Helios - Open source, low cost DAC

[Grix]

Hi,
I'd like to share my newest project: Helios Laser DAC. It's an USB to ILDA DAC, with fully open source hardware and software. It's low cost but still powerful with good specs: 12bit XY channels, 8bit RGBI channels, up to 65+ kpps, up to 4000+ max points in frames, plug and play automatic driver installation, aluminium enclosure etc.

http://pages.bitlasers.com/helios/

I've prepared a batch and will be selling them for $79.90 each. I think that makes it the cheapest full color ILDA DAC available right now. Even cheaper than a sound card and correction amp! I'm hardly making money on these really
Buy here: http://bitlasers.com/opencart/index.php?route=product/product&path=68&product_id=87

It currently supports LasershowGen natively of course, but the drivers are built to be compatible with cfavreau's Open Laser Show Controller API, so by renaming and replacing a DLL file, it is already supported by several third party programs, including Dr Lava's LaserShow Xpress (actually I have only tested it with ILD SOS because I don't own LSX but I believe they use the same device interfacing system, right?).
I will be contacting various other developers soon too, in hope that it will be supported by even more software. If you're a developer reading this, post in this thread or PM me if you are interested, or just go straight to the github repo: https://github.com/Grix/helios_dac

The device uses the open source libusb library for USB connectivity, which supports many platforms including Mac and Linux. I haven't built binaries for these environments yet but I will soon, if no one else does first.

 

[Benm]

That looks pretty good!

I guess it is as good as going the soundcard-dac + compensation amp route, but all in a very compact package and at a decent price point as well.

Does it do the full differential -10/+10 volt ranges for the galvos? And is there any way to set gain in software? It looks like there is a nice dc/dc converter on there to derive these voltages from usb though

 

[Grix]

That looks pretty good!

I guess it is as good as going the soundcard-dac + compensation amp route, but all in a very compact package and at a decent price point as well.

Does it do the full differential -10/+10 volt ranges for the galvos? And is there any way to set gain in software? It looks like there is a nice dc/dc converter on there to derive these voltages from usb though

Yes it has the full signal range. No way to set gain in firmware but most programs can do this anyway on the computer side so there's really no need, also since it's open source you could make a custom firmware with this functionality yourself if needed anyway.

 

[Benm]

I guess you're right about that, just adjusting things in the control software seems adequate as long as it can output the full ilda voltage range.

Kudos for the nice design anyway

Are you doing all the DAC work with the atmel chip and are the two chips to the right opamps (i can't really read their part numers of the photos)?

 

[Grix]

Are you doing all the DAC work with the atmel chip and are the two chips to the right opamps (i can't really read their part numers of the photos)?

There are two 12bit DAC channels on the atmel chip for XY, then there's an external DAC for the 8-bit color channels, the little IC to the right of the MCU. Yes, all the three rightmost ICs are opamps, one handles XY signals, one handles color signals and one handles misc. signals like the shutter, and biasing for the XY channel amp.
Full part numbers etc can be found in the BOM file, in the github repo somewhere

 

[Benm]

Very good stuff

Perhaps in a future version some kind of fault detection could be built in that switches off the lasers in some failure modes, for example the processor locking up with the lasers on for some reason (programming error, voltage spike/dip, cosmic radiation or whatever).

Not that this would put me off from buying one of these though, i think it's a great deal.

One thing is was wondering about though: Is there room to make 3mm-ish holes on the corners so you could bolt the board down (on pillars) in a different enclosure (say one that contains the entire projector system) so you don't need an ILDA cable and separate box but get a unit that plugs straight into USB?

 

[Grix]

Very good stuff

Perhaps in a future version some kind of fault detection could be built in that switches off the lasers in some failure modes, for example the processor locking up with the lasers on for some reason (programming error, voltage spike/dip, cosmic radiation or whatever).

Not that this would put me off from buying one of these though, i think it's a great deal.

One thing is was wondering about though: Is there room to make 3mm-ish holes on the corners so you could bolt the board down (on pillars) in a different enclosure (say one that contains the entire projector system) so you don't need an ILDA cable and separate box but get a unit that plugs straight into USB?

A failsafe is a good idea yes. I will see if I can make something for a future firmware version or I would have to make hardware changes..

As for holes. I had originally planned to add mounting holes but it got de-prioritized when I found the solution with the enclosure. But I think there should be room to drill some holes manually in strategic positions. You can see here on the copper layout to see where there is room: http://i.imgur.com/pqFPd6e.png

By the way in other news, as posted on Photonlexicon, Spaghett*, Maxwell and possibly LSX will support the Helios natively soon!

 

[Benm]

Looks like it's possible indeed, the outer edge of the pcb is all groundplane right?

I guess failsafe features would require hardware changes. For the best protection you'd need some feedback from the galvo's to assert the mirrors are still moving, but for a simpler version you could just assert that the processor in the helios is working properly.

The latter can be quite simple and rely on a pin changing state regularly, generating a AC voltage that is monitored by some simple analog circuitry. If that signal stops the laser DAC's should all turn to zero output, or you could simply cut the power supply resulting in the same effect.

 

[Grix]

The DAC can now be bought at Innolasers as well: https://innolasers.com/shop/index.php?id_product=23&controller=product

Also, it might be available for purchase as a bundle deal with LSX in the future.

 

[Benm]

Is there any difference between the original $79 and this $99 product?

 

[Grix]

Is there any difference between the original $79 and this $99 product?

There's not, no, except their store is US based so Americans would get the product faster, and probably save shipping costs to even it out a bit, etc. I originally priced the product with direct distribution in mind so unfortunately I couldn't justify absorbing the extra price margin. It's still a good deal I think

 

[Benm]

Sure thing, i was just curious.

For about $100 including shipping i think it'll be a great product for many, especially those that want to have the ILDA connector plug and play.

I hope you don't get plagued by chinese clones though, as they often figure out how to make similar devices down to a cost even if they are less reliable. I reckon this is specific enough to avoid the attention of those types.

 

[OldNo7]

Looks like it's possible indeed, the outer edge of the pcb is all groundplane right?

I guess failsafe features would require hardware changes. For the best protection you'd need some feedback from the galvo's to assert the mirrors are still moving, but for a simpler version you could just assert that the processor in the helios is working properly.

The latter can be quite simple and rely on a pin changing state regularly, generating a AC voltage that is monitored by some simple analog circuitry. If that signal stops the laser DAC's should all turn to zero output, or you could simply cut the power supply resulting in the same effect.

Like a watchdog timer. You could add a watchdog timer chip and tied it's reset to both the micro and the DAC. However that being a hardware change maybe it would be best suited for version 2.

The atmel does have an internal watchdog that can be used to reset the micro but I don't know if that drives the reset pin low. In that case you could end up with output on the DAC still on and not moving the galvos while the micro is reseting.

 

[Grix]

Like a watchdog timer. You could add a watchdog timer chip and tied it's reset to both the micro and the DAC. However that being a hardware change maybe it would be best suited for version 2.

The atmel does have an internal watchdog that can be used to reset the micro but I don't know if that drives the reset pin low. In that case you could end up with output on the DAC still on and not moving the galvos while the micro is reseting.

Yeah in the newest firmware v3 there is already a watchdog that resets the uc and thus shuts off the output. But it's just set to the default 16 seconds right now, so it won't do much good if it ever needs to be urgently shut off. Decreasing the timer value shouldn't be hard though, and either way I believe most projectors already have some safety measures that shuts off the output if no change is detected on the ilda port signal.

 

[AngelG]

I'd like to see a video of this in action.
Have you used it or the sole purpose is to sell it ?

 

[Benm]

You can probably set the watchdog timer to a much shorter interval, info on the registers for that, well, read the manual

I'm not sure using the same chip to do all the work and check if everyting is okay is the best of ideas though. If something (like a big nasty voltage spike on the supply rail) caused the main uC to misbehave, isn't it likely that this could affect the watchdog timer too?

Then again scan-fail detection can go pretty far. You can look at galvo feedback signals to assert the mirror is actually moving, and detect if it was not moving with correct drive signals intact. Such a failure could be caused by someone holding the mirror in place by hand, but also mechanical lockup due to wear and what not.

Being uber-safe this should probably also trigger some kind of mechanical shutter blocking the output since you cannot rely on the laser control circuitry to actually shut down when commanded too.

The latter two things go beyond the scope of a DAC though, there are standalone modules sold for those purposes. I think a safety measure that shuts down the lasers in a condition where they are powered on but the mirrors are not moving would be enough.

This should also have some override (jumper?) to allow adjusting the lasers to line up with the mirrors in line with the beam without any audience present.