Analog Echo Sound Effects Box (Eli SL-3535)

Recording Techniques, People Skills, Gear, Recording Spaces, Computers, and DIY

Moderators: drumsound, tomb

Post Reply
User avatar
ubertar
ears didn't survive the freeze
Posts: 3775
Joined: Wed Feb 18, 2004 7:20 pm
Location: mid-Atlantic US
Contact:

Analog Echo Sound Effects Box (Eli SL-3535)

Post by ubertar » Mon Feb 06, 2012 6:02 am

Just picked up this thing on the cheap off ebay... the make/model is Eli SL-3535. According to the inter-tubes, it's a MN3005 bucket brigade circuit. I'll have to open it up once I get it to see how many of those chips are in there. There's a clip on youtube of someone playing a Hammond through one. Sounds nice.
I used to have a Boss DM2, which I liked. Got it at a yard sale in the late 80s for $15. Didn't use it enough to justify keeping it when I discovered they were going for $200. I miss it now and then-- I hope this is similar. They use the same chip.
Anyone here have one of these?
I don't think it's meant for guitar level input. Maybe I should put an input transformer in it?

User avatar
ubertar
ears didn't survive the freeze
Posts: 3775
Joined: Wed Feb 18, 2004 7:20 pm
Location: mid-Atlantic US
Contact:

Post by ubertar » Sat Feb 11, 2012 3:11 pm

This thing came today. It sounds great! Sounds just like my old boss dm2. The delay times aren't very long, so I put in a bigger value pot for the delay control. That worked, but it causes a high pitched squeal, so I put in a little low pass, and a bigger one (more highs cut) on a switch. The squeal gets lower as the delay time gets longer, so the low pass becomes ineffective after a certain point. With really long delay times, weird stuff starts to happen, an odd distortion along with high pitched sounds echoing lower, while low sounds remain unaffected.

I've heard it's possible to chain together multiple MN3005 chips to increase delay time. Maybe that'd be a better way to do it. The DM2 wasn't any longer, from what I remember.

User avatar
leigh
carpal tunnel
Posts: 1636
Joined: Wed May 07, 2003 11:16 am
Location: Portland, OR
Contact:

Post by leigh » Mon Feb 13, 2012 12:51 am

From what you're describing, the squeal is the clock signal itself leaking through. You could look for designs with a variable lowpass filter, so as you turn the delay time to be longer (and the clock's frequency lowers), the lowpass cutoff also lowers. Might be a fun mod to try.

Leigh

User avatar
leigh
carpal tunnel
Posts: 1636
Joined: Wed May 07, 2003 11:16 am
Location: Portland, OR
Contact:

Post by leigh » Mon Feb 13, 2012 1:00 am

ubertar wrote:With really long delay times, weird stuff starts to happen, an odd distortion along with high pitched sounds echoing lower, while low sounds remain unaffected.
And this is probably aliasing, where the high pitched sounds are exceeding the Nyquist frequency when you've got the delay clock running really slow. So you'd need to drop the frequency of the lowpass on the input side to prevent this (as opposed to what I was talking about in the previous post, which was dropping the lowpass on the output side to cut out clock noise). Or maybe it sounds cool and you leave it in!

User avatar
ubertar
ears didn't survive the freeze
Posts: 3775
Joined: Wed Feb 18, 2004 7:20 pm
Location: mid-Atlantic US
Contact:

Post by ubertar » Mon Feb 13, 2012 4:47 am

Thanks for the explanation of what's happening.

Is it possible to do the low pass in a way that it affects the squeal, but not the signal? I have the low pass on the output, in two stages-- one is permanent; the highs it cuts are too high to affect the tone. The other one is on a switch-- it gets rid of more squeal-- I can turn it on in longer delay settings when the squeal is a little lower. It also gets rid of the slight hiss the unit makes. It affects tone somewhat, but not in a bad way. It just takes off some of the brightness I might not want anyway. It's good to have it on a switch so it can be on or off. Any lower and we're really cutting into the signal. With the really long delay times, the squeal isn't a squeal anymore-- it's a mid-range tone, so low pass isn't really an option at that point.

If there's a way to get rid of that sound and the distortion without killing the signal tone, ideally that's what I'd like to do, but I'm not sure it can be done. The distortion and pitch change in really long delay settings could be cool for some things... it has an old-school low-budget sci-fi robot-like quality to it.

If I were to use it live, I'd just use it on the lower settings, where it doesn't squeal or where the low pass cuts it. For recording, I can use a section of the squeal on its own as a noise sample and filter that out with noise reduction (software).

User avatar
ubertar
ears didn't survive the freeze
Posts: 3775
Joined: Wed Feb 18, 2004 7:20 pm
Location: mid-Atlantic US
Contact:

Post by ubertar » Mon Feb 13, 2012 4:53 am

leigh wrote:From what you're describing, the squeal is the clock signal itself leaking through.
Maybe I'm taking "leaking" too literally, and thinking in terms of plumbing, but maybe there's a way to plug the leak, rather than filtering out what's in it? Or does that not make sense at all? I'm imagining the delay chip in a mu-metal container, like a transformer. This is probably crazy-talk, I know.

User avatar
leigh
carpal tunnel
Posts: 1636
Joined: Wed May 07, 2003 11:16 am
Location: Portland, OR
Contact:

Post by leigh » Mon Feb 13, 2012 10:27 am

ubertar wrote:Is it possible to do the low pass in a way that it affects the squeal, but not the signal? I have the low pass on the output, in two stages...
That is a great setup for lowpass, to have that extra switchable stage.

There is a mod you can do to eliminate some of that clock noise, without using more lowpass stages: balance the level of the BBD chip's two output pins (pins 3 and 4 of an MN3005) with a trimpot. Or possibly, this is already part of this effect's circuit. Do you see a trimpot anywhere on the circuit board?

BBD chips actually contain two parallel delay lines, offset by half a clock pulse. One line outputs on the rising edge of the clock, the other on the falling edge (see the "Circuit Diagram" on the MN3005 datasheet). These two outputs must be at the same amplitude, or else the offset will show up on the output as a square wave at the clock frequency.

So look for a trimpot off pins 3/4, or add one if it doesn't exist, and adjust for min clock noise.

A ton more info on BBD's in this GS thread - here's a direct link to a post about the output balancing: http://www.gearslutz.com/board/6693261-post27.html

cheers,
Leigh

User avatar
ubertar
ears didn't survive the freeze
Posts: 3775
Joined: Wed Feb 18, 2004 7:20 pm
Location: mid-Atlantic US
Contact:

Post by ubertar » Mon Feb 13, 2012 11:19 am

Thanks Leigh, I'll check that out. I'll have to open it up again to check for the trimpot. Any suggestion as to value if it doesn't have one? Maybe I should read the GS thread first before asking...
I'll let you know how things go.

User avatar
leigh
carpal tunnel
Posts: 1636
Joined: Wed May 07, 2003 11:16 am
Location: Portland, OR
Contact:

Post by leigh » Mon Feb 13, 2012 12:06 pm

ubertar wrote:Thanks Leigh, I'll check that out. I'll have to open it up again to check for the trimpot. Any suggestion as to value if it doesn't have one? Maybe I should read the GS thread first before asking...
I'll let you know how things go.
Have a look at the datasheet for their example circuit. Looks like they've got a 5k trim in there, with 100k resistors off each output to ground. So, depending on the circuit in your effect box, that would be a good starting point. If they followed the datasheet verbatim, you're set - otherwise I'd guess a trimpot whose value is 5% of the "terminating resistors" would work.

Also note on there the lowpass filters they're using in the example circuit - pretty steep. 24dB/oct on the input, 36dB/oct on the output. You didn't say how steep the filter you added was, but if you're still fighting clock noise after adjusting the trimpot, you could add a couple extra poles to your lowpass!

cheers,
Leigh


PS: also, if all else fails, you could always try this circuit: http://xkcd.com/730/

Post Reply

Who is online

Users browsing this forum: No registered users and 52 guests