My_Ref RevO

Status
Not open for further replies.
From Mauro Penasa home page:

http://www.webalice.it/mauro.penasa/Archivio_Progetti_audio.html

Apart Evolution presentation (Google translated) It's possible to read some interesting documents about the My_Evo.

In addition the Italian magazine 'Costruire HI-FI', issue #116 published a simplyfied schematic (no compensation network, no limiter schematic, no values).

Obviously I can't post it but the My_Evo main topology is identical (apart the DC-Servo and the second pump) to the My_Ref Rev C one.

So, probably, the My_Evo is basically a RevC with a different compensation network, better LM318 PS (14V), a 28V voltage limiter, a second current pump, a DC-Servo and a new and improved PCB.

So, is it possible to made a My_Ref RevC that takes advantage of some of the feaures presented by the My_Evo without infringing Mauro's IP?

IMHO it could be possible for quite all features:

  • a more stable amp using SMD
  • raise LM318's PS voltage with voltage limiter
  • LM318 compensation
  • Better PS for LM318
  • DC-Servo
  • Parallel current pump (already done and tested by a forumer)
The most difficult to design is the DC-Servo so, for now it would be left out (as the second pump) but obviously we should start from something so I've designed (starting from Russ White's Eagle schematic) a brand new (and untested) PCB with these features:


  • bigger caps (16mm diameter) for C1, C2, C9
  • large use of SMD components
  • all small caps and some critical resistors are still through-hole
  • the feedback resistor (R7) can be through hole or MELF
  • very small paths thanks to SMD
  • new and untested TL431 14V shunt PS for LM318
  • zener limiter based on schematic 5b from this link (which seems pretty similar to My_Evo one)
  • a double diode bridge like in most gainclones (tested by Suburra with great results)
  • If we find a better compensation scheme the PCB will still be valid (only values, not positions can change)
If you look at the attached schematic a resistor in the feedback net has been removed (R42), like Mauro himself suggested.

Both PCB and schematic are untested and prone to errors so any comment, fix, suggestion on both is welcome and appreciated.
 

Attachments

  • RevO - Schematic.jpg
    RevO - Schematic.jpg
    181.4 KB · Views: 801
  • RevO - PCB.jpg
    RevO - PCB.jpg
    325.5 KB · Views: 771
Last edited:
the original myref thread is such a dinosaur than i would rather look the part of a bum asking it here but... would it lower stability or raise the temp of any component when r42 is replaced with a jumper?

btw, i'm very interested in this pcb, clave. will definitely be making it my next amp when it becomes available. keep up the good work! :)
 
Dario,
I understand the goals of this project in terms of amp stability and physical component changes. Can you give some information about the anticipated sonic differences , if any, between this design and the current version of the MyRef?

Also, I noticed you are proposing the same 10,000 uF caps, Would increasing that value provide more head room?
 
Last edited:
Member
Joined 2005
Paid Member
would it lower stability or raise the temp of any component when r42 is replaced with a jumper?
...
keep up the good work! :)

You're welcome PreSapian :)

In fact, according to Mauro, removing R42 should improve stability.

Double diode bridge similar to this? It is useful with 2 separate secondary windings, and see how the ground is taken from the bridges.

Exactly, and like you pointed out my schematic is wrong!

Thanks, that's the kind of feedback I'm looking for...:cool:

I'm updating the schematic and I'm redesignig the associated part of the PCB...

Can you give some information about the anticipated sonic differences , if any, between this design and the current version of the MyRef?

If the PCB will be correctly designed we should exepct a lower THD and noise floor.

According to Suburra the double bridge gives more contrast, bass is more articulated, a general refinement.

I really don't know what to expect from the new LM318's PS, I hope to test it on the RevC next week.

If and when the new compensation scheme will be fully investigated and tested we should expect much more refined highs and a much tighter bass, like in the My_Evo.

Also, I noticed you are proposing the same 10,000 uF caps, Would increasing that value provide more head room?

It's possible but also the My_Evo has 10000uF smothing caps for each pump.

The PCB is already tight and full with 30mm caps, in this moment I can't imagine how to stretch in 35mm ones (and bigger value caps are mostly 35mm...)
 
Last edited:
Why have you adopted a dual bridge rectifier for use with a centre tapped transformer? They don't work together !
What is VR2? It can't be a +ve regulator turned upside down !

If adeqaute smoothing cap values result in too large a footprint, then move the main capacitance off board and put in adequate decoupling.

What about replacing the mechanical relay with the mosFET relay described recently? The space left over after changing to smaller footprint caps can be used to fit the mosFET triggering circuits.
 
Hi Andrew,

Why have you adopted a dual bridge rectifier for use with a centre tapped transformer? They don't work together !

It's not so, if you look at the schematic there are two complete AC input (AC1/PGND1 and AC2/PGND2) for a dual secondary transformer.

GND is for safety ground/CL-60 to safetyground/resistor to safety ground.

Maybe the names are not the best ones, any suggestion is welcome.

But maybe I've implemented it wrong, if you can check it would be nice. ;)

What is VR2? It can't be a +ve regulator turned upside down !

It's a TL431 voltage reference, I've taken the schematic here for both the positive and negative regulators.

And actually I've made an error also here, now checking and fixing, thanks for pointing me on it. :)

If adeqaute smoothing cap values result in too large a footprint, then move the main capacitance off board and put in adequate decoupling.

It can be an alternative but I'm trying to keep all on single PCB.

What about replacing the mechanical relay with the mosFET relay described recently? The space left over after changing to smaller footprint caps can be used to fit the mosFET triggering circuits.

Can you point me to it, please? :)

Thanks in advance for any help/suggestion.

In the meanwhile I've revised polygons and power traces...
 

Attachments

  • RevO - PCB.jpg
    RevO - PCB.jpg
    338.5 KB · Views: 159
Last edited:
I still don't see it.

The amplifier PSU shows only two AC connections (AC1 & AC2) from the transformer.
For a dual bridge rectifier, the transformer must be a dual secondary and each secondary has two taps/ends. I expect to see AC1a, AC1b, AC2a, AC2b, i.e. 4 tapas/ends.
That is not shown.

VR2 is still wired differently from VR1.

OK, settle for inadequate smoothing caps after you have stated that the PCB size is more important to you than your realisation that other designers have increased the smoothing cap values
the My_Evo has 10000uF smothing caps for each pump.
 
Last edited:
I still don't see it.

The amplifier PSU shows only two AC connections (AC1 & AC2) from the transformer.
For a dual bridge rectifier, the transformer must be a dual secondary and each secondary has two taps/ends. I expect to see AC1a, AC1b, AC2a, AC2b, i.e. 4 tapas/ends.
That is not shown.

Andrew, as I've already said they're there (AC1+PGND1 and AC2+PGND2)

I do agree that names are probably wrong (in fact I've changed them as you can see in the attached updated version) but they still are the four connection you're expecting...

BTW with a different arrangement I've managed to re-implement holes for bulk wire.

VR2 is still wired differently from VR1.

Exactly, so they must be.

Please take a look on the schematic (there's also a PS detail) and point me to the error.

OK, settle for inadequate smoothing caps after you have stated that the PCB size is more important to you than your realisation that other designers have increased the smoothing cap values

No, Andrew, in the Evolution there's space for four 10000uF per channel, 2 for pump #1 and 2 for pump #2.

If you have a single pump, like in the MyRef, there are only 2 big caps...

An externally hosted image should be here but it was not working when we last tested it.


I can't really understand if you want to help or only criticize.. :scratch1:
 

Attachments

  • RevO - PCB.jpg
    RevO - PCB.jpg
    328.4 KB · Views: 273
  • RevO - Schematic.jpg
    RevO - Schematic.jpg
    206.9 KB · Views: 290
  • RevO - PCB (PS).jpg
    RevO - PCB (PS).jpg
    192.5 KB · Views: 107
  • RevO - Schematic (PS).jpg
    RevO - Schematic (PS).jpg
    93 KB · Views: 106
That's an improvement. You have removed ground references from the secondary connection points.
VR2 is different from VR1. You have said it need be different
Exactly, so they must be.
why?
Don't they supply the same voltage, i.e. 12Vdc?
It's possible but also the My_Evo has 10000uF smothing caps for each pump.

The PCB is already tight and full with 30mm caps, in this moment I can't imagine how to stretch in 35mm ones (and bigger value caps are mostly 35mm...)
Maybe I can't read, but I interpret your statement as: Two 10mF need two 35mm diameter footprints, but I only have 30mm diameter spaces available, so I'll just fit smaller value capacitors.

If two pumps need four 10mF, then why should one pump get two smaller capacitors? This does not make sense to me.
 
Last edited:
That's an improvement. You have removed ground references from the secondary connection points.

That's only a different components arrangement and more clear naming, schematic is the same...

VR2 is different from VR1. You have said it need be different why?
Don't they supply the same voltage, i.e. 12Vdc?

Yes but one is a positive regulator the other is negative...

Take a look at the attached comparison, schematic is identical.

Obvioulsy since TL431 (datasheet) it's not a simple zener it needs two additional resistors for voltage setting.

Maybe there's something wrong?

If so please tell me what is.

Maybe I can't read, but I interpret your statement as: Two 10mF need two 35mm diameter footprints, but I only have 30mm diameter spaces available, so I'll just fit smaller value capacitors.

If two pumps need four 10mF, then why should one pump get two smaller capacitors? This does not make sense to me.

No, what I've said is that, as the RevC PCB, mine has space for two 30mm smoothing caps so 2x 10000uF 63V is the maximum (and suggested) value.

Bob was asking if there is need for caps bigger than 10000uF.

2 per 10000uF are plenty for the MyRef... and hundreds of kits, including yours, perfectly works with those caps.
 

Attachments

  • RevC Vs REvO.jpg
    RevC Vs REvO.jpg
    140.4 KB · Views: 159
R29 and R30 should be swapped to get correct voltage

Yep!

Right, thanks Amycheng :)

Andrew, probably you were pointing the same error... just be a bit more clear next time, it was not wired differently, it was wrong ;), thanks :)

I've attached the fixed schematic and PCB.

BTW, adding an elco parallel to R27 and R29, according to this article should lower and linearize impedance, any opinion on it?
 

Attachments

  • RevO - PCB.jpg
    RevO - PCB.jpg
    328 KB · Views: 139
  • RevO - Schematic.jpg
    RevO - Schematic.jpg
    180.7 KB · Views: 108
Has anyone tried switching power on these?

Hi Soongsc,

not yet, I'm still in the design phase...

I've tried so far (on MyRef RevC):

- the TL431 PS (it works but I'm not sure it's better than zeners)
- different changes in the compensation scheme (all of them works)

I'm investigating some better (but simple, low component count) LM318 PS at the moment.

I've yet to try:

-the voltage limiter
-the PCB itself

Regarding the PCB I still need help on design and best practices, but so far no comments on it... :(
 
I recall Mauro provided a good list on layout practices, as usual the devil is in the details. Not too long ago working on a different design, I discovered that any two components placed close together will interact with each other, due to the electromagnetic effect of the small amounts of current. So it's also a good practice to check the direction so that these sort of run perpendicular to each other.
 
Status
Not open for further replies.