Yeah, that will work in this case. However, in general, you'll have to analyze the effect of the load impedance (here it's fairly high) as well as the source impedance (fairly low, depends on your CDP or DAC source) when you change the values of a passive filter.
Clarky,
The RC filter is simple and follows that formula. You can keep the same frequency with any combination of component values where the RC time constant is kept the same.
Your original RC time constant for 3k3 & 220pF is 0.7us. (3300 * 0.00000022)
F-3dB=219kHz
BUT,
that RC filter requires the load impedance to be close to infinitely large and requires the source impedance to be close to infinitely small. The unity gain opamp after the RC satisfies the near infinite load.
The R value should be adjusted to take account of the source impedance (Rs).
eg if you know the output impedance of the previous active stage was 5ohms and you had a 220r resistor tacked on to isolate the active stage from the cables and load that followed, then the effective source impedance would be Rs= 5+220 = 225ohms.
R = 2k2 = Rinsert+225ohms. Note that adding the 220r also swamps variations in the 5ohms output impedance.
Rinsert is thus 2k2-225=1k975
Use 2k and recalculate you RC using Rinsert+Rs.
The RC filter is simple and follows that formula. You can keep the same frequency with any combination of component values where the RC time constant is kept the same.
Your original RC time constant for 3k3 & 220pF is 0.7us. (3300 * 0.00000022)
F-3dB=219kHz
BUT,
that RC filter requires the load impedance to be close to infinitely large and requires the source impedance to be close to infinitely small. The unity gain opamp after the RC satisfies the near infinite load.
The R value should be adjusted to take account of the source impedance (Rs).
eg if you know the output impedance of the previous active stage was 5ohms and you had a 220r resistor tacked on to isolate the active stage from the cables and load that followed, then the effective source impedance would be Rs= 5+220 = 225ohms.
R = 2k2 = Rinsert+225ohms. Note that adding the 220r also swamps variations in the 5ohms output impedance.
Rinsert is thus 2k2-225=1k975
Use 2k and recalculate you RC using Rinsert+Rs.
Last edited:
Decided to add volume and source selector board with remote control to fe (love this amp!). But it is a bit too tight in a box for everything. What if I put this boards to separate box, call it pre? What I d like to do, since it is remotely controlled, I d use pre as a head unit which Controls on off status of fe? Power cable to pre----vol,source,on off, ---- power cable to fe. So fe will be powered via pre. In this case no extra wires in amp, Less interference and neat solution.
Is anything to consider like implementation technique or something ?
Is anything to consider like implementation technique or something ?
I would definitely *not* power the MyRef through a trafo in a pre-, if that's what you're proposing. Use separate trafos, rectifiers, filter caps in each of the boxes to power each circuit (FE and pre-) separately. There's still a risk of hidden ground loops, but that can be tackled with some care in grounding and ground isolation.
I just tried a polystyrene axial at C12 (220 pF) in Rev C and there appears to be no sonic regression - it actually appears to improve micro-detail slightly, but YMMV. I have K71-5 Russian polystyrene at C13, so I have a hunch that it helps to have the same dielectric at both C13 and C12, i.e. it makes the audible sonics more natural/cohesive.
According to this hunch/theory:
With any Polypropylene film/foil at C13, Wima FKP2 (default) should be optimal at C12.
With any Polyphenylene Sulphide at C13, Panasonic ECHU should be optimal at C12.
With any Polycarbonate film/foil (say KC1850) at C13, Evox CFR or Wima FKC2 should be optimal at C12.
With PETP Russian K73-16 at C13, Wima FKS2 or similar should be optimal at C12.
I haven't yet tried most of these combinations, but may give them a shot in some future builds - I just received my EBay 1.8uF K73-16s yesterday, which I'll evaluate and give an update later.
Hi all
I have just finished my my_ref , and it is sounding amazingly good. Too early to say anything definitive.
The reason for this post is that I am a little puzzeled over the big fuzz about C13.
Why not use the best capacitor ever made??
A short piece of wire!
The input offset voltage is in the mv range and the input bias current is max 500nA
Of course your preamp must be @ 0 DC voltage, but I suppose all the preamps we are using are just that.
Just for the record, I have just tried it and there is no more DC at the output than with C13.
Koldby
I have just finished my my_ref , and it is sounding amazingly good. Too early to say anything definitive.
The reason for this post is that I am a little puzzeled over the big fuzz about C13.
Why not use the best capacitor ever made??
A short piece of wire!
The input offset voltage is in the mv range and the input bias current is max 500nA
Of course your preamp must be @ 0 DC voltage, but I suppose all the preamps we are using are just that.
Just for the record, I have just tried it and there is no more DC at the output than with C13.
Koldby
Hi all
Why not use the best capacitor ever made??
A short piece of wire!
The input offset voltage is in the mv range and the input bias current is max 500nA
Sure, it's been done and it works fine as long as all your sources have 0V DC offset. If you accidentally connect one with say 50 mV offset, it gets multiplied by 28 = 1.4V and your woofer may make some sounds like onions frying on a skillet and get cooked within a few minutes. The speaker DC protect is designed to prevent accidents like these, but there's a grey area with output offsets in the region of 0.5 to 1V where it may not trigger.
Last edited:
Cooked woofer with 1,4 V???
That is 0,245 Watt.
What woofer cannot withstand that?
Seriously the amount of DC from your pre has to be significant in order to fry anything.
There is a lot of DC coupled power amps out there and if you want the best sound, I personally would remove C13.
Koldby
That is 0,245 Watt.
What woofer cannot withstand that?
Seriously the amount of DC from your pre has to be significant in order to fry anything.
There is a lot of DC coupled power amps out there and if you want the best sound, I personally would remove C13.
Koldby
Fine
Two reasons:
- safety: you don't know in advance what will be branched to amp's inputs, if a high enough amount of DC is presented at input you would fry your woofers
- according Penasa (the My_Ref designer) the input filter concur to set current on LM318's input, I don't have the competence to agree or deny but I choose to follow designer's lead.
I still feel the potential for frying you speakers is non existing. There is a dc protection circuit.
If missing the input filter would change the input current and change the sound one would hear it, yes? I have not yet listened to the amp without C13, only checked if it worked.
I will report back.
The reason I build this amp was a freindly challenge by fellow member bcmbob as a response to my ravings about a Golmund Telos 250 clone in another thread:
"
koldby - You have traveled a path quite similar to mine. I started with the RC version of Dario Inserra's MyRef Fremen Edition and built several amps for comparison. Though not as extensive as your collection, my experiences dictate that I must propose you haven't yet heard "The Best" (whatever that means ) of what DIY can produce.
If you still have some time, interest and leftover solder, I invite you to keep an eye on the Final Edition Build Thread and consider participating in an upcoming second group buy. So far, that phrase "Best I Have heard" has come from all the current builders of Mr. Inserra's efforts.
Would love to hear your impressions if and when you take up this friendly challenge
Well I did excatly that, and I am glad I did.
My first impression is that this is the only other amp that can challenge the GM.
Not that it sounds like it, but it has some very nice features.
Actually to my ears and on my speakers it sounds more like a very good traditional tube amp. Very sweet and with a very big soundstage especially depth.
It cannot match the GM in shear power of course given its >250 Watt rating.
The My_Ref clips here on very demanding material.
It cannot match the exquisite highs of the GM and is a not as focused as the GM.
On the other hand it is so sweet and warm and has better depth and the room is easier to hear.
If you listen to a violin group in a symphony orchestra the GM separates the violins better than the My_Ref and the air is more prominent.
But the My_Ref gets the romantic character of wood and string even though it paints with a broader pensel.
Because of the good summerweather here, I am not using much time on listening inside, but I am very much looking foreward to hear a lot more music on these amazing amps. Looking at the schematic, I wouln´t have thought it could be this good.
I would call it a draw between the to champs at this point!
Thanks for the challenge bcmbob!!
Koldby
If missing the input filter would change the input current and change the sound one would hear it, yes? I have not yet listened to the amp without C13, only checked if it worked.
I will report back.
The reason I build this amp was a freindly challenge by fellow member bcmbob as a response to my ravings about a Golmund Telos 250 clone in another thread:
"
koldby - You have traveled a path quite similar to mine. I started with the RC version of Dario Inserra's MyRef Fremen Edition and built several amps for comparison. Though not as extensive as your collection, my experiences dictate that I must propose you haven't yet heard "The Best" (whatever that means ) of what DIY can produce.
If you still have some time, interest and leftover solder, I invite you to keep an eye on the Final Edition Build Thread and consider participating in an upcoming second group buy. So far, that phrase "Best I Have heard" has come from all the current builders of Mr. Inserra's efforts.
Would love to hear your impressions if and when you take up this friendly challenge
Well I did excatly that, and I am glad I did.
My first impression is that this is the only other amp that can challenge the GM.
Not that it sounds like it, but it has some very nice features.
Actually to my ears and on my speakers it sounds more like a very good traditional tube amp. Very sweet and with a very big soundstage especially depth.
It cannot match the GM in shear power of course given its >250 Watt rating.
The My_Ref clips here on very demanding material.
It cannot match the exquisite highs of the GM and is a not as focused as the GM.
On the other hand it is so sweet and warm and has better depth and the room is easier to hear.
If you listen to a violin group in a symphony orchestra the GM separates the violins better than the My_Ref and the air is more prominent.
But the My_Ref gets the romantic character of wood and string even though it paints with a broader pensel.
Because of the good summerweather here, I am not using much time on listening inside, but I am very much looking foreward to hear a lot more music on these amazing amps. Looking at the schematic, I wouln´t have thought it could be this good.
I would call it a draw between the to champs at this point!
Thanks for the challenge bcmbob!!
Koldby
There is one other secondary reason for using a high-quality C13. You may not have too much control over the DC blocking cap (if any) in an unmodified source CDP, DAC, Preamp, etc. Often it's electrolytics (say a pair of back-to-back 220uF Silmics, as in many Marantz CDPs). By putting a small (~1 uF) high-quality film/foil cap in the MyRef, a series combination of a large electrolytic (~100 uF) and the small film/foil cap (~1 uF) is created in the signal path.
What is the sonic character of this series combination governed by? Almost entirely by the *smaller* cap, which is the high-quality film/foil cap. If you had a wire at that location, the sonic character would have been entirely that of the back-to-back Silmics in the source - which is much worse than almost any film/foil cap (check the recommended mods to the CD63/67 - almost everybody recommends getting rid of the Silmic DC blocking caps in the signal path).
A lot of basic physics is non-intuitive, including the notion that a feather and a ball of lead will accelerate identically due to gravity in a vacuum.
A series connection of a small-valued cap and a much larger-valued cap, is essentially the same as the small-valued cap *alone*. You're missing the fact that the same current charges the large cap to a much smaller delta-V, and the smaller cap to a much larger delta-V. For a ratio of 100:1 in capacitance, the voltage change is in the ratio of 1:100.
However lousy the large cap, it almost completely drops out of the picture when a small high-quality cap is in series with it. Non-intuitive, but true nonetheless.
That may be true, but has nothing to do with caps...
ANYWHERE I red about dc blockin caps, always had different soundwhen 2 different kind was in series. And they actually tested, not just thought like you.
To add something to the metaphore, they may fall the same speed, but you need a totally different amount of force to stop and turn them around. Just like the sound is not a simple 1 way movement, it constatnly changes.
Last edited:
Yes, it sound different.
The DC coupled My_Ref have excellent HF and MF but it's somewhat 'wrong' in the medium/bass, bass regions.
I usually use DC coupling as a reference when evaulating caps for C13.
Wow! Incredible review
Which BOM did you use?
Can you post some pics?
Well I have just reused a chassis that formerly housed the Pass F-5 so it is not its final home.
In the rack you see from below:
GM Telos 250
Aikido preamp with Marconi B65 metalbase tubes.
TDA1541a NOS DAC with audio widget USB to I2S converter.
My_ref_Fremen
Improved Bow Technolgies ZZ-1
I used mostly the Mouser bom and I will be looking for parts to upgrade.
What about K71-4 as C7? Or NOS phillips KP types?
Have you tried Silver Mica as C32 and C30?
If C13 stays in have you tried to bypass C9?
I know the DC gain wil raise to 28, but perhaps the DC fault from LM 318 is small enough. Alternatively you could inject DC at LM318 pin 2.
Koldby
In the rack you see from below:
GM Telos 250
Aikido preamp with Marconi B65 metalbase tubes.
TDA1541a NOS DAC with audio widget USB to I2S converter.
My_ref_Fremen
Improved Bow Technolgies ZZ-1
I used mostly the Mouser bom and I will be looking for parts to upgrade.
What about K71-4 as C7? Or NOS phillips KP types?
Have you tried Silver Mica as C32 and C30?
If C13 stays in have you tried to bypass C9?
I know the DC gain wil raise to 28, but perhaps the DC fault from LM 318 is small enough. Alternatively you could inject DC at LM318 pin 2.
Koldby
Attachments
Problem, guys
Been away for some time.
Last week I slammed the door on the tv stand and suddenly heard weird high pitching noise.
Switched off unit (I am using b1 for now, till ldr based pre is ready). Disconnected everything.
Few days ago I got back and just turned on amp with small test speaker - hum on both channels.
Today I rewired signal with teflon isolated wire, but hum is still on.
I have checked dc with cheap mm (tomorrow will bring fluke and test it with another). Put mm on 200m, showed right channel 0.4 and fluctuating, goes down to 0 and up to 0.4. Switched off, mm peaked at 13 .
Left channel is on and goes about 0.5 and 0.4 ( 4 or 5 mv ? ) and on turn off peak at 30...
did not expect anything like that, wanted to finish up with on off switch and grounding and close box and forget... but it seems it is not the end...
Been away for some time.
Last week I slammed the door on the tv stand and suddenly heard weird high pitching noise.
Switched off unit (I am using b1 for now, till ldr based pre is ready). Disconnected everything.
Few days ago I got back and just turned on amp with small test speaker - hum on both channels.
Today I rewired signal with teflon isolated wire, but hum is still on.
I have checked dc with cheap mm (tomorrow will bring fluke and test it with another). Put mm on 200m, showed right channel 0.4 and fluctuating, goes down to 0 and up to 0.4. Switched off, mm peaked at 13 .
Left channel is on and goes about 0.5 and 0.4 ( 4 or 5 mv ? ) and on turn off peak at 30...
did not expect anything like that, wanted to finish up with on off switch and grounding and close box and forget... but it seems it is not the end...