Aron,
I started another thread to avoid having to scrolling down such a long discussion.
I am back from holidays now and will be checking the schematics and publishing updates. Give me a day or two because we are having a heatwave at the moment and it is 35°C in my shed (no aircon, this is Britain!) so I have only managed to do a few of the tests that I planned to verify Z80 timings.
I love that Pololu SM regulator: it has great dropout characteristics and should power the drive easily. A bit pricey but good. My only concern is that I could not find any information on ripple and switching noise (upstream and downstream) and I could not see any filter inductors on the its board. It may be wise to add an LC network on its input and have two on the output to stop switching noise from propagating: one for the W2I and drive electronics and the other to the drive motor. Unfortunately, that would mean more board space. I understand that the two +5V pins on the 44-pin connector are split between logic and motor. I assume that the motor includes the head positioner because Peter makes a good point about current "gulps". Encouragingly, in my shed I use a 3A lab PSU which has the +5V supply to the W2I and the drive connected together and I am not seeing anything odd happen with my 3.5in or 2.5in drives. Also, the regulator needs to be away from sensitive circuitry such as the oscillators and the 75107/75110. Remember the fun you had when your SMPS was too close to your 7710!
Great idea to use oscillators to replace two (all?) of the ugly HCT04/xtal circuits. That could result in eliminating U44 altogether by using the spare gates in U47 & U38.
If you do manage to fit a 40-pin header on the board then 3.5in drives (and other adaptors) can use this but will need their own external 5/12V supplies.
Peter answered your question about where pin 27 on the RAM went.
Although I like the suggestion of using a more modern part for the RAM, the 62256 appears to be slower than the 61256. According to datasheets, the former is available in 70-150ns types whereas the latter is 12-25ns. To keep the design simple, I do not specifically decode the Z80 M1 cycle but just use MEMREQ. I measured the MEMREQ time this morning and it is 78ns for an M1 cycle and 120ns for a normal memory cycle. Unless the 62256 is available in a faster variant we might need to keep to the old chip.
Finally, I am happy to pay for the prototype board along with postage (ie reimburse you). I can populate it and debug it before others are ordered.
Mike
--
You received this message because you are subscribed to the Google Groups "Cromemco" group.
To unsubscribe from this group and stop receiving emails from it, send an email to cromemco+u...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/cromemco/0d2cda05-35c7-4059-abe7-de7797eeb0f2n%40googlegroups.com.
I do not have a problem with SMD ( apart from the fact that I don't have the technology to do it). I am not too hot on SMD chips but the few that I have looked at (a few 74 TTL) seem to have the same pin outs. More exotic ones may differ. Mike
To unsubscribe from this group and stop receiving emails from it, send an email to unsub...@googlegroups.com" target="_blank">cromemco+unsub...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/cromemco/0d2cda05-35c7-4059-abe7-de7797eeb0f2n%40googlegroups.com.
--
You received this message because you are subscribed to the Google Groups "Cromemco" group.
To unsubscribe from this group and stop receiving emails from it, send an email to unsub...@googlegroups.com">cromemco+unsub...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/cromemco/CAD9iBDLaGVNSv8n1ZHWsYP3rooaJqGMLEOmQaZ0ubmgu5DH4%2BQ%40mail.gmail.com.
To unsubscribe from this group and stop receiving emails from it, send an email to cromemco+u...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/cromemco/5c089b29-e497-47f3-98e3-5b608e06b372n%40googlegroups.com.
Yes with JLCPCB you can order as few as 5 PCB's with assembly so it's not like we'd need to make huge investment. I'm guessing we can get the boards made + shipped for under $30 ea and that's with the SMD components.
To view this discussion visit https://groups.google.com/d/msgid/cromemco/88fc78d1-43b9-4dcb-880e-8e7936415d19n%40googlegroups.com.
To unsubscribe from this group and stop receiving emails from it, send an email to cromemco+u...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/cromemco/bfe1d737-acc2-4d29-9d19-015ba6a5f038n%40googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/cromemco/0cb5baff-51ef-4662-9364-c95f9558324cn%40googlegroups.com.
Aron
I have done a review of the circuits and also corrected the items highlighted in my master diagrams. I attach the three that have changed.
During my review I noted the following:
Processor: the legends for U5-14, 7, 13, 8 should be TxA, TxB, RxA, RxB. This is not a wiring issue just the naming.
WDI I/O: trivial point but suggest remove the Schmitt trigger symbol from U13, the HCT part lacks hysteresis.
S-P conversion: U24-11 should connect to U39-5.
S-P: U25-9 (and associated connections) should connect to +WGATE.
S-P: The legend for U29-5 should read -CMDSEL1.
S-P: On U29 & U42 are the references to -CMDSEL0, CMDRW, -CMDSEL1, -CMDSTR, -CMDBUS01 and -CMDBUS00 connected through to the relevant pins on the 34-way WDI connector?
IDE interface: U30-6 should connect to U43-12.
IDE interface: the pulldown on the IDE connector pin 31 should be 10k.
Display: the timing capacitors for U102A/B and U103A should be 1uF. That was due to me telling you to change all the 1uF to 0.1uF.
I feel that we are moving forward at quite a pace on this project.
I would be pleased to hear if Peter has spotted anything.
Mike
If you use a 22.5792MHz oscillator with a 3.3V supply then this should work because the Vout-high should be ~3V which is within the TTL thresholds.
To view this discussion visit https://groups.google.com/d/msgid/cromemco/d64982af-e438-4cd6-aab2-a7cfb00274e3n%40googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/cromemco/a2d31968-4d3f-4b52-ab4a-c8c166211d24n%40googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/cromemco/c2f5b500-e0bc-422e-9296-35e52e4eb015n%40googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/cromemco/CAD9iBDJYgHGA2sQW2xyikhi7j%2BCYxn%3DKtxcZw8k3eekvrSx0Ww%40mail.gmail.com.
To view this discussion visit https://groups.google.com/d/msgid/cromemco/e6847f9d-380f-448b-b027-45bf1189454cn%40googlegroups.com.
Just want to make sure we're looking at the same thing, I am showing the SOIC part is 19 for $8.81:I think the one that is 14 for $9.10 is a SSOP part:
Also, it appears that JLCPCB stock the Z80 -> https://jlcpcb.com/parts/componentSearch?searchTxt=z84c0020
To view this discussion visit https://groups.google.com/d/msgid/cromemco/26cddad7-1955-422d-b31a-3b366a805d3cn%40googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/cromemco/bbb92c39-cb38-4b98-b74d-89b632d57351n%40googlegroups.com.
Hi,
That looks like an interesting board. I haven’t been reading all the details about this, but couldn’t there be more SMT chips on the board and just be ready to plug ‘n play?
Not that I would need one, but for the future, I think taking advantage of the SMT is better than having a mix on the board if the SMT is all available … or at least cover more of the chips etc.
==============================================================================================
C U L8r, °|° Walt Perko °|° "Kids ... teach them the good stuff, and they still learn the bad stuff on their own."
http://www.R2Pv1.com/ RoboGuts™ Intelligent content for 3D printing making S.T.E.A.M. education better, easier and more affordable
Experiments to learn how to use various Electronic Components, Structured Computer Programming, Phonemes for Speech &Song in any language, and Art.
"The World Needs a New Economic Model"
==============================================================================================
Sent from the Cyber7
From: crom...@googlegroups.com On Behalf Of Aron Hoekstra
Sent: Wednesday, July 16, 2025 9:40 PM
To: crom...@googlegroups.com
Subject: Re: WDI to IDE adaptor (2)
Ok so this is just me playing around - I am able to get all the components on a single side.. The chip layout doesn't make any sense at the moment, just wanted to see if it'd fit.. Should I put it all on one side like this or put all the SMD components on the back side?
Any comments on the positioning of the headers, drive or LEDs? For LEDs I was using through-hole right angle LEDs, but they do make "side view" SMD LEDs.. should I use some of those?
To view this discussion visit https://groups.google.com/d/msgid/cromemco/CAD9iBDK5sHgBnptbh0k-WzmUt%3DOkcMQrJWG-i3JPOLQnsPMGLg%40mail.gmail.com.
Just a thought: should we run pRESET from the S100 bus into the W2I?
I don't know of any issues or interactions for power up/power down. I have not had any problems. The only irritation is that the PIOs on the WDI have rather random outputs when power is first applied but I handle this in the software.The time it takes for the IDE to come ready means that you cannot boot immediately (5-30 seconds delay depending on model) but the IMI 7710 takes 15s to spin up and everyone lived with that.
To view this discussion visit https://groups.google.com/d/msgid/cromemco/9f4b06f1-d2ba-43a6-b573-26845960db15n%40googlegroups.com.
7438s are specified in the 7710 manual (albeit for the controller) and are used in the drive itself. My datasheet (TI rev March 1988) says that the max sink current for LS38 is 24mA whereas 7438 is 48mA. U17 & 18 are only driving terminators at the WDI so will have to sink 22mA whereas U19 & 20 drive terminators at both ends so need to sink 44mA.
Well, I heard back from LCSC on the Z80 SMD:" Sadly the item is not available for quotation currently."Good thing we checked first. But this also begs the question, weren't there a few other parts we selected that were "pre-order"?
Hi,
It is possible that if you buy the chips from another distributor/vendor, you could probably send the chips to JLCPCB to be installed.
JLCPCB was allowing that a few years ago, but I don’t know if they continue that service.
==============================================================================================
C U L8r, °|° Walt Perko °|° "Kids ... teach them the good stuff, and they still learn the bad stuff on their own."
http://www.R2Pv1.com/ RoboGuts™ Intelligent content for 3D printing making S.T.E.A.M. education better, easier and more affordable
Experiments to learn how to use various Electronic Components, Structured Computer Programming, Phonemes for Speech &Song in any language, and Art.
"The World Needs a New Economic Model"
==============================================================================================
Sent from the Cyber7
From: crom...@googlegroups.com On Behalf Of Peter Higgins
Sent: Thursday, July 17, 2025 11:33 AM
To: Cromemco <crom...@googlegroups.com>
Subject: Re: WDI to IDE adaptor (2)
On Thursday, July 17, 2025 at 11:11:10 AM UTC-7 Aron Hoekstra wrote:
Sent from the Cyber7
From: crom...@googlegroups.com On Behalf Of Aron Hoekstra
Sent: Wednesday, July 16, 2025 9:40 PM
To: crom...@googlegroups.com
Subject: Re: WDI to IDE adaptor (2)
Ok so this is just me playing around - I am able to get all the components on a single side.. The chip layout doesn't make any sense at the moment, just wanted to see if it'd fit.. Should I put it all on one side like this or put all the SMD components on the back side?
Any comments on the positioning of the headers, drive or LEDs? For LEDs I was using through-hole right angle LEDs, but they do make "side view" SMD LEDs.. should I use some of those?
To view this discussion visit https://groups.google.com/d/msgid/cromemco/ad43c0fa-7050-4d6c-af82-cfc3b3cfc602n%40googlegroups.com.
Hi,
I figured they stopped that. It’s extra work for them now days. Back a few years they didn’t have access to such a large catalog of SMD components.
==============================================================================================
C U L8r, °|° Walt Perko °|° "Kids ... teach them the good stuff, and they still learn the bad stuff on their own."
http://www.R2Pv1.com/ RoboGuts™ Intelligent content for 3D printing making S.T.E.A.M. education better, easier and more affordable
Experiments to learn how to use various Electronic Components, Structured Computer Programming, Phonemes for Speech &Song in any language, and Art.
"The World Needs a New Economic Model"
==============================================================================================
Sent from the Cyber7
From: crom...@googlegroups.com On Behalf Of Peter Higgins
Sent: Thursday, July 17, 2025 12:09 PM
To: Cromemco <crom...@googlegroups.com>
Subject: Re: WDI to IDE adaptor (2)
I think there is a FAQ question about that on the JLCPCB web site, and the answer is no.
On Thursday, July 17, 2025 at 11:37:47 AM UTC-7 Walt Perko wrote:
Hi,
It is possible that if you buy the chips from another distributor/vendor, you could probably send the chips to JLCPCB to be installed.
JLCPCB was allowing that a few years ago, but I don’t know if they continue that service.
==============================================================================================
C U L8r, °|° Walt Perko °|° "Kids ... teach them the good stuff, and they still learn the bad stuff on their own."
http://www.R2Pv1.com/ RoboGuts™ Intelligent content for 3D printing making S.T.E.A.M. education better, easier and more affordable
Experiments to learn how to use various Electronic Components, Structured Computer Programming, Phonemes for Speech &Song in any language, and Art.
"The World Needs a New Economic Model"
==============================================================================================
Sent from the Cyber7
From: crom...@googlegroups.com On Behalf Of Aron Hoekstra
Sent: Wednesday, July 16, 2025 9:40 PM
To: crom...@googlegroups.com
Subject: Re: WDI to IDE adaptor (2)
Ok so this is just me playing around - I am able to get all the components on a single side.. The chip layout doesn't make any sense at the moment, just wanted to see if it'd fit.. Should I put it all on one side like this or put all the SMD components on the back side?
Any comments on the positioning of the headers, drive or LEDs? For LEDs I was using through-hole right angle LEDs, but they do make "side view" SMD LEDs.. should I use some of those?
To view this discussion visit https://groups.google.com/d/msgid/cromemco/5505fbe2-d508-49c6-9d53-4fbcb6fd4692n%40googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/cromemco/ad43c0fa-7050-4d6c-af82-cfc3b3cfc602n%40googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/cromemco/50435d1f-dac1-4d89-b6a9-294ee6509e57n%40googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/cromemco/d4b12a78-91f3-48d9-b079-5320e03bf2e9n%40googlegroups.com.
I was thinking again about -RESET from the S100 bus and I agree that it is probably not necessary but then came up with a thought. If there were two pads 0.4in spaced with one going to -RESET (S100 bus pin 75) and the other to U45-1 then it would be possible to install a 1N4148 if it were ever needed. However, that depends on how the layout is going.
Another silly question since I've been out of the loop: I know it only works with the 7710, but can it emulate two drives?And does it work with a WDI-II as well as a WDI?
--
You received this message because you are subscribed to the Google Groups "Cromemco" group.
To unsubscribe from this group and stop receiving emails from it, send an email to cromemco+u...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/cromemco/68f0aa20-a03c-4ad3-bc44-4afd3366788en%40googlegroups.com.
--
You received this message because you are subscribed to the Google Groups "Cromemco" group.
To unsubscribe from this group and stop receiving emails from it, send an email to cromemco+u...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/cromemco/5bb68ee9-5f81-405c-a619-27ce302189c3n%40googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/cromemco/04bc2824-5283-4978-bdb6-42f366ac5b00n%40googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/cromemco/d4b12a78-91f3-48d9-b079-5320e03bf2e9n%40googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/cromemco/3c43c953-b11b-4ce3-b2a0-74d739df5e9dn%40googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/cromemco/48428be2-e45b-444c-b233-6f812e477482n%40googlegroups.com.
Curt, thanks for the warning. I have an idea! Aron, would it be possible to also put an SMD pad layout on the board for the Z80 so that if we cannot source the DIP then we could manually solder (!!!) a QFP/PLCC in?
To view this discussion visit https://groups.google.com/d/msgid/cromemco/05fc31bf-647f-4c16-bf82-e910168d5fe6n%40googlegroups.com.
Also I don't think I've added the external power headers to the schematic yet. What should I add inputs for: +8V, -16V, and GND?
To view this discussion visit https://groups.google.com/d/msgid/cromemco/dbe78322-7344-47d4-a344-808d9be4bb1bn%40googlegroups.com.