Naked Vishays (TH) or Susumu RGs (SMD).
Apart the cost factor (RGs less than a € Vs 14€ the Vishays) the RGs sounds almost as detailed and neutral with added musicality.
In R7 RGs are fantastic too.
But RGs are not so good for shunt, I've preferred KOA RK73H (SMD).
As shunts also MK132 are good but a bit closed and strangely less full.
A nice (lesser) alternative for shunt is a plain carbon film.
I think you can understand why I choose to go for full SMD signal path resistors for final boards, isn't it?
Now it's time to sleep for me...
If the slight difference in Vsupply is affecting the apparnet volume and also affecting the measured values of SPL, then that suggests that one or both the amplifiers are clipping some or many of the transients.
That is not the way to compare amplifiers.
Much more likely is that the Vsupply is NOT affecting the volume levels.
The difference is likely due to differences in gain between the two amplifiers.
Check that accurately before wasting any more time fiddling with other causes.
That is not the way to compare amplifiers.
Much more likely is that the Vsupply is NOT affecting the volume levels.
The difference is likely due to differences in gain between the two amplifiers.
Check that accurately before wasting any more time fiddling with other causes.
If Rikkens are coloured like all my resistors then why does using extra ones make the performance worse?
I suspect you mean the Rikkens are introducing some form of distortion.
If this is the case then no wonder you want to minimise the number of Rikkens used.
If I were finding the same results using Rikkens I would be throwing them in the bin.
I suspect you mean the Rikkens are introducing some form of distortion.
If this is the case then no wonder you want to minimise the number of Rikkens used.
If I were finding the same results using Rikkens I would be throwing them in the bin.
McBob,
if you are finding that a small change in a small cap is changing the sound you hear, then that tells me that it is time to properly measure and compensate the amplifier.
I would not be playing the guessing game you are following.
Clave has done the guessing game from the start. Now it appears he even recommends what to most of us would be "badly specified" resistors to deliberately create an inaccurate amplifier.
if you are finding that a small change in a small cap is changing the sound you hear, then that tells me that it is time to properly measure and compensate the amplifier.
I would not be playing the guessing game you are following.
Clave has done the guessing game from the start. Now it appears he even recommends what to most of us would be "badly specified" resistors to deliberately create an inaccurate amplifier.
Hi Andrew,
I'm not talking about performance but about tonal balance/timbre.
What I mean is that Rikens have a full tonal banlance, towards being slightly closed.
If you add more the effect tends to be cumulative (so they're coloured).
Too much of them and sound could became closed.
Sure...
In fact I'm recommending the opposite...please read my posts...
What I'm recommending is the use of high quality, very low ppm industrial precision SMD resistors (Susumu RG series) and high quality thick films, low ppm (for a thick film) industrial resistors for shunt position (KOA RK73H).
In fact final boards will have all SMD signal path resistors.
This is libel... repeated.
That is deliberately FALSE and malicious.
You affirm that:
- I'm recommending a resistor which is 'badly specified'
- I'm doing it to deliberately create an inaccurate amplifier
2 is false too since my recommendation is a different resistor and you pretend to assign me a goal that it's not mine.
Rikens are not bad resistors and used sparingly in TH designs they can correct other unbalanced parts.
For sure I prefer more transparent resistors but sometimes Rikens can be great.
I do agree and it was one of my goals (transparence) from the start.
It's for this reason I'm not recommending Sonicaps (musical but coloured and they lack detail).
It's for this reason that final boards will have only SMD signal path resistors, since I've found some excellent industrial ones that measure and sound great with little or no colouration.
It's for this reason I'm trying a lot of different components and, when possible, I compare them to the same circuit without them.
Hi Andrew, as stated in post #356, I agree there has to be some interaction/influence caused by the extra components in the integrated build. By next week I will have a simpler setup with all four toroids more closely matched.
I would be willing to, and interested in using a standardized method/process in this beta stage (as mentioned - even including speakers). That's why there is a little disappointment that more veterans didn't join the beta, and many of those with professional lab equipment who did sign on haven't progressed as hoped.
As you can tell from my posts, I feel Dario has designed something that is a cut above what I have heard from many LM3886 amps. Technical/tweaking issues being an ongoing process - more hands-on time by more builders in surely needed. Siva seems to also be on the verge of a major development and I suspect even with all his excellent work, a lot of field testing will need to be done. I don't mind some trial and error as long as the results warrant the costs. I've made some major "blind purchases" during this beta but so far they have all been worth the time and money.
As of now, my "lab" is limited to a DMM, RS level meter, heat sensor, a flashlight and my ears. If there are suggestions for methods and/or equipment to help dial in more professionally (and don't require winning the lottery) I'm all ears - pun intended.
I would be willing to, and interested in using a standardized method/process in this beta stage (as mentioned - even including speakers). That's why there is a little disappointment that more veterans didn't join the beta, and many of those with professional lab equipment who did sign on haven't progressed as hoped.
As you can tell from my posts, I feel Dario has designed something that is a cut above what I have heard from many LM3886 amps. Technical/tweaking issues being an ongoing process - more hands-on time by more builders in surely needed. Siva seems to also be on the verge of a major development and I suspect even with all his excellent work, a lot of field testing will need to be done. I don't mind some trial and error as long as the results warrant the costs. I've made some major "blind purchases" during this beta but so far they have all been worth the time and money.
As of now, my "lab" is limited to a DMM, RS level meter, heat sensor, a flashlight and my ears. If there are suggestions for methods and/or equipment to help dial in more professionally (and don't require winning the lottery) I'm all ears - pun intended.
Last edited:
Thought I'd comment on some listening going on. In most cases, if the gain is the same, the device that sounds clearly louder has more sound coloration due to the energy from additional harmonics. What should happen in the improvement process is that the sound will gradually seem reduced in SPL, while clarity in sound image should gradually increase. If the device that sounds less loud does not have enough micro details, it will sound dry an a bit lifeless. As improvements occur, the life will gradually grow until the image focus just seems live.
soongsc, in general I agree with your comments. It is just in this case, though maybe not conveyed accurately by the numbers I posted, (as you know, db changes are not linear) there has to be something really unbalanced in the test method/platform I used. The volume differences were too large to be the result of only individual on-board component choices. I'll have a better and more direct setup next week.
The reduced SPL and improved clarity/musicality are exactly what I heard with the alternate compensation I reported on yesterday. I hope some other MyRef builders will try those substitutions and report back.
The reduced SPL and improved clarity/musicality are exactly what I heard with the alternate compensation I reported on yesterday. I hope some other MyRef builders will try those substitutions and report back.
Last edited:
Would it measure nearly 5dB different?
That was the point that Bcmbob was making it sounded louder and when he measured it the measurements showed it was louder.
I much prefer Mcbob (sounds so Scottish).
Well, bcmbob is using an SPL meter for acoustic measurements, with that method, considering the slight difference, you never know whether the ambient noise had any effect or not. I basically measure the frequency response of the amp, they are the same.
Would there be a reason for it to measure different if you just change power supply caps on the DC rail in a power supply?
Of course there could be other things that effect loudness, some could be determined by looking at the schematic of both.
Last edited:
I knew about the reduction to 47ohm (i'm using this value) but not completely shorted (0R). Thanks anyway!
Regarding mods i followed Soongsc sugestion to remove all decopling capacitors around 3886 and installing one 2.2uf capacitor between rails and i really liked the improvement.
Regarding mods i followed Soongsc sugestion to remove all decopling capacitors around 3886 and installing one 2.2uf capacitor between rails and i really liked the improvement.
Last edited:
State of the art
I'm still refining some choices but so far:
C30 became again a FKP2 or possibly a KP1830
C7 will probably be a KP1830
PS and input shunt resistors will stay TH, Caddock MK132.
I've spent a lot of time determining parts best direction (if you don't trust it, simply ignore this part) and, while subject to change, you can find attached what I've found so far.
I've finally managed to try Audyn True Copper input caps... they sound fantastic, as detailed as Zn with added musicality and soundstage depth.
They'll be the recommended cap for top performance, Zn for high performance and I've found a cheap cap that, while inferior, is balanced and good sounding (Vishay/Sprague V-730P/V-735P) for tight budgets.
I've updated the PCB (preview attached) to suit this choices:
Is there any interest?
I'm still refining some choices but so far:
C30 became again a FKP2 or possibly a KP1830
C7 will probably be a KP1830
PS and input shunt resistors will stay TH, Caddock MK132.
I've spent a lot of time determining parts best direction (if you don't trust it, simply ignore this part) and, while subject to change, you can find attached what I've found so far.
I've finally managed to try Audyn True Copper input caps... they sound fantastic, as detailed as Zn with added musicality and soundstage depth.
They'll be the recommended cap for top performance, Zn for high performance and I've found a cheap cap that, while inferior, is balanced and good sounding (Vishay/Sprague V-730P/V-735P) for tight budgets.
I've updated the PCB (preview attached) to suit this choices:
- SMD input and feedback resistors
- Changed LM318's compensation components to 0805 SMD format
- adapted the input cap to accomodate True Copper/Zn/V-73xP
- created a library with Caddocks resistors (MK132, MP915, MP930)
- Output resistor is now a MP930
- Shunt PS resistors are now all Caddock (MP915, MK132)
- Shunt input resistor is now a Caddock MK132
- increased space between ground planes, pads and traces to ease manifacturing.
- PCB now supports directly Rev A, B, C, alternate compensation.
Is there any interest?
Attachments
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.