One more MegaDrum firmware to test

[fuzzysnuggleduck]

Two reason I haven't tested yet:

1. Outlined above. If MCT would work on my Mac for save and load all settings without trouble, I would do this right away.
2. I'm still using BFD1.5 which doesn't have the ability to do positional sensing and I'm on the fence about upgrading.

[kimouette]

fuzzysnuggleduck,
I'm also using a Mac and have the same problem with MCT.
But to do this test you only need your MD.
You can set your parameters using Left-Right-Up-Down buttons, save your settings in MD and simply look at the LCD to get the values needed to test this firmware.

[Kurtsa]

Light hits
1. M=33 m=7 A=15
2. M=49 m=15 A=21
3. M=53 m=17 A=27 (occasionally M=67 m=0)
4. M=65 m=21 A=27 (sometimes M=71 m=0)

Medium hits
1. M=41 m=7 A=19
2. M=67 m=19 A=29
3. M=79 m=19 A=57
4. M=81 m=19 A=60?

Hard hits
1. M=27 m=13 A=19
2. M=69 m=27 A=40
3. M=75 m=7 A=60
4. M=67 m=25 A=60

The results were quite mutch varying so not exactly scientefic but I tried my best . I have just put these things together, this is the first version of the pad and the wiring is temporary etc, so there might be extra error sources, so pls, do not let me be the only one doing this feedback practice...

I also took some oscilloscope photos (see file comments for clarification). When the pad was connected to Megadrum the results were varying more than when disconnected, so that's why also pictures without Megadrum. The waveforms were very stable and looked similar all around the pad. When it was connected to Megadrum there was some variation from hit to hit. These are taken with my 12" tom. I quickly looked also my 10" pad and the results seems to be somewhat similar.

In addition to my messy wiring, the varying input load of the Megadrum _might_ cause some inconsistency to the oscilloscope measurement. I was thinking that if during some part of the waveform the snareH input is read more often by MD then the voltage drops more... and the sharp edges of the "reading pulses" are filtered out by the capacitance of the piezo... don't know just a guess (at least with dc voltage applied to a input the reading interval seems to vary, but DC is totally different case though). The reading interval can perhaps be abnormal due to my inproper settings of the MD parameters also...

_
Kurtsa

[dmitri]

Kurtsa, many thanks for the first results and feedback! They are pretty much similar to my results with a Pintech snare except in my results I have troubling figures drop when hitting dead centre probably due to a wide foam cone.

Can you give more details about the pad you did the test with? Can you tune mesh's tension?

[dmitri]

Can I also ask you to repeat the test with a 1k-1k5 resistor in parallel with a piezo and adjusting Threshold/Gain/Highlevel (to achieve the same sensitivity and dynamic range)?

[Kurtsa]

Can I also ask you to repeat the test with a 1k-1k5 resistor in parallel with a piezo and adjusting Threshold/Gain/Highlevel (to achieve the same sensitivity and dynamic range)?

I'm soon hurrying to work, but I'll be playing with these at the evening. Were you thinking to make the resistance seen by the piezo more constant by adding a resistor parallel? If so, I can also try adding a buffer opamp... (if I can find a one that operates down to gnd).

_
Kurtsa

[dmitri]

Can I also ask you to repeat the test with a 1k-1k5 resistor in parallel with a piezo and adjusting Threshold/Gain/Highlevel (to achieve the same sensitivity and dynamic range)?

I'm soon hurrying to work, but I'll be playing with these at the evening. Were you thinking to make the resistance seen by the piezo more constant by adding a resistor parallel? If so, I can also try adding a buffer opamp... (if I can find a one that operates down to gnd).

_
Kurtsa

No need for an opamp, just a resistor will do.

[Kurtsa]

Can I also ask you to repeat the test with a 1k-1k5 resistor in parallel with a piezo and adjusting Threshold/Gain/Highlevel (to achieve the same sensitivity and dynamic range)?

I used 1.7kOhms (2x3k3). The signal level droped so that the sensitivity is worse than without resistor even with gain=8,threshold=2 (without resistor gain=0 threshold 990. The lightest hits were not triggered sometimes. But anyway here you go:

Light hits
1.-3. M=7 m=3 A=5
4. M=21 m=3 A=5

Medium hits
1. M=15 m=3 A=7
2. M=15 m=3 A=7
3. M=15 m=3 A=7
4. M=29 m=7 A=15

Hard hits
1. M=7 m=5 A=7
2. M=9 m=5 A=7 (a 47 alsoi)
3. M=13 m=5 A=7
4. M=29 m=13 A=21

You wanted to know about my triggers. Those are made in a quite typical manner (I think), except I don't use cross-bar. I replaced it with a plate that is fixed to the drum shell from six points (all of the lugs). I did this just for the the look... the cross-bars are always in a "wrong" angle . The cone dimensions are pretty mutch like the Roland cones. The material I took from the 3M sanding blocks. The tip of the cone is a bit smaller, diameter about 5mm and it is a bit taller (if I cut it to the Roland height then the dimensions are just like Roland's). I use mesh heads I found from local store... don't know the manufacturer, but somekind of dence net, wire-to-wire distance is about 0.7mm. A damper ring on top of the mesh head (not in the pictures), next to the rim. It is to prevent the ringing the head was having. The damper is fixed with thin douple sided tape. Piezo is 23mm/35mm, attached with 1.5mm (I think) double-sided soft foam tape on top of the rubber thing you can find in the pictures. Cone is attached to the piezo using thin double sided "red devil" tape.

_
Kurtsa

[dmitri]

Looks like low value resistor is a wrong direction. If it's not too much trouble, can you try with a 10k-22k resistor?

[Kurtsa]

Looks like low value resistor is a wrong direction. If it's not too much trouble, can you try with a 10k-22k resistor?

Errmmmh... Could you explain a bit what are you after. I'll do that, but I just disassembed the trigger and made some changes (different fixing arrangement for the piezo) and the location of the piezo is not exactly in the center anymore (a mistake). I have read that the position of the trigger is very critical for positional sensing. ...but don't know for sure might be a rumor. I'll fix the position anyway tomorrow.

<edit> After my modifications to the piezo attachment (made it softer) the behaviour of the pad is now more clearly as described in the Roland positional sensing patent. And I think the reversing of center piezo now makes perfectly sense. If you look the pictures of my earlier post... the narrow (200-300us) negative peak just in the beginning of the pulse seems to be actually very well representing the strength of the hit. So, to measure this more easily, the polarity should be reversed. A very fast and steady way to measure the strength of the hit I think. Then the position of the hit can be seen from the width of the following positive pulse (mayby with some filtering). Previously I was thinking that the position could be more accurately detected from the voltage difference of the beginning and end parts of the positive pulse. But now after my modifications the beginning of the pulse is pretty mutch just a slope, not a voltage level proportional to the hit strength as it seemed to be previously. The modification affected pretty mutch to the positive pulse, but the negative pulse seems to be guite unsensitive to tuning of the head and attachment of the piezo...

_
Kurtsa