Power supply part II

[eslapion]

Just to be sure I wasn't dreaming all this I unpacked my VIC Cr today and tried the following. Note that for this the datasette was replaced by a link for the switch and an 8V torch bulb for the motor as my datasettes are both in storage.

Powered if from it's own PSU to make sure it was working.

Powered it with just +5V from an AT supply.

Powered it with +5V and +12V from an AT supply.

Everything worked as I expected. With the +12V DC on pin 6 the lamp lit when the datasette motor would have been running and nothing exploded. In fact it is still working now, some four hours later.

There's even a picture of my VIC running on a single +5V supply on my site on this page.

That is incorrect for all the reasons Eslapion points out.

Unfortunately for Eslapion his teenage error does not make what I can demonstrate to be true false.

Lee.

What you've demonstrated is that you can power a VIC Cr is a VERY awkward manner using only 5V and 12V.

If you made a power supply like that, the userport devices would no longer work and it would be impossible to power a 64 with the same power supply.

What you've tried is NOT a 5Vdc with 9Vac powered by the same transformer. The reason your setup works is because both the 12V and 5V you feed are rectified by the same bridge which gives them a common ground level. In this setup only one out of 4 diodes conduct in the CR2 bridge as the voltage drop on the negative side is near zero.

BTW, this is essentially how the 2 Prong VIC works internally... single bridge, single ground, no problem.

Power a VIC Cr with 5Vdc and 9Vac with a power supply that uses the same transformer to generate the 5Vdc and 9Vac and you WILL get an explosion inside your VIC. In a case like that, the 5Vdc is rectified by the bridge in the PSU and the higher voltage fed to the datasette comes from a different rectifier inside the VIC which imposes its own ground level.

Two different bridges each trying to impose their own ground level and there you have it... kaboom!

I don't know for other people but frankly, a power supply that only works with my VIC Cr and not the 64 is quite useless to me, especially if the userport become useless.

As for what I said concerning the TOD, I saw on the 64's schematic that the 9Vac is rectified and connected to the TOD line of the CIAs so I expected something similar to be present in the VIC. Apparently not.

Still, if you feel comfortable powering your VIC Cr with only 5Vdc and 12Vdc... I sell nice switching PSUs for 8$ each which provide only these 2 voltages. You could have a single power supply powering you VIC Cr, a 1541-II and/or a 1581 as these drives also use 5Vdc and 12Vdc. A triple power supply in one single box.

[Leeeeee]

What you've demonstrated is that you can power a VIC Cr is a VERY awkward manner using only 5V and 12V.

It's not at all awkward as it lets me use any PSU I have handy. Usually a PC PSU.

If you made a power supply like that, the userport devices would no longer work

That's not strictly true. Any userport device that rectifies the 9VAC to make a DC supply will work fine with the +12VDC on both 9VAC lines without changes. This includes things like the C= RS232 adapter.

and it would be impossible to power a 64 with the same power supply.

Not a problem, I don't have a C64.

What you've tried is NOT a 5Vdc with 9Vac powered by the same transformer.

But I have tried that and it does work. I used to use an arcade machine PSU which has just the one winding to provide +5V via an LM323K and the ~11VAC lighting supply.

Two different bridges each trying to impose their own ground level and there you have it... kaboom!

No you don't. It does not matter one bit whether you use one or two rectifiers, all that happens with two rectifiers is that the diodes with their cathodes towards the transformer end up in parallel. What does matter is that you don't try to use a 9V - 0V - 9V transformer as this will put about 24VDC across C39, a 16V electrolytic.

I don't know for other people but frankly, a power supply that only works with my Cr VIC and not the 64 is quite useless to me,

Which is fine. I don't have a C64 to worry about.

especially if the userport become useless.

Which it doesn't. Only devices that need to multiply the 9VAC to get more than 12VDC will be affected and they are few and far between. I certainly have none.

Still, if you feel comfortable powering your VIC Cr with only 5Vdc and 12Vdc... I sell nice switching PSUs for 8$ each which provide only these 2 voltages. You could have a single power supply powering you VIC Cr, a 1541-II and/or a 1581 as these drives also use 5Vdc and 12Vdc. A triple power supply in one single box.

Thank you but I have plenty of AT and ATX supplies that will do just fine.

Just to sum up.

You can run a VIC Cr on just a 5V DC supply. If you want to use a datassette you can replace the 9VAC with 12VDC on either or both AC supply pins. You can run a VIC Cr on 5VDC and 9VAC derived from the same winding on a transformer. I have tried all these configurations today and they all work as I expected.

Also, just in case you ever wondered, you can run a two prong VIC on 12VDC.

Lee.

[Leeeeee]

As impressed as I am about you running an 8 bit PC card with your Vic 20, I still think that FM radio is a waste of the Vic 20's clock cycles. Even back in the early 80's, FM radios were very common and cheap.

I wanted to see if it could be done and it can, it's not even a particularly good radio.

The sound, VGA, network and multi I/O cards are much less wastefull of the VIC's cycles though. Prehaps one day I'll get round to completing all the pages for them.

Lee.

[eslapion]

Still, if you feel comfortable powering your VIC Cr with only 5Vdc and 12Vdc... I sell nice switching PSUs for 8$ each which provide only these 2 voltages. You could have a single power supply powering you VIC Cr, a 1541-II and/or a 1581 as these drives also use 5Vdc and 12Vdc. A triple power supply in one single box.

Thank you but I have plenty of AT and ATX supplies that will do just fine.

I hope you have considered the minimum load requirements of your power supplies.

Usually, the load of a VIC is not enough and you get higher voltages than expected.

Just to sum up.

You can run a VIC Cr on just a 5V DC supply. If you want to use a datassette you can replace the 9VAC with 12VDC on either or both AC supply pins. You can run a VIC Cr on 5VDC and 9VAC derived from the same winding on a transformer. I have tried all these configurations today and they all work as I expected.

I stand corrected...

What does matter is that you don't try to use a 9V - 0V - 9V transformer as this will put about 24VDC across C39, a 16V electrolytic.

Precisely the capacitor that blew up in my old VIC... I guess that's what I might have done.

Also, just in case you ever wondered, you can run a two prong VIC on 12VDC.

That I knew.

I suspect you might run into problems if you power your 2 prong VIC with a 12Vdc source from the same PSU as the one feeding a 1541-II or 1581 attached to it.

The diode bridge inside the 2 prong VIC will cause a 0.7V raise in its internal ground level relative to the PSU and drive. Because of that, the ground line will try to return throught the IEC serial connector instead of the power connector.

[PaulQ]

As impressed as I am about you running an 8 bit PC card with your Vic 20, I still think that FM radio is a waste of the Vic 20's clock cycles. Even back in the early 80's, FM radios were very common and cheap.

I wanted to see if it could be done and it can, it's not even a particularly good radio.

The sound, VGA, network and multi I/O cards are much less wastefull of the VIC's cycles though. Prehaps one day I'll get round to completing all the pages for them.

Lee.

Yes, those would be worthy of Vic 20 clock cycles.

[Leeeeee]

I hope you have considered the minimum load requirements of your power supplies.

Usually, the load of a VIC is not enough and you get higher voltages than expected.

Another 'documented and known' fact from the internet that may have had some truth in the past when the cheap PC clones first proliferated but that hasn't been true for years.

Take a look at this schematic. This is pretty typical of the ATX PSUs on the market.

Look at R50, R52, R15 and R59. Guess what they're for. Every clone AT and ATX PSU that I've ever disassembled has these resistors or their analogue, the 'minimum load' feature was resolved a long time ago.

And, just to make sure I'm not dreaming again, I've pulled one random ATX supply from the pile, and my bench AT supply, which wasn't specially selected either, and have an old MFM hard drive and my VIC to use as a loads.

The voltage was measured on an unloaded disk drive lead in both cases to avoid also measuring the voltage drop on the loaded lead.

• AT +5V - with no load 5.126V, with hard drive 5.104V, with VIC 5.097V
  ATX +5V - with no load 5.241V, with hard drive 5.354V, with VIC 5.116V

Looks like the VIC is a bigger load on the 5V than the drive is and that the ATX PSU regulates the +12V as well as the +5V rail which causes the +5V to rise with the load on the +12V.

Hmmm, I just tried the ATX PSU with my 19" LCD, ~5A @ 12V, and the +5V rail goes up to 5.431V with just that loading the +12V, still that's less than 200mV change in output for a 60W change in load.

Now I don't recommend using any old AT or ATX PSU without checking that it works properly first, but I wouldn't be concerned over providing any minimum load.

Lee.

[Leeeeee]

Yes, those would be worthy of Vic 20 clock cycles.

Well there is this and this.

I also have some code that wedges in to replace the tape device with the internet for loading so you could do LOAD "http://www.somehost.com/programs/demo.prg",1 or similar but that has some unresolved memory conflicts with the screen print routines and the ARP code doesn't work reliably yet.

Lee.

[eslapion]

...

• ATX +5V - with no load 5.241V, with hard drive 5.354V, with VIC 5.116V

Looks like the VIC is a bigger load on the 5V than the drive is and that the ATX PSU regulates the +12V as well as the +5V rail which causes the +5V to rise with the load on the +12V.

Hmmm, I just tried the ATX PSU with my 19" LCD, ~5A @ 12V, and the +5V rail goes up to 5.431V with just that loading the +12V, still that's less than 200mV change in output for a 60W change in load.

That's not as bad as the 8V Ral-Clan got last year with his 1541-II/1581 power supply but its more than 5% over. Probably not enough to cause any damage but close to the official max rated voltage of 5.5V of the CMOS SRAM chips.

To the people who do use PC PSUs to power their 64/VIC, I always recommend having both the 64 or VIC and a 1541-II or 1581 attached to it and turned on to prevent unwanted voltage rises. Maybe its overcautious but it feels safer.

Now I don't recommend using any old AT or ATX PSU without checking that it works properly first, but I wouldn't be concerned over providing any minimum load.

I know there are some AT or ATX PSUs which will not even start if there isn't enough load on the +12V output. I guess that's a problem that's relatively easy to fix with a small 12V lamp or resistor.

I would tend to say if you're going to go with an old PC PSU, go for an AT one if you can considering how much easier it is to turn on/off.

I believe this solution is really worth considering if you have access to a free PC PSU as it usually is much cheaper to get a new 5/12 V PSU rated for 4-5 A than it is to buy a new AT/ATX PSU.

One of the reasons I am usually opposed to using PC PSUs to power old 8 bit Commodore computers is because some people assume that because it can be okay to power your VIC/64 with them, they consider these to be "repairable" as the fuse is easy to replace internally.

I have seen people make internal modifications to these PSUs and one particular case where a person actually took it out of its steel case and exposed the high voltage rectifier/doubler. I think that's a very unsafe thing to do.

[eslapion]

While at the moment I'm not likely to build this kind of power supply, it would be interesting to hear if my friend was thinking in the right direction. The hardest part, to convert a high voltage to 12V would already be solved.

To sum it up, yes, your friend is thinking in the right direction.

However, as it is, apparently, this PSU is okay for a VIC Cr but not for a C64.

Alterations required:
- make C1 a 2200uF capacitor instead of 0.33uF
- Compatibility with a C64 requires that the 12Vac used for the 5Vdc output is either from a different transformer or from a totally independent winding of the same transformer. Do NOT use a 12-0-12 volts or 9-0-9 volts transformer. (Thanks for pointing this one out goes to Leeeeee)
- Optionally, to minimize heat and improve efficiency considerably, replace the 7805 voltage regulator with a PT78ST105 switching regulator.

[carlsson]

I see. Perhaps a transformer which delivers 2x12VAC 1500 mA (total load) would be a better choice than one which does 1x12VAC 2000 mA and split that voltage?

As a matter of fact, there are cheap transformers which do 2x9VAC 500 mA too but I figure that won't be enough power if the VIC/64 likes to get up to 1A on the 9V line. At least that's what the spec is on the original power supplies.

But as I wrote, it is mostly of academic interest. My biggest motivation was whether one could build your own heavy duty PSU strong enough to power at least 1-2 computers and one disk drive. I know about the one Protovision (?) sells, which basically is a PC AT(X) PSU with an additional 9VAC transformer on top.

Your link to Lemon seems interesting but I haven't evalutated that other design.

[eslapion]

I see. Perhaps a transformer which delivers 2x12VAC 1500 mA (total load) would be a better choice than one which does 1x12VAC 2000 mA and split that voltage?

I assume here when you say 1500mA total load that you want to say 2 12Vac outputs with each having a maximum load of 750mA. That's not enough. The C64 with nothing attached draws about 1.1A from the 5Vdc and 800mA from the 9Vac.

Also, a single 12Vac transformer of 2000mA (2A), assuming the C64 would not require two independent sources, would be very much borderline for the same reason.

The original Commodore brick is rated for 1.7A on the 5Vdc and 1A on the 9Vac source. That's a total of 2.7A if you calculate only the current load. Obviously, a "heavy duty" power supply would have to provide more than that, not less.

As a matter of fact, there are cheap transformers which do 2x9VAC 500 mA too but I figure that won't be enough power if the VIC/64 likes to get up to 1A on the 9V line. At least that's what the spec is on the original power supplies.

Nope, its not enough.

But as I wrote, it is mostly of academic interest. My biggest motivation was whether one could build your own heavy duty PSU strong enough to power at least 1-2 computers and one disk drive. I know about the one Protovision (?) sells, which basically is a PC AT(X) PSU with an additional 9VAC transformer on top.

The Protovision desing seems good to me. It uses a transformer to provide a completely floating 9Vac source and the ATX PSU is unopened as the transformer is added externally.

Quite frankly, however, I think it is very expensive and grossly overpowered for the job. Most ATX PSUs are rated for 20A or thereabout when the 3.3V source is unused. That's enough to feed about 15 C64s or VICs...

Your link to Lemon seems interesting but I haven't evalutated that other design.

Its essentially the same as yours but using two different (bigger)transformers and a switching regulator.

It uses very inexpensive parts and it is about the same rating as a C128 power supply.