition, but also more problematic to modify to a magnet bearing, so I left this also as is. At the beginning, I had flutter issues(10Hz spectrum), but I realized Idler is touching the motors next speed wheel. Setting idler higher, Wow&Flutterr is 0.05%. Video. Using 45RPM ratio with 33.69Hz gives the best result. 67.5Hz in 33.33 mech ratio to get 45RPM makes the motor too noisy due to high RPM. Actually 50Hz(1500RPM) is already close to being noisy… The wow spectrum shows, to get lower(better) results, need to work on the plate (0.5hz…) which is the next target. Also Plinth is under design to surpress vibrations and motor noises. Soundwise beats my earlier modded Dual which is beyond Garrard 301, Lenco PTP5, Micro Seiki RP1500. So at the moment an dragon slayer. Happy to compare it to anything if somebody willing to bring his gear for a test. My friend, owner of 72audio modded similar way an 50H. Result also spectacular.
Wanted to keep note of great systems, components, which I found good enough to spend a word on it. I visit München HES, Klangbilder successor, Budapest audioexpo or HES, so being continuously inspired. I do not recommend Ag9015 or Dual or other DIY things here as their standard condition is not that good. So only as is components are recommended here:
Amplifier:
-Musical fidelity A1. As the old, also the new is a great amplifier. I found also the A2 really exciting. Is a nice budget choice.
-By the way I recommend only tube amp. Sad, that I have no idea which is good beyond the many DIY- one’s I know.
Loudspeaker:
–Alsyvox. Is simply woww. Is not a budget speaker. It does many things simply properly. I love it.
-Popori speakers. For my taste not warm enough, but is clearly a remarkable distortion less wideband speaker. Please never buy electrostatic or magneplanar with a separate bass speaker!!! Popori is great in performing also in the bass range.
-LS3/5A. Bass is always weak in the bottom, but hey!!! Such a stage, so much fun. I can always just recommend them. Mostly with transistor amps…
-Canton LE900… price/delivery is great. It does everything well. Since I met Braun, I tend to omit Canton. Lack of warm, has not enough air. For the typical price still a best buy.
-Braun L810, L710. Is fabulous. Super cheap and beats big players. Is what I also own.
-Fyne audio loudspeakers. Very interesting. A speaker who’s products do sound like my Tannoy HPD315. LOL. While Tannoy’s do not sound like old Tannoys. What a world we live in.
Turntable:
-Lenco L75. For this price hard to get better. Nice DIY roadmap with the PTP which peak high end.
-Garrard 401. Was never a weak part of any chain. Is on my bootlist to do deep A/B comparison to my Dual DIY.
CD:
Philips TDA1540, CDM0… what else.
Phono(RIAA:)
This system AEONOR was the best in München 2024. Main reason the LR RIAA from Michael Ulbricht. Actually is a DIY thing…
I always admired the Philips EL3501 strobe disc, so when I saw Sepea offering one, I immediately purchased one. Using it I saw, that somehow I fail to stabilize it, while the speed is for sure correct.
Making calculations revealed the secret, that having 66/33 marks at 50Hz strobe light (real 100Hz) will have -0.5% of sleep. If Sepea would have added one more mark, then +1% would be the result. So not that bad decision, but they forgot to tell about this sleep. I think I will have to make an 99.592Hz strobe light with my https://turntablepower.wordpress.com/. Calculations:
Philips AG9015 
You must be logged in to post a comment.