Friday, November 24, 2006

Developing the Dada 405 amplifier-boards

Joost, who is the development-engineer (and CTO) of Dada/JP Engineering, wanted to give you a first impression of the starting-points why we are developing the new 405 High-end boards.

The boards will be launched in the beginning of 2007. We are developing & testing the prototypes for the moment but we can not publish any pictures yet.

Joost's paper is below.

Stefaan

There are five sources for serious 405 modifications:

  • The Quad 405 service manual
  • The text from Bernd Ludwig
  • The DC-Daylight (Keith Snook) Web pages
  • The designs of the Quad 240/306/510/520 and 606 which were all developed during the life cycle of the 405.
  • The designs of Tiberiu Vicol

The Quad 405 service manual.

In this manual the history of the 405 in terms of official factory modifications is fully documented. Also are most of the diagrams, part lists and printed circuit board layouts present.

It can give you ideas for retrofitting modifications to the 405-1 and early 405-2 models.

(the 405 service manual is available for free on simple e-mail demand from info@dadaelectronics.eu)

The Bernd Ludwig paper.

Bernd Ludwig was (as far as I know) the first one who described the 405 circuit in full detail, suggested three upgrades and shared a lot of (his) general knowledge about the 405 circuit. The paper covers the following subjects.

  1. Changes for the input Opamp circuit
  2. Modification of the Current Limiter
  3. Extensive output circuit modification
  4. General tips and hints concerning the 405 circuit

DC- Daylight Web pages.

Using the work of Bernd Ludwig as a basis, Keith Snook suggested two important modifications:

  1. Changing the input Opamp circuit to a non inverting one.
  2. Retrofit (my words) the Current Dumper Bridge design of later Quad models to the 405 circuit.

Also Keith suggested more practical ways to implement Bernd Ludwig's modifications, completed the set of diagrams of the 405, introduced his own fully modified 405-3, ( my number) and discovered a fault in the Service Documentation concerning the diagram of the 405-2 current limiter.

The designs of the Quad 240/306/510/520 and 606.

This give you some ideas how a fully developed 405 could look. Quad themselves did only one substantial upgrade, the 405-2 modification.

The designs of Tiberiu Vicol.

Although he made very extensive redesigns, some parts will fit as an upgrade to a more or less normal 405. The quality of his work is very good, so keep track!

The JPE / Dada design.

Instead of producing an endless number of prototypes, during the design stage I decided to built the two circuits, the Non and Inverted input circuit, in a computer simulation tool. I modelled the full input circuit including the mid section filter stage. The rest of the circuit is simulated by a non inverting Opamp. The latter one providing the necessary gain and the DC feedback circuit. See the diagram for details. Both circuits emulate the standard 405 with an input sensitivity of 500mV, and a overall gain of 56.

Some observations:

The gain of the input Opamp circuit is 15, the input impedance is 20K at 1kHz. When changing the gain, you also have to change the time constants of the local and overall feedback loops. When using the model you can play with the values of the components and see the results on the spot. To preserve the overall frequency behaviour, follow the recipe of Bernd Ludwig. If you change some values, i.e. to increase the DC suppression, it also has effects on the LF response. The behaviour of the circuit is plain Opamp technique, no magic or surprises.

With the non inverted input circuit, which is a far better idea as a first stage than a inverted one, you only have to replace two components, to change the gain and the time constant of the overall feedback circuit. In a inverting circuit the noise is higher and in general the input impedance is low. By changing the power supply voltage (asymmetric), testing of the DC feedback circuit can be done. The non inverted model performs ok in this respect. I did not build a complete 405 circuit, I believe the current dumping principle works fine!

Power output of the 405.

Instead of adding power to the 405 by extensive modifications, we advise to combine the two channels of the amplifier, either in serial or parallel mode. You can run our modules without current limiters, but we advise only to do this when the thermal switches are installed. In normal operation the current limiters, based on the 405-2 circuit, will have no audible effect.

The model and our own input gave the following functional/technical demand for the prototype

  • Onboard phase inverter for connecting the two channels in serial mode
  • Onboard pot for balancing the channels in block mode
  • Non inverting input circuit (lower noise)
  • Regulated power supply for the Opamps
  • Opamps of your choice, starting with OPA604
  • Optimized Current Dumper Bridge configuration
  • Current limiter on separate PCB’s, can be removed!
  • Active ground DC protection circuit instead of Clamp circuit, requires separated PSU
  • Input sensitivity jumper select
  • Voltage limiter jumper select
  • Based on the latest 405-2, without all the layout issues
  • Thermal switches to protect the modules from overheating

The power supply

The upgrade that Bernd and Keith suggest is replacing the PSU capacitors with fresh bigger ones and bypassing the capacitors with small caps to reduce the HF impedance. Although this is good advise in itself, we have to take the development of the PSU a step further.

The active ground DC protection will only work with a fully separated double power supply. But the big sonic reason for building a double supply is the following: Crosstalk! A standard 405 have a channel separation of only -60dB at 10KHz. So the crosstalk signal is 20dB higher than the distortion. The double supply in combination with a careful internal cable layout and building the amplifier as a four-pole device, so no ground connections, can reduce the crosstalk level to at least -90dB across the audio spectrum. The four pole approach also reduces ground loops, with a standard 405 there is always a ground loop, formed by the Cinch or Din input cables. For safety reasons the power supply cord must be connected to a wall socket with a proper earth connection, so touching the case is not a hazardous job.

Joost



6 Comments:

Anonymous vzs said...

Hi!
First of all congratulation for developing the amplifier board.

I have a question regarding the four-pole device approach. You mentioned "building the amplifier as a four-pole device, so no ground connections... The four pole approach also reduces ground loops, with a standard 405 there is always a ground loop, formed by the Cinch or Din input cables".

So, I'm having a double PSU with their grounds connected togheter, what do I have to modify to use this four pole approach? Would you please detail me this.

Thank you,
Zsolt

12:13 PM  
Blogger Joost Plugge said...

Hello Zsolt,
I suppose you using the NetAudio supply. I don't know the exact circuit board layout; maybe it is possible to split the ground from the two supplies. But there is extra work to be done, before the two amplifier modules are really floating. You have to isolate the centre earth screw on the heatsink, by using a To3 isolating ring, of the modules and use that point as the new central common point per channel, also the input grounds should be disconnected from the chassis, and the speaker return leads should be connected to the new common point on the heatsink. On PCB's 12368 and earlier you also have to connect the signal input ground to this central point, because of a layout issue. We will publish the fitting instructions very soon, and then it will be clear to you what to do. You can also reduce, not avoiding it, the earth loop of the cables by removing the outer plastic shield, use large shrink tube to connect the two cables together. You need a brave heart to suggest this to owners of cables of 500 dollars per meter! But I stick to my opinion, most cable/amplifier systems perform badly in this respect.

Joost Plugge
JP Engineering

5:34 PM  
Anonymous vzs said...

Hello Joost,

It is clear to me now how to make the floating ground in the amplifier.

Thank you,
Zsolt

10:30 AM  
Anonymous dust.path said...

hai to the quad-revive-crew!

i heard/read about this developing and get totaly flamed up! i will build a quad "clone" with this boards by my own, and now i can't wait for the "deadline" of this project. is there a startig date for shipping fix? ... just a house number will keep the fans happy :-)

very big thanx for doing this fine stuff for all the quad-fans!

bye, dusty.path /trier(ger)
[dustman.path(at)gmx.net]

8:36 PM  
Blogger Joost Plugge said...

Hello,

Yes indeed with the modules and the power suppply you can build your own amp or restore a 405 to its old glory. Beginning of February this year seems a realistic launch date.

Regards Joost Plugge

11:39 AM  
Blogger kostiska said...

What's the update on this ??

9:28 PM  

Post a Comment

Links to this post:

Create a Link

<< Home