No New Projects?

[buffo]

I knew I couldn't sleep unless this was up.

Sigh... Very well. Even though I should also be asleep right now (working nights), I hooked up the 'scope and ran two simple tests.



This image above shows a Flexmod P3 running at 12V, powering a pair of A140 diodes in series. The max drive current is set to 1 amp. Top trace is the input modulation signal (10KHz), bottom trace is the output of the Flexmod.



This image shows the same Flexmod P3 still running at 12V, but with the max drive current cranked up to 1.5 amp. Also this time the modulation signal is 20KHz (still on top), and the output is on the bottom.

Please forgive the mess; I had to set this up in a hurry as I really NEED to be sleeping right about now. This is not an ideal test to demonstrate all the artifacts I have seen. But it is enough to show that there is an issue here, and it can be re-created at will in just a few minutes, by just about anyone.

If you want to re-create this, simply build a test rig with a pair of 445 diodes in series and hammer away. Better yet, a quad layout (2 x 2 array in series-parallel) might be even easier, since you could push the current quite a bit higher and see the artifacts better while minimizing the chance of cooking a diode or two.

I admit that at low currents, these artifacts are not present at all. Up to 700 ma or so, the output looks very clean (even at realistic modulation frequencies of 20-30KHz, though it does get a little messy when you get close to 100KHz), so you'd never have any reason to think something was wrong. I suspect that most of the testing done by others was performed at relatively low currents, which is why this has never been identified before.

But the Flexmod P3 datasheet states that it can handle up to 4 amps of current and DC to 160 KHz modulation. I haven't tested anything above 100KHz so far, and I haven't gone much beyond 2.5 amps in my testing yet, but from what I have seen it seems pretty clear that at higher currents and higher modulation rates there is a problem.

Adam

[norty303]

As a short diversion for the thread, I've been posting a pictoral blog of my recent 2 new 3W RGB builds (more medium power RGB solid state builds....*yawn* 10 a penny....)

Here's the link to the album on my FB page if anyone is interested in looking

https://www.facebook.com/media/set/?...1&l=e331976b82

[planters]

Drlava,

I own 12 of your Flexmod P3's and despite multiple projects and some pretty hard outputs in the 3A range, I have not seen any issues with catastrophic diode or driver failure. I have not tested the modulated output integrity as both you and Adam have demonstrated. However, the problem with your company's repeated periods of non-communication and out of stock status led me to adopt the Badpip driver for my latest projects. This driver has also preformed well, and it was due to the limited current capacity that I became excited about this new driver. I hope it turns out to have even better diode protection (these IR diodes are not cheap) and higher frequency modulation accuracy. That would be nice.

Adam,

Current flow will be an issue with the green builds. At a minimum and in it's lowest power configuration this laser needs 12 A. and in its highest power form, 48A. I am assuming that these drivers will be able to run in parallel with some load sharing adaptation and if used as independent diode drives the 5V. is perfect.

[Robin]

I've been very satisfied with the performance of the Flexmod P3. I've destroyed one by being absolutely stupid, but I've done all kinds of other dumb stuff that should have resulted in a dead driver or diode but didn't due to the P3's functional fault detection.

[chad]

Considering that you are measuring a negative spike and there is presumably no negative rail perhaps it is a wiring Inductance issue and not a flaw in the driver. Or a probing error.
In other words user error. If it is a Inductance from the current shunt on the driver, by the looks of your new driver you are using probably the same dale resistor.
Yours will suffer the same fate.
And I really hope that isn't the final driver. For a 'professional' driver that layout is a disaster.

chad

[swamidog]

pb,

there's a new gb starting for them:


http://www.photonlexicon.com/forums/...hlight=flexmod

Drlava, will the P3 be available again anytime soon? your site has had them out of stock recently.

Thanks

[buffo]

Considering that you are measuring a negative spike and there is presumably no negative rail perhaps it is a wiring Inductance issue and not a flaw in the driver. Or a probing error.

There is a positive spike visible on the rising edge as well, which is far more pronounced at higher currents and higher frequencies. There is also some ringing which gets much worse as you push the current and frequency up. As I said, this was a quick and dirty test to substantiate what I posted earlier.

As for probing/inductance issues, if that were the case it would be evident on both traces. And before you ask, yes, I have swapped channels and even used different leads just to rule out this sort of interference. It is only present on the output of the P3. I should also point out that I'm not the first person to notice these artifacts.

by the looks of your new driver you are using probably the same dale resistor

Not even close. There are actually 3 different current-sensing resistors we are testing now, and the largest is less than half the size of the one on the P3. The smallest is an order of magnitude smaller.

While I admit that this is not readily apparent from looking at the board, if you had read my post above about the specs of this driver, you would understand immediately that using the same current sensing resistor found in the P3 would be impossible. .2 ohms times 10 amps squared (in the full-current configuration) would yield 20 watts across that thing. Even if we have to de-rate it to 8 amps, that's still nearly 13 watts. Far too much...

And I really hope that isn't the final driver. For a 'professional' driver that layout is a disaster.

It's not the final driver. The image above is a prototype for final testing. As such, concessions had to be made with the layout to make it easy to swap components. Also, many parts that will be surface-mount in the final design are through-hole on this one. Finally, there are some components that will not even be present in the final design. Only 3 of these proto-boards were made, and there are no plans to make any more.

Adam

[planters]

Depending on the timing of the availability of these new drivers, unless the need is very high pressure, why begin to participate in a new production round for the P3?

I hope this isn't a replay of the EMS 7000's vs the P. scanners. Here, promised capabilities with uncertain future availability and cost vs. a commitment to a group buy of a known product.

Chad,
You may be right, but why do you assume no negative rail? I can't see the test set up at all.

Buffo,

The current issue?

Bill,

Can they provide this driver as an OEM component?

[buffo]

Buffo,
The current issue?

I was told that your IR pump diode needed a maximum of 7.5 amps, and that your design would use at most 2 of them. This was back in December, and I re-confirmed it with Steve in January. That's what we designed around. 48 amps is a HUGE leap...

That being said, if there is enough demand, we could design a one-off that could handle more current. We already have a design for a 15-amp version that is as-yet untested, and we've also kicked around a model with a pass-bank to reach higher currents. But we wanted to get this one out the door first.

There are two main problems with really high current drivers:
1) Traces on the board can't handle that kind of current, so you need point-to-point wiring for the output, and
2) Current sensing becomes very difficult at that level of current.

In the short-to-medium run, I think the better option would be to parallel the drivers. But once we get this out the door, I'm willing to look into other options.

Adam

[White-Light]

... vs the P. scanners. Here, promised capabilities with uncertain future availability and cost

The reason why no exact speed has been given by Pangolin is because no driver exists on the planet currently that can drive the new scanners to their limits.

Also, how anyone doubt their ability to finally deliver the product is beyond me given the proven company record and existing manufacturing contacts / clout.

Its all going to be about patience. Rome wasn't built in a day, but with running water and aqua ducts it was revolutionary when it finally appeared.

Whose scanner will be better, who knows at this stage, its all speculation but it will become clear in the end to those who are really bothered. For the rest of us I imagine it will more about performance vs price than whose is best.

PS Nice cases Norty.