diff --git a/doc/7-systems/sms.md b/doc/7-systems/sms.md
index a560ce065..7991186a8 100644
--- a/doc/7-systems/sms.md
+++ b/doc/7-systems/sms.md
@@ -13,3 +13,19 @@ the original iteration of the SN76489 used in the TI-99/4A computer, the SN94624
   - `y` controls whether to select noise or thin pulse.
     - `0`: thin pulse.
     - `1`: noise.
+
+# chip flags
+## SN7 versions
+SN7 was extremely popular due to low-cost. Therefore, it was cloned and copied to no end, often with minor differences between each other. Furnace supports several of these:
+- SN94624, can only produce tones as low as 100Hz, and is clocked at 447 KHz.
+- SN76494, which can play notes as low as 13.670 Hz (A -1). It has a different noise feedback and invert masks.
+- SN76489, identical to SN94624, just without a clock divider
+- SN76489A, identical to 76494, just with a /8 clock divider
+- SN76496, literally identical to former. Why is it even here?
+- SN76496 with a Atari-like short noise. The chip of many legend and rumours, might be a result of inaccurate emulation.
+- Sega Master System VDP version has a different, characteristic noise LFSR.
+- Game Gear SN7, identical to the above, but with stereo
+- NCR8496, different noise invert masks
+- PSSJ3, literally identical to the former, it justs swaps "high" and "low" signals in the output, which results in no audible difference
+
+TODO: all these checkboxes