search

gzweigle / DIY-Grand-Digital-Piano

Building a hybrid MIDI digital piano
https://www.youtube.com/@gzpiano88
GNU General Public License v3.0
44 stars 3 forks source link

HW: sca part inconsistence #66

Closed davidedelvento closed 2 months ago

davidedelvento commented 3 months ago

The BOM lists U1 as REF6025 but kicad lists it as REF6030

gzweigle commented 2 months ago

Good question. The part comes in variants 2.5V, 3.0V, 3.3V, 4.1V, 4.5V and 5.0V. Probably when I did the schematic I was thinking 3.0V would work. However, that made VCE across the CNY70 too small. All my documentation in stem piano repo is for 2.5V. For your sensors you may be able to use a 3.0V or even 3.3V (check the amp range) as the reference value. The higher the reference the better for signal.

I added a comment to hardware README files that the bill of materials wins if different than schematic.

davidedelvento commented 2 months ago

I am not sure I understand the problem of Vce becoming too small when using an U1 which provides a higher voltage.

Assuming that Vce is 5V minus what U1 provides and comparing Fig.3 of https://www.sparkfun.com/datasheets/Robotics/QR_QRE1113.GR.pdf with Fig.8 of https://www.vishay.com/docs/83751/cny70.pdf it appears indeed that 3.0V might be marginally better.

But looking at the characteristics tables I don't see any other differences. Given the complexities of asking the vendor to manufacture some boards with one part and some others with another part, I'm tempted to let everyone take the 2.5V version. Do you think that would be a problem? I guess in the worst case I could try to desolder and replace it if necessary (it would be a painful thing to do, but not impossible, as far as I can tell...)

gzweigle commented 2 months ago

Yes, for HPS, vdd=3.3 and with vref = 2.5 the difference (vce) is 0.8 volts when hammer is at stop. If vdd > 3.3, vref can be bigger. I agree vref=2.5 for all boards is simpler. Select ACE resistors so the output is 0 to 2.5v as hammer goes from rest to stop.

davidedelvento commented 2 months ago

Just to make sure I don't misunderstand what you are saying, can you please confirm the following for HPS?

  1. Vdd is the voltage difference between pins 1 and 2 of the J0 connector, and it's fixed at 3.3V with the current design (nor you are suggesting to change it)
  2. Vref is what the output (namely pin 3 of the J0 connector) will be compared to in the ADC and with the current design that is 2.5V but it could be changed by simply swapping REF6025 for REF6030
  3. The change in 2 did not work for CNY70 without also the (not possible) change mentioned in 1
  4. The change in might work for the QRE1113 without the change mentioned in 1

If this is correct, I now understand how to correctly select resistors values for ACE.

Thanks a lot for your patience!

gzweigle commented 2 months ago
  1. Yes.

Vdd is an input to the HPS. For HPS 0.8, Vdd is the voltage difference between pins 1 and 2 of J0.

The HPS 0.8 is designed to receive 3.3V for Vdd.

  1. Yes.

For example, with the resistors in the HPS 0.8 parts list, HPS 0.8 will output a voltage near 0 when the hammer is at rest and a voltage of near 2.5V when the hammer is at the stop bar.

Therefore, I selected REF6025 for the ADC reference since 2.5V matches the maximum output of the HPS 0.8.

  1. This one is a little more complicated to answer:

REF6030 will also "work" with HPS 0.8. In this case, when HPS 0.8 is at its maximum output of 2.5V, the ADC will output 2.5/3.0 of full scale. The full scale of ADC is not used. This gives up some noise margin. But the piano will "work".

What didn't work was sizing the resistors on HPS 0.8 so that it outputs 3.0V when hammer is at the stop bar. In this case, the voltage drop across the CNY70 was too small (3.3-3.0).

  1. Not sure I understand but if the interest is to use REF6030 for all boards, it will also "work" for the HPS 0.8, with the caveat that these sensors will not use the full range of the ADC.
davidedelvento commented 2 months ago

Thanks, this is extremely helpful, now I understand everything (and I have been able to run some tests).

I now also understand better why you wanted to put potentiometers in some of these designs. There is a key-to-key variability which (upon inspection and testing) is due to the different reflectivity of the hammers. That, compound with the different brightness that can be achieved with different currents at the LEDs make settling to a definitive answer a bit tricky. The QRE1113 can easily reach 3.05V when energized around 3.3V if the target is "shiny" enough (e.g. white paper, new hammers). On the other hand, a less reflective target (e.g. old hammers, even after filing them to make them very white) can stay easily under 2.5V

Puzzlingly, that maximum value does not change when changing the value of the bias resistor on the phototransistor (the non maximum achieved with a old hammer does). As such, I am tempted to have a 6030 installed -- but I fear they would get confused and ship those to Europe and give me the 6025 instead, so I am settling on the 6025 to begin with and then I will figure out some way of "voicing" the hammers to achieve the maximum range.

Thanks