Jump to content
retro_doog

New power supply project for TI-99/4a

Recommended Posts

Looks like I figured out how to allow for lower than 12V input adapters without increased cost or complexity...

 

My original parameters for the 12V was an input voltage range from +6V to +18V. This was suggesting Step-Up/Step-Down combo regulators which are complex and expensive. They also typically have low efficiencies in the 75% to 80% range(still better than the old linears, I'm sure, but...)

 

Then I realized the way to make 12V is actually from my plentiful 5V switcher which is 90+% efficient! Also, a 5V to 12V 400mA boost regulator will be 86%-90% efficient depending on which IC controller I use(for about 20% more BOM cost, I can get the more efficient design). The cascade of near 90% efficiencies is pretty close to the lesser efficiency of the Step-up-down design, but much simpler and lower cost. :) In practice the combined efficiency of the 12V rail should be around 86% assuming there's not going to be much current drawn from it and no new hardware save oddball stuff that one might plug into the video connector is likely to increase the load by much, if anything.

 

Normally, If I thought I needed a very clean +12V, I'd boost to, say 13.5V, and then put an low dropout, ultra low noise, linear regulator to 12V from that. However, it looks like the +12 is largely being used by fairly noisy digital circuits anyway(The CPU and RAM), so I suspect as long as I design for low current ripple(reasonable output capacitance should cover this), the 12V rail will be happy. Usually a rail like this is used in Audio/Video circuits for OpAmp bias and the like, but it appears that the bulk of the AV circuits use a balanced +/-5V dual supply, and the console creates filtered copies of those rails for the analog "branches" of the console circuitry. Again, I just need to cover current ripple since the 5V rail in particular is subject to constant switching loads(albeit at much lower frequencies than modern computers ned to handle). Again, a reasonable amount of bulk capacitance should handle the 3.6MHz switching transients...

 

So, I've chosen the final candidates for all three rails and, thanks to TI's handy "WebBench" tool, I have the reference schematics for designs that should "just work" the first time as well! I may need to tweak the WebBench designs as they sometimes call out inductors that are not commonly stocked at DigiKey or Mouser, but that's usually not too big of a deal.

 

I'm still a bit busy with other things, but I suspect once I get started, the schematic will only take a couple of hours and the layout should go fast as well. I'll just need to find a free weekend day to get this thing knocked out. Also, I want to get some reference measurements on my other two related projects to ensure I can reuse these designs on those as well.

Share this post


Link to post
Share on other sites

So I just watched a video on the TiPi and thought "That's pretty cool. Don't know if I'd actually use it once the novelty wears off. But, what a mess of wires, stacked boards, external power adapters!"

 

I feel like I could probably make a single board that combines the 32K sidecar, TiPi with whatever R-Pi zero "hat" circuitry is needed, and just has the Pi-Zero plug into it. So a single board that you just pop the Pi-Zero module onto and put it in a nice case. Also, with my new supply everything would just be powered when you switch the TI on. Of course I'm assuming the hardware and firmware of the TiPi adapter board is "open" and freely available. I could probably even make a high performance version that you pop that DIMM looking R-Pi "compute" module onto. Of course I'd change all of those large DIP components into much more compact, but not impossible to hand solder SOICs and possibly have components on both sides of the PCB. Still... I'm not saying I'm going to make this, but I'm not saying I wont :P Is this a thing that people(enough people) would want? Or is the existing, possibly saturated market for TiPi users already content with the current solution?

  • Like 2
  • Confused 1

Share this post


Link to post
Share on other sites

 

 

 

9 minutes ago, retro_doog said:

I feel like I could probably make a single board that combines the 32K sidecar, TiPi with whatever R-Pi zero "hat" circuitry is needed, and just has the Pi-Zero plug into it. So a single board that you just pop the Pi-Zero module onto and put it in a nice case.

 

You didn't watch the right video.  It's already been done by @J-Data

 

  • Like 1

Share this post


Link to post
Share on other sites
10 minutes ago, Omega-TI said:

 

 

 

 

You didn't watch the right video.  It's already been done by @J-Data

 

Close, but I'd use surface mount flash and SRAM. No need for those huge DIP packages. Also I'd bring more of the PI-Zero down to the carrier board and put an Enet connector and possibly other useful stuff on the main carrier board(HDMI? Haven't checked to see what all the Pi-Zero makes available on the header connector). Also, more pins will make for a more stable Pi-Zero mounting. I'd probably also have standoffs to the Zero so you can "bolt it down". Alternatively, It could incorporate a DIMM connector for an R-PI Compute module as people seem to complain about the Zero's slow-ness.

 

All Moot points as it appears there is no real market left and its unlikely people will want to keep buying the same thing over. And I need to stick to the small subset of things I actually want for my own machines so I can move on to that "next big thing" I alluded to above ;)

Share this post


Link to post
Share on other sites

I skimmed through this thread, but am unsure if you intend to use frequencies above 60Hz, or a flyback back based design. They can be really tough on the P/S's filter caps.

 

I'm not a big fan of voltage boosting designs either. To be honest... the less electronics in the P/S, the better. Although, the Vectrex sys. had a decent GFCI IC protecting things somewhat. Curiously, I just looked through a Vectrex service manual, and could not find it.

 

If you could get a small ferroresonant transformer design to work well, I might be impressed with that!:cool:

 

:)

Share this post


Link to post
Share on other sites
28 minutes ago, HOME AUTOMATION said:

I skimmed through this thread, but am unsure if you intend to use frequencies above 60Hz, or a flyback back based design. They can be really tough on the P/S's filter caps.

 

I'm not a big fan of voltage boosting designs either. To be honest... the less electronics in the P/S, the better. Although, the Vectrex sys. had a decent GFCI IC protecting things somewhat. Curiously, I just looked through a Vectrex service manual, and could not find it.

 

If you could get a small ferroresonant transformer design to work well, I might be impressed with that!:cool:

 

:)

I'll be using modern switching controllers complete with internal protection circuits. These things work with very small inductors as they operate at very high frequencies. The -5V inverting switcher is something like 2Mhz. I believe the boost regulator is 500KHz. Trust me, electronics in the power supply is a better solution than wirewound/heavy core transformers. We're only talking about a few hundred milliamps for the boost regulator. Most of these modern switchers are stable with ceramic filter caps as well due to the high switching frequency. You'll be surprised at how compact these modern supplies will be. The main variable for me will be how much Bulk capacitance to put on the board. I'll probably have pads for an overkill number of caps and only stuff a few of them in the final shipping supply.

 

Just for reference, here's a 500mA 5V switching supply from one of my designs. The footprint of this circuit is roughly 0.75" x 0.5" (as referenced by the 0.1" headers the circuit is tucked between).That Common Cathode diode in the upper right corner isn't even officially part of the switcher. It's just a "diode OR" of two possible input connectors.

 

 

Screen Shot 2020-07-13 at 12.29.04 AM.png

Edited by retro_doog
Clarity Needed
  • Thanks 1

Share this post


Link to post
Share on other sites

I used to have one of these units... 

 

Sola-63-13-150-Power-Conditioner-Input-95-130-61A.jpg.c7f19e003fc88880b93a53f624a588fa.jpg

 

...used in tandem with an adjustable self energizing contactor, on it's output. Overcurrent range, shutdown and restore times, could be tightly controlled!

 

I had the smaller 8-10 Amp unit. It was difficult to burn things out, even while doing destructive tests! Since losing it, some 25 years ago, I've lost a small room full of equipment to overloads and lightning! The smaller units provide better protection for light loads. I've always thought that a tiny one, would be so cool!!!:grin:

Edited by HOME AUTOMATION

Share this post


Link to post
Share on other sites
15 hours ago, retro_doog said:

Indeed it appears it does! That was hiding near the middle of the schematic page where the scanner misses the binding edges of the pages in the middle... I don't own a modulator to test, but I suspect it does not draw a significant additional load, and the attach rate is going to be pretty low on these. I'm guessing the "purist" crowd that still attaches these via the RF modulator is going to prefer the original TI power brick as well...

It only matters for europeans, since they have to mod the modulator to get composite, so they need that +12

 

Edited by Badaboom

Share this post


Link to post
Share on other sites
13 hours ago, HOME AUTOMATION said:

I used to have one of these units... 

 

 

 

...used in tandem with an adjustable self energizing contactor, on it's output. Overcurrent range, shutdown and restore times, could be tightly controlled!

 

I had the smaller 8-10 Amp unit. It was difficult to burn things out, even while doing destructive tests! Since losing it, some 25 years ago, I've lost a small room full of equipment to overloads and lightning! The smaller units provide better protection for light loads. I've always thought that a tiny one, would be so cool!!!:grin:

I... Don't know quite what to do with that. ;) Looks like a Bench test supply, which are designed with a lot of extra features to ensure the supply isn't damaged when testing faults on the unit under test(ie destructive testing). I'm sure a smaller, solid state version with the same features exists. Of course that is an AC power supply. I'm designing a modern DC to DC supply board that fits inside console. There will still be an external off the shelf AC/DC supply used, but the choices now are much smaller for equivalent power as they use semiconductor switching instead of large wire-wound transformers.

Share this post


Link to post
Share on other sites
9 hours ago, Badaboom said:

It only matters for europeans, since they have to mod the modulator to get composite, so they need that +12

 

Ah, of course! I'm designing in a 300-400mA capable 12V boost supply, so that should leave plenty of additional capacity for an external regulator. I just don't have one on hand to actually measure the exact amount of additional current needed, but I suspect it is only 50-100mA...

Share this post


Link to post
Share on other sites
3 hours ago, retro_doog said:

I... Don't know quite what to do with that. ;) Looks like a Bench test supply, which are designed with a lot of extra features to ensure the supply isn't damaged when testing faults on the unit under test(ie destructive testing).

Sorry, if I seemed mysterious. I have an odd habit of ...continuing(unceremoniously) right at the sentence, where I left off. Sometimes, even decades later!:twisted:

 

These do have an uncharacteristic bench type design, they are more commonly installed in the "mains" line. Usually at some distance, as they tend to buzz sharply under light loading.

 

verysimilarinside.jpg.4ace68ac52904add5afb81f62b105fad.jpg

 

Inside "the box" is a simple ferroresonant-transformer with a capacitor. These have a fairly/seemingly high impedance output, isolation, and can TANK quite a bit of power.

 

ferroresonant-transformer.jpg.d7c85cf32808ea68d6ec16b366635d80.jpg

 

You seem intent on pursuing semiconductor solutions. So, maybe this wont be particularly compatible with your ideas. However, you will have to go from AC to DC at some point. So you might consider, working this in.

 

Your form factor sounds quite tight. That should be something to see!

 

   P.S. I'm available for testing!:grin:

Edited by HOME AUTOMATION

Share this post


Link to post
Share on other sites
3 hours ago, retro_doog said:

... which are designed with a lot of extra features to ensure the supply isn't damaged when testing... 

 

When it comes to "Good Engineering", Scotty offers good advice!

 

  • Haha 1

Share this post


Link to post
Share on other sites
49 minutes ago, HOME AUTOMATION said:

Sorry, if I seemed mysterious. I have an odd habit of ...continuing(unceremoniously) right at the sentence, where I left off. Sometimes, even decades later!:twisted:

 

These do have an uncharacteristic bench type design, they are more commonly installed in the "mains" line. Usually at some distance, as they tend to buzz sharply under light loading.

 

 

 

Inside "the box" is a simple ferroresonant-transformer with a capacitor. These have a fairly/seemingly high impedance output, isolation, and can TANK quite a bit of power.

 

ferroresonant-transformer.jpg.d7c85cf32808ea68d6ec16b366635d80.jpg

 

You seem intent on pursuing semiconductor solutions. So, maybe this wont be particularly compatible with your ideas. However, you will have to go from AC to DC at some point. So you might consider, working this in.

 

Your form factor sounds quite tight. That should be something to see!

 

   P.S. I'm available for testing!:grin:

I was be facetious in my post :P Actually, you'll want to read my first post for my opinion of anything large and brick-like for the first stage supply. The entire point of my project is to replace the large, heavy floor brick with something more like one of these:

 

https://www.digikey.com/products/en/power-supplies-external-internal-off-board/ac-dc-desktop-wall-adapters/130?FV=57|318132%2C57|318134%2C57|318137%2C57|318143%2C57|318171%2C57|318172%2C57|318173%2C57|318175%2C57|318178%2C57|318180%2C57|318182%2C57|318183%2C57|318185%2C57|318187%2C57|318191%2C57|318193%2C57|447538%2C57|447539%2C1371|379518%2C1989|0%2Cmu10W|2187|0%2Cmu30W|2187|1%2C-8|130%2C48|103542%2C48|116509%2C48|234830%2C48|235004%2C48|270755%2C48|305478%2C48|74515%2C48|87718%2C48|89177&quantity=0&ColumnSort=1000011&page=1&stock=1&k=ac+adapter&pageSize=25&pkeyword=ac+adapter

 

You can get up to 24W in quite a small wall mount package!

 

The internal supply I'm designing accommodates these compact external supplies in two ways:

1. The internal supplies are now all DC/DC not Low voltage AC to DC like the original.

2. The internal supplies are substantially more efficient, thereby further reducing the wattage requirement, and thus the size and cost, of the external supply.

 

Also, in the world of power supplies, the simplest design(Transformer+cap, for instance) are usually the least efficient. For example, I've read that the ferroresonant XFMR is actually very inefficient at low loads. Take your Bench Supply, for example. Does it not get quite warm/hot? If so, that is all wasted energy not being presented to the output. Also, the "buzzing" noise you hear from transformers is also wasted energy. Electricity being converted into mechanical energy + heat. Modern semiconductor controllers can adapt to the load ensuring high efficiency across a range of loads. Not terribly important for our TI consoles, but super important in modern machines. Did you know that the core typically 1.0V power supply for, say, an intel processor running at a few GHz will draw a maximum of 50 Amps or more? And yet, it has many power modes that draw dramatically different amounts of power, so the current draw is actually violently "trashing". Imagine the power supply that has to handle that! And yet, it can be implemented in a couple square inches of PCB space and run quite efficiently across the wide range of load currents.

 

Despite the older, higher voltage, technology in our TI Consoles, the power requirement is quite low. A bit less than 10W, I'd say. The semiconductor supplies I'm designing in will be small, cool, and efficient. When you get a look at the finished product, you may have a hard time believing it is not only powerful enough to run the machine, but will actually have about 150-200% of the potential capacity of the original, especially if you're used to looking at old hardware. 😎

 

It's a bit early for testing, but thanks for the offer :) I'll probably be recruiting a handful of beta testers to help out. They'll have to do a little work for me though, like measuring under different loads etc. I'll probably want people with specific external equipment to help me test like the TiPi, RF-Modulator, and so forth. I just have a console and a Speech Synth myself...

Share this post


Link to post
Share on other sites
32 minutes ago, retro_doog said:

The video link doesn't seem to work for me. I'm guessing this is the ST:TNG Scotty and LaForge scene?

 

Really?  It's only a 4.91mb imbedded MP4 not linked to another site.  Weird.

Share this post


Link to post
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

Loading...

  • Recently Browsing   0 members

    No registered users viewing this page.

×
×
  • Create New...