SAD4096 MN3205 retrofit

milosch · December 30, 2007, 10:07:46 AM

I have a pedal that used an SAD4096 originally. I would like to retrofit it using a small board and replace it with an MN3205 due to availability. Has anyone done this, and if so, how much support circuitry is actually required? I realize that the clock circuit may need to be changed, but otherwise how difficult will it be to do this?

The idea is to keep the pedal with its original look, but actually have it work. There is room for a small board in there.

oldschoolanalog · December 30, 2007, 02:30:46 PM

http://i175.photobucket.com/albums/w121/oldschoolanalog/DSC00456.jpg
DOD 680. 4096 to 3205 retrofit.

milosch · December 30, 2007, 04:45:59 PM

Thanks for the push. I had started working on this earlier and finally got it working today. The trick was voltage loss from the input to where I had been tapping in. Now that my 5V reg can supply 5V I got my bias and Vgg pots setup. The sound is pretty noisy, but still usable. Looks bad, but it works:

theehman · December 30, 2007, 05:06:25 PM

Please post the schematic and info on what you did. I think a lot of people will find this useful. I have an MXR Omni that needs 2 SAD4096 replaced.

milosch · December 30, 2007, 05:22:06 PM

Screen capture from DipTrace:

Note the pin numbers on U1, the MN3205. I had to change them to what their values would be for a 16-pin DIP to get the board layout to behave. IC1 is a 78L05, TO92 package despite the square image there.

VggSet is adjusted to roughly 4.65V (14/15 * 5V).
BIAS is adjusted to approx 2.5V, but a wide range works here. Basically, you could center the pot or replace with identical fixed resistors.

I installed this into an MXR Time Delay. I adjusted the onboard clock trimmer for the most delay (all the way CCW). I can get probably 400ms or more out of this, of course not super clean. The delay knob will adjust from that down to unnoticeable delay below reverb.

milosch · December 30, 2007, 05:48:54 PM

Once I am feeling less lazy tomorrow, I'll try some pico across the 5V supply. There's probably some clock noise in there. Of course, having IO buffers would be nice as well...

milosch · December 30, 2007, 10:54:49 PM

Here's a sample - Whoa, I'll say there's some clock noise:
http://www.groupwhere.net/guitars/timedelay.mp3

milosch · April 14, 2008, 08:30:10 PM

Sorry I haven't gotten back to this in awhile. Short version is that this is a nice retrofit. However, something closer to a true voltage setting perhaps with a zener for Vgg would work better, but I haven't taken a stab at it yet. In my case, the operation would come and go depending on my power supply choice.

milosch · January 11, 2011, 09:22:59 AM

Getting back to this again now. I've redone the circuit to use the MN3102 and hope to be testing this week. That should remove the need for the 5V regulator and the pot attempting to give the right Vgg to the MN3205.

theehman · January 11, 2011, 09:59:05 AM

We're all rooting for you.

milosch · January 11, 2011, 10:10:06 AM

Surely this must have been done already. Maybe there just isn't that much call for it.

milosch · January 11, 2011, 10:17:04 AM

milosch · January 16, 2011, 01:01:47 PM

Current board layout. Still waiting on board making supplies so I can try this out.

milosch · January 22, 2011, 01:18:21 PM

That layout is bogus. THe MN3205 pattern is wrong and of course doesn't fit. Don't ask me how I found out.

milosch · January 25, 2011, 04:58:33 PM

In the meantime, here is a circuit for the R5101. Not yet tested:

Thomeeque · January 26, 2011, 02:49:10 AM

Quote from: milosch on December 30, 2007, 05:22:06 PM

Hi!

May I ask, did you see some impact when adjusting V_GG trimmer? I always wondered what influence it has (because way it's set across different circuits vary - some use 14/15, some 15/15 etc.).

Thanks, T.

EDIT: In R5101 circuit you should consider putting some resistor (typically around 100k) between BBD IN and central lug of BIAS trimmer - this way you load input signal too much (plus you cut lot of low frequencies thanks to C2=10nF)

EDIT2: Or use bigger value for BIAS trimmer - 500k~1M could be OK.

milosch · January 26, 2011, 05:44:38 AM

Quote from: Thomeeque on January 26, 2011, 02:49:10 AM

Hi!

May I ask, did you see some impact when adjusting V_GG trimmer? I always wondered what influence it has (because way it's set across different circuits vary - some use 14/15, some 15/15 etc.).

Thanks, T.

EDIT: In R5101 circuit you should consider putting some resistor (typically around 100k) between BBD IN and central lug of BIAS trimmer - this way you load input signal too much (plus you cut lot of low frequencies thanks to C2=10nF)

EDIT2: Or use bigger value for BIAS trimmer - 500k~1M could be OK.

I didn't have much consistency with the first circuit, I think because of Vgg. This is why I elected to use the MN3102 since it has the correct voltage division. However, since we don't really need the clock portion of this chip there, I guess a good fixed precision resistor setup for 14/15 (per the spec for the 3205) should work.

In the R5101 circuit, I changed this at the last minute to 20k only due to availability in my shop. I'll test that, thanks.

milosch · January 29, 2011, 06:39:42 PM

Success! I had to short the coupling cap on the input, and bias adjustment definitely makes a difference. The hardest part has been the proper way to connect it to the the board. As it is, I'm using two 8=pin headers on this board with a 16-pin socket underneath plugged into the socket on the board. I'm happy. I'll post the final layout, etc. shortly.

milosch · January 29, 2011, 07:39:00 PM

Installed in the MXR Time Delay:

milosch · January 29, 2011, 07:43:10 PM

News:

SAD4096 MN3205 retrofit