My_Ref Fremen Edition - Build thread and tutorial

Rest assured your attention to detail, long history of participation and success in optimizing the MyRef output - and willingness to be completely open in you comments was and still is why so many of us were thrilled that you eventually approached this version. What is surprising - not disturbing - is the relatively wide swing from the reports of your early impressions. I for one, will be anticipating further comments on your future mods/discoveries/suggestions with great interest.
 
Last edited:
I do not want to insult anyone or their taste, so please do not insult mine.

Hi Tom,

I guess you're referring to my post but I can't figure out where I've insulted you... BTW it was not my intention and I'm sorry if you feel so.

I am putting my efforts into the compensation network because those are the only components that have changed value between the Rev C amps and the FE's I now have playing.
...
I think changing the compensation from Mauro's carefully selected values is an attempt to fix something that wasn't broken.

As I've already wrote you in private if you have such a doubt you should try Rev C compensation first and verify if problems you ear are still there or not.

Obviuosly without changing other parts, one thing at a time, or your result will be impaired...

Until you don't make such test playing the guessing game with compensation is useless.

My opinion is not infallible, but I have a pretty good record of selecting components and mixing them effectively to make the MyRef sound better.

It's the same for my opinion...I've reverted it several times when I've found it wrong and I've also a good track of parts selection but this is not a race...

You already know that I have great consideration for your opinions and earing abilities but I would like to read them expressed less boldly and with less (wrong!) speculations on my intentions...

With the FE amp, I am not sure why a part such as the Riken resistor that obscures some of the signal sounds "better" than parts that do not.

Tom, did you try to reverse them?

In one direction they sound closed and muffled and much more open on the other.

Nevertheless they are less detailed and more musical (like all carbon film) than a thin film resistor and also a bit colored.

The MK132 I've suggested you to try is more neutral and a bit more detailed but a bit less musical.

Also the carbon film in the industrial BOM fares well, in fact all three sound good (but different).

I've selected the Riken as the suggested audiophile part for its musicality at a reasonable price, and being smoother than plain carbon films, not for (supposed) masking abilities.

That indicates to me that the amp is producing sound that needs to be flavored to sound good. I would prefer to fix whatever is wrong than try to cover it up. I'll liken using the Riken to covering everything with a layer of maple syrup that obscures natural flavors.

This is speculation, I've already said that it was not the case, in practice is like saying I'm lying...

I've spent hundreds of hours and euros trying to find the most transparent parts and excluding 'nice' colored ones.

The hard edge you ear is present in some parts combination, I've had the same problem and after trying to change pretty all parts except True Coppers, I've finally changed them, in my system they were the (main source of) the problem.

You say that if TCs sounded good on Rev C they should sound equally good on FEs... not necessarily, the FE are much more detailed and transparent so you're earing problems that were previously masked, IMHO.

Tom, your opinions are welcome and I read them with interest, I just would like, at least, to be trusted when I explain my choices...
 
The cause of lack of high frequencies in the FE is the Riken resistor at R10. It obscures a lot of high frequency detail. ... With the Riken at R10, it sounded more musical but muffled and flat. ... If you really want high frequency clarity and resolution with expansive and deep soundstage, try a Caddock MK132 at R10.

I tend to agree - much the same effect of the Rikens can probably be obtained at lower cost using an Allen-Bradley CC or Panasonic ERDxx CFR at R10, which I've tried in Rev C. To regain the micro-detail, any non-magnetic metal-film like a PRP, Dale RLR or Holco does OK.

I can already say with significant confidence that 27pF is too high a value for C34. I have installed 22pF on one channel, and it is far more musical, with less edge than the other.

With C34 at 22pF, one can increase R43 to 27k to keep RC constant. What I've tried in a Rev C is C34=15pF and R43=27k, without any obvious regression in subjective overall audible sonics (but possibly slightly improved bass/mid-bass definition, compared to 22pF/22k). Something similar is probably worth trying in an FE also.

Also, the quality of C34 is not as important as the quality of R43 - a fully non-magnetic resistor like a PRP seems to improve presentation.
 
Siva, have you tried the Caddock MK132 at R10?

Nope - presently, my only source for anything boutique-oriented is EBay. I'll give it a shot if there's some consensus that it's audibly better than good non-magnetic MFRs.

Related: In addition to 15mm pitch FKP3 at C4 (which is audibly more detailed than MKP10), there's also 5mm pitch ERO MKP1837 film/foil available in 100nF/100V - which is the next thing on my to-do list.
 
A lot of tweaks to consider. For instance , how big change would be if I change C2 in my speaker's crossover ? Currently there is clarity cap esa.
It all could change and the "cooking ingredients" approach is valid for sure.
 

Attachments

  • ER18DXT-final-crossover.gif
    ER18DXT-final-crossover.gif
    9 KB · Views: 322
Dario, I tried to subtly suck you in but you didn't take the bait on post 1375.🙄 Do you think it would be helpful to create a small Google doc where significant choices in key areas could be listed? Not a "what's best" rating - just a collection point for valid alternatives. There is a lot of information passing by quickly and a small spreadsheet might clarify things a bit.

I just did a few searches for some of Siva's substitutions and found, for example, the eBay listings for the Allen-Bradley pieces don't include the exact value. Some components on the FE most likely have an acceptable value range which could also prove helpful if posted on that document.

Just a little clarification on my Subbu/BrianGT comment. Using the same source (Jriver/PC) I'm able to switch between three DIY DACs completed over the past few years. The differences in presentation vary enough that what I might consider necessary to adjust on a particular amp is significantly tied to the character of the signal chain that feeds that amp. If one has settled on or prefers a particular source this might not matter so much. However, the impact of the new DAC in that combination convinces me I will have to make that choice/preference before I proceed to any "tuning " of the upcoming FE final build. Otherwise I could end up chasing my tail - even more than is my normal pattern.😉
 
Last edited:
A lot of tweaks to consider. For instance , how big change would be if I change C2 in my speaker's crossover ? Currently there is clarity cap esa.

A Mundorf Supreme (plain ones, not fancy oily ones) will be a nice upgrade for a reasonable price.

It all could change and the "cooking ingredients" approach is valid for sure.

This 'Cooking' metaphor is abused, it has some merit but unless you tend to choose the most neutral part it will lead to disaster...

I tend to agree - much the same effect of the Rikens can probably be obtained at lower cost using an Allen-Bradley CC or Panasonic ERDxx CFR at R10, which I've tried in Rev C. To regain the micro-detail, any non-magnetic metal-film like a PRP, Dale RLR or Holco does OK.

Hi Siva,

nice to see you on this thread 🙂

I'm not sure your comment is spot-on, though.

Tom was lamenting that he thinks Rikens mask HF content so an AB would be even worse...

And, BTW, I would never use a carbon composition resistor...they drift too much for my taste.

Do you think it would be helpful to create a small Google doc where significant choices in key areas could be listed? Not a "what's best" rating - just a collection point for valid alternatives. There is a lot of information passing by quickly and a small spreadsheet might clarify things a bit.

Maybe but, sadly, I didn't take notes during my listening tests so I'm mostly using my memory.

Some components on the FE most likely have an acceptable value range which could also prove helpful if posted on that document.

Carbon composition resistors like the ABs are not suitables for the FE, IMHO.

However, the impact of the new DAC in that combination convinces me I will have to make that choice/preference before I proceed to any "tuning " of the upcoming FE final build. Otherwise I could end up chasing my tail - even more than is my normal pattern.😉

Absolutely and it's for this reason that I've always tried to select the most neutral part and not the one that sounded 'nicer'.

Obviously this is not always possible... 😉
 
Rikens Vs Caddocks

Just for the sake I've just resocketed R10 and compared again Rikens and Caddocks (both correctly oriented).

Both sound great and while with some (tiny) difference in timbre I can't ear any lack of HF content on Rikens.

Nevertheless MK132s are more detailed and also a bit less grainy (and so less harsh) than Rikens.

Also, soundstage is deeper with Caddocks.

Pretty coherent with what I was remembering.

I still think that Rikens are an excellent choice with a good price/perfromance ratio but probably was an error not to indicate clearly MK132 as a more detailed and costly alternate.

So I've just updated BOM (Comment row) and new boards will have also 5mm lead spacing to better accomodate MK132.
 
To Bob and Dario: thanks for the kind words of support and encouragement. I know you guys appreciate my efforts, but it's still nice to read that. Your experience and suggestions are just as important to me. Dario, your patience and pool of knowledge has been essential at every stage. This is not the first time we've disagreed, but we always respect each other's opinion. Cool that you will add pad for MK132, although bending leads is not so difficult. If you're making changes, you might consider shrinking the hole for the outboard C13 ground wire. I doubt that anyone will ever use a lead that size.

I appreciate everyone's response and input. Siva, thanks for the technical stuff, too. Our brotherhood of MyRef's continues to grow and prosper with a larger body of knowledge and experience.

Lots of things to try. Right now, I'm waiting on parts from HiFi Collective (Amtrans resistors for R10) and Mouser (more values of C10 and C34). I also have a couple candidates for C9, including one nobody has tried yet. It's so cold here that there is nothing else to do but fool with stereo and read books. The air is so dry that I'm a little afraid of catastrophic static discharge.

No one has answered my question about amp instability. Since that is one major concern of fiddling with comp values, how would I recognize it without test equipment? Will sound suffer? Chip overheat? Something explode?

Voxx, any Mundorf cap is appropriate in series tweeter circuit. They are colored, but in a good way for tweeters. The quality does improve as the price goes up, but in smaller increments. You can also use a smaller value as a bypass for a cheaper large cap. Five to ten percent of the total required value is a good starting point. Don't spend more on the cap than the tweeter!

Peace,
Tom E
 
Jaw dropping prices ! 5.88 £ Mouser ! What ?

yes, it was the reason why I initially excluded them.

BTW now you can have them from PCX at same price of Rikens, 3.95 US$

It seems Rikens price increased by a not small margin, it were 2.70 US$.

At this prices maybe Rikens, also out of production, should leave BOM...

Amtrans, not cheap but a possible R10 contender for madisonears issue, it's supposed to be better than Riken Amtrans AMGR carbon film Resistors,0.75 watt and 2 watt

Amtrans and Rikens are very similar, the newer ones are a bit less colored but they have the same level of detail.

In My DCB1 I've tried PRP, Riken, Amtrans, Caddock, Vishay bulk foil and the winning combo was Vishays on signal and Mk132 as shunts, all the other were not as detailed or transparent.

To Bob and Dario: thanks for the kind words of support and encouragement. I know you guys appreciate my efforts, but it's still nice to read that.

You're welcome 🙂

Cool that you will add pad for MK132, although bending leads is not so difficult. If you're making changes, you might consider shrinking the hole for the outboard C13 ground wire. I doubt that anyone will ever use a lead that size.

With those new prices it's starting to be necessary...

I've followed your suggestion on pad size.

No one has answered my question about amp instability. Since that is one major concern of fiddling with comp values, how would I recognize it without test equipment? Will sound suffer? Chip overheat? Something explode?

Sincerely I don't have the knowledge to answer.. the only thing I expect from an oscillating amp is (much) more heat.
 
Sincerely I don't have the knowledge to answer.. the only thing I expect from an oscillating amp is (much) more heat.

I'm no expert either, but I agree that without instrumentation, heat would be the best way to find an oscillation. My question for any experts is, "Which element is likely to get hot?" It could be the LM318, it could be the LM3886, maybe even power supply elements. Maybe even R3.

It would be nice to be able to see the oscillation, but there are scopes that don't measure high enough frequency to show many oscillations. My old tube scope isn't effective at high frequencies.
 
Question for Siva, or anyone else who has more technical insight than I:

What is the purpose of adjusting R43 along with C34 to keep the RC constant the same? What then is being changed, other than the respective component values? I thought the purpose of changing C10 and C34 values was to improve/change the sound.

Just when thought I had a cursory grasp on this compensation stuff...

I'm going to start with 27pF C10 and 18pF C34, which is midway between original Mauro comp values and those derived by Dario and his advisors, but leave R43 at 22k.

Peace,
Tom E
 
Long term plan, I want to slowly populate my second set of boards with the best components you guys can identify. The Black Gate FK for C9 is no longer made and becoming hard to get now, any alternative suggestions?

PCX still has the Black Gate PK in 220 uF for about a dollar. The issue with the PK is that it is rated at only 4 volts which is enough for normal operation at C9, but too low if a fault is present. They do sound very nice. I like Dario's suggestion that you buy a few extra, in case a fault develops. Also, use a generic but higher voltage cap for initial start up, just to make sure you don't have a fault.

Dario has looked at many alternatives. I've looked at a few as well and I come to the conclusion that the BG PK is better than most available alternatives.

Jac