FEMTOVERB /Wavefrontsemi/ PCB layout

[MetalGuy]

The encoder is connected as per schematic.
I spent couple of hours staring at the encoder's output table:

http://catalog.sensing.honeywell.com/printfriendly.asp?FAM=encoders&PN=510E1A48F416PC

Also I found this:

0110 Hall 1 Bright hall reverb for drums, guitars, and vocals.
0010 Hall 2 Warm hall for acoustic guitars, pianos, and vocals.
1010 Room 1 Hardwood studio for acoustic instruments.
1110 Room 2 Ambience for acoustic mixes and synth sounds.
1111 Room 3 Warm room for guitars and rhythm instruments.
1011 Plate 1 Classic plate reverb for lead vocals and instruments.
1001 Plate 2 Sizzling bright plate reverb for vocals and drums.
1101 Plate 3 Short vintage plate reverb for snares and guitars.
1100 Chorus Stereo chorus for guitars and pianos.
1000 Flange Stereo flanger for jet wash effects.
0000 Delay 1 125ms slapback delay for vocals and guitars.
0100 Delay 2 190ms delay for percussive arpeggios.
0101 Chorus/Room 1 Chorus with reverb for guitars, synths, and pianos.
0001 Chorus/Room 2 Auto-wah guitar effect with reverb for lead instruments.
0011 Vocal Cancel Removes lead vocals from many stereo recordings.
0111 Rotary Speaker Rotary speaker emulation for organs and guitars.

These codes are valid only if you count encoder's pins /from the output table/ as follows : 8 - 1 - 4 - 2  /physically 5 - 1 - 2 - 3/. This exactly according to the schematic.
On the other side AL3201 pins 5, 6, 7 and 8 are marked as  Prog0, Prog1, Prog2 and Prog 3. Why the first bit is assigned to pin 8 /Prog3/ then? Shouldn't bit 1 go to Prog0, bit 2 to prog1 and so on?
Otherwise everything is working only the effects order is somewhat messed up.

[toneman]

That's a Grey Code Encoder.
Binary would be easier to relate.

[bassmeister]

That's a Grey Code Encoder.

The idea behind Gray code is that you only toggle one bit at a time. In ordinary binary code the worst case scenario is when you switch all bits at one time, for example 1111 -> 0000 (of course one, two or three bits can switch simultaneousy as well). In between states there can be short periods of undefined "middle" states (don't know the technical term) when the physical contact surfaces for each bit don't line up exactly, a problem which might get worse with time when the encoder wears out. What happens is that don't jump from for example 0111 -> 1000 (four bits toggled, binary code) but 0111 -> nnnn -> 1000.

Quote from first search hit at Google using "Gray code" as search words:
"Gray codes are particularly useful in mechanical encoders since a slight change in position only affects one bit. Using a typical binary code, up to n bits could change, and slight misalignments between reading elements could cause wildly incorrect readings."

[RobB]

So, what's it sound like?  Any good?

[MetalGuy]

I would say very good and this is expected at 24-bit/ 48kHz. The sound is nice and fat. The biggest "drawback" is everything is preset but it provides several useful presets that I needed. And let's face it - there's no way you can get several professional quality types of reverbs /and other effects/ in one DIY project for this money. This is maybe the first  DIY real reverb/s project. And let's not forget how big is a single reverb tank...
So is it perfect? Hardly. Does it worth the money and the effort? You bet!

[Michael Allen]

How did you guys go about getting the chips fram Wavefront? Just call and order them? It seems like you guys had a heck of a time getting them. Nice job though! I'm making this my next project...and first attempt at digital stuff..

[MetalGuy]

I got the chips directly from them - it took me almost 3 weeks and couple of angry emails.
AFAIK currently these chips  can't be found anywhereelse. Steve, where are you?

[Peter Snowberg]

I got my chips from them directly too... but that was when they were called Alesis-Semiconductor. I spent about 15 minutes on the phone with the CEO and he was very helpful. Their sales guy was very helpful too and while he kept the engineers fairly insulated from the outside, they get all my questions answered very quickly.

[gaussmarkov]

I got the chips directly from them - it took me almost 3 weeks and couple of angry emails.
AFAIK currently these chips  can't be found anywhereelse. Steve, where are you?

i got mine direct and it took much more than 3 weeks.  persistence works and, though they
were slow, they were also very nice.

still haven't built this yet.     it will be my first pnp blue project and i tend
to approach the "big moment" slowly.    and i have had almost no bench time
for the last several months.    i was really pleased to see that you got it going, metalguy.
nice work!

[MetalGuy]

Everything works fine except for the problem mentioned above  - the efffects order is somewhat different.
I would really appreciate if any one could solve this "mystery"?

[Peter Snowberg]

If your encoder outputs Grey code, then the problem is going to be in mixing up the data bits.

[MetalGuy]

I'm using the same encoder as suggested on the schematic, ordered from Allied. It's connected as per schematic.
I'm just curious if anyone is using the same encoder and has similar problem?

[MetalGuy]

I spent several hours trying to figure out why the effects order is messed up because there's one thing for sure - if you connect the encoder as per schematic you can't get that effect order. Below are some observations concerning the matter.

When  rotating the encoder clockwise depending on where you start you're going to get the following output /pin 1,2,3 and 5 - pin 4 is common or gnd/:

1   1   1   1
0   1   1   1
0   1   0   1
1   1   0   1
1   0   0   1
0   0   0   1
0   0   1   1
1   0   1   1
1   0   1   0
0   0   1   0
0   0   0   0
1   0   0   0
1   1   0   0
0   1   0   0
0   1   1   0
1   1   1   0

where 1 is +5V and 0 is 0V or ground.

When comparing the original effects order table from my previous post and this one you'll notice that that effects order using this encoder is not possible. Why? Let's take a look at the position of 1111 and 0000 in both tables. In the first table 0000 is 6 positions away from 1111, while in the encoder's table it's 10 positions away /in both cases clockwise/ which means that no matter how you combine those bits you can't get that effect order. The only way you can get that is to substitute 1 for 0 and 0 for 1 and connect as 5123, not 5312 as suggested. I don't know if that is possible - maybe some of the experts should chime in.
Next was to check if any effects order is posssible and how. After spending couple of more hours combining bits the only combination  that promises a plausible effect order turned out to be 2135. This way you should get the effects in the following order:

2   1   3   5

1   1   0   0   Chorus
1   0   0   0   Flanger
1   0   1   0   Room 1
1   1   1   0   Room 2
1   1   1   1   Room 3
1   0   1   1   Plate 1
1   0   0   1   Plate 2
1   1   0   1   Plate 3
0   1   0   1   Chorus/Room 1
0   0   0   1   Chorus/Room 2
0   0   1   1   Vocal
0   1   1   1   Rotary
0   1   1   0   Hall 1
0   0   1   0   Hall 2
0   0   0   0   Delay 1
0   1   0   0   Delay 2

which means that encoder's pin 2 should connect to AL3201's pin 8, 1 to 7, 3 to 6, 5 to 5. I haven't tried this yet but the weekend is near and it will be done. I hope it works.
If you see any flaws in my observations please let me know so we can finally put the cherry on top of this project. Also if anyone has any suggestions concerning this problem I would appreciate if you share them with us.

[MetalGuy]

It's now confirmed that if you connect the encoder in the way described above /2135/  you're getting that effects order.

[gaussmarkov]

It's now confirmed that if you connect the encoder in the way described above /2135/  you're getting that effects order.

thanks for working this out!

[A.S.P.]

encoder "pinout"?

2135 is backwards/mirrored 5312 ...

[MetalGuy]

Yes, it is. Any other combination I've tried works only partially - at some point the effects get mixed.

[A.S.P.]

oops - my last post missed a wink/hint-smiley... 

[QSQCaito]

One question about these guys, is there any layout with components values, or any parts list.. any sound sample.. difficulty something to have an idea, how it sounds etc.


thansk a lot


bye bbye

DAC

[Peter Snowberg]

I don't know of any sound samples on-line, but the capabilities are almost exactly the same as the Alesis Picoverb, but the FemtoVerb is more suited to guitar signals.

See a review of the picoverb here: http://www.soundonsound.com/sos/Mar03/articles/alesispicoverb.asp

For a layout with component values, read this whole thread.