Ideal sound card for multichannel scan and modulations.

[James Lehman]

I still like the +/- 12 or 15 VDC chip. That's just cool! It gives you a lot more options, I think.

James.

[mixedgas]

[QUOTE=James Lehman;51271]I still like the +/- 12 or 15 VDC chip. That's just cool! It gives you a lot more options, I think.

ICL7660 or its more modern clones by maxim or TI. Two of them would give you +/- ~ 10V from your 5V.

Its easy to generate the opamp supply from the +5 without a switcher or expensive module.
6 2n2222s and a few caps and diodes. You build two simple little multivibrators driving cockroft-walton style doublers. The 7660 just puts it in all one package. You probably could get by at +/- 8V or so, but 12V would be ideal.

As for going rail to rail @ +/- 5, dont. You loose slew rate near the edges, right where you need it.

The other thing to keep in mind is the length of cable your driving has significant capacitance if your out in the field. figure 5 to 20 pF a foot depending on what your customer uses. even in just a SELEM situation we had a few 40-50 foot runs. So a micropower rail to rail chip may not be able to slew that. Some chinese scanner amps have 10K input impedance, some have 600 ohms. Keep that in mind too. Also design it to withstand a shorted output, or do what Bob Ash (now retired) did, use a socketed "disposable" opamp as the output buffer and keep its feedback and input resistors around 10-100K. Your gonna be selling to a lot of newbies, and back terminating the two lines with a series 600 ohms and driving them half a volt or so higher to compensate might be a "good thing". "Good Thing" is unfortuanely a trademark of a Felon known as Martha and is NOt used with permission.

you have about 90 ma @ 5V to play with on a good USB powered hub, right around the permissable 100 many usb chips shut down , some I've found clip at 90 during soft start and I have a laptop that clips at a very nonstandard 40 or so ma. That comes from designing a usb joystick with HID enumeration and tesing it across a range of systems., not just goofing around.

Steve Roberts

[tocket]

you have about 90 ma @ 5V to play with on a good USB powered hub, right around the permissable 100 many usb chips shut down , some I've found clip at 90 during soft start and I have a laptop that clips at a very nonstandard 40 or so ma. That comes from designing a usb joystick with HID enumeration and tesing it across a range of systems., not just goofing around.

Huh, I thought the USB standard specifies 500mA @5V? There's no way you could run a coffee mug warmer on just 90mA.

[drlava]

Even though slew rate is lost at the edges, the slew rates required for scanning are slow compared to the rates available on many r-r amps. I haven't tested it but for 10k input impedances it probably wouldn't be a problem.

However, given the 600 ohm termination requirement, I agree we need more supply voltage. This probably won't be doable on 90mA alone, to be on the safe side an external power source should be available in needed on weak laptops. The CM106 based card (for example) requests 500mA mode from the hub, max draw is 350mA according to the spec sheet.

Also, slow charge pump converters like the ICL7660 have very noisy output under load unless you use a huge filter.

[heroic]

Huh, I thought the USB standard specifies 500mA @5V? There's no way you could run a coffee mug warmer on just 90mA.

USB specifies that a device is allowed 100 mA until it has negotiated more, and then up to 500 mA if the hub port has that much power available.

[drlava]

After some searching I have found some parts that should meet our requirements and stay within the 500mA typical USB power allotment.
Switching DC/DC converter operating at 85% efficiency which produces +/- 6V Rails up to 120mA each.
Quad op-amp 7.5V/us Can drive capacitive loads 2nf stably, 50mV from rails drive at +/-10mA with a 1.3us settling time, when used with a 85 ohm series protection drive resistor, can provide +/-5.2V drive into 600 ohm load. Will probably need a small compensation cap for driving long lines.

[James Lehman]

Can we speculate as to the arragement of things connected together?

I mean, if the idea is to use USB, are we going to limit the cable length to 2 meters?

Are we necessarily going to have long leads from the USB device to the laser projector? Or is the USB device going to be part of the laser projector?

What I'm getting at is quite simple. There's plenty of current in the PSU for the scanners. That's how both drlava and barold are powering their op-amp correction circuits now.

I take that back: barold is using a PCI card and he got his +/- 12VDC just like I did, from the card.

James.

[drlava]

That's true, there are perfect supplies on the scanner boards. However the 'standard' is to have the DAC outside. It comes down to cable runs. Early on, all the dacs were probably in the computers, so it wasn't even a question. With a USB DAC, the max-pushing-it USB cable length is 5 meters, where the cable length for the lower fequency DAC output signals can be much longer. Even for home use, if one were inclined to mount their projector permanently, it's likely it would be further than 5 meters from the computer.

theoretically we could tap the 'user defined signals' on the ilda connector http://www.laserfx.com/Backstage.Las.../ISP-DB25.html to pass +/- 15V to the DAC, but relying on that is sure to cause potentially damaging issues in terms of flexibility.

[carmangary]

Maybe you could do it sort of like the way I suggested the optional balanced drivers and have an optional power supply section so that if you want to drive it from an external power supply you could just leave the power supply components off and hook into to it external. Or, you could add the components and get power from the USB power. That gives you the best of both worlds so you can either use it externally or internally but without wasting money on components. It just takes a little clever board design.

[James Lehman]

I don't like the idea of putting voltages into the line.

It's a shame there isn't some kind of [external] power take off standard for computer power supplies. There are so many peripherals that could run on just +12VDC, like scanners, printers, speakers, cameras, etc. That would sure clean up the power strip and save some copper.

Do you think it's commercially feasible to expect someone to find +/- 12VDC & GND on a computer power supply? Probably not!

You could make a blank PCI card with a power socket on the spine, or like I said before a fake PCMCIA or PCexpress card.

James.