9AM Wednesday morning:
My client, ArrayMaster, is announcing a new member of their popular FPGA family aimed specifically at developers of drum-playing monkey toys. The new PercussivePrimate ™ series of FPGAs can reduce the development time for the average drum-playing monkey toy (DPMT) by up to 67% while increasing battery life and allowing the developer to add key, product-differentiating features such as user-programmable beats. The new hard-wired animatronic IP blocks with built-in rhythm generator PLLs allow developers to have DPMT development platforms literally up and drumming within minutes.
If you have a few minutes, our DPMT product-line manager would like to pre-brief you this week for next Monday’s announcement.
Don the PR Dude
Hmmm… As editor of FPGA Journal, I have had very little experience with DPMTs. I do own a set of congas, and I understand some basics of robotics, but I’ll have to do some research to find out if ArrayMaster’s claims are reasonable. However, our fictitious demographics show that 29.2% of our readers are involved in building or prototyping primate percussionist simulators, so I guess the first step is to take the briefing.
I can take a briefing tomorrow morning at 10. Do you have any early customers who would be willing to talk about their experiences with PercussivePrimate?
We are set for 10AM tomorrow with our product-line manager. I’m checking on a customer, but they’re all very secretive about their use of PP FPGAs. They see them as a competitive advantage in the fast-moving DPMT market.
I do some pre-briefing research on DPMTs but find little of help. Wikipedia has a small entry on cymbal-banging monkey toys (http://en.wikipedia.org/wiki/Cymbal-banging_monkey_toy), but my real-world experience with percussion sections tells me that cymbal players are at the lower end of the hierarchy while battery percussion (drum) is held in higher regard. I assume that the same must be true for battery-powered battery percussion.
An hour before the briefing, the PowerPoint slides arrive. The first two slides have a description of ArrayMaster as a company, the next six talk about forecasts for growth in the DPMT market, and the final few go into specifics on the new PercussivePrimate family.
10AM Thursday morning:
After some scrambling about the proper version of the press presentation, we finally get started on the pre-briefing. The product-line manager emphasizes (twice) that the embargo is 8AM Monday – Taiwan time. Apparently DPMT designers are heavily into corporate espionage, and even an hour advance notice of a major technological breakthrough such as this in DPMTs could give an unscrupulous competitor an almost insurmountable advantage. I take the editor’s pledge to keep all information encrypted and under lock-and-key until the critical time… when is that? Let’s see, Taiwan is GMT+8 and I’m on Pacific Standard time is GMT-8 which is…
“…as you know, is a global leader in the production of toy-automation FPGAs, serving a market estimated at 126 billion by 2012…”
“Wait, 126 Billion?!”
“Yes, the global market is forecast to reach 126 billion by the year…”
“The global market for what?”
“Toys. The global toy market – or GTM as we refer to it.”
Oh, this is the find-the-biggest-number-you-can-associate-with-your-product strategy – often used for raising venture capital. I understand now where this guy is coming from.
“So, what percentage of the, uh, GTM is DPMTs, and what percentage of the typical DPMT BOM is FPGAs?” (As an editor, it’s important to learn and use the acronyms quickly — it puts the presenter off their guard.)
“Well, nobody has actually used an FPGA in a production DPMT yet. This is a breakthrough technology.”
“Ah, of course.”
For the next 40 minutes, he continues through his PowerPoint deck, making sure I understand that it’s the hard IP that really differentiates PP from the run-of-the-mill FPGA that other vendors might be pushing as a DPMT solution. The summary slide brings home, once again, the claims of 67% faster development cycles and adds a number to power reduction – 3X less power. OK, this marketing-speak always trips me up. Does 3X less power mean 67% less power? Or, if it previously consumed 50 Watts does it now produce 100 Watts?
Regardless, the power reduction is apparently due to the use of hard IP blocks compared with implementing the same logic in FPGA fabric. I ask a follow-up question on this: “Are people already using FPGAs in DPMTs and just using the fabric for these functions? Or, how does this compare with non-FPGA solutions on power consumption?”
“Well, of course we don’t have any data on other vendors’ FPGAs in this area, but we don’t think any of them are suitable. On the other part of your question, we’ll have to get back to you.”
“And, did you guys find out if a customer would be willing to talk with me?”
“No, we’re still working on that. These guys are really protective, and most of them can’t get permission to talk about their development projects with the press.”
Of course not.
So now the deadline looms, we have only the vendors’ PowerPoint as an information source, and the prospect of talking to a real customer with real experience is looking grim. This is a fact of life in the high-technology trade press. We don’t have the luxury of the “normal” media in walking up to a random guy on the street with a camera crew and recording, “Well, yeah, I used one of them FGPAs in my personal monkey-toy, and, frankly, I think it smelled like chocolate!”
OK, maybe it isn’t such a luxury after all.
I check my usual contacts in the design community, but none of them have any experience in “DT..P… what was that? No, I didn’t even know that was a big market. Monkeys?”
Friday afternoon brings a break. One customer, presumably on the promise of free FPGAs for life, is willing to talk about his DPMT design project. I begin the conference call with anticipation.
“Thanks for speaking with us today. So, have you finished an FPGA-based DPMT yet?”
“DP.. what? Oh, we call them ‘monkey-toy robot – drum edition’ or MTRDE. We have a broad line of monkey-toys including the cymbal-playing…”
“Yeah, I’m up to speed on that. So, how far along are you on your project?”
“Well, we’re still in the early prototyping stages. We’ve got the DPMT development board from ArrayMaster, and we’ve got the servo-arms on there playing a basic beat…”
“So, is this going faster than your previous development projects?”
“Well, there’s no way to tell at this point. We’re just becoming familiar with the ArrayMaster tools and the ArrayMaster IP and integrating the ArrayMaster DPMT development board into our development flow…”
Perhaps he just gets a free development board for each time he works “ArrayMaster” into the briefing…
“What are you seeing in terms of power consumption compared with other solutions you’ve used?”
“Well, I’m not really at liberty to discuss power consumption. Today, I’m just supposed to talk about our use of the ArrayMaster IP with the ArrayMaster…”
“OK, I understand. Well, thanks for your time today.”
Monday morning arrives and we’re in no better shape. I’ve written an article with all the technology claims in it, editorially distanced with prefixes like, “The company claims that power consumption is…” Still, I’m not happy with the information we’re providing our audience. We want to give our readers true, objective insight into the products they’ll be using, and short of putting together a primitive percussive primate myself, I’m left relying on my intuition and wits instead of hard data and user experience.
The article publishes on Tuesday, and the “letters to the editor” start rolling in:
“Kevin, are you crazy? People have been using our CPLDs in DPMTs for years, and these guys are claiming it’s a new market? Nobody with a lick of sense would try to put an FPGA in one of these devices.”
“Kevin, I read with interest your article on FPGAs in DPMTs. We’ve been successfully using another vendors’ FPGAs in our line of mallet-playing toy monkeys (MPTMs) for quite some time. We plan to try out these new devices. Thanks for the info!”
“Dear editor, I am designing DPMT as graduate project and would appreciate you forward all relevant VHDL and application notes for completed DPMT with FPGA. Also, if you could help me get free development board I would thank you very much.”
Is it just me, or does anybody else think this situation could use some improvement?
This year, we’re breaking this mold. We want to include more of your experience in FPGA Journal. If you’ve designed something interesting or novel that takes advantage of the unique capabilities of programmable logic, we want to hear from you. If you’ll drop us a line at firstname.lastname@example.org, we’ll set up a brief e-mail or phone interview where you can tell us the key points of your project. We don’t want you to give away anything proprietary, of course, and we’ll also keep you anonymous if you wish. Your experiences, good or bad, with specific products or companies can give us valuable editorial insight. If your perspective or project is interesting enough, we may dedicate a whole feature article to it. In any event, you’ll be contributing to our collective community, and your peers will appreciate you more for it.
Of course, we’ll still be brining you our normal features, webcasts, and news stories over the course of this year, along with some exciting new features. 2009 should be a great ride!