From 2b523257cac1ccefa3bc835af5b6227fe01e3ccc Mon Sep 17 00:00:00 2001
From: Couriersud <couriersud@users.noreply.github.com>
Date: Sun, 6 Jul 2014 16:53:47 +0000
Subject: [PATCH] Document recent measurements done on the AY8910. [Couriersud]

---
 src/emu/sound/ay8910.c | 123 +++++++++++++++++++++++++++++++++++++++++
 1 file changed, 123 insertions(+)

diff --git a/src/emu/sound/ay8910.c b/src/emu/sound/ay8910.c
index 513d528bcbe..3b5b563ef73 100644
--- a/src/emu/sound/ay8910.c
+++ b/src/emu/sound/ay8910.c
@@ -1,3 +1,126 @@
+/*
+ * Couriersud, July 2014:
+ *
+ * This documents recent work on the AY8910. A YM2149 is now on it's way from
+ * Hong Kong as well.
+ *
+ * TODO:
+ *
+ * - Create a true sound device nAY8910 driver.
+ * - implement approach outlined below in this driver.
+ *
+ * For years I had a AY8910 in my drawer. Arduinos were around as well.
+ * Using the approach documented in this blog post
+ *    http://www.986-studio.com/2014/05/18/another-ay-entry/#more-476
+ * I measured the output voltages using a Extech 520.
+ *
+ * Measurement Setup
+ *
+ * Laptop <--> Arduino <---> AY8910
+ *
+ * AY8910 Registers:
+ * 0x07: 3f
+ * 0x08: RV
+ * 0x09: RV
+ * 0x0A: RV
+ *
+ * Output was measured on Analog Output B with a resistor RD to
+ * ground.
+ *
+ * Measurement results:
+ *
+ * RD      983  9.830k   99.5k  1.001M    open
+ *
+ * RV        B       B       B       B       B
+ *  0   0.0000  0.0000  0.0001  0.0011  0.0616
+ *  1   0.0106  0.0998  0.6680  1.8150  2.7260
+ *  2   0.0150  0.1377  0.8320  1.9890  2.8120
+ *  3   0.0222  0.1960  1.0260  2.1740  2.9000
+ *  4   0.0320  0.2708  1.2320  2.3360  2.9760
+ *  5   0.0466  0.3719  1.4530  2.4880  3.0440
+ *  6   0.0665  0.4938  1.6680  2.6280  3.1130
+ *  7   0.1039  0.6910  1.9500  2.7900  3.1860
+ *  8   0.1237  0.7790  2.0500  2.8590  3.2340
+ *  9   0.1986  1.0660  2.3320  3.0090  3.3090
+ * 10   0.2803  1.3010  2.5050  3.0850  3.3380
+ * 11   0.3548  1.4740  2.6170  3.1340  3.3590
+ * 12   0.4702  1.6870  2.7340  3.1800  3.3730
+ * 13   0.6030  1.8870  2.8410  3.2300  3.4050
+ * 14   0.7530  2.0740  2.9280  3.2580  3.4170
+ * 15   0.9250  2.2510  3.0040  3.2940  3.4380
+ *
+ * Using an equivalent model approach with two resistors
+ *
+ *      5V
+ *       |
+ *       Z
+ *       Z Resistor Value for RV
+ *       Z
+ *       |
+ *       +---> Output signal
+ *       |
+ *       Z
+ *       Z External RD
+ *       Z
+ *       |
+ *      GND
+ *
+ * will NOT work out of the box since RV = RV(RD).
+ *
+ * The following approach will be used going forward based on die pictures
+ * of the AY8910 done by Dr Stack van Hay:
+ *
+ *
+ *
+ *              5V
+ *             _| D
+ *          G |      NMOS
+ *     Vg ---||
+ *            |_  S Vs
+ *               |
+ *               Z
+ *               Z Resistor Value for RV
+ *               Z
+ *               |
+ *               +---> VO Output signal
+ *               |
+ *               Z
+ *               Z External RD
+ *               Z
+ *               |
+ *              GND
+ *
+ *  Whilst conducting, the FET operates in saturation mode:
+ *
+ *  Id = Kn * (Vgs - Vtn)^2
+ *
+ *  Using Id = Vs / (RV + RD)
+ *
+ *  Vs = Kn * (RV + RD) * (Vg - Vs - Vtn)^2
+ *
+ *  finally using Vg' = Vg - Vtn
+ *
+ *  Vs = Vg' + 1 / (2 * Kn * (RV + RD)) - sqrt((Vg' + 1 / (2 * Kn * (RV + RD)))^2 - Vg'^2)
+ *
+ *  and thus
+ *
+ *  VO = Vs * RD / (RV + RD)
+ *
+ *  and this can be used to re-Thenevin to 5V
+ *
+ *  RVequiv = RD * ( 5V / VO - 1)
+ *
+ *  The RV and Kn parameter are derived using least squares to match
+ *  calculation results with measurements.
+ *
+ *
+ *  FIXME:
+ *  There is voltage of 60 mV measured with the EX520 (Ri ~ 10M). This may
+ *  be induced by cutoff currents from the 15 FETs.
+ *
+ */
+
+
 /***************************************************************************
 
   ay8910.c