Soundchips SN94624N vs SN76494N

[Sid1968]

Hi folks as you can see on this pictures my TI-99/4A 1981 V2 uses the SN94624N Soundchip while my TI-99/4A 1983 V2.2 uses a SN76494N Soundchip (marked with a red dot).

Can someone explain the differences between these chips and why they were changed?

 

Kind Regards

Sid1968

 

 

 

TI-99/4A 1981 V2

 

 

TI-99/4A 1983 V2.2

Edited September 21, 2019 by Sid1968

[Tursi]

The Wikipedia page breaks down the major variants of this chip: https://en.wikipedia.org/wiki/Texas_Instruments_SN76489

 

As for why, y'got me.

 

What's interesting is their table appears to be wrong, since as far as I know all the consoles use the sound chip's audio in line to mix external audio (and they say the 624 doesn't have it). I have the 494 datasheet (which covers four variants - 494, 494A, 496 and 496A - the 494s are the 500kHz version and the 496s are the 4Mhz versions, it's not clear the A is anything more than a revision. The datasheet mentions the 489A as the version without an audio input). I also have the 489A datasheet which explicitly states that it's identical to the 494N except for the input clock lacking a divide-by-eight stage (thus the max frequency).

 

But the variants have always been a bit confusing. As you've seen on your boards, there are more than just those in use.

 

 

[HOME AUTOMATION]

#9

[Sid1968]

30 minutes ago, Tursi said:

The Wikipedia page breaks down the major variants of this chip: https://en.wikipedia.org/wiki/Texas_Instruments_SN76489

 

As for why, y'got me.

 

What's interesting is their table appears to be wrong, since as far as I know all the consoles use the sound chip's audio in line to mix external audio (and they say the 624 doesn't have it). I have the 494 datasheet (which covers four variants - 494, 494A, 496 and 496A - the 494s are the 500kHz version and the 496s are the 4Mhz versions, it's not clear the A is anything more than a revision. The datasheet mentions the 489A as the version without an audio input). I also have the 489A datasheet which explicitly states that it's identical to the 494N except for the input clock lacking a divide-by-eight stage (thus the max frequency).

 

But the variants have always been a bit confusing. As you've seen on your boards, there are more than just those in use.

 

 

Thanks... cant you correct that in the wiki? There are still questions about the differences between the SN94624N and the SN76494N. What would you call the most advanced soundchip for the TI-99/4A?

Edited September 21, 2019 by Sid1968

[Tursi]

15 hours ago, Sid1968 said:

Thanks... cant you correct that in the wiki? There are still questions about the differences between the SN94624N and the SN76494N. What would you call the most advanced soundchip for the TI-99/4A?

I stopped messing with Wikipedia a long time ago.

 

From a software point of view, the chips in the TI console all operate the same way, except /maybe/ if you load a frequency count of 0. Reportedly the later sound chips will output a flat line rather than a low frequency tone (on the consoles I have, a count of 0 equals a count of 1024 - lowest possible tone). I've seen in photos those chips have been placed on TI motherboards, but I've never tested the assertion myself.

 

There's a "FourTI" board that puts four sound chips in the expansion box for more complex sounds, and a SID Blaster that puts a C64 SID chip there, but otherwise there's not been a lot of work on TI sound...?

 

[Sid1968]

On the C64 the different Soundchips and different Revisions of it sound more or less different. Sometimes the differents are huge. Would be interesting to find out, how the different soundchips sounds on TI-99/4A.

[Tursi]

27 minutes ago, Sid1968 said:

On the C64 the different Soundchips and different Revisions of it sound more or less different. Sometimes the differents are huge. Would be interesting to find out, how the different soundchips sounds on TI-99/4A.

That is true for the SID, but not noticeably for the SN chips. The difference on the C64 chips is caused by the analog sections, and the TI chip has only barely got an analog section.  Even comparing to the later chips in the Master System, which changed the noise patterns, you can really only tell the difference if you're using user-defined shift rates on noise, the tones still sound pretty much the same.

 

There's only so much that can be done with a square wave.

 

 

[+OLD CS1]

What about Ninerpedia?

[Sid1968]

2 hours ago, OLD CS1 said:

What about Ninerpedia?

You mean that the answer is there? Please show it to us. ?

Edited September 22, 2019 by Sid1968

[+OLD CS1]

8 hours ago, Sid1968 said:

You mean that the answer is there? Please show it to us. ?

No.  I mean using Ninerpedia instead of Wikipedia.

[wierd_w]

Considering the Ti has a Centronics parallel port addon, an LPTOPL3 is possible as well. (It's a Yamaha FM synth hooked to an 8bit LPT port, for vintage laptops without sound hardware.)

 

You would have to write software to control it, but it could be tacked onto a Ti, at least in theory.  With clever use of its tone generators and volume control, you CAN do 6bit mono PCM. Some old dos games that suppored adlib cards did that. The LPTOPL3 would basically be an adlib card, just tacked on the Ti.

[Tursi]

Bah, if you're going to do PCM through the parallel port, just use a resistor array and get the full 8 bits.

 

I actually wanted to put that into Dragon's Lair, but I ran out of time

 

[+FarmerPotato]

OPL3. I'm not sure I could love it.

 

Here is a table of data I made of the Yamaha chips, for Geneve2020. 

 

 

YM2149 SSG  3 ch                    clone of AY8912  MSX2
YM2151 OPM  8 ch 4 op          $1   Altered Beast, Marble Madness, Rastan. accept DX7 algos. CX-5MU.
YM2164 OPP                    $14   DX-100, CX-5M-II. Use YM3012 DAC. Never as good as DX-7.
YM2203 OPN  3 ch 4 op               
YM2413 OPLL 9 ch                    MSX-2, SMS
YM2608 OPN2 6 ch 4 op          $3   2612 parent + ADPCM, Rhythm, AY-3-8910, GPIO. SDIP-64.  I2S like.
YM2610 OPNB                         NeoGeo. Rastan Saga II. SDIP-64
YM2612 OPN2 6 ch 4 op          $2   Megadrive/Genesis. multiple "favorite". Analog stereo out. 2x2203. DIP24
YM3812 OPL2 9 ch 2 op or 6+5   $1   Adlib, Soundblaster.   use YM3014 DAC, 1-pin data. DIP24
YMF262 OPL3 18 ch 2 op/6 4op   $1   Soundblaster AWE32.  2x3812.  100-PQFP
YMF278B OPL4 18 ch 2 op/6 4op  $4   Moonsound for MSX. Most use only PCM. I2S like. 80-QFP 0.8mm

I'm still considering Yamaha chips above rolling my own FM synthesis, on the principle of "use real vintage chips." FPGA cores are kind of soulless.

 

One inspiration is the failed Yamaha CX-5M (MSX) music computer I saw in Compute! They said MSX would "invade" the US market (a prediction that failed. I kept a copy of the book "Here come the MSX Computers!"

I read that the last great era of FM synthesis was on the YM2612 OPN2 on Megadrive/Genesis. It has analog out, but the parent YM2608 has digital out, plus ADPCM and more.  I'm looking for something vintage with recognizable character. This could be it.

 

The YMF278B OPL4 is the end of the line? From what I read about its use in the MSX Moonsound card, the community used PCM almost exclusively, ignoring the FM core. So those MSX machines entered the modern all-sampled-sequenced era.

 

Choosing between the OPL4 and OPN2, I'm leaning toward OPN2 parent YM2608, though I haven't studied the FM programming manual to see what's really different between OPN2 and OPL-type. I haven't done any FM programming since high school, and then I had no manual or nothing. It also had ADPCM, unlike the OPL3 chip. 

 

The DX7 had sine wave carriers and 6 operators. You used up an operator to transform a sine into a sawtooth or square (with modulation and feedback). Later chips had up to 4 operators but other waveforms. I'm skeptical that the character is really reproduced by a canned waveform in a 4op in place of the sculpted one in the original 6op. Which was never released as a chipset.

 

The Yamaha chips are all dirt cheap now, though I remember seeing OPL chips advertised in Computer Shopper for more money than I had as a teenager.
 

Looking over all of these, and hearing crappy FM patches, makes me sad. I need to hear more good patches. None of them have the 6 operators of the DX-7 and anything more diminished than that seems not a good investment.

 

 

 

 

References

https://web.archive.org/web/20160516042639/http://www.dtech.lv/techarticles_yamaha_chips.html

https://en.wikipedia.org/wiki/List_of_Yamaha_Corporation_products#Sound_chips

http://www.atkinsoft.com/soundgenerators.html

https://electronicmusic.fandom.com/wiki/CX5M

https://web.archive.org/web/20160414051547/http://gendev.spritesmind.net/forum/viewtopic.php?t=386&postdays=0&postorder=asc&start=0

https://web.archive.org/web/20160414051547/http://nemesis.hacking-cult.org/MegaDrive/Documentation/GenesisSoftwareManual.pdf

http://nemesis.hacking-cult.org/MegaDrive/Documentation/YM2608J Translated.PDF

https://web.archive.org/web/20160516042639/http://www.dtech.lv/files_ym/ym3812.pdf

http://www.msxarchive.nl/pub/msx/docs/datasheets/opl4.pdf

https://www.reddit.com/r/synthesizers/comments/4mwyq9/how_limiting_is_4_operators_compared_to_6_in_fm/

 

The Best - YAMAHA DX7 - Sounds - EVER - Paul Eastham. REVERB added.

 

 

 

 

[wierd_w]

5 hours ago, Tursi said:

Bah, if you're going to do PCM through the parallel port, just use a resistor array and get the full 8 bits.

 

I actually wanted to put that into Dragon's Lair, but I ran out of time

 

disney soundsource for the win?

 

but then you miss out on the FM synthesis...

[Tursi]

37 minutes ago, wierd_w said:

disney soundsource for the win?

 

but then you miss out on the FM synthesis...

Essentially, yeah

 

FM synth is good but without someone who knows how to make it sing, it doesn't really do much good. I have a personal bias against adding hardware that will never see software. (Doesn't mean other people can't, just that I'm biased against doing it myself  ).

 

[matthew180]

Probably the only way to finally know for sure how each version of the SN sound chips actually differ, would be to get each one and test them.  You would need to be able to adjust the input clock frequency to verify the divider input stages, have an audio input source to see if it mixes with the output, and a way to send programming data to the chip to set the count to 0 (zero) and see if the output flat-lines.  Any volunteers?  You could *almost* use the TI console for the test, other than being able to modify the input clock frequency (easily).

 

[Sid1968]

Nobody dares to answer.. ;-)

[Tursi]

You got 16 answers

[Sid1968]

On 9/27/2019 at 11:23 PM, Tursi said:

You got 16 answers

I applied to matthews last suggestion...

[Tursi]

12 hours ago, Sid1968 said:

I applied to matthews last suggestion...

Well, on your bike.

 

[Sid1968]

Just started... what do you folks in the meanwhile? ?

Edited September 29, 2019 by Sid1968

[+FarmerPotato]

On 9/25/2019 at 4:41 PM, matthew180 said:

Probably the only way to finally know for sure how each version of the SN sound chips actually differ, would be to get each one and test them.  You would need to be able to adjust the input clock frequency to verify the divider input stages, have an audio input source to see if it mixes with the output, and a way to send programming data to the chip to set the count to 0 (zero) and see if the output flat-lines.  Any volunteers?  You could *almost* use the TI console for the test, other than being able to modify the input clock frequency (easily).

 

I'm up for it. I have one on a breadboard with an AVR that plays a Parsec explosion. I'll probably rebuild it though!

Clock frequency is solved now by an FGEN.

 

[Sid1968]

On 9/29/2019 at 5:24 PM, FarmerPotato said:

I'm up for it. I have one on a breadboard with an AVR that plays a Parsec explosion. I'll probably rebuild it though!

Clock frequency is solved now by an FGEN.

 

 

Do you make progress or do you need support?

Edited October 6, 2019 by Sid1968