ARP Odyssey Oscillators

Broadwave · Posted: Sat Aug 30, 2008 9:14 am Post subject: ARP Odyssey Oscillators

Sorry, me again...

I've just built a clone of the ARP Odyssey Oscillators (MkIII), including the Ring Mod - All works great, very pleased.

...except that the oscillators are locking together when I get close to unison.

I've heard of this problem before, especially when two Osc's are built onto one board, and particularly when (for example) a quad op-amp is shared by both osc's.

I've removed any flux off the PCB, it's very clean and I know the PSU is fine.

Unfortunately, I just don't know enough about the theory of electronics to know what I can do about this locking.

As I mentioned before, it's an exact copy of the Oscillator section of the PCB, and no component substitutions have been made, apart from using good quality WIMA 680pf timing caps instead of polystyrene and multiturn trimmers.

Any suggestions?

Andy

PS. You may be wondering why the hell I've built Oddy Osc's... It's all to do with the sync and ring mod combo, I've tried several VCO's/RM circuits in the past and never been able to get that distinctive Oddy sound.

ARP Oddy Osc.jpg

Description:

Filesize:

207.18 KB

Viewed:

1987 Time(s)

This image has been reduced to fit the page. Click on it to enlarge.

urbanscallywag · Joined: Nov 30, 2007 Posts: 317 Location: sometimes

Congratulations, that's a cool project. I have an Odyssey and would like to build some modular Odyssey VCOs. They do have a unique sound.

It doesn't look like there are any decoupling capacitors around your opamp.

Broadwave · Posted: Sat Aug 30, 2008 10:58 am Post subject:

Ah, now then... this is the problem you see : )

There are no op-amps - there's a couple of 3086 tranny arrays, and a CD4011 which is used for the osc's integrators and the ring mod.

All the de-coupling (not that there's much... if any!!!) on the PCB is identical to the original. I can only presume that its all the switching that the discrete FET's are doing that's causing the lock-up.

Will have to have a long hard think, and see if I can somehow de-couple the FET's.

Andy.

urbanscallywag · Joined: Nov 30, 2007 Posts: 317 Location: sometimes

Ah, I see. I saw the 2 chips on the sides weren't op amps, I couldn't tell what the center chip was (and then you mentioned quad opamp...).

Hmmm. Can you just add an electrolytic and ceramic decoupling capacitor to each rail on the actual board? (never mind I see the tantalums)

urbanscallywag · Joined: Nov 30, 2007 Posts: 317 Location: sometimes

You could still decouple the 4011 right?

Peake · Posted: Sat Aug 30, 2008 11:37 am Post subject:

Is the overall layout the same as the original?

The SEM uses a 3086 (IIRC, the Prodigy as well) for a pair on the same board, and they have both avoided the locking issue. I have no idea how. If someone can open an Odyssey and send you a photo of the oscillator section it might help.

Very nice board, by the way.

Broadwave · Posted: Sat Aug 30, 2008 12:14 pm Post subject:

bubblechamber · Posted: Sat Aug 30, 2008 3:21 pm Post subject:

when i've encountered this problem it was often coming from using the same ground for both oscillators. and there's always the possibility that they are meant to do that and it's a "feature"...

Broadwave · Posted: Sat Aug 30, 2008 3:32 pm Post subject:

I've decided to look at the power and ground rails to each osc, maybe try to separate them a bit better.

If that fails, I'll re-design the board so each of the 3 parts (2 oscs and the ring mod) has it's own supply.

It's late I'm off for some shut eye.

Andy.

Broadwave · Posted: Sun Aug 31, 2008 2:00 am Post subject:

Hey Bubblechamber - Thanks, that's what the problem was!

I thought I'd be clever and extend the ground trace to make it easier to wire up the pots... I didn't realise that I'd created a loop between the two oscillators.

Working really well now Very Happy

I thought about uploading all the details, but I'm not sure what the legal aspect of this would be as I guess someone, somewhere still has a copyright on the design?

Andy.

julianw · Joined: Jul 30, 2007 Posts: 78 Location: UK

synthmonger · Posted: Sun Aug 31, 2008 7:39 am Post subject:

so how did ya make a separate ground?

Broadwave · Posted: Sun Aug 31, 2008 8:44 am Post subject:

creekree · Posted: Sun Aug 31, 2008 9:02 am Post subject:

rjd2 · Joined: Sep 02, 2007 Posts: 236 Location: philly

2 questions:

1-which vco version did you clone off of? i understand there are slight differences in the topology of the odyssey vco's.

2-so....anyone wanna make a run of these pcb's? i'd be in for 4-8, at least... Very Happy

urbanscallywag · Joined: Nov 30, 2007 Posts: 317 Location: sometimes

Wow, glad the power routing solved the problem. Cool

Broadwave · Posted: Sun Aug 31, 2008 3:29 pm Post subject:

Zodiak · Posted: Sun Aug 31, 2008 5:13 pm Post subject:

Peake · Posted: Sun Aug 31, 2008 8:20 pm Post subject:

I've never cared for the sound of any ARP oscillators after the 2600, but this thread has caught my attention.

Service Manual, page 45, 7.7 Board B-II:

2N5485:
Central Semiconductor at Mouser: 610-2N5485, .55 singly.
Fairchild at Mouser: 512-2N5485, .12 singly.

2N5910/PN5910: Eep, 6 in stock at americanmicrosemi at $6.25. Doesn't show up at my favorite places. Mouser crosses the PN to the NTE106. I've never had much luck with substitute components. $15.90 on ebay for the PN version, quantity of six. Try:

H&R ENTERPRISES, 1060 in stock.
All Parts Sales, 200 in stock.
Digitron Electronics, 300 in stock.

2N:

WE Express United States 2N5910 MOTOROLA 625
Perfect Electronics Ltd United Kingdom 2N5910 MOTOROLA 3500
Digitron Electronics United States 2N5910 FAIR 127

Can't identify Q5, and 6,7,13,14. You'll have to figure out how they selected the various transistors...

Sample and Hold, etc:

2n3958 dual FET- Didn't think that I'd see this outside of Buchla design (it's in the 259 waveshaper and the 291 dual BPF). I believe that Mark Verbos replaced it in his 291 clone with an opamp: http://www.simple-answer.com/bandpass.jpg
There's another 291 version on the net, link here on E-M in this forum, with another substitution, but this time a tranny. BF245 replaces 2N3958.

BF245 (c version): Mouser 512-BF245C, .13 each. a and b versions, same price.

2N5459: Mouser 512-2N5459, .15 singly.

LM1458: Mouser 512-LM1458N, .33 singly. Very slow opamp; replace if used to buffer audio outputs.

LM301AN: Mouser 512-LM301AN, .30 singly. Slow, but clips nicely.

Board A, if anyone is interested:

2N4392: Mouser 610-2N4392, $1.64 singly, but the Fairchild PN version is 512-PN4392, .13 singly.

2N6076: Mouser 512-2N6076, .09 singly.

Broadwave · Posted: Mon Sep 01, 2008 12:11 am Post subject:

Broadwave · Posted: Mon Sep 01, 2008 5:02 am Post subject:

OK, It's pretty rough and ready... but it works Smile

There's plenty of info around the web if you need the schematics.

Enjoy,

Andy.

ARP Odyssey Dual VCO.pdf

Description:

PCB Layout and Component Placing.
Build at your own risk :)

Download (listen)

Filename:

ARP Odyssey Dual VCO.pdf

Filesize:

203.06 KB

Downloaded:

4237 Time(s)

Peake · Posted: Mon Sep 01, 2008 8:09 am Post subject:

Is this the oscillator board? Looks like the heat-sunk pair have some level of matching/selection:

http://www.arpodyssey.com/odybfrnt.jpg

1.87K Tempco? And people chide me for liking the uA726 Wink

Thanks for the project!

loss1234 · Posted: Mon Sep 01, 2008 9:33 am Post subject:

wow

thanks a lot!

Broadwave · Posted: Mon Sep 01, 2008 10:08 am Post subject:

/mr · Posted: Mon Sep 01, 2008 3:21 pm Post subject: