Various System Clocks and generation of 10MHz NuBus Clock in IIsi & SE/30

[Trash80toHP_Mini]

For starters: Macintosh IIsi @25MHz: NuBus Adapter Complications?

Found some old info on speeding up (Oscillator Replacement Method?) the Macintosh IIsi and am curious if anyone has done this and experimented with the Apple NuBus Adapter thereafter? Very interested to hear about any problems that may have arisen from the modification. Apple's adapter might exhibit strange behavior or might work just fine if it was designed to run on the IIsi at either the 20MHz underclock as released or its 25MHz design spec? The Apple adapter has a 40MHz Crystal on board, the same as the 40MHz Can for the System Clock on the Logic Board. If that's swapped out for a 50MHz can to "chip" the IIsi to 25MHz, does the 40MHz can on the NuBus Adapter somehow stay in sync with the faster system clock logic?

Curious about that as we've been playing with the bared to the bones NuBus Adapter for the IIsi from Radius. Having no crystal on board, that adapter simply halves the 20MHz System Clock on the IIsi 030 PDS to generate the required 10MHz NuBus clock. The Radius adapter would have NuBus running at 12.5MHz in a IIsi at 25MHz as it clocks NuBus at 8MHz when used in the SE/30.

IIsi Clocks:

40.0000MHz - divided by two for System Clock?

31.3344MHz - divided by two for 16MHz reference Clock on 030 PDS?

30.2400MHz - Pixel Clock for Vampire Video 640x480@66Hz Output - per Wikipedia Table

57.2823MHz - Damfino?  :huh: Pixel Clock for 640x870 Portrait a/o 512x384 RGB output?

Very interested in hearing from anyone with a Hot IIsi and Apple NuBus Adapter. Wondering about wonky video output refresh freqs from a NuBus VidCard on an over/underclocked NuBus expansion bus for one thing? 8-o

Background:

Thanks to @nickpunt finding a Radius IIsi NuBus Adapter on eBay some time back and letting me know it might perhaps be a new path toward NuBus on SE/30, we were able to tempt @joethezombie into buying it! joe quipped that he officially hates us for that, but he's had the schematic for his DOA adapter done for some time now. Recently, from joe's detailed pics of his card, @Bolle identified the solder side of another (unmentioned in the listing) Radius NuBus Adapter installed where it didn't really belong in a Mac IIsi listed on eBay. @maceffects generously acquired and send same to Bolle, where he got the goods on the formulas for all logic on board. As Bolle said in the eBay Finds thread, it "just works in the SE/30," albeit at only 8MHz.

Goal:

Gotta find a way to synchronize a 10MHz NuBus clock with the16MHz System Clock of the SE/30 which is tied to that line and the 16MHz reference line on its 030 PDS. Generating 10MHz from 16MHz for the SE/30 NuBus adapter should also allow a IIsi version of the adapter to run off the 16MHz reference clock in a IIsi running anywhere between its 20MHz Stock and an FDD torquing 28MHz Clocks.

Phase Lock Looping 16MHz down to 62.5% yields 10MHz. but it can't be that simple? In the first rev. discrete component NuBus implementation Macintosh II, it's a bit more complicated than that.

Macintosh II Clocks:

40.0000MHz - divided by four for NuBus?

31.3344MHz - divided by two for 16MHz System Clock as is the case in the single clock SE/30?

View attachment 25374

How are the two clocks synchronized and what the heck is that beaded-loopy-thingie at L6?

Suggestions? Answers? WAGs?

 

[joethezombie]

The beaded-loopy thing is an inductor, usually used to filter high frequency noise.

I haven't overclocked my IIsi, but it sounds fun.

As far as clock synchronization, if the differing clocks leading and trailing edges are aligned properly to avoid hold time violations, it may be just a matter of utilizing D flip-flops and simple REQ / ACK handshaking.  Still learning these techniques.  It would be interesting to try and shoehorn the IIcx implementation onto a SE/30 main board.

EDIT:  Looking at the IIcx schematics, indeed the clocks are fed into a 74F240 line driver and 74ALS374 D flip flop. Hmm....

 

[Trash80toHP_Mini]

Interesting you should say that. I've been visualizing a board carrying the 31.3344MHz Can from the SE/30 LB along with the 40MHz Can as in the Mac II piccie. 10MHz RFI generation antenna connecting ClockCard to Radius SE/30 NuBus Adapter with the whole shebang sitting in a clock socket (or better) installed in the vacated thru-holes of the SE/30's can to drive the MoBo.

 

[bibilit]

One of my IIsis is overclocked to 25 Mhz using a 50 Mhz oscillator

https://68kmla.org/forums/index.php?/topic/48615-clocking-the-iisi/

 

[Trash80toHP_Mini]

I'll bet that's a fun one. Have you tried any NuBus Cards out in it yet?

Figuring out whether the 40MHz Crystal on Apple's IIsi NuBus Adapter stays synced with the faster CPUCLK line on your machined will be an interesting data point. 20MHz Shipped spec vs. 25MHz Design spec of CPUCLK set up as an either/or proposition for the adapter would be a PITA, limiting overclocking the IIsi to only the 50MHz Crystal swap you have done, not to mention being wholly unhelpful in terns of NuBus in SE/30.

However, if Apple's NuBus Adapter syncs to the 16MHz PDS reference clock at C16M, we might be able to simply add a Crystal to Radius' IIsi NuBus Adapter in like manner.

@joethezombie The IIcx setup is the same as in the Mac II pic above where 33MHZ and 40MHz clocks are tied together/harmonized in close proximity as they are on the IIsi board as well. Might the 40MHz crystal on the Apple NuBus Adapter be (re?)harmonizing its 10MHz NuBus clock output with C16M on the PDS? My question would be how localized does the coupling of the clocks need to be, as in would it be feasible to jumper a 10MHz clock line to a Radius Adapter clone from a dual clock bodge on the SE/30 logic board?

Is there a IIci schematic to compare against the IIcx schematic? No board handy, what's the clock assortment on the IIci board? 50/40/33?