Seeking Your Feedback on a Beat Our Board Design Challenge - Grand Prize $5,000

It is very cool but too hard, I think. Maybe it is hard just to me, but I think it is very hard in comparison with standard design challenge, when you can take just a Arduino and do something. I consider it very hard also because DER-953Q design is realy nice. Power Integrations spend lot of time with designing it and datasheet shows all evaluation results. I think there is very low chance that someone make something better.

I do not understand the point "Cannot use InnoSwitch". At begining you wrote that we can win by improving their design, but later it seems that we should create (new) design using competittor products practicaly showing that it is impossible to make better design. That is interesting.

You can also consider sending member result (final board) to Power Integration for propper (and equal/fair) evaluation under same terms (high voltage tests, tests at thermal limits, ...). In other words sponsoring shipping from member to Element14 (or Power Integration), basicaly the reverse way you send packages to us.

I think 10 sponsored kits are too much. Because of complexity I gues that no more than 5 kits are necessary. Instead you can consider sponsoring PCB manufacturing.

Sorry, the workscope is not in my talent quiver and too much like real work. 

Here's another idea, just to brainstorm a bit;

It could be attractive to customers to see the actual use of the PI product, for instance, what it is like designing for that unusual IC package, and what real-world transformer (or custom one) works, and what is practical in terms of PCB size and PCB technologies, and what measurements were taken. This would (I hope) be interesting to the manufacturer too; they could learn a lot since nobody will know their product better than customers who get their hands dirty with it.

Trying to compete with other manufacturer parts is for sure a valid challenge, but it's a type of challenge that might help in initial device selection but won't help customers in using the PI product.

Since it's a lot of work, and 10 contestants who will do it justice with the custom work (PCBs, etc that will be required) will be hard to find, perhaps a method without a contest could be better, that way only one person (or say two persons) need to be found, and let them have lots of latitude in what they design, and providing they do produce the material, then for them to be awarded the full amount (it would easily consume a couple of months of work, so instead of June 2024, material could be available sooner this way, and if customers benefit from it, then PI could run a second campaign of some sort - so it's quite efficient. It would cost the same - might as well let both people (if there are two) be awarded the identical full amount if they produce the content; let them not feel they are competing against each other so that they can share material as they work.

If it were me (I am not saying I would apply, but I might be tempted if I could concentrate on the things that I feel would interest me and customers, rather than try to meet a specification that doesn't meet my needs) I'd be interested in making a reliable power supply (perhaps several of the PI devices, OR'ing them together) with the benefit of higher output current at lower input voltages.

It might not be what all their customers might want (but I bet manufacturers only know what a subset of customers do with their products; and if PI really wanted precise alignment then they might benefit from looking for a case study with the customers that they might already be working with). So, in this case, perhaps it is not a negative thing that it will be hard to find 100% alignment; if the application is still likely to reveal relevant stuff - for instance, I would say that's quite realistic from a heat generation point of view, having more heat sources on a board than just a single power supply.

Others might be interested in (say) using the chip for a different output voltage, perhaps, and maybe that might reveal valuable information to customers, about how versatile that chip is (or not). 

As others already stated,I also think that could be better decreasing the overall kits number while increasing design/development support for challengers (for example increased initial budget or sponsored PCBs).
This design challenge is very cool but also quite hard. I like challenges (and this one is definitely outside my confort zone) but this one I think it will be too hard for me to carry out in a proper way. Also, the test equipment needed to evaluate such design is no joke. Time availability could also play a big role here (and for me it will since I've already scheduled other important things for next year).

I voted for:
- "After reading the information provided above, are you interested in joining this challenge?" : Maybe. I'm interested? Yes, I like this challenge. Will I apply? Probably not. (but good luck to all future challengers!)
-"Are you working as an Automotive Power Design Engineer?" No. I have little experience with some switching converters and motor power stage but I would not call myself a power electronic engineer
-"On your design workbench, do you have access to at least 40VDC": Yes, my power supply can go up to about 60V if I remember correctly. Still, I don't think that's enough to evaluate this specific design request. A lot of other equipment will be needed too.

There is a trick or typo in the target spec - PI's board accepts 150 - 500V DC but the challenge is for 40-500V DC.

But on to the substance.

If I were to accept this task as a commercial project (which I probably wouldn't as it stands because it's definitely and Automotive spec project and I haven't worked in that domain for  along time).

But if I were to take it on I would be looking at charging of the order of £25k for the work - might be double - that was a 5 minute quote with an asymmetric 200% tolerance !

So it looks a bit like an engineering equivalent of those dance marathons that desperate people entered in the 1930s because they needed the prize money.

en.wikipedia.org/.../Dance_marathon

Misaz summed it up in 7 words:

misaz said:

It is very cool but too hard

BTW, you need some pretty good and fancy test gear to develop this - like the 900V power supply.

I think the basic idea of a challenge is OK - but this one is much too hard for a fun event.

A much better challenge might be to do something useful with  a handful of PI's Innoswitch chips.

MK

I voted NO. The last poll should ask "do you have access to at least 500VDC"

This is not a design challenge, this is project outsourcing. I agree with michaelkellett 's quote estimate. 

Just a 6 layer PCB with 2 oz outside copper from one of the far east proto vendors is US$ 167.

I do not think it is project outsorcing, but showing that it is impossible to beat board is very crazy marketing campaign, so in my opinion it is more like converting marketing effort to enginnering effort which is interesting.

I think this approach may work with different kind of board where the several upgrades are possible. For example imagine upgrading Arduino competition. You have plenty of options like adding onboard debugger, sensor, some different connector, power consumption monitor, changing chips, .... But in case of purely analog and power evaluation boards it is quite hard because there are very limited options how to make it better.

Additionaly there is high voltage nature of the project and related safety and legal issues. For example, in my country, it is prohibited to work with live parts with voltage higher than 25VAC or 60VDC without qualification which this competition limit to professionals only.

This is an intriguing challenge. Designing a power supply is not an impossible task, especially since a working schematic is provided. For sure it will be tough to beat the performance of the reference design, but that is an objective to strive for, not a requirement in order to win. You are not competing with the manufacturer's design, you are competing against other contestants.

You could or would essentially clone the reference design, except substituting other GaN devices and try little tricks to improve efficiency, like heavier copper. 

Given that there is a free shopping cart to buy components, it would seem to be a good, low-cost, opportunity to try your hand at a power supply design challenge.

I am kicking the idea around a bit, but so far not too keen on spending 6 months on this. I think I have a 2 KW 150V supply in storage, but I can't get near 500V. My normal bench supply only goes to 60V and it can only supply 2A.

Like Michael, it has been a long while since I did heavy duty automotive compliant design.

I have mixed feelings about this type of challenge.

Some companies prohibit their employees from using their talents in this type of event as they might unwittingly reveal company trade secrets.

This is not a trivial challenge, to succeed you will have to have some very good engineering skills and access to some sophisticated test equipment.

There is also the issue of intellectual property.

If someone comes up with a better design, will the sponsor own the rights to manufacture the solution or variants using ideas from the challenge?

Yes, you could win $5,000 , but you could miss out on potentially high value royalties from the sale of the product.

Even if I had the skills, I would not enter this type of challenge without first consulting a lawyer on my IP rights.