My new multichanel passive balanced pre-amp, battery/solar powerd, remotecontrol.

[Brian Guralnick]

Almost done here, Gerbers should be going out tomorrow night. Can’t wait!!!

Here are the architectural outlines, they will explain everything:

The attenuator module.
http://pages.infinit.net/helloftp/attenattenuator.png

The main processor & display / keypad PCB module.
http://pages.infinit.net/helloftp/attenprocessor.png

Here is the latest schematic of the attenuator. This should give you an idea of how the relay pot is wired.
http://pages.infinit.net/helloftp/attennewsmall.png

 

[x-pro]

Hi Brian,

What is the attenuation range and steps on your design?

Regards

x-pro

 

[Brian Guralnick]

Listed in the first illustration in the top middle white box.

64 steps, 96Kohm, 48Kohm, 24Kohm, or 12Kohm, Constructed with 0.1%, or 0.5%(econo) resistors.

Current draw is listed in the bottom left hand corner box.

Module sizes are in the bottom right hand box.

If you are having problems seeing the fine text, it's Internet Explorer, you just need to un-zoom the image, or, download it & view with a picture viewer.

************************************************
HELP!!! I notice some over-sized image posts do not make
Internet explorer crunch up the image, how do I do that???
************************************************

 

[cowanrg]

if the picture is shrunk to fit, hold the cursor over it, and you will see icons appear, in the bottom right, click that icon, will goto full size.

 

[x-pro]

Hi again,

thanks for a quick reply, I didn' t read the text properly, sorry. There is stated clearly that the pots are linear. I did somewhat similar design couple of years ago, however I did use 7 relays per channel to give me 80 steps of 1 dB = 79 dB max attenuation + one more relay for a complete mute. My design had logarithmic curve for volume, otherwise it looks quite familiar .

Good luck!

x-pro

 

[Brian Guralnick]

Well, if you look at the software options on the processor page, I have a Volumetric, or Logarithmic software emulation.

However, that just makes a nicer volume curve with 24 steps, 24 sending the full 63 level on the relays. It does not simulate the surface resistance charactersist of a real logarithmic pot, or does it? Man, now my head is now spinning off in the wrong direction.

 

[Brian Guralnick]

Hi again,

I did somewhat similar design couple of years ago, however I did use 7 relays per channel to give me 80 steps of 1 dB = 79 dB max attenuation + one more relay for a complete mute. My design had logarithmic curve for volume, otherwise it looks quite familiar .

Good luck!

x-pro

Thanks!

You should see the new relays I'm using, Omron G6KU-2G. They latch & they are the puniest thing i've ever seen. They are 10mmX6.5mmX5.2mm that's SMALLER than a DIP8. I even broke 1 open to see the distance on the contact switches, they are so short and small, you almost need a magnafying glass to get a good look at the contacts.

 

[Brian Guralnick]

Here's a better view of the attenuator's schematic.

http://pages.infinit.net/helloftp/attennew.png

 

[x-pro]

Hi Brian,

I did use exactly the same relays - G6K, thought not the latching type. These are not just small, they've got very good contacts quality as well and very low current consumption. Excellent choice . I did look at the larger schematics and have to say that the way you've organised the switching is quite different from what I did.

Cheers

x-pro

 

[fmak]

Here's a better view of the attenuator's schematic.

----------------------------------------------------------
Brian

Looks very good. Are you making pcbs and building plans available and at how much pl?

What are the attentuator steps and the reistors types used?

 

[Brian Guralnick]

Hi Brian,

I did use exactly the same relays - G6K, thought not the latching type. These are not just small, they've got very good contacts quality as well and very low current consumption. Excellent choice
x-pro

If you thought the G6K had very low current consumption, the G6KU has 0. Man, I mean 24Ma for 1/20 th of a second, then 0.
As for the switch organizement, it was intentionally laid out so that the lower impedance side of the signals connect to the center pin of the relay & the higher to the smaller pin contact. The layout is also set up to minimize channel crosstalk.

Brian

Looks very good. Are you making pcbs and building plans available and at how much pl?

What are the attentuator steps and the reistors types used?

64 steps, originally, I was going to have a choice of 128, or 32, but, when I switched from 3 IC2 peripheral 8bit IO ICs to a PIC On the attenuator PCB, the 22 IOs was just right for 64 steps X 2, and, 4 line ins.

For the resistors, I'm using SUSUMU, or, Panasonic ERA series precision 0.1%, 0603 smd resistors right under the relays.

My PCBs & their manufacturing will be contracted out to commercial manufacturers. Could you imagine me mounting the 175 smd parts which go on both sides of the PCB, & thats for 1, I personally want an 8 channel XLR system, that's over 1400 components. If it was a 1 sided PCB, I may have done the assembly locally.

PCb & building plans will be made available, however, because of the bulk quantity of relays & 0.1% smd resistors I'll be using, & I'll be making them modules available at near cost to DIYaudio members, it will cost you 1/2 to 1/3 rd as much just to buy completed PCBs from me.

Gerbers are going out tonight, or, tomorrow morning. Using an econo route, a completed PCB usually gets back to me within 3-5 weeks.

 

[fmak]

64 steps, originally, I was going to have a choice of 128, or 32, but, when I switched from 3 IC2 peripheral 8bit IO ICs to a PIC
For the resistors, I'm using SUSUMU, or, Panasonic ERA series precision 0.1%, 0603 smd resistors right under the relays.

----------------------------------------------------------
Brian

Are these 1 dB step for 64 steps?

And are the resistors metal film non magnetic?

Thanks

 

[Brian Guralnick]

They are metal film, I'm still researching non-film 0603 0.1% precision. There's a local company called K-tronics who have a specialized resistor division made just for these 'insane measure' devices.

The relays are organized into a true linear 64 step pot.

This means:
______________________
Step 0 = vout = vin * 0/63.
Step 1 = vout = vin * 1/63.
Step 2 = vout = vin * 2/63.
...
...
...
Step 61 = vout = vin * 61/63.
Step 62 = vout = vin * 62/63.
Step 63 = vout = vin * 63/63.
_______________________

The control module can create diferent gain curves by dividing out the steps differently.

I am now considering going back to the 8-bit 256 step version. This will change the step# & 63s in the last table to 255. This way, I can simulate a 90 step log pot very closely, &, for tube users, there will be a really fine balance control.
However, a 12Khom version of this pot has a ridicously small LSB resistors.

 

[jan.didden]

Well, if you look at the software options on the processor page, I have a Volumetric, or Logarithmic software emulation.

However, that just makes a nicer volume curve with 24 steps, 24 sending the full 63 level on the relays. It does not simulate the surface resistance charactersist of a real logarithmic pot, or does it? Man, now my head is now spinning off in the wrong direction.

Brian,

How is your software configured? Can I control, say the level in each pair for use in balancing passive speaker systems, then also up/down the volume in a full 2 or 3 channel system to maintain the balance which I set before?

Jan Didden

 

[Brian Guralnick]

Brian,

How is your software configured? Can I control, say the level in each pair for use in balancing passive speaker systems, then also up/down the volume in a full 2 or 3 channel system to maintain the balance which I set before?

Jan Didden

YES! On everything! The module is configured using the front panel buttons. Just hold down all 6 buttons for 10 seconds straight. Then, the display will switch over to config mode.

In config mode, the pre-set options you set up here will be stored in flash, so even power outages after that will retain your settings.

New improvements!:
- I've already improved the com->attenuator modules, now it only needs 3 conductors (2 power + 1 com), & all the modules may now be driven in parallel using the same 3 wires passed board-board.
- Now you can run up to 32 XLR channels. Needed for the new comercial 26 (yes, 26) channel + 6 channel subwoffer true 3D sound card technology which is being developed.
- All attenuators will now be 256 steps & they are now even smaller.

 

[x-pro]

Hi Brian,

I did post my attenuator circuit here - have a look :

http://www.diyaudio.com/forums/showthread.php?s=&postid=246390#post246390

Cheers

x-pro

 

[grataku]

In the link you posted there is a patent mentioned
I am actually curious to know what is patented in this design?

 

[Brian Guralnick]

New final draft of the attenuator.
Really cleaned up schematic!
Lest than 50 components on a 1 side PCB.

http://pages.infinit.net/helloftp/attschemnew.png

Main text in schematic:

The listed values create a 47.8125 Kohm pot, well, almost 50 Kohm. The 8 relays move the 'Actuator' between 'Min / -' and 'Max / +' in 256 equal steps.
- All resistors are 0.1%, 0603. Values:
24.00k -> standard
12.00k -> standard
6.00k -> non-standard, 2 x 12.00k will be used in parallel.
3.00k -> standard
1.50k -> standard
750.0r -> standard
375.0r -> non-standard, 2 x 750.0r will be used in parallel.
187.5r -> non-standard, 3.00k & 200.0r will be used in parallel.

- This is a 'true pot emulation', it's resistance remains constant, a volume of 0 shorts the 'Actuator' to the 'Min / -', this produces a true 0 ohm short. Same for a volume of 255, it authentically shorts the 'Actuator' to the 'Max / +'.
- The added vactrol can be used to 'softly' ramp in & out the audio during a step to remove the 'zipper' effect of a switched attenuator.
- The shield has not been connected to any of the attenuator circuitry to allow compatibility with some DAC I/V stage schematics. This is also why the vactrol's outputs are open on separate terminals, in different designs, you may want to use them differently.