Sunday, December 26, 2010

The best CD4066 switch for a Quad 44 is.........

Instead of looking for the right 4066, I build a relay solution. The two relays (4 contacts) are placed on a small Pcb, piggybacked on the copper side of the Psu board.

The only awkward thing to do, is soldering a 14 pin Ic socket to this side of the board, but Quad made large soldering "islands", so with a steady hand and a small soldering tip it went ok. The relays are driven by transistors in so called “open collector mode”. The Pcb’s are degraded to prototypes because of a small non critical (repairable) layout error. How can you make an error in such a simple Pcb? Don’t ask!

In a 44 there are two switches in series with each signal path, lucky for me all the CD4066 are configured the same, so the layout of the Pcb is standard for all positions. The relays are high quality ones with gold plated contacts.

I measured THD (non weighted) and IM (Din) distortion with the Virtins analyzer (left and right channel). This is a 44 MK I with single OPA’s. All the OPA’s are BB OPA 627.

With all the original CD4066 in place:

THD 0.0055 and 0.0062 IM 0.013 and 0.015

The selector switch replaced by a relay:

THD 0.0032 and 0.0039 IM 0.009 and 0.010

The output switch replaced by a relay:

THD 0.0008 and 0.0007 IM 0.0017 and 0.0012

So a reduction of 17 dB or so of THD and IM, the relays operate click free in the signal path, but the mechanical clicks reminds me of the good old past (I am a sentimental old man!).

Altough the results of the relays are very good, the original performance of a 44 with new Op Amp's is still outstanding, considering the age of the design!

Joost Plugge

Update 8 march:

New PCB's have arived, also a connector sytem instead of a DIL socket, watch this space for the anoucement and pricing.

DaDa Electronics


Blogger EJP said...

That indicates that it is the output switch that is critical? as THD was reduced by 75% when you applied it?

3:05 AM  
Blogger Stefaan said...

Hello EJP,

From the numbers, your conclusion is valid. But I did not test if the two 4066 were identical or differant. When the production boards are ready, I will also test the differances between the 4066 switches.

Joost Plugge

5:57 PM  
Anonymous Anonymous said...

When a relays kit could be available ?

Thanks for information, Ronan, Paris

11:55 PM  

Post a Comment

Links to this post:

Create a Link

<< Home