Is my Electric Mistress working properly???

[Stefan]

Hello,

a few weeks ago I decided to build the JD Electric Mistress (GeneralGuitarGadgets). I ordered all the required stuff, even one of the "holy" SAD1024's from Small Bear. This weekend I built all the things together and did my first test.

The circuit is working    It modulates and produces that basic sound that I expected, so far so good.....

But the depth and intensity of the effect is not that much that you can hear at the ToneFrenzy example. The clock, gain and the bias trim-pots are working properly, but the balance and feedback pots deliver no difference to the sound.....

Is there anybody out there who has an idea about what is going wrong in my circuit?? Is my SAD 1024 defect? I ordered an LM311 but got an LM 211, but it should work also......doesn't it? All the other parts that have been used match the requirements.

Thank you very much in advance for all the hints....  

Greetings from Bavaria, Germany

[Bill Bergman]

Sorry, cant help much other than saying you should notice quite a bit  of difference at max setting. I used one of the trim pots as a surface mount control  (the clock,I think) to be able to dial in more flanging weirdness when needed.

[Stefan]

Thanks Bill! Yes, I will replace the clock trim-pot and mount it on the surface. I already recognized the impact of changing the clock rate.....cool mod... But the depth of the effect on my mistress is not that much than expected... (I don't know how to describe it exactly, I am not a native english speaker...sorry...)

Greetings

Stefan

[Bill Bergman]

Stephan, your english is very good. The depth should go from almost no effect to almost a bird chirping noise and the other one controls the resonance sound .. from an open to almost a ring sound. Maybe someone like Mark Hammer could help. I'm not to good at trouble shooting.

[Mark Hammer]

I heard my name called... :wink:

Although the SAD-1024 and other BBD chips *are* static-sensitive, they are much hardier than people think.  I bought a bunch of MN3007's from a place that kept them loose ni a drawer without any static protection at all, treating them like TTL chips or op-amps.  When I bought them, the guy stuck them in a normal plastic bag with a bunch of other components, no conductive foam, no anti-static coating on the bag, not even any concern about discharging static electricity from his body before handling them.  When finally I got around to using them, my first project failed to work, and I thought I had bought a bad batch.  Not only that, I had encouraged Steve Daniels to buy the rest of them at that store, and I felt guilty that maybe I had unintentionally lured him and all my DIY buddies into a "static-damage trap".  

As it turned out, the trouble with my project had nothing to do with chip damage, and it works fine, as do all the other MN3007's I bought at the time.  As it also turns out, Steve tests each and every BBD that comes his way (so THAT's why he had to leave his day job!  :wink: ).  Certainly you should be careful about static, but the risk is a bit like SARS, West-Nile Virus and Lyme Disease; just foloow the rules and everything should be fine.

Having said that, make sure that all your electrolytic caps are oriented the right way.  Make sure the regeneration trimpot is set appropriately (turn the regen pot up full and adjust the trimpot just until you start to hear a bit of iscillation, then turn it down a bit).  Make sure all your pots are functioning properly.  (Last night, I was having difficulty with the Speed pot on a Ross-Phaser based Univibe I made for someone.  I took my pliers and pinched the rivets where the lugs make contact with the conductive element so that they were secure, and that fixed the problem.)

I'm not sure what the "balance" control is that you refer to, but the balance trimpot adjusts the level of the two complementary signals coming out of the SAD1024.  Each of those signals carries noise from the clock as well as audio.  However, the clock signals they carry are out of phase.  When the balance is adjusted right, the two out of phase clock signals cancel, leaving mostly audio and less clock.  Playing with this trimpot will not provide a different wet/dry balance.

The "gain" trimpot adjusts the wet/dry balance, by adjusting how much delayed signal reaches the mixing stage.  You get maximum cancellation and notch depth (effect) when the gain trimpot is adjusted for a 50/50 balance between wet and dry.

If you look at the schematic at JD Sleep's generalguitargadgets site, you will see that the dry signal reaches the mixing stage via C18/R40.  The wet signal reaches the mixing stage via a combination of R14/R15.  In some delay-based effects, the regeneration/feedback signal is tapped BEFORE the wet and dry signals are mixed together, and in others, the feedback is derived AFTER the mixing.  The DEM uses the second kind of arrangement.  You can see that the feedback signal comes from the output (pin 7) of the final mixing stage.

This arrangement produces a richer effect, and is chiefly responsible for a lot of what folks like about the DEM.....when it works.  The problem is that if the wet/dry balance is not optimal, the feedback won't work very well.  In the first sort of feedback arrangement (which Boss uses in the BF-2), you can feed back the delay portion separate from the mixing.  If the wet and dry signals are poorly mixed, you will hear a strong feedback/regeneration sound, but it will be in the background.  In the case of the DEM, because of how it is designed, you probably won't hear much effect of the regeneration/feedback control UNLESS the wet/dry balance is set right.

Hope this clears up the problem.

[Bill Bergman]

Thanks Mark!
There wasn't much I could do to help.

[Jason Stout]

I took my pliers and pinched the rivets where the lugs make contact with the conductive element so that they were secure, and that fixed the problem

That is a good tip!! I have a few pots that this may help.
Thanks Mark!

[Peter Snowberg]

Just to add some about static....

The vast majority of static damage will not take a chip out right away, but it will degrade traces on the interconnect layer, and cause "holes" in the oxide under that metal layer.

When I took my Apple service certification classes, they had a nice bunch of electron micrographs showing the damage. Most of the zaps took out 20 to 50% of a trace.

After a zap, a 500,000 hour useful lifspan can be reduced to just a couple thousand or a few hundred hours. That's a big issue when you have to replace parts under warranty.

Just because it still works doesn't mean it isn't terminally broken.

Take care,
-Peter

[Mark Hammer]

That's interesting and useful information, Pete.  Does the impact of static depend on how "V" a VLSI chip is? I.E., in terms of this unseen erosion of traces, is the impact of static on an MN3005 that much greater than say a CD4007, or is trace size fixed regardless of how much is packed under that 14-pin epoxy covering?  I have a sense that trace size is maximized to what the available standardized chip volume permits, but I may be entirely wrong in that regard.

[Stefan]

Hi all,

thank you very very much for your kind assistance!!!  

Mark, your little adjustment guide is brilliant. I am working on my DEM now and it seems that my only problem is (and hopefully was) the adjustment. I will let you know later on/tomorrow if and how my DEM works.

Thanks a lot again!!!

Greetings

Stefan

[Peter Snowberg]

Does the impact of static depend on how "V" a VLSI chip is?

The interconnect trace size depends on the fab process in use at the moment. The more "V", the easier it is to damage things, but the less"V" you have in the process (implying older equipment), the more registration error there is between layers, which makes things more fragile.

As time goes on, older processes are retired and generally replaced with larger wafers and a "shrink" of the old die design. I assume there are problems in direct scaling a BBD because of the cap value for each stage, but I just can't see a company like Matsushita not doing everything they can to maximize wafer yield. I don't know anything about the fab process used for BBDs, but I would assume they're decently robust. This is just an assumption however.

If you have huge feature sizes, a static zap is like nicking the insulation and just grazing the wire. With small features, it's like taking a chunk out of the wire, increasing the resistance at the point which generates heat. That localized heat will case migration in the semiconductor which will kill it over time. If the current is great enough, the spot that was zapped will just blow like a little aluminum fuse.

I would like to note that this failure mechanism has nothing to do with gate oxide damage which is a much faster acting failure mechanism.

ESD damage pics: (from a google image search)
http://www.analog.com/Analog_Root/static/corporate/quality/papers/rci_ebic/image012.jpg
http://www.analog.com/Analog_Root/static/corporate/quality/papers/rci_ebic/image016.jpg
http://www.micromagazine.com/archive/00/06/0006m131a.jpg
http://www.micromagazine.com/archive/00/06/0006m131b.jpg
http://www.assurtech.com/images/1k.jpg
http://www.assurtech.com/images/1l.jpg
http://www.assurtech.com/images/1m.jpg


Take care,
-Peter

[mattv]

Now Peter, is this something I can look for through my lighted magnifying glass?  

[Mark Hammer]

Pete,

Yeesh!!  Those pictures aren't very reassuring.  I think I'm going to stop wearing the cashmere sweater and phentex slippers when I work with BBDs or CMOS!  :lol:

This whole unseen-effects aspect of working with semiconductors has me intrigued now, so I am particularly glad you raised the issue.  I'm also curious about the sorts of static discharges that would produce such damage.  Are we talking about the sort of thing that makes you afraid to use take the laundry out of the dryer (or drink from a metal water fountain), or something much subtler?

Some sort of FAQ with respect to static and device safety/longevity is probably in order.  The pictures have convinced me that a lot of what I took as "useful experience"  was akin to one's 65 year-old uncle mumbling something about "I been smoking since I was 11, and it ain't killed me yet (hack, hack, cough, cough, hork, hork)".  It would seem that the temporal disconnect (pardon the pun) between exposure to static charges and chip functioning can be misleading.

[Peter Snowberg]

Stefan,

Sorry to hijack your thread with all this talk about static electricity.


Mattv,

Can you see it under your lighted magnifying glass? Sure, just hold the magnifier up to the screen of the scanning electron microscope.


Mark,

At the time I was obediently absorbing the Apple service mantra (read: module swaps without understanding the guts), 4M SIMMs (30 pin) were just starting to appear. 1M SIMMs, GALs, and 68000s can be killed by static that is much to low power for humans to detect. Those dryer static shocks are a couple or a few orders of magnitude larger. Once you look at the volume of energy being dissipated within a couple square microns, you suddenly go from too little to feel it to being able to vaporize material.

Good idea about adding a static FAQ.


Take care,
-Peter

[Stefan]

Hello again,

sorry for the delay..... I put my DEM circuit in a Hammond 1590 and adjusted it following Mark's instructions. Well, what should I say......it sounds great!!! I tried to use it for several styles, it doesn't matter if you use it while your're playing funky licks or 70's Rock style guitar.....always brilliant!!!    

Thank you very much again for all the hints (even if the thread was hijacked..... :wink: )

I strongly recommend building the Deluxe Electric Mistress to anybody who is searching for that special sound (see Tonefrenzy.com for sound sample).

If somebody is interested in some pictures of mine, no problem, just reply to this thread....

Greetings

Stefan

[smoguzbenjamin]

You know I can't resist pics. Post 'em!