I'm seeking advice on bench top power supplies

[ChuckT]

Help me choose a bench top power supply. I might want something for 6502 work but I'm also mindful of other more modern microprocessors or microcontrollers. I don't want to spend a bundle but I think they are safer than building my own power supply and some microcontrollers are going down to 1.8 volts.

http://www.mpja.com/Bench-Supplies/products/2/

[Dajgoro]

I bought myself some cheap supply, it broke twice already, and it does not have a 5V output mode. As for the laboratory supplies, every singe one that i encountered in school or in the university's lab was broken. I've seen many people building their own supplies with lm385, and it works, and it is much cheaper also. But if you need precise voltages, i suppose you should buy one of toes...

[GARTHWILSON]

A page in the 6502 primer:
http://wilsonminesco.com/6502primer/BasicWkbench.html

[cjb]

GARTHWILSON wrote:

A page in the 6502 primer:
http://wilsonminesco.com/6502primer/BasicWkbench.html

> The LM7805 [..]
"Stopped reading there..."

I'll suggest getting the power-points-with-USB-recharger-sockets, a suitably chopped-up USB cable, and you'll have a regulated 5VDC supply that's good for up to typically 500mA to 1A. See http://i.imgur.com/nRafp.jpg

For the other voltages, I'm using a recent' car-lighter-socket to USB adaptor that uses a high-efficiency DC-to-DC converter IC inside (I bought another this morning for $5..) and replacing the "R1" resistor that controls the Vout on the DC-DC conversion IC inside with a suitable potentiometer. The STM MC30463 IC I've got here has datasheets available for knowing what value is needed.

[GARTHWILSON]

Quote:

> The LM7805 [..]
"Stopped reading there..."

Um, is that to say you have something against it? I've designed it into aircraft products that have been flying for nearly 20 years without a single problem--ever. I think the LM340T is supposed to be the updated version. The bench power supply I showed on the page uses LM317T's (which are adjustable), and We've used them in our products too, again with never a single problem. Global, the maker of that workbench power supply, told me personally (at an electronics industry trade show) that they don't have any trouble with them either, even with long-term dead shorts. We have had more-modern Maxim switching-regulator control ICs go out occasionally though. The old linear ones are definitely more dependable, and don't require the heroic efforts the switching regulators do to get the output really quiet if that's important to your application.

[cjb]

> Um, is that to say you have something against it?

In designed products, they're maybe acceptable, but the suggestion of using a LM7805 for ad-hoc power provision usually brings out their worst-- not being adequately heatsinked (the proverbial prototype supply that then gets used 24/7.. and then there's all that heat next to an electrolytic cap..) It's also the 21st Century, you don't need to waste watts of electricity for a milliwatt mcu..
Linear semis also have a significant piracy problem, in that it's good practice now to just test any of their devices you have for being on-spec.

[GARTHWILSON]

Quote:

not being adequately heatsinked

Well, they have internal short-circuit current limit, plus internal thermal overload protection, plus SOA (safe operating area) protection. Letting them get hot can shorten their life somewhat, but not end it quickly. I've seen them just shut themselves down, and in fact oscillate between on and off because turning off makes the die cool, which again turns it back on, and so on, with some hysteresis. As for how hot a semiconductor can work though, I have had transistors operating under an infra-red microscope to take the temperature of any spot 0.001" diameter, and actually seen them operate at over 350°C, about 200°C above the maximum at the end of the specified derating curve. I didn't try to see how long they would live at that temperature, but I suspect it would not be much over an hour. At the temperature the 7805 (or better, the LM340T-5) shuts itself down though, I don't think you'll outlive it.

Quote:

and then there's all that heat next to an electrolytic cap..

There's no need to have an electrolytic capacitor close to it. Tiny monolithic ceramic ones are good to have close to it though.

Quote:

It's also the 21st Century, you don't need to waste watts of electricity for a milliwatt mcu..

Switching regulators don't do well with light loads. They go into pulse-skipping mode and make a lot of electrical noise. Many of them cannot regulate at all at less than 10% of their maximum rated output current. But for a milliwatt MCU, let's say you only need 2mA for the entire computer (like one I mentioned making for a product in the 1980's), and you have a 9V battery to power it (which we also did). I used a linear LM2950 LDO for that one in a TO-92 package (the same kind you see 2N3904 transistors in). If the battery is at 7V, the power dissipation of the linear regulator is (7-5)V*(2+.05)mA=4.1mW. There's a place for switching regulators, but there's still a place for linears. I use both.

Quote:

Linear semis also have a significant piracy problem, in that it's good practice now to just test any of their devices you have for being on-spec.

There we buy from reputable distributors who get the product direct from the manuafacturer, so there's no chance someone sneaked in some fakes. If there's a chance of fakes, you might get it from Jameco (which probably sells mostly to hobbyists) but not Mouser or Digi-Key or any of the major industry suppliers. I buy from Jameco sometimes, and have put some Jameco links in, and have not had any trouble; but I would never buy our production parts from them.

[BigEd]

Getting 5v power from USB is so convenient (for hobby purposes) that I'd like to know if there are any downsides to it.
Cheers
Ed

[Dajgoro]

BigEd wrote:

Getting 5v power from USB is so convenient (for hobby purposes) that I'd like to know if there are any downsides to it.
Cheers
Ed

The 5V output of an usb charger/adapter is good, but when used from a pc usb port it is not 5V at all, if i am lucky i get 4.7V, with lots of interference.

[cjb]

Dajgoro wrote:

The 5V output of an usb charger/adapter is good, but when used from a pc usb port it is not 5V at all, if i am lucky i get 4.7V, with lots of interference.

Ha, exactly this. The 68HC11 board in my photo above stopped working, and it was only recently when I got the time to sit down and test out what the problem was that I discovered the ~0.3v difference in the usb bus power from my current system's motherboard, and the USB1 PCI card I had been using. (The extra 1.2v drop across a power polarity protection diode didn't help either..)

[BigEd]

Thanks - worth knowing. Things which run at 3V3 or less can evidently get away with deriving their power from a PC USB, but I suppose that means they still need a regulator. Also, you only get 100mA from a host port until you request more - not so for chargers.
http://www.maxim-ic.com/app-notes/index.mvp/id/3241
(actually, you might get more, but that's all you're supposed to take.)

[ChuckT]

GARTHWILSON wrote:

Quote:

> The LM7805 [..]
"Stopped reading there..."

Um, is that to say you have something against it? I've designed it into aircraft products that have been flying for nearly 20 years without a single problem--ever. I think the LM340T is supposed to be the updated version. The bench

I have nothing bad to say about your choice of product.

I think if I spend the money on one, that will be my final purchase for the year. I don't mind throwing away a chip or two that gets damaged because it is cheaper than buying a top of the line benchtop supply. I just wanted the next best one from my list and hopefully under $100. For aircraft safety, the best is critical. For me, expense is the difference between being able to afford a project and failure.

[BigDumbDinosaur]

I use old PC power supplies for situations that can get by on fixed voltages. The three most widely used voltages with computer hardware are available in copious amounts, and with minimal noise. In fact, I have an old ATX power supply running my POC unit, which is powered from one of the 5¼ inch floppy drive connectors. The current (!) output voltages (I just checked) are 5.04 volts and 12.03 volts, according to my Fluke DVM. That's less than a one percent error on both outputs. Close enough for me!

[BillO]

cjb wrote:

> -
Linear semis also have a significant piracy problem, in that it's good practice now to just test any of their devices you have for being on-spec.

Do you have any evidence of this? Not trying to be confrontational, but it seems to me to be counter productive for a company to copy items that sell for so little, even from sources like TI and such.

1) Setting up production of such items is not cheap. Payback at a few cents per item would be measured in terms of years.

2) The design and production of linear regulators are well understood. It is difficult to imagine a company that did go to the trouble of getting into production of these, intentionally screwing it up to save a few milli-cents.

3) I've not heard of this before now. Have I had my head in the sand?

Besides, the regulation tolerance of LM78XX regulators is about +/- 5%. If you want really close tolerances you'll need to be selective even from first line suppliers

[BillO]

As for a great bench supply, I have a Systron-Donner / Trygon TL8V-3-0V. It's huge and old, but the performance is beyond reproach. I got mine on eBay for about $90 four years ago, including shipping. It seems they are still in circulation: http://www.ebay.com/itm/SYSTRON-DONNER- ... 0340602873

This is but one of several currently for sale on eBay.

They are truly awesome. I doubt you could get a modern, made in china alternative that would perform as well for so little. or weigh as much.

If anyone does ever get one, I have the service manual and would be willing to scan it for you.

Mine has a General Dynamics asset tag on it. I wonder what it was used to help design?